BattlePenguin

Liberating a PlayStation 5

Fri, 31 Jul 2026 00:00:00 +0000

I’ve previously discussed how I modified my PlayStation 4 to give me control over the hardware I own. Sony has been increasingly anti-consumer. Not to be undone, Microsoft requires linking a Microsoft account to a Sony account for gamers who want to play the first Halo game released for PlayStation, even in single-player mode¹. Sony is also requiring customers purchase additional PlayStation accounts simply to play local, split-screen multiplayer². These increasingly terrible decisions make me glad I took the time to learn how to liberate my PlayStation 5.

Planning

My PlayStation 5 is running firmware 10.40, installed two years ago in 2024. It’s difficult to find decent guides for jail breaking a PlayStation 5, with most of the public exploit code having very vague documentation. This is likely intentional to prevent them from getting flagged and removed from hosting sites like GitHub.

PlayStation 5 Running 10.40 Firmware

From what I could determine, breaking firmware 10.40 still requires exploiting the PlayStation 4 Lua engine found in certain games. Most of these games are only digital downloads, many of which have been patched, so the exploitable version can no longer be purchased from the Sony store.

Digital Game Exploits

Sony also started placing expiration times on new digital purchases, so your console has to connect to the Internet every 30 days to renew those licenses³⁴. This expiration hasn’t been applied to existing downloaded purchases⁵. Sony made a statement to GameSpot claiming that players have nothing to worry about⁶, indicating that the license would change to a permanent one so long as the player checks in once during the 30-day window⁷. I haven’t found anyone confirming if this is true. Even so, I doubt it was the original intent. My guess is that Sony implemented this policy to force game updates in case future exploits are discovered in currently purchased games.

Star Wars Racer Revenge

As far as I could tell, the only viable exploit for my current firmware revision involved using a physical copy of Star Wars: Racer Revenge. Limited Run Games released a PS4 port of this game. Although copies used to sell for around $50 used, after the potential exploit was discovered in January, prices skyrocketed to nearly $500 due to Limited Run only pressing approximately 8,500 copies of this disc⁸.

Star Wars: Racer Revenge Premium Edition

I had avoided going this route for a while, hoping I could wait out for another exploit. Last month, I bit the bullet and purchased this copy used for $280 ($305.90 with taxes and shipping).

The copy of this game was in pristine condition, in what I assume was the original shrink wrap. Most normal people would keep this game untouched in their collection or use it for an unboxing video and playthrough. I opened this perfect copy in hopes of getting administrative access on my PlayStation 5. Would my efforts pay off, or would I be left with an expensive rare copy of a money grab sequel to a game I played on my PC in the 90s?

Lua Save Game Exploit

At first, I tried using the instructions in the GitHub repository for remote_lua_loader by n0llptr⁹. I discovered that save games are encrypted based on your PlayStation user ID, making it difficult to exchange save games with others. Most of the instructions I had found involved uploading a zip of your saves to a cloud provider like Google Drive and then sharing them with some random Discord bot. Although I could have installed Apollo Save Tool¹⁰ on my PS4, I found an online service called Garlicsaves¹¹.

I loaded Racer Revenge on my PS5, created a save game, exported it to a USB drive and decrypted it using Garlicsaves. I then added the payload from the remote_lua_loader project and used Garlicsaves to encrypt and sign the saves. I tried loading them on my PS5 and… nothing happened.

Looking more closely at the remote_lua_loader exploit, Star Wars: Racer Revenge wasn’t on the list of supported games. It turned out the exploit I needed was Luac0re¹². Its repository contains a save for the game in a much larger 44 MB image format to contain all the payload data. Resigning it with Garlicsaves would result in a load error from within the game, as it would truncate the save to the default size and discard the extra data. I found another online save editor called Savely¹³¹⁴ that successfully resigned the larger save game file and kept the payload intact. From there I simply started the game, selected the Options > Hall of Fame menu option and watched the exploit code execute.

Luac0re 2.4 Waiting for Remote Lua Payload

I do hate that I couldn’t find a local tool for unpacking and signing PS4 save files. After all, websites can go offline and disappear from the Internet. Thankfully, it’s an operation that only needs to be performed once.

Send the Payload

Luac0re’s documentation referred to remote_lua_loader for save game signing instructions, and so I also used that repository for the Lua Patience to Jailbreak (P2JB) payload. However, the payload delivery Python script would just hang. Hitting ctrl+c would close the connection, and the PS5 would pop up a dialog acknowledging the payload had been received, similar to my use of nc on the PS4, but nothing would happen. Video tutorials I found showed the payload script also reading status information from the running application¹⁵, which made me wonder if the Python script was designed for Windows and if the Linux socket implementation was preventing full payload delivery. However, it turns out Luac0re has its own, much simpler, payload delivery script written in Python¹⁶, which is what I should have been using. This delivers and executes the Lua payload correctly.

The P2JB exploit can take between one and two hours to run. It’s helpful to go into Settings > System > Power Saving and adjust the power-down time when running this. I initially made the mistake of letting the Lua script sit at the opening dialog, thinking it was running. I wondered why it was taking so long, looked at the source code, and noticed the shell code didn’t start running until after this dialog was dismissed. 🤦

Blu-ray Exploit

Once P2JB ran successfully, I now had administrative access on my PS5. However, it’s not persistent. There is another exploit that allows for loading custom code from a jar file off a Blu-ray disc, but it was patched in firmware after 7.61. The bdj_unpatch payload¹⁷ needs to be run to undo this firmware patch.

Luac0Re 2.4 ELF Loader

After applying bdj_unpatch, the system can be rebooted. Now it no longer takes an hour and multiple steps to jailbreak, it simply requires a Blu-ray disc. It must be a BD-R. Not a DVD-R or CD-R, only a BD-R because it works using the Java subsystem specifically for Blu-ray players¹⁸. Luckily I have an optical drive in one of my machines I use for ripping audio CDs. I decided to get a full Blu-ray burner years ago, just in case I needed it. I sold my last optical writer when I left the country for a few years in 2012. I bought a 10-pack of BD-Rs for $14 from a local store and realized this was the first time I had burned a disc in at least 12 years! Thankfully, cdrtools still exist, and I was able to use the classic cdrecord command to find my optical drive and burn a BD-R image.

Blu-ray Recorder

# Determine the correct device

cdrecord --scanbus
Cdrecord-ProDVD-ProBD-Clone 3.02a09 (x86_64-pc-linux-gnu) Copyright (C) 1995-2016 Joerg Schilling
Linux sg driver version: 3.5.36
Using libscg version 'schily-0.9'.
scsibus6:
        6,0,0   600) 'HL-DT-ST' 'BD-RE  WH14NS40 ' '1.05' Removable CD-ROM
        6,1,0   601) *
        6,2,0   602) *
        6,3,0   603) *
...
...

# Write the image
cdrecord dev=6,0,0 -eject -sao driveropts=burnfree ./BD-UN-JB_1.1.iso

At first, I tried to use BD-UN-JB and deliver payloads to it. I had issues with sending the jar files to the loader, and instead I decided to use PS5 BD-JB Autoloader¹⁹, a tool I had seen in a video tutorial¹⁵. PS5 BD-JB Autoloader is a fork of BD-UN-JB that automatically launches a Payload Manager when it starts. It can copy executables in ELF binaries from a USB device and manage their execution.

cdrecord dev=6,0,0 -eject -sao driveropts=burnfree ./ps5-bd-jb-autoloader-v1.4.1-2d36e16.iso

Recommended Payloads

ShadowMountPlus - Can load programs from UFS images with the .ffpkg extension
Kstuff-lite - Needed for spoofing signature checks in order to run homebrew applications and packages
Lapy JB Daemon - Standalone homebrew payload for PS5-Xplorer and similar apps
ftpsrv - FTP server to transfer files from/to your PS5

Creating ffpkgs

ShadowMountPlus can load images in UFS, a FreeBSD file system. This can be done using native tools in FreeBSD, but I wanted to be able to build packages on Linux. UFS2Tool is a cross-platform tool written in C# that makes it easy to create UFS images on Windows, macOS or Linux²⁰. If images are created ending with the .ffpkg extension, they can be placed in one of the default scan paths for ShadowMountPlus (such as /data/homebrew) and be loaded automatically.

Tutorial for Firmware 10.40

I mostly went by the PS5 FW 12.70 Jailbreak Tutorial by Alex Free²¹. It was well laid out, but some of the tools and instructions were outdated or weren’t intended to be used with games other than Racer Revenge. I would highly recommend looking around for better options first before trying to source this obscure Star Wars game. This scene moves rapidly, and by the time you are reading this, there may be an easier or free way to gain administrator access on the PS5 10.40 and above firmware. For reference, here are the steps I took.

Obtain a physical copy of Star Wars: Racer Revenge
Go to Settings > System > Power Saving to temporally increase the sleep timeout for your PS5 (optional, helpful for P2JB)
Download and extract the Luac0re Save file and Payload
Digitally sign the save file for your game’s region (CUSA03474 for the US or CUSA03492 for Europe) using Savely Editor
- Savely will need your PlayStation Account ID and Username
- Your Account ID can be found by exporting a PS4 game to a USB drive
- Go to Settings > Storage > Console Storage > Saved Data > PS4 Games
- Select any save data and run Copy to USB Drive
- You will see PS4/SAVEDATA/xxxxxxxxxx/CUSAnnnnn on the USB drive.
- The xxxxxx is your Account ID.
Copy the signed save to a USB stick and insert it in your PS5
Go to Settings > Storage > Console Storage > Saved Data and Game/App Settings > Saved Data (PS4) > Copy Or Delete From USB Drive > Copy From USB Drive and copy the save files to your console
Start Star Wars: Racer Revenge and go to Options > Hall of Fame
You should see the payload screen displaying an IP address and port
Within the Luac0re project, run python payload_sender.py <PS5 IP Address> 9026 payloads/p2jb.lua on a computer that’s on the same network as your PS5 to load P2JB
Close the dialog to start the exploit process
Wait 1~2 hours. A dialog will tell you when the process is complete with an IP and port for the ELF loader
Download the bdj_unpatch ELF binary from BD-UN-JB Releases
Send the bdj_unpatch payload using cat bdj_unpatch_1340.elf | nc <PS5 IP Address> 9021 to unpatch Sony’s Blu-ray exploit fix
Download the latest ps5-bd-jb-autoloader-v<version>-<hash>.iso from PS5 BD-JB Autoloader and burn this image onto a Blu-ray disc (it must be a BD-R. It will not work from a DVD-R or CD-R¹⁸)
Restart the PS5
Insert the PS5 BD-JB Autoloader disc
Hit R1 on your controller from the PS5 main menu to navigate to media
Start the Autoloader disc
Copy the following ELF payloads onto a USB drive
- ShadowMountPlus - UFS images (.ffpkg) handler
- Kstuff-lite - Spoofs signature checks for homebrew and packages
- Lapy JB Daemon - homebrew payload for PS5-Xplorer and similar apps
- ftpsrv - FTP server
Insert the USB drive and use Payload Manager to copy these payloads to your console and run them on startup
(Optional) Go to Settings > System > Power Saving and restore your sleep timeouts for your console

Now, all you need to do if your PS5 is power cycled is simply run the Blu-ray disc from the media tab. It will automatically load the payloads you’ve selected.

PS5 BD-JB Autoloader

Closing Thoughts

Modding a console in the early 2000s typically involved soldering in an integrated circuit and adding a few bodge wires. Modern consoles have gotten considerably more sophisticated with layers of security including signed updates and hypervisor isolation, all designed to prevent people from running their own code on hardware they purchased. The World Wide Web of the 90s was also filled with hacking and security related search engines, making it easier to find other people who were modifying their own hardware and making amazing demos.

I don’t think enthusiasts in that world would even consider buying most of the hardware that’s available today. The first iPhone opened a market to consumers for high end compute that users didn’t have administrative rights to. That started a trend of normalizing software-as-a-service and hardware that feels more like something you lease than you own.

Most consumers just accepted buying programs that are locked to accounts controlled by companies. It’s designed to feel frictionless, until things go wrong. One user spent three years in a legal battle with Sony regarding a permanent ban on his stolen account, which had over $3,000 worth of purchases attached to it²²²³. A similar situation arose over an Xbox account in Brazil, resulting in another long legal battle in which the courts sided with the consumer and forced Microsoft to restore the account²⁴²⁵.

For several years, I’ve exclusively bought physical media for the few consoles I own. I’ve refused to go the route of digital downloads except in the case of PC games via stores such as Steam and GOG. I’ve participated in the system and even compromised on the PC. Yet, consumers like myself cannot vote with our money against a horde of people who do not care at all about their privacy.

In the end, Sony and game developers are getting rid of physical media for the last few holdouts like myself. Yet, thanks to the work of countless people poking around and discovering holes in consoles, it’s possible to liberate these systems and not be subject to a world where Sony controls everything about a device I purchased.

Halo: Campaign Evolved on PS5: How to Link Your Microsoft Account and Fix Sign-In Errors. 23 July 2026. SteelTron. GamerFuzion. ↩
Halo Campaign Evolved PS5 Split-Screen Requirement. 20 June 2026. Bhatnagar. Phrasemaker. ↩
Hugely terrible DRM has now been rolled out to all PS4 and PS5 digital games. Every digital game you buy now requires an online check-in every 30 days. If you buy a digital game and don’t connect your console to the internet for 30 days, your license will be removed.. 25 April 2026. @manfightdragon. X. ↩
Sony Just Killed Game Ownership: New PlayStation DRM Locks Your Games. 1 May 2026. Karan. Techwiser. ↩
PS4/PS5 Jailbreak News: PSN Game Expiry, PS5 13.20, Linux Update, PS5 Xplorer 1.05 & More!. 24 April 2026. MODDED WARFARE. ↩
PlayStation Users Report New Online License Checks, But Sony Quells Concerns. 29 April 2026. Gurwin. GameSpot. ↩
Sony responds after 30-day owned game timer spotted on PlayStation - Dexerto. 30 April 2026. Warby. Dexerto. ↩
Star Wars: Racer Revenge PS4 Price Skyrockets After PS5 Jailbreak Rumor. 3 January 2026. Hansen. Star Wars Gaming. ↩
GitHub - n0llptr/remote_lua_loader. GitHub. Retrieved on 29 July 2026. ↩
Apollo Save Tool (PS4). bucanero. GitHub. Retrieved on 29 July 2026. ↩
garlicsaves. garlicsaves.com. Retrieved 29 July 2026 ↩
Luac0re: Luac0re is a mast1c0re variation that uses Lua scripting for easier exploit development.. Gezine. GitHub. Retrieved on 29 July 2026. Mirror ↩
[Free] Online PS4 Save Resigner and Save Editor Tools. 28 Feburary 2026. unknownv2. GBAtemp. ↩
Savely. savely-editor.com. Retrieved 29 July 2026. ↩
Unpatching the Blu-Ray Jailbreak with P2JB. 14 May 2026. MODDED WARFARE. ↩ ↩²
Luac0re payload_sender.py at main. Gezine. GitHub. Retrieved on 30 July 2026 ↩
Releases · BD-UN-JB. Gezine. GitHub. Retrieved on 30 July 2026. ↩
DVD instead of Blu-ray?. 25 May 2026. NeirX. r/PS5_Jailbreak. Reddit. ↩ ↩²
ps5-bdjb-autoloader: An automated payload loader for exploited PS5 consoles. itsPLK. GitHub. Retrieved on 30 July 2026. ↩
UFS2Tool: A FreeBSD UFS1/2 Filesystem Manager for Windows, macOS & Linux.. SvenGDK. GitHub. Retrieved on 31 July 2026. ↩
PS5 FW 12.70 Jailbreak Tutorial. Free. 31 May 2026. Retrieved 30 July 2026. ↩
To the folks still defending Sony’s pathetic decision to remove physical media from their production, here’s a solid reminder why I will never support and you shouldn’t either. I had to literally sue Playstation because my OG account was wrongfully banned and it took me 3 years…. 13 July 2026. @PrimitiveAK. X. ↩
After a 3 year battle with PlayStation regarding my account being stolen, then Permanently banned, PlayStation Moderation team from BBB gave me a call today to resolve the issue and unbanned my account and gave it back to me. Beyond blessed. This account is my childhood. https://t.co/mn9vVxLFn1. 14 June 2021. @PrimitiveAK. X. ↩
Sounds similar to what another gamer had to do with Xbox after they lost their account and had to sue. Players need to stop defending a digital only future, it benefits only the corporations and not you, the player….. 13 July 2026. @PenguinLamont. X. ↩
Microsoft loses Brazilian court case after telling hacked Xbox user to re-purchase games. 13 July 2026. PCN Staff. Pop Culture News. ↩

Game Reviews: Firewatch, Hitman Go and Leisure Suit Larry 6

Tue, 28 Jul 2026 00:00:00 +0000

I haven’t been playing a whole lot of games lately. Over the past year I finished two titles on my PlayStation 4. The first was the narrative walking adventure Firewatch, and the second was the relaxing turn-based puzzle game Hitman Go. To finish things out, I took a trip down memory lane and played the raunchy 1990s classic Leisure Suit Larry 6: Shape Up or Slip Out! using ScummVM on Linux. It was a good mix of relaxing and engaging games with varying degrees of puzzles and story.

Firewatch

In the deep wilderness of Wyoming, you play Hank, a hipster who’s taken a job in a fire tower. The game begins with a series of text choices where you build out Hank’s backstory. Between this text adventure, you are also given scenes of hiking deeper into the woods. When the backstory completes, you’re in a fire observation tower, talking to your supervisor, whose tower you can see in the distance. The game is a narrative walking simulator with a lot of exploration and mystery, but it’s not intense with puzzles. It’s an engaging, yet short and relaxing, game.

Your character talks to himself very little and has no inner monologue. All the details of the story are expressed by talking to your supervisor over a radio. She is usually giving you instructions or making conversation from a distant fire tower. Throughout the game, you’ll find notes, clipboards and pieces of story. Several odd occurrences lead to a creepy mystery, which occupies much of the game’s plot.

There are very few other character interactions throughout the adventure. The writers really try to push the beautiful nature of being alone in the wilderness and wrestling with one’s inner dilemmas. There are themes of escape, conflicts, guilt and a journey of coming to terms with difficult life situations. When the story is all said and done, it does appear to tell a complete and compelling story with major growth arcs. When you further scrutinize the narration, there are a couple of issues and holes throughout, and some of the themes feel a bit heavy-handed. Many of the conversation decisions likely lead to the same place, and the story is fairly linear. Still, it feels about the right length and is generally well-rounded.

I realize this is an independent title, but I still feel where the game really suffers is the visuals. The game designers try to build a fairly large world. Finding tools helps open up paths that are carefully arranged to be unlocked as the story progresses. There is a map, which you have to manually bring up, but it does thankfully have a convenient dot indicating your location (even though the story takes place in the 80s, long before cellphones and GPS technology were generally accessible). The developers could have added a hard mode where there is no visual indication of your position on the map, forcing the player to use visual cues and signs. However, I don’t think that would have worked well.

There are some generally pretty scenes in this game, but overall the scenes don’t have a high level of detail or fidelity. The game pushes exploration and enjoying the beauty of the world, but it hardly has the artwork for the immersion I feel the game designers wanted. Where the game really shines is the dialogue and character building. The focus on the protagonist and his supervisor, and the challenges they overcome, is really well-developed storytelling. The underlying mystery is a good plot driver, but honestly doesn’t feel as well constructed.

Overall, I enjoyed this relatively short walking simulator. It doesn’t really have any major puzzles, with most of the game advancement based around figuring out transversal. The 3D world is a coaster for the story and narrative. It’s not a great story, but it is a decent mystery with excellent human elements and enjoyable voice acting.

Summary

Wilderness storytelling and an interesting dialog format, suffers slightly from mediocre graphics, but still provides a well-acted, short adventure in the form of a transversal simulator.

68%

Hitman Go

Hitman Go has the same main character as the classic first-person stealth combat game, but with a very different format. This game is a puzzler, where the player must guide an assassin, move-by-move, through dozens of scenarios to eventually assassinate a target. It’s a much slower and more casual counterpart to its namesake. The puzzles are challenging, the artwork is creative, the settings are well constructed and the overall game is a fun little companion for anyone who loves classic puzzles with board game mechanics.

I really like the level design of Hitman Go. All the art is designed to make this video game feel like a board game. Each level is like an intricate chessboard. All the enemies are portrayed as figures that automatically move around on each turn. Each level has three goals, presented as a set of cards. Several of the goals are mutually exclusive, so the levels must be played multiple times to meet every challenge. There is no dialog or explicit story, but each scene is beautifully crafted to the point where the player can easily create their own narrative for the various levels.

Each set of levels is portrayed as a board game box. Completing enough challenges unlocks different sets of levels. I’m usually not a fan of games that require a certain amount of challenge points to unlock further stages, like Sackboy: A Big Adventure. However, few of the additional challenges are that difficult once you’ve already made it through a level. I never found myself feeling stuck, having to go back and replay several scenarios just to open up new areas.

The puzzles themselves can get pretty challenging, yet none of them felt insurmountable. Most of the time, if I got stuck, I could just come back after a few hours or a day and easily find a solution. There is a built-in hint system, but I only used it three or four times out of the dozens of stages I played.

Hitman Go is a great, relaxing puzzle game. The design is creative, the mechanics are interesting and the game increases difficulty at a reasonable pace. Most of the challenges follow a straightforward logical progression. The music and sound add an excellent ambiance to a very complex, assassination-style chess game. There were some frustrating moments, but overall this is a well balanced, casual puzzle game.

Summary

Assassination as a board game, Hitman Go is a challenging yet fun puzzler with amazing art and level design.

78%

Leisure Suit Larry 6: Shape Up or Slip Out!

Leisure Suit Larry 6 was the first installment of the series to have full voice acting on CD-ROM. I was probably too young when I played this game, and the jokes certainly hit way different now that I’m older. It has the same problems as many classic point-n-click adventure titles, but LSL6 still has some creatively absurd art and whimsical music. I don’t think it’s quite as funny as the original creator/writer envisioned, but it still has a weird and raunchy adorableness as your creepy protagonist from a bygone era attempts to make it with a woman way too young for him in a health spa prison.

The introduction has our protagonist, Larry Laffer, compete in an embarrassing 90s era dating show called Stallions. His participation wins him a visit to the La Costa Lotta resort, where the entire game takes place. The graphics have the haphazard art-deco style that’s been present since the first remake of LSL1. The music is cute and whimsical for the era. The story revolves around Larry trying to make it (often failing) with several of the women at the resort, ultimately building up to the climax of the one woman he’s able to successfully woo under his spell. It’s pretty silly, but still a good laugh.

Unlike other Sierra games, the toolbar is always present. The ability to pick up items versus using items has been separated into two different actions, and the inventory is always visible. Other than that, it suffers the same way all adventure games of this era did: requiring the user to click everything on everything until finally giving up and looking up the answer on a pre-Internet bulletin board or ordering a hint book. Replaying these games is fun because you can see how much you remember before getting stuck. I scored over 600 points before I broke down and looked up hints. I only needed hints for two puzzles and finished the game with over 900 points.

Honestly, a lot of the jokes from this game weren’t that funny when they came out. Although there are only a few hotel room doors you can enter, you can knock on several of the resort suites and hear funny little bits. Most of them are lame, but some are really quite funny or have references to long-forgotten cultural events from the 90s. There are some really great moments that honestly couldn’t be replicated at all due to modern cultural norms, such as the reveal with Shablee. Ultimately, this game revolves around fetching items or solving puzzles for each of the women you encounter, leading to the ultimate goal of a woman in a penthouse suit. Its emphasis is on sexual conquest, with each woman moving Larry closer to the endgame. It’s a classic girl-in-the-tower idea with a softcore-pornographic theme, wrapped in a game that’s not very different from LSL1 or any other game in its genre.

I really enjoyed the trip down memory lane with this game, both in trying to remember all the puzzles, and also in all the little aspects of life from this era. The gags aren’t great, but there are some that are hilarious and amazing in the sense they couldn’t be replicated in video games today. The music is catchy, the artwork is weird but still fun and unique. Larry is a creepy dude who is somehow able to score women way above his status. It’s a fun adventure that might have screwed me up a bit as a teenager, but it certainly hits a bit too hard today. It was fairly short and definitely worth the modern-day replay.

Summary

Raunchy point-n-click adventure game that pushed the limits for its era with a few classic jokes and crazy art direction, yet still has all the shortcomings of the adventure game genre.

65%

The End of Physical Media and The PS4 Pirate Pro

Tue, 21 Jul 2026 00:00:00 +0000

Although I’ve bought some digital titles on PC (via Steam, GoG and other stores) and a few titles on a very old PS Vita, starting with my PlayStation 4 Pro, I decided to only purchase games on physical discs. Same with my Switch and PlayStation 5. Last year I purchased Hogwarts Legacy for the PS5 and was horrified, after playing the introduction, that an Internet connection was required to download the actual game. It was clearly written on the box. I had missed it, and now I could no longer return the game. I kept my consoles disconnected from the Internet, and the frustration with Hogwarts Legacy was the final straw I needed to push me toward finally liberating my PlayStation 4.

Sony being Anti-Consumer

Sony recently announced they were ending the production of physical discs for new games on their consoles in 2028¹²³, which has greatly upset some gamers⁴. This news came immediately after Rockstar Games announced that Grand Theft Auto VI will not come on disc either, with retail versions simply having a download code in their box⁵. These actions have brought up the classic Sony advertisement from 13 years ago, where Sony showed that sharing a game for the PS4 was as simple as sharing a disc⁶, an ad that has not aged well.

On top of all of this, back in 2023, Sony removed many shows they lost licenses for after the Discovery Warner Brothers merger. They deleted these shows off of people’s devices without offering any type of refund⁷. This happened again this year with Sony deleting movies from accounts after they had “bought” them, with no refunds offered⁸. The average consumer is slowly being boiled by the concept that they don’t own any of their digital purchases, unlike actual discs or media that can be played offline and preserved. Sony has faced zero repercussions for these insane actions, showing once again that, in the era of digital rights management, piracy is completely justified⁹¹⁰.

Taking Ownership of my PS4

I had kept my PS4 disconnected from the Internet for quite some time. A quick firmware check indicated my system software was at version 8.03, old enough to make use of exploits in order to run my own code on the device I had purchased.

PS4 Pro with 8.03 Firmware

Exploiting my firmware was a multistage process. First, I had to upgrade to version 9.00 of the firmware to run the homebrew enabler I wanted to use¹¹. I didn’t want to connect to the Internet, so instead I put the official Sony firmware on a USB drive to perform the update. After that, I connected a Raspberry Pi directly to the PS4 via Ethernet. The Pi had a DHCP server to assign my PS4 an IP address. It also had a web server running on it, allowing me to use the PS4’s web browser to run the pOOBs4 exploit¹². This exploit contains a web component, but also requires the user to plug in a USB drive with a custom corrupt FAT filesystem on it after the web component executes. This allows a hole through which a payload, or custom code, can be sent and run on the PS4. I then used netcat (nc) on the Raspberry Pi to deliver and load GoldHen, an enabler that allows for running custom code and homebrew packages on the PlayStation¹³.

This isn’t a permanent jailbreak. If I ever lose power or reset the PlayStation, I have to go through the entire dance of opening a web browser, loading the exploit page, plugging in the USB stick and running cat goldhen.bin | nc 192.168.15.16 9020 to load GoldHen on my console (I also have to hit control+c on the nc command after a few seconds to force it to close the connection as I couldn’t figure out how to get it to close on its own after sending the payload). It’s not really a complicated process once you know it, and there are several developers who have created firmware for embedded devices that handle the web server and payload delivery.

Newer Raspberry Pis can run in USB Gadget mode, where they can emulate USB devices such as cameras or mass storage. I wanted to write my own all-in-one solution that allowed me to select a payload on the PS5’s web browser, run the web exploit, emulate the corrupt USB storage image and deliver the payload all in one easy flow.

The PS4 Pirate Pro

Creating such an application is not very complicated, but it still requires time I didn’t want to invest just to make starting up my PlayStation 4 a little more convenient. I ended up spec coding this application. Anthropic’s current models, Fable and Opus, refused to work on the specs for this program due to the reference to pOOBs4, even though it was a four-year-old exploit. Instead, I ended up using GLM-5.2 as the model for creating the PS4 Pirate Pro. It took several iterations in order to get it to properly display error messages to the user and successfully deliver the GoldHen payload, but overall it works well and has greatly simplified the jail breaking process on my PS4.

You can use the PS4 Pirate Pro with any Raspberry Pi Linux distribution. I decided to go with Void Linux. The onboard Ethernet is connected to my local network, and the USB-C is connected to the PlayStation’s USB-A ports (make sure it’s a data cable and not just a power delivery cable). I use an external network adapter to connect to the PlayStation 4, but my previous version had the onboard adapter connected to the PS4 and Wi-Fi connected to my local network.

PS4 Pro Connected to Raspberry Pi 5

Networking

In Void Linux, static network configuration is just done in /etc/rc.local. I use this to establish the network the PS4 will live on. Void also uses standard Ethernet names like eth0 and eth1. Other distributions may use unique adapter names or recommend you put static network settings in specific configuration files.


# /etc/rc.local
ip link set dev eth1 up
ip addr add 192.168.15.1/24 dev eth1

I installed dnsmasq in order to give my PS4 an IP address via DHCP. You can use any DHCP server, but the configuration for /etc/dnsmasq.conf, assuming the PS4 is connected to eth1 and your LAN is on eth0, would look something like this (with the MAC address replaced with the one on your PS4).

except-interface=eth0
dhcp-range=ps4,192.168.15.10,192.168.15.20,12h
address=/ps4pirate/192.168.15.1
dhcp-host=c8:63:f1:XX:XX:XX,192.168.15.16,ps4pro
no-resolv

USB Gadget Support

USB gadget support is only available on certain Raspberry Pis, such as the Pi3 Model A+, Pi4, Pi5 and various Pi Zero boards¹⁴. For Void Linux, I installed the rpi5-kernel package. I’m not sure if this was entirely necessary.

# Install the Raspberry Pi kernel specific to your device
# (in my case, the Pi5). Not sure if this is necessary

xbps-install -Sy rpi5-kernel

I then added configfs to /etc/fstab, which will be used by PS4 Pirate Pro to enable USB gadget and mass storage mode.

echo "configfs  /sys/kernel/config  configfs  defaults  0  0" >> /etc/fstab

The libcomposite and g_mass_storage kernel modules are required to emulate USB mass storage devices. They can be added to a configuration in /etc/modules-load.d so they are loaded at boot:

echo -e "libcomposite\ng_mass_storage" > /etc/modules-load.d/usb-gadget.conf

Finally, add the correct overlays to /boot/config.txt for the Pi’s configuration (this file might be in different locations depending on which Linux distribution you use):

echo "dtoverlay=dwc2,dr_mode=peripheral" >> /boot/config.txt

You can have multiple dtoverlay statements in config.txt (Void Linux comes with one to enable the 3D graphics systems), so appending a new one to the end of the file will simply load it after the others. The overlay name, dwc2, is a reference to the DesignWare Core drivers originally used for USB-OTG mode¹⁵.

Installation

Now that all the pre-configuration is complete, the PS4 Pirate Pro can be installed. This is a Python application that serves the website needed to jailbreak the PS4, emulates the USB mass storage device for the second stage of the exploit and finally delivers our payload. It does need to run as root for now in order to create the USB storage device. I’m going to place it in /opt/ps4piratePro/. I’m using git to clone the repository directly on the device, but you can use rsync, scp or other tools to sync it from a local device to the Pi if you prefer.

# Install tools we need on Void Linux
xbps-install -Sy git uv

# Clone Repository
cd /opt
git clone https://forge.sumit.im/djsumdog/ps4piratePro.git

# Build the Python virtual environment
cd ps4piratePro
uv sync

There will be an example configuration file at /opt/ps4piratePro/config.yaml.example. Copy this to /opt/ps4piratePro/config.yaml and adjust if needed (e.g. if you used a different IP address for your PS4).

web:
  bind: 0.0.0.0
  port: 80
ps4_payload:
  address: 192.168.15.16
  port: 9020 
payloads: /opt/ps4_payloads

In the above configuration, the service will expect all the possible binary payloads to be in /opt/ps4_payloads. It will present the user with every .bin file in that directory. This is where you can put releases for tools like GoldHen¹³.

Finally, the service files to start the program need to be created. Void uses runit, and the following creates and enables runit service files. Other distributions may require systemd or openrc versions of these initialization scripts.

# Create the Service and run scripts

mkdir -p /etc/sv/ps4piratePro

cat << 'EOF' > /etc/sv/ps4piratePro/run
#!/bin/sh
exec 2>&1
cd /opt/ps4piratePro || exit 1
exec uv run python -m ps4pirate_pro web
EOF

# Set the script to be executable

chmod +x /etc/sv/ps4piratePro/run

# Enable the service

ln -s /etc/sv/ps4piratePro /var/service/

# Create and enable logging  for the service

mkdir -p /etc/sv/ps4piratePro/log
cat << 'EOF' > /etc/sv/ps4piratePro/log/run
#!/bin/sh
exec svlogd -tt /var/log/ps4piratePro
EOF
chmod +x /etc/sv/ps4piratePro/log/run
mkdir -p /var/log/ps4piratePro

Usage

If the service started successfully, navigating to the hostname (which I configured as ps4pirate previously in dnsmasq.conf) or IP address of the Raspberry Pi from the PS4’s web browser will present a payload selection screen and move through the process of jail breaking the PS4.

PS4 Pirate Pro Payload Selection

PS4 Pirate Pro USB Exploit

GoldHen Menu

The PlayStation 5

PlayStation 5 Running 10.40 Firmware

I’ve kept my PlayStation 5 disconnected from the Internet ever since downloading Hogwarts Legacy. Even games on physical media will contain system updates, both for the Nintendo Switch and Sony consoles, which can patch valuable yet-to-be-discovered exploits. Currently my PS5 is still on firmware 10.40, and there are some options for jail breaking such devices. My next project is to do the same thing with my PlayStation 5, using some of the known exploits for my firmware version currently circulating in the wild.

Physical disc production ending in January 2028 for new games releasing on PlayStation consoles. 1 July 2026. Shuman. PlayStation Blog. ↩
Sony PlayStation Will Stop Releasing Games On Discs In 2028. 2 July 2026. BeauHD. Slashdot. ↩
HW News - DRAM Antitrust Lawsuit, ID Verification, USA Wants 5% of OpenAI, Sony Drops Discs. 10 July 2026. Gamers Nexus. ↩
Does This Make You Mad Sony?. 7 July 2026. videogamedunkey. ↩
GTA VI is a worrying sign for the future of physical games. 24 June 2026. Peters. The Verge. ↩
Official PlayStation Used Game Instructional Video. 10 June 2013. PlayStation. ↩
PlayStation To Delete A Ton Of TV Shows Users Already Paid For. 1 December 2023. Gach. Kotaku. ↩
Sony Deletes More Movies From Accounts of People Who ‘Bought’ Them. 17 July 2026. BeauHD. Slashdot. ↩
Sony stole what you purchased AGAIN: piracy is COMPLETELY JUSTIFIED. 29 June 2026. Louis Rossmann. ↩
Steal This Comic. 13 October 2008. xkcd. ↩
PS4: GoldHEN 2.4b17.2 released, adds support for Firmwares 10.00 / 10.01. Support for 9.60 is next. 21 May 2024. wololo. Wololo.net. ↩
pOOBs4. ChendoChap. Github. Retrieved on 18 July 2026. ↩
GoldHEN: PS4 Homebrew Enabler. GoldHEN. GitHub. Retrieved on 18 July 2026. ↩ ↩²
USB gadget mode in Raspberry Pi OS: SSH over USB. 21 January 2026. Oberosler. Raspberry Pi. ↩
What does “dtoverlay=dwc2” REALLY do? 29 December 2017. Sanders. Raspberry Pi Stack Exchange. ↩

The Systematic Removal of Security in Consumer Operating Systems

Mon, 20 Jul 2026 00:00:00 +0000

Slide to Unlock by Solstad

This year we’ve seen cracks in Windows 11 BitLocker, VeraCrypt developers losing their ability to distribute their software on Windows, and Apple removing encryption support from the HFS+ filesystem. We’ve also seen some cases where public outcry and awareness were able to reverse some of the corporate decisions that removed the freedom of users to run secure software. There was a time when people in tech stood against government mandated keys under doormats¹². Yet in today’s surveillance age, every big company seems like a willing participant in the coming technocracy.

BitLocker

Earlier this year, a security researcher who goes by Nightmare Eclipse found a means to bypass BitLocker encryption on Windows 11 and Server 2025, claiming it looks suspicious enough that it was likely an intentional backdoor by Microsoft³⁴. Some people claim using a Trusted Platform Module (TPM) and boot pin prevents this exploit⁵. The original author has claimed TPM+PIN was exploitable, but he didn’t publish the proof of concept⁶. A few weeks later, Nightmare Eclipse published another BitLocker bypass that affected users who used Defender Offline Scan⁷⁵.

Nightmare Eclipse was criticized for not responsibly disclosing such a major vulnerability, to the point where Microsoft threatened legal action against the developer⁸. In response, Eclipse claimed Microsoft had ignored his previous disclosures and deleted the account he had used to report them⁹. Several other researchers and developers also spoke up about how Microsoft Security Response Center (MSRC) had a history of closing legitimate reports and silently patching vulnerabilities, not giving credit to the original reporters and refusing to pay out bug bounties¹⁰¹¹¹².

VeraCrypt and WireGuard

VeraCrypt, an open-source disk encryption system based on TrueCrypt, had their publishing account for Microsoft Windows terminated in March, removing their ability to publish updates¹³. Tools like VeraCrypt require kernel-level implementation, and users get scary red boot screens if they try to disable verification for these low-level drivers.

This policy change by Microsoft affected other open-source security projects, such as the popular VPN protocol WireGuard¹⁴. Microsoft claimed this was due to new requirements of mandatory account verification for partners in their hardware drivers program¹⁵, but developers claimed they completed these new verification requirements and were still locked out of their accounts anyway¹⁶. Microsoft VP Scott Hanselman publicly responded, stating he was actively working to restore the accounts for WireGuard and VeraCrypt¹⁷ and that “Not everything is a conspiracy sometimes it’s literally paperwork¹⁸.”

VeraCrypt was forked from the TrueCrypt project after it was very suddenly discontinued by the original authors. TrueCrypt’s project page put up a very cryptic message encouraging users to switch to their operating system’s native encryption¹⁹ and included instructions for enabling BitLocker²⁰. Some have wondered if shutting down the development was a type of warrant canary; a response for a National Security Letter²¹. A TrueCrypt author claimed at the time that forking the project was “impossible²²²³²⁴,” and yet here we are over a decade later with VeraCrypt being fully maintained and actively developed.

MacOS and Extended Filesystem Encryption

Apple announced that macOS 28 will drop support for HFS+ encrypted volumes²⁵²⁶ and posted a support article indicating users should back up their data and migrate to APFS encryption²⁷. What’s interesting is that they’re not dropping support for their legacy filesystem entirely, just versions of it that are encrypted. It begs the question, is support being dropped because the previous filesystem had strong encryption? Is Apple trying to move users to formats that aren’t really secure, allowing for their own backdoors to be used at government requests?

AMD Hardware

AMD recently removed support for Transparent Secure Memory Encryption (TSME), also known as Memory Guard, from their consumer processors. These came via updates to their processor firmware packages used on various operating systems, including Windows and Linux. Due to public outcry over the removed feature, AMD did eventually reinstate memory encryption²⁸²⁹.

Conclusions

Each one of these stories individually could have a completely benign explanation. Many companies are trying to tighten security in the new era of large language models and tools that make it easier to find security issues and vulnerabilities. Some developers and companies may want to drop unmaintained code to reduce attack footprints, such as the case with Apple and legacy filesystems. However, all of these stories coming so close together, back-to-back within weeks of each other, suggest a different picture.

I don’t believe a Microsoft VP when he claims the VeraCrypt and WireGuard situations were a “paperwork” issue¹⁸. Microsoft has aggressively made it impossible for normal consumers to set up Windows 11 without an Internet connection or Microsoft account during their initial setup. Local accounts on Windows 11 require looking up specific terminal commands as if they were game cheat codes. Meanwhile, companies like Discord are requiring government ID for some of their users³⁰, and many governments are pushing legislation for mandatory ID verification for web services and operating system use under the lie of protecting children³¹³²³³. It’s obvious these laws have nothing to do with children and are really designed to uniquely identify and track Internet usage for everyone, tying your identity to every action taken both online and offline.

I think we’re seeing security in major parts of the tech industry intentionally being eroded. Strong open-source encryption systems are being stifled in exchange for broken tools from big tech companies that have zero interest in user privacy. There seems to be a concerted effort to remove all real encryption, security and privacy from consumer software. This doesn’t seem like incompetency mistaken for malice. This seems intentional, and I predict we’ll continue to see more stories in the coming months and years. I expect more exploits and backdoors in major tools specifically geared around mass data collection, removing data privacy and removing anonymization. It’s been happening for years and now it’s accelerating.

The Risks of Key Recovery, Key Escrow, and Trusted Third-Party Encryption. 27 May 1997. Abelson, Anderson, Bellovin, et. al. ↩
Keys under doormats: mandating insecurity by requiring government access to all data and communications. 17 November 2015. Abelson, Anderson, Bellovin, et. al. Journal of Cybersecurity, Volume 1, Issue 1, September 2015, Pages 69–79. https://doi.org/10.1093/cybsec/tyv009. ↩
YellowKey. NightmareEclipse. Project NightCrawler. Retrieved on 18 July 2026. Original Archived from Github ↩
Two more public disclosures, it will never stop. 12 May 2026. deadeclipse666 ↩
Re: BitLocker. 23 June 2026. madbrain. Slashdot. (comment) ↩ ↩²
We’re doing silent patches now huh, also a quick note about YellowKey. 13 May 2026. deadeclipse666. ↩
GreatXML a bitlocker bypass that seems to only work if you ever had Defender Offline Scan. 10 June 2026. deadeclipse666 ↩
Microsoft under fire for threatening security researcher with criminal investigation. 29 May 2026. Franceschi-Bicchierai. Yahoo Tech. ↩
July 14th. 23 May 2026. deadeclipse666. ↩
“We will ruin your life” -Microsoft. 1 June 2026. ThePrimeTime. ↩
Last time I dealt with MSRC I found a command injection vulnerability present for a decade in context menus, not highly critical but still exploitable. (see my talk Shift Happens) MSRC did not reward a bounty nor did they attribute a CVE to this finding because this “doesn’t…. 28 May 2026. @podalirius_. X. ↩
My last submission to MSRC was for a Device Guard bypass. I learned my lesson from prior drawn-out submissions, so I included a 90 day window this time. MSRC responded saying that it met their bar and they would fix it, but asked me to withhold disclosure well past 90 days because they needed a few extra months to fix it. I agreed on the condition that they issue a CVE, to which they agreed. After the agreed-upon Patch Tuesday a few months later, I couldn’t find any mention in the CVE list, so I reached out to MSRC to inquire. It turns out - they changed their minds, deciding it did not meet their bar for servicing, yet they patched it anyway… 28 May 2026. @GabrielLandau. X. ↩
VeraCrypt / Forums / General Discussion: Project Update. 30 March 2026. idrassi. SourceForge. ↩
@garymarks: Thank you for the feedback. Judging by the fact that other projects like WireGuard are affected, I don’t think my move to Japan has anything to do with it. There seem to be an internal Microsoft policy that triggered this issue at large scale and there lack of communication and feedback made things worse.. 9 April 2026. idrassi. Sourceforge. ↩
Action Required: Account Verification for Windows Hardware Program Begins October 16, 2025. 1 October 2025. NatachaC. Tech Community Microsoft. ↩
Microsoft account verification fiasco halts critical OSR updates: tech giant scrambles to restore access. 10 April 2026. Naprys. Cybernews. ↩
Microsoft responds after WireGuard, Windscribe, and VeraCrypt accounts get locked. 9 April 2026. Cubbins. PiunikaWeb. ↩
Hey I love dumping on my company as much as the next guy, because Microsoft does some dumb stuff, but sometimes it’s just check emails and verify your accounts. Not every “WTF micro$oft” moment is a slam dunk. I’ve emailed VeraCrypt personally and we’ll get him unblocked. I’ve…. 8 April 2026. @shanselman. X. ↩ ↩²
TrueCrypt. 29 May 2014. SourceForge. Archive ↩
Enabling BitLocker. TrueCrypt. SourceForge. Retrieved 19 July 2026. ↩
Encryption canary or insecure app? TrueCrypt warning says use Microsoft’s BitLocker. 29 May 2014. Smith. CSO Online. ↩
TrueCrypt Author Claims That Forking Is Impossible. 19 June 2014. timothy. Slashdot. ↩
Here is the note I sent to the only (alleged) Truecrypt dev still answering emails, along with his/her response. pastebin.com/RS0f8gwn. 16 June 2014. @matthew_d_green. X. ↩
I am sorry, but I think what you’re asking for here is impossible. I don’t feel…. 16 June 2014. Pastebin. archive ↩
macOS 28 Will Drop Support For Encrypted Mac OS Extended Volumes - Slashdot. 10 July 2026. BeauHD. Slashdot. ↩
PSA: macOS 28 will drop support for encrypted Mac OS Extended volumes - 9to5Mac. 8 July 2026. Mendes. 9to5Mac. ↩
About support for encrypted Mac OS Extended disks in macOS 28 or later. 7 July 2026. Apple Support. ↩
Following User Outcry, AMD Reinstates Memory Encryption In Consumer CPUs. 22 June 2026. BeauHD. Slashdot. ↩
Following user outcry, AMD reinstates memory encryption in consumer CPUs. 22 June 2026. Goodin. Ars Technica. ↩
Discord’s Disturbing Ties to Global Surveillance - ID Verification, Palantir, & Thiel. 14 Feburary 2026. Gamers Nexus. ↩
HW News - DRAM Antitrust Lawsuit, ID Verification, USA Wants 5% of OpenAI, Sony Drops Discs. 10 July 2026. Gamers Nexus. ↩
Brazil Law: All OS’s Have 13 Days to Add Age Verification. 4 March 2026. The Lunduke Journal. ↩
New Federal Law to Require Age Verification on All Operating Systems. 15 April 2026. The Lunduke Journal. ↩

Experimenting with Lazy Spec Driven Development

Tue, 30 Jun 2026 00:00:00 +0000

I’ve talked a bit about large language models (LLMs), and I do have friends who’ve been at odds with me for my reluctance to call these predictive algorithms “artificial intelligence”. Still, I do use LLMs frequently for coding both professionally and on my own projects. For things I care about, I often use them as a scalpel, either using chatbot output to implement my own design or heavily refactoring generated code to be maintainable. I’ve limited pure prompt-based development, or vibe coding, for simple one-time use tasks or throwaway scripts. Recently, I’ve started using LLMs for workflow projects using only lists of text specifications. In a radical departure from my normal engineering flow, I’ve let agents just write software without individually verifying the code. I do often get usable tools, although they often need a considerable number of refining steps. However, looking under the surface reveals confirmations of many of my apprehensions about heavy dependence on these tools for critical or production systems.

The Spec

I hesitate to call what I’ve done true spec-driven development. I’ve seen projects that embrace this practice create detailed documents defining languages, unit test creation and frameworks. They’ll treat the specifications as the source of truth, often using tools like Spec-Kit, and only allowing code updates via specification documents.

For my experimentation, I use a skeleton project with the base layout, libraries and tools I’ve come to use on other Python projects. I’ll add in an AGENT.md or CLAUDE.md that specifies all my coding conventions and suggested best practices. Then I create a specification of what I want to build, listing each requirement in as much detail as I can. I include database layouts, framework choices and specific details catering to how I would implement the project. I then tell a coding agent to read the base convention file and implement the spec.

After testing it out and making sure the base functionality works, I create a git commit with the spec in it. Usually I try to list all the bugs, issues or further improvements in a new spec. Small changes I might just ask in the existing prompt before committing, so the entire project may not be entirely reproducible using just the committed specs. If it’s something as small as a style sheet or visual change, I might just find the code myself so I’m not wasting tokens or compute power on something as stupid as changing margins on a box or the size of an icon.

If I use a new context window, I may simply tell the LLM to load AGENT.md, tell it that specs 01 - 05 have already been implemented by previous sessions, and to implement spec 06. Although this approach may prevent the previous specs from being loaded into the current context window, I haven’t seen any noticeable regressions by running my agent this way.

Cosszy

Back when I was in college, I wrote Isszy, a simple image sorting program written in Java. Over the years, I’ve still used this program, with light modifications so it could also sort RAW images from my cameras. I haven’t updated the original release, as most of my modifications were kludged together for my specific photography workflows. I decided to see if I could get a full rewrite of this tool using an LLM and specs. The result is Cosszy, a web-based image and video sorter.

Cosszy Image Sorter

I was pretty surprised by just how well this application worked from the start. There were some bugs and UI issues, of course, but I mostly had the LLM fix them with further iterative specs. Still, the initial specs were very detailed with the exact technology, project layout, libraries and tooling I wanted the project to be based around. It used my current skeleton Python web project, which had a clean platform from which to build from.

Image Metadata Window

There were some interesting things the tool did that I didn’t specifically ask for, such as writing unit tests and a README file. Curious as to what would happen if a file with a duplicate name was moved to an existing folder, I discovered the LLM had implemented a process where the file gets an _1 appended to its name without any user intervention or indication. I then used a spec to implement a full overwrite and rename dialog.

Rename Duplicate Image Dialog

After several iterations, the program does work fairly well. I’ve only hand-tweaked a few issues that were easier to do myself or that the LLM would not fix correctly, mostly around visual positioning and style sheets. Parts of the underlying code are actually quite clean and fairly well organized, while other pieces have a lot of repeated elements and are a pile of spaghetti mess. It is something that anyone can run thanks to the included README, and works well as a part of my general workflow.

Comf-meta

ComfyUI is a tool used for image and video generation. It can send prompts to cloud providers, but I only use it to generate media on my own hardware using models I download and run locally. Comf-meta started off as a tool I created for extracting metadata from generated images to help me organize media and workflows. I used a skeleton project as a basis, but a lot of the commands were flushed out using prompt-based coding.

Comf-meta Image Workflow Screen

ComfyUI is honestly not that comfy. I found it more powerful than alternatives like AUTOMATIC1111, but working with it for any amount of time quickly leads to the realization most of the tool is likely AI developed. Looking up long-standing bugs and going through its issue tracker really makes this apparent.

Comf-meta Image to Video Workflow Screen

I wanted to make a console or text user interface (TUI) for my ComfyUI workflows. I took several workflows for images and video I’ve found and refined, exported them (both UI and API JSON), and then tasked the LLM with converting them so the requests could be submitted in a console application that used the Textual and Rich libraries for Python.

Comf-meta Job Queue

Although I’ve posted the source code for Comf-meta, it’s heavily dependent on a customized Docker container built specifically for running ComfyUI on an AMD video card using ROCm. The workflows depend on a lot of custom nodes and assume the installation of specific models. At some point in the future, I might update it with a fully reproducible and distributable installation process. For now, it’s just a demo of a custom workflow I created using specs.

The Code

This approach focuses on the end result and forces me to not care about the code at all. When I do occasionally look at what the weighted code prediction machine has produced, some of it is surprisingly clean and well engineered. Some of it is repeated blocks. A lot of it uses function names and patterns I would never use. Most of these LLMs love to use blanket exception catching to hide issues that really should be errors. They tend to assume a lot of defaults as well, and generate things that are sometimes suboptimal or clearly wrong.

Using this approach is incredibly powerful for small local workflows, so long as they’re carefully tested along the way. However, I would be very hesitant to use anything like this for production grade applications or anything in a mission-critical environment.

This approach also requires a lot of understanding of all the underlying technology needed to implement the project. I specifically requested tools like Vite and Vue in my specs, as well as defining what tables and data structures I wanted. People without a software engineering background may simply get applications written in a generally popular software stack, such as Node and React. It could lead to a stagnation of newer technology adoption as the models will simply default to the technologies popular at the time they were trained.

Add GPLv3 Poison to Everything

I’ve previously licensed a fully generated project I used for fixing Sonatype’s license bait and switch as CC0 (public domain). Without writing any of the actual code, I considered licensing it as anything else to be dubious. However, I am concerned with just how much open-source code has been used to train these models. So far the court systems seem to be siding with the producers of LLMs¹, while several lawsuits from publishers are still being litigated²³. This is a pretty stark contrast to how the music industry convinced a college student to settle for $12,000, his life savings in 2003, for simply writing a search engine that happened to index music files⁴⁵.

Large tech companies may claim to support open source, but they do not like the General Public License (GPL). Apple actively removed most GPL-licensed software from their operating system years ago. The GPL requires that when you release modified software to the public, the modified source code needs to be released as well. Big tech prefers licenses like MIT or Apache, which allow them to take existing open-source projects and integrate that code into their products without contributing those changes back to the community.

With Meta’s Llama 3.1-70B model being able to reproduce close approximations of nearly half of Harry Potter and the Sorcerer’s Stone⁶⁷, I’m not convinced recent court cases are going to hold long term in regards to LLM-generated output and copyright. A higher court decision could radically shift the rules around products with generated content. Still, for now, I think it’s important to release all generated code under version 3 of the GNU Affero General Public License (AGPLv3). If another developer gets the LLM to generate the exact same code, yet it is already published as AGPLv3, does the generated code become AGPLv3 as well? It’s an interesting legal question, and adding an AGPLv3 license to every generated project ensures that question will have to be addressed eventually.

Cognitive Decline

There is considerable evidence to suggest using LLMs reduces our cognitive abilities when it comes to research, especially after returning to human or search engine-based research after prolonged exposure to LLMs⁸. Heavy use of chatbots can also lead to delusional spiraling⁹, and there are a seemingly endless number of studies coming out showing how these tools can lead to AI psychosis and cognitive issues¹⁰.

I find these studies compelling, and I’ve found that it’s easy to get disconnected from software engineering when using a lot of these tools in the manner I’ve described so far. I do think that it negatively affects my engineering abilities. However, I can’t deny the fact that they’ve helped me make tooling I’ve put off for weeks or months, some that I may have never started at all. For local workflows, homelabs and small, specific tasks, they’re incredibly useful. Yet they also require a considerable amount of “handcrafted” design and programming experience in order to tell the machine specifically what needs to be generated and avoid getting totally unusable output.

Owning Your Tools

I also hate that the models needed to perform these tasks require cloud services. As someone who self-hosts nearly everything, including my own chat services, social media, text message gateway and even e-mail, I really do not like using a paid or cloud service for critical software work. Although local models are getting better and some can work on consumer hardware, they still don’t quite compare to the large and computationally expensive models designed by Anthropic and others.

Anthropic the Drug Dealer

Anthropic has also been notorious for baiting and switching their customers. In May, they tried to disguise moving some Claude usage over to API billing as ‘free credits,’ hiding a potential 25x price hike¹¹¹². They’ve also been found offering different features and pricing selectively on their product page in order to determine how much potential customers were willing to pay for their services¹³¹⁴; a type of targeted pricing that’s also been happening with Instacart in the retail space as well¹⁵¹⁶¹⁷.

Recently, Anthropic announced policy changes starting on July 8th, stating they could ask for government ID and facial scans in order to maintain access to their platform for all free, pro and max subscribers (business and enterprise customers are exempt)¹⁸. As someone who doesn’t even allow restaurants and music venues to scan my ID, there’s no way in hell I’m ever uploading my ID or facial biometrics to Antrhopic. We might see a rise in purchasing tokens from grey or black market transit stations (which might be fraudulent and stealing all your chat session data¹⁹).

It’s not clear if this policy will be selective or apply to everyone. I do think there are a lot of software engineers who, despite privacy concerns, will gladly give over their government ID and biometrics in order to either maintain access to Opus or gain access to Anthropic’s mythical fable of a model. There are people who have become so dependent on this technology that writing software by hand again would seriously lead to withdrawal and require weeks to rebuild lost cognitive skills.

Final Thoughts

I have a good friend, a fellow software engineer, who I’ve known and have kept up with since graduating university over 20 years ago. He’s very “bullish” on AI and has told me I should engage with LLMs more while coding, while also challenging many of the concerns I’ve brought up over data centers, sustainability, the growing cost of personal consumer hardware and the potential decline of our own cognitive abilities.

This article’s original intent was to document my experiences experimenting with specification and agent-based coding, without paying attention to the generated output. However, the acceleration of events in the LLM industry in just the past few weeks has also pushed me to elaborate more on my concerns as to where we are headed with these new tools.

The projects mentioned here, as well as others I’ve driven using specification documents, both at home and at work, are limited to non-critical workflows with very specific requirements. They are amazingly powerful and have saved me a considerable amount of time, but they also don’t feel like they’re mine at all. I don’t feel like I have personal ownership of these programs. Although I post their git repositories publicly under my own account, I don’t think I would ever package and distribute them. They still feel like an open-source program I found and ran, or a piece of software I commissioned a contractor to write, rather than something I created myself.

I’m sure at some point I will lose access to these frontier models, most likely due to an unwillingness to submit government ID. I’m okay with having to return to using a shovel instead of a backhoe to dig out a trench. Unlike an LLM, using a backhoe doesn’t atrophy one’s cognitive ability. I may also look at open coding models again because owning your own tools is far better than having to rent those tools from people with acceptable use policies that could ban your ability to rent those tools forever.

I am greatly concerned about the future, not just of software engineering, but many of the other professional fields that LLMs are being shoved into where they don’t really fit. The danger of AI is not what’s presented in the movies: a killer super-intelligence that decides to either enslave or end humanity. The real dangers are the people who become dependent on these machines, thinking they’re some kind of authoritative source, oracle or god, when in reality they’re just very good next-word prediction algorithms. Sometimes they are magically impressive, which makes us forget or dismiss the other times when they string together the wrong set of tokens, telling us to eat rocks, use glue as a pizza ingredient, or dismiss our own cognitive abilities in favor of the weighted word generator.

Meta Wins Blockbuster AI Copyright Case—but There’s a Catch. 25 June 2025. Knibbs. Wired. ↩
HW News - 12VHPWR vs 5090, Cyberpower Responds to GN, AMD RX 9070, Meta’s Piracy, & 5060 Ti. 16 Feburary 2025. Gamers Nexus. ↩
Kadrey, et. al. vs Meta Platforms Case 3:23-cv-03417-VC - Document 598. 25 June 2025. ↩
RIAA Grabs Student’s Life’s Savings. 10 June 2003. CowboyNeal. Slashdot. ↩
RIAA Wrath Hits Teen. 9 June 2003. Martell, Tech Live. ABC News. (Archived) ↩
Study: Meta AI model can reproduce almost half of Harry Potter book. 20 June 2025. Lee. Ars Technica. ↩
Extracting memorized pieces of (copyrighted) books from open-weight language models. 1 May 2026. Cooper, Lemley, Casasola, et. al. Computation and Language (cs.CL). ↩
Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task. 10 June, 2025. Kosmyna. Hauptmann. Yuan. et. al. MIT Media Lab. arXiv:2506.08872. ↩
Sycophantic Chatbots Cause Delusional Spiraling, Even in Ideal Bayesians. 22 Feburary 2026. Chandra. Kleiman-Weiner. Ragan-Kelley. et. al. Computer Science: Artificial Intelligence. ↩
We’re Already Living in an Alien Invasion Movie. 17 June 2026. Truthstream Media. (15m:20s) ↩
I need to rant, Anthropic is worse than my high school’s drug dealer. 16 June 2026. Louis Rossmann. ↩
Claude’s ‘Free Credits’ Hide 25x Price Hike. 13 May 2026. Clawd.rip. ↩
Louis loses mind on Anthropic speedrun towards enshittification. 29 April 2026. Louis Rossmann. ↩
For clarity, we’re running a small test on ~2% of new prosumer signups. Existing Pro and Max subscribers aren’t affected.. 21 April 2026. @TheAmolAvasare. X. ↩
Same Cart, Different Price: Instacart’s Price Experiments Cost Families at Checkout. 9 December 2025. Groundwork Collaborative. ↩
Same Product, Same Store, but on Instacart, Prices Might Differ. 9 December 2025. Casselman. The New York Times. ↩
Who Wants to Live in This Digital Prison. 26 June 2026. Truthstream Media. ↩
Claude Identity Verification Starts July 8: What Facial Data Anthropic Collects. 21 June 2026. Owens. Tech Times. ↩
The Secret Token Underworld. 29 June 2026. The PrimeTime. ↩

The Fable of Mythos

Sun, 21 Jun 2026 00:00:00 +0000

Fables de La Fontaine - Epinal (public domain)

Earlier this year, Antrhopic announced a large language model (LLM) called Mythos, a model they claimed to be so powerful for finding software exploits that it was too dangerous to release to the general public¹²³. Instead, they provided a limited release of this model to select organizations as part of Project Glasswing⁴. Then, last week, Antrhopic released a version of this model called Fabel⁵. Current subscribers were allowed an entire three days to work with this model before the US federal government issued an unpublished export control order for Anthropic, citing concerns of model jail breaking and weaponization, to restrict access to the model from any foreign nationals living in America or abroad⁶. Anthropic denied the government’s finding, but complied with the order by revoking access to the model for all customers⁷. There has been a massive roller coaster of press over this “frontier model,” much of it seeming to be media-hype and exaggerated claims. However, I question if this latest intervention by the US government is really about this coding model. The names of the products, the timing and the general story surrounding technology leads me to believe this is a narrative propaganda tool being used to increase government regulation, cripple competition and push for increased government surveillance.

The Myth and the Magic

Anthropic’s models have names like Haiku, Opus and Sonnet, all which seem related to songs or poetry. But this most recent model started as Mythos, and the released version was Fable. Both of these names invoke ideas of fantasy or mystical fictions. As a developer, my peers and I got to play with Fabel for all of three days, before it disappeared over a weekend. Think about that: software engineers got to use it for work, and by the weekend it was no longer available, even on our personal accounts.

I used Fable pretty heavily on some complex features while it was available. I was impressed by some of its planning outputs and did find it somewhat cleaner than previous models. My friends and colleagues have given me various types of feedback, some amazed by it while others seeing it as a disappointment or incremental improvement.

The Call for Regulation

Earlier in June, Anthropic had called for a “global pause” on further AI development due to risks of “self-improvement⁸.” So, they called for a pause and then turned around and made a new product release? What I find most concerning is a post from Anthropic’s founder calling for government regulation⁹.

It is time to go beyond transparency to more serious and binding regulation of AI. I believe the best analogy, at least at the current stage of the exponential, is to cars, airplanes, or drugs—powerful technologies essential to the modern economy, but capable of killing large numbers of people if designed or operated poorly. I therefore believe we should model AI regulation on agencies like the Federal Aviation Administration (FAA). Frontier AI models, like airplanes, should be required to go through technical testing and auditing, and their release should be blocked or reversed as a threat to public safety if they do not meet high standards of safety. I am grateful to see the Trump administration’s Executive Order move incrementally towards a greater role for government in AI, though Anthropic’s proposal recommends even further action. -Dario Modei⁹

Whenever a company within an industry calls for more regulation, it’s rarely in the public’s interest. It’s usually a means of preventing competitors from entering their market. What would regulation even look like in this case? Would developers be required to submit any software that multiplies matrices together to the United States government? Will linear algebra software fall under export regulations? Will there be more cases similar to Roman Storm and Tornado Cash, where simply writing open source software could get someone arrested for evading sanctions?

People may say, why didn’t you just disable access to foreign nationals? Why did you disable access to everybody? And this comes back to an issue we talk about a lot … How are you going to keep foreign nationals from using it? When I sign up, all I need to provide is an email and my credit card information … That’s all I need to sign up for one of these services … but how do they know if I’m a foreign national, US citizen or anything else? Maybe my ID will show that. Maybe my ID is not enough. Maybe I got to show my passport. Maybe I show my ID, my passport, and my birth certificate to show proof. You see where we’re going with this? This is moving the internet in the direction of you need to show your passport, your birth certificate, and your ID to sign up for a service … And to do that would require that you put all this infrastructure in place … I’ve spoken about a lot of the age verification laws that have been popping up recently and why I’m annoyed by them … This is bad. I don’t like this. And this kind of speed-runs you further in that direction. -Louis Rossmann¹⁰.

The entire story with the release and redaction of Fable, followed by calls for regulation, reeks horribly of surveillance-state propaganda. It creates a narrative of fear, which could provoke citizens to call for protection in a way that stifles innovation and pushes us further into surveillance-state policies.

Scary Security Stories

Anthropic claimed that their Mythos model detected 23,000 potential vulnerabilities across a thousand open source projects¹¹¹². This included a “27-year-old bug in OpenBSD¹³,” which was patched¹⁴, although I can’t find any official statements on it from OpenBSD developers regarding the severity. FFmpeg did thank Anthropic¹⁵ for patching a vulnerability in their H264 decoder¹⁶. The developer of cURL reported that a scan with Mythos resulted in it finding a single vulnerability¹⁷, while also stating that Mythos was “an amazingly successful marketing stunt for sure¹⁸.”

The Fable of Defying Trump

Anthropic has allegedly been at odds with the Trump administration. They initially tried to take the ethical high ground, refusing to allow their models to be used for fully autonomous weapons. This caused the Defense Department to label Anthropic a supply-chain risk¹⁹, which then resulted in Anthropic suing the federal government²⁰. This moral grandstanding was short lived as shortly afterward, people discovered that Anthropic was experimenting with displaying different features to customers randomly on their Claude Pro product, in order to experiment with seeing how much people would pay for additional feature sets²¹²².

So, What Just Happened?

One of the supposed jailbreaks came from “Pliny the Liberator,” a person who continually tries to prompt inject every newly released model²³. Another person was able to bypass the “review this code for security issues” guardrails by using the prompt “fix this code²⁴²⁵.” TechCrunch argued that the issues should have never triggered export control violations and that the government ban was never about jailbreaks²⁶.

Going by Dario’s statements on practically begging for regulation⁹, it’s likely this entire fiasco is about regulatory capture. As of writing, AI is profitable for absolutely nobody except Nvidia²⁷. Theories have also emerged that the LLMs are so unprofitable that the dangers of Mythos were intentionally exaggerated as an attempt to force the US government to use eminent domain to purchase Anthropic at fair market value²⁵.

The Future of A.I.

I’d like to think many people are weary of the push for Large Language Models into every aspect of our lives²⁸. Still, I cannot deny that, for software engineers and other professionals, they are incredibly powerful tools for rapid prototyping or creating workflows for specific, well defined tasks. There is a capture taking place, among the engineering-class of society, that wants to alleviate the painful details we often have to exercise for meticulous tasks. This might be all well-and-good for people who write billing or HR software. However, I do wonder if at some point we will see a bridge collapse, and the blame will be placed on an inference model that incorrectly text-completed (or as some say “hallucinated”) the tensile strength needed to hold the expected load; something that might have been caught by a human engineer vested in the safety of their product and the real humans walking across it.

In the recent Anthropic fiasco, I do think we are watching the complexities of an early onset market collapse around the LLM industry. The names of Mythos and Fable invoke a fantastical storytelling narrative, which I think were intentional. Anthropic has political motivations combined with exponential spending. Their fearmongering and calls for regulation could be a strategy for regulatory capture, increasing the costs for competitors to enter their product space. We are speeding forward toward a tipping point, because the subscription rates people pay likely don’t remotely begin to pay for data center costs²⁹³⁰.

We are being sold a myth right now³¹. It’s right there in the name. The entire situation with the technology sector today feels way too close to the 2008 financial collapse, combined with the 2001 dot-com bubble, and then multiplied by an order-of-magnitude coefficient large enough to collapse the global economy.

We don’t know if Anthropic is acting independently, or if they are actively colluding with a government (or multi-government) narrative. The names, timing and media stories makes me skeptical of everything surrounding the entire Mythos/Fable shipwreck. We are literally living through the mythology of a software engineering fable, and I’m very weary of where this all leads.

OpenAI To Limit New Model Release On Cybersecurity Fears. 10 April 2026. Slahsdot. ↩
Scoop: OpenAI plans new product for cybersecurity use. 9 April 2026. Sabin. Axios. ↩
Claude Mythos is too dangerous for public consumption…. 4 June 2026. Fireship. ↩
Project Glasswing: Securing critical software for the AI era. 7 April 2026. Anthropic. ↩
Claude Fable 5 and Claude Mythos 5. 12 June 2026. Anthropic. ↩
Legal Considerations Related to the Anthropic “Export Controls Directive”. 15 June 2026. Egan. Just Security. ↩
Statement on the US government directive to suspend access to Fable 5 and Mythos 5. 12 June 2026. Anthropic. ↩
Anthropic Urges Global Pause in AI Development, Flags ‘Self-Improvement’ Risk. 5 June 2026. BeauHD. Slashdot. archive ↩
Policy on the AI Exponential. June 2026 Amodei. archive ↩ ↩² ↩³
Anthropic CEO learns meaning of “life comes at you fast”. 13 June 2026. Louis Rossmann. ↩
Mythos Detected 23,000 Vulnerabilities Across 1,000 OSS Projects. 27 May 2026. BeauHD. Slashdot. ↩
Anthropic: Mythos Detected 23,000 Potential Vulnerabilities Across 1,000 OSS Projects. 25 May 2026. Kovacs. SecurityWeek. ↩
The significance of Claude Mythos Preview for cybersecurity. 7 April 2026. Frontier Red Team. Anthropic. ↩
OpenBSD 7.8 errata 025, TCP packets with invalid SACK options could crash the kernel. 25 March 2026. OpenBSD.org. ↩
Thank you to @AnthropicAI for sending FFmpeg patches. 7 April 2026. FFmpeg. X. ↩
avcodec/h264_slice: reject slice_num >= 0xFFFF. 14 March 2026. carlini. FFmpeg. Github. ↩
Mythos finds a curl vulnerability. 11 May 2026. Stenberg. haxx.se ↩
Anthropic’s Bug-Hunting Mythos Was Greatest Marketing Stunt Ever, Says cURL Creator. 12 May 2026. BeauHD. Slashdot. ↩
Pentagon formally designates Anthropic a supply-chain risk. 5 March 2026. Bordelon. Politico ↩
Anthropic sues the Pentagon after being labeled a threat to national security. 9 March 2026. Nolan. Fortune. ↩
Louis loses mind on Anthropic speedrun towards enshittification. 29 April 2026. Louis Rossmann. ↩
For clarity, we’re running a small test on ~2% of new prosumer signups. Existing Pro and Max subscribers aren’t affected.. 21 April 2026. @TheAmolAvasare. X. ↩
🚨 JAILBREAK ALERT 🚨 ANTHROPIC: PWNED 🫡 FABLE-5: LIBERATED 🦋 …. 10 June 2026. @elder_plinius. X. ↩
The Fable 5 Export Controls Harm US Cyber Defense. 14 June 2026. Moussouris. LutaSecurity. ↩
Fable Deserved It. 19 June 2026. The PrimeTime. ↩ ↩²
The US government’s Anthropic models ban was never about an AI jailbreak. 15 June 2026. Whittaker. TechCrunch. ↩
Is AI Profitable Yet?. Retrieved 20 June 2024. Archive ↩
We’re Already Living in an Alien Invasion Movie. 19 June 2026. Truthstream Media. ↩
Recently, we purchased one of each Anthropic/OpenAI subscription plan and randomly ran long horizon coding tasks until we exhausted the weekly limit. It’s widely believed that a $200/month plan maxes out at ~$2000/month worth of tokens (assuming API pricing). However, we found that the subscriptions are actually far more generous. (2/4). 10 June 2026. @SemiAnalysis_. X. ↩
Subscription plans are massively subsidized. And by massively, I mean absurdly. 11 June 2026. @kimmonismus. X. ↩
I Think They Are Lying To You. 12 June 2026. The PrimeTime. ↩

When was the last time RAM was this expensive?

Fri, 27 Mar 2026 00:00:00 +0000

Memory has become insanely expensive over just the past few months. Micron cut production of Crucial, their consumer brand of memory, to focus on data center products at the end of 2025¹. This was after Micron received massive amounts of taxpayer money from the CHIPS Act². Crucial’s exit has radically changed the market. We’ve seen big shifts in RAM prices over the past few decades. It has always been volatile, but it’s never been this bad. In this unprecedented time of memory pricing, I started going back through all my previous purchases to see how far into the past I’d have to go to see the current-day cost per byte of memory.

I built a desktop in 2016 after living out of two bags for nearly a year. It was the first desktop I had put together in years, and I spent $190 for 32GB of DDR4-3200 memory. A year later, I ended up selling that desktop before driving across the country. When I settled down in Chicago and decided to build another desktop, I paid $330 for 32GB of the same DDR4-3200 memory, a nearly 60% increase.

RAM prices spiked massively in 2018³⁴. Supply may have been constrained due to the Samsung Galaxy Note 7 recall⁵. There are only three major manufacturers of DRAM chips: Samsung, Micron and SK Hynix. Small supply and demand changes can have volatile effects, as we’re seeing today.

Recent Jumps

In September of 2025, I purchased this 32GB (2x16GB) RAM kit for $86. As of March 1st of this year, it was $370.

A 64GB DDR5-6000 set of RAM cost me $210 in July of 2024. In August 2023, it was around $300⁶. By the end of March 2026, it is currently $950.

Date	Capacity	Type	Cost	Price Per GB
2025, September	32GB (2x16GB)	DDR5-6000	$86	$2.69
2026, March	-	-	$370	$ 11.56
2025, February	64GB (2x32GB)	DDR5-6000	$90	$1.41
2026, March	-	-	$950	$14.84

The current data center and AI-mania has pushed most memory manufacturers to switch to production of enterprise high-bandwidth memory, leaving the consumer market in panicked short supply.

Price per Gigabyte

These are obviously massive increases in prices, but just how bad? Typically, consumer storage tends to gain capacity while reducing cost over time. How far back do we have to go before we see a regression to today’s prices? I decided to go back through decades of ancient e-mails, searching for old DDR orders from both large and small online retailers.

Date	Capacity	Type	Cost	Price Per GB
2005, May	512MB	DDR PC3200	$38	$76
2006, January	2GB (2x1GB)	DDR-266 PC2100 (Registered ECC)	$160	$80
2006, February	2GB (2x1GB)	DDR2-667 PC2-5400	$156	$78
2009, April	1GB	DDR2-667 PC5400 (SODIMM)	$20	$20
2009, May	2GB	DDR2-667 PC5400 (SODIMM)	$29	$14.50
2009, July	4GB (2x2GB)	DDR2-1066 PC8500	$140	$35
2016, November	8GB (2x4GB)	DDR4 (SODIMM)	$44	$5.50
2016, December	32GB (4x8GB)	DDR4-3200	$190	$5.94
2018, December	32GB (2x16GB)	DDR4-3200	$330	$10.31
2020, September	16GB (2x8GB)	DDR4-3200	$57	$3.56
2020, December	64GB (2x32GB)	DDR4-3600	$230	$3.59
2021, November	16GB (2x18GB)	DDR4-3000	$50	$3.13
2022, July	64GB (2x32GB)	DDR4-3200 (SODIMM)	$160	$2.50
2024, July	64GB (2x32GB)	DDR5-6000	$210	$3.28
2024, November	64GB (2x32GB)	DDR5-5600 (SODIMM)	$140	$2.19
2025, February	64GB (2x32GB)	DDR4-3200	$90	$1.41
2025, September	32GB (2x16GB)	DDR5-6000	$86	$2.69

Of course, this isn’t a fair comparison, as I’m only looking at the price per GB. I’m also only looking at my purchases, not overall market data. Memory has changed over the years, both in DDR standards and in bus speed. Many of these purchases came in matched kits as well, to take advantage of dual-channel configurations. A 2x16GB pair might run at a lower cost per GB than a single 32GB due to chip density, but for the sake of simple comparison, we’re going to ignore that as well.

The Height of Memory Mania

Once again, let’s compare the capacity cost of the current market, ignoring everything else.

Date	Capacity	Type	Cost	Price Per GB
2026, March	64GB (2x32GB)	DDR5-6000	$950	$14.84
2026, March	32GB (2x16GB)	DDR5-6000	$370	$ 11.56

I would have to go back to 2009 to get to our current cost-per-gigabyte market price. The largest capacities I could find for consumer-grade memory during this time period were 6GB sticks of DDR3, which ran for around $180, or $30 per GB⁷. Getting anywhere near current memory capacity back then likely required server hardware and enterprise motherboards.

Pressure Release

We have regressed a decade and a half when it comes to RAM prices. 4K editing, software development, local AI models, gaming and the ever-increasing bloated requirements from web browsers and Electron applications have all pushed up memory requirements a lot for general computer use. Yet, due to massive data center expansion, RAM scarcity is forcing prices outside the range for most normal consumers.

There eventually must be a pressure release. It could be the impending and easily predictable collapse of the AI bubble, or it could be the United States government finally removing sanctions from ChangXin Memory Technologies (CXMT), allowing for another supplier to compete in the market.

The danger of continued high RAM prices is that it could push more services into data centers and away from local machines and homelabs. Even businesses may push developers to utilize more cloud architecture rather than provide their staff with sufficiently powerful laptops for their technology stacks. Decades of technological innovation that have pushed more compute and storage power to individual engineers could regress due to the price of chip-based storage.

When will we see reasonable prices for consumer RAM, as well as for NVMe storage that’s also risen dramatically? The DRAM industry is a difficult market to enter into, and has long been held by a small cartel of chip manufacturers⁸, some of whom have been charged and pleaded guilty to illegal price fixing⁹. Gartner is predicting we won’t see slowdowns in the AI markets until 2027¹⁰, which could lead to entry level gaming PCs becoming affordable by 2028¹¹.

The availability of affordable consumer compute has radically changed what people have been able to create and do for the past several years. There’s been a large democratization of technology¹², and constricting supply only benefits those who offer cloud services and software subscriptions. It’s a terrible hit for enthusiasts that may go unnoticed by those who are fine with renting and not owning hardware.

WTF Just Happened? - The Corrupt Memory Industry & Micron. 5 December 2025. Gamers Nexus. ↩
Trump’s big bill boosts Micron’s tax break for Clay chip plants by about $5 billion. 26 July 2025. Weiner. Syracuse. ↩
Why RAM Prices Are Through the Roof. 30 January 2018. Hruska. ExtremeTech. ↩
RAM Price Report: DDR4 Same Price as Initial Launch. 22 January 2018. Lathan. Burke. GamersNexus. ↩
Samsung Galaxy Note 7 recall: Here’s what happens now. 22 January 2017. Dolcourt. CNet. ↩
This 32GB kit I bought in September, for $86 … is now $370. 1 March 2026. @djsumdog@djsumdog.com. mirror ↩
Memory - DDR3 RAM. TigerDirect. Archive from 8 December 2009 ↩
The DRAM Cartel - Price Fixing, Anti-Consumer Collusion, & Corporate Conspiracy. 6 March 2026. GamersNexus. ↩
Japanese Company Agrees To Plead Guilty To Participating In Dram Price- Fixing Conspiracy. 30 January 2006. US Department of Justice. ↩
Gartner Says Surging Memory Costs Will Reduce Global PC and Smartphone Shipments in 2026l. 26 Feburary 2026. Gartner. ↩
Gaming PCs will be too expensive for ordinary people by 2028 predicts report. 5 March 2026. Beckwith. Metro. ↩
An Army of Davids. Reynolds. (2006) ↩

80 TB of RAIDZ Storage

Fri, 20 Mar 2026 00:00:00 +0000

4x 30TB Seagate Hard Drives

Over the course of my life, hard drives have continued to get denser as far as their data capacity. Back when I was traveling for years, I could comfortably put most of my data on a single hard drive. Over the years, when I reached capacity, I’d copy everything to a new drive and ship the old one as an offsite backup to a family member halfway across the world. In recent years, I’ve switched from smaller custom NAS servers to a custom rack-mounted solution with ZFS storage. I was able to upgrade to a ZRAID array before storage prices started skyrocketing. This post talks about the new 80TB ZFS RAID array I built, as well as how I’ve converted my old hard drives to backups using ZFS snapshots.

The Homelab

In 2022, I purchased a Sysracks rackmount in order to build out a small homelab. It’s an entry-level “gamer rack¹,” which would struggle to properly fit a lot of enterprise rails and hardware. It’s good that I’ve mostly used commodity hardware and reused consumer parts. The NAS I built used a Plink case², containing the ITX motherboard and parts from my older NAS. I then added two sets of 3 X 5.25 hot-swap tray-less hard drive enclosures.

The Old Rotation

Even with this new server chassis and its large number of drive bays, I didn’t move to a full RAID array. Instead I kept volumes for three general classifications of data: main, storage and archive. I’d rotate backup drives into the server every few weeks and transfer incremental ZFS snapshots to them. I kept a complicated spreadsheet of which drive contained which volume.

Data pool tracking spreadsheet

The backup volumes would go into a fire-resistant box. This is one of the off-the-shelf boxes that can be found at most hardware stores. I don’t know how fire resistant they truly are and hope I never find out.

Hard drives in fire resistant box

This is a terrible way to handle backups. In the case of a data failure, I could be weeks or months behind, depending on how well I’ve kept up my schedule. However, now that I have this new array, I also have several old drives that I could turn into backups. I could keep two sets of backup drives. One could be kept online and automated, while the other could be traded out occasionally into the same fire resistant box.

4x 30TB Seagate Drives

On December 3rd, 2025, I purchased four 30TB hard drives directly from Seagate for $2,451.54, including sales tax. By January, these drives were only available from 3rd parties at $900 each! Prices and supply for large capacity hard drives have fluctuated heavily. Had I waited, I could have purchased Seagate’s newer 32TB offerings. I’ve seen them for around $700 recently on several websites, but they do tend to go out of stock very quickly. Pricing and supply for these hard drives isn’t as bad as RAM, but it’s still volatile.

In February, Western Digital stated they had already allocated their entire hard drive manufacturing capacity for the year³⁴, meaning we could see further supply constraints in the hard drive market.

4x 30TB Seagate Hard Drives

The New RAIDZ Pool

I created this as an encrypted RAIDZ pool using the following settings.

zpool create -f -o ashift=12 \
                   -O compression=lz4 \
                   -O acltype=posixacl \
                   -O xattr=sa \
                   -O relatime=on \
                   -O encryption=aes-256-gcm \
                   -O keyformat=passphrase \
                   -O canmount=off \
                   -O mountpoint=none \
                   -f arcadia raidz1 /dev/sde1 /dev/sdf1 /dev/sdg1 /dev/sdh1

This created a RAIDZ volume with one drive of redundancy and a total usable capacity of 79.1 TB. In reality, it’s quite a bit shy of 80 TB due to RAIDZ overhead⁵, but the round number sounds cooler, so I’ve kept it as the title of this post. 😇

The Backup Server

The last of my spare compute units consisted of an AMD Ryzen 3 3200G on an Asus Prime A320I-K ITX motherboard with 16 GB of RAM. This particular motherboard does support PCI-E cards waking it from the sleep state, allowing me to equip it with a Wake-on-LAN (WoL) network card.

Since it will be in a low power state most of the time, it can live silently in my home office rackmount. I can remotely wake it up to perform backups and immediately suspend it again.

Backup Server in Server Rack

I wanted to connect this server to my 10 Gb fiber network. Unfortunately, it doesn’t appear that any cards with SPF+ cages support WoL, most likely due to lacking the power needed to keep the fiber transceiver active while the machine is suspended. I found this network card (NIC), which used a Marvell AQC113 chip, that offered 10Gb over copper with WoL for $114.

NICGIGA 10Gb Ethernet Adapter

Unfortunately, this NIC operates using PCI 4.0 by x1. The slot on this motherboard is PCI 3.0. So even when placed in the motherboard’s only PCI-E slot, lspci shows the card’s link status as Speed 8GT/s (downgraded), Width x1 (downgraded). In this configuration, I seem to be limited to transfer speeds of ~6 Gb/s over the network.

iperf3 -c chloe
Connecting to host chloe, port 5201
[  5] local 2605:8600:x:x:x:x:x:x port 45560 connected to 2605:8600:x:x:x:x:x:x port 5201
[ ID] Interval           Transfer     Bitrate         Retr  Cwnd
[  5]   0.00-1.00   sec   731 MBytes  6.12 Gbits/sec   67   1.12 MBytes
[  5]   1.00-2.00   sec   729 MBytes  6.12 Gbits/sec    1   1.23 MBytes
[  5]   2.00-3.00   sec   728 MBytes  6.11 Gbits/sec    0   1.25 MBytes
[  5]   3.00-4.00   sec   729 MBytes  6.12 Gbits/sec    8   1.27 MBytes
[  5]   4.00-5.00   sec   729 MBytes  6.11 Gbits/sec    3   1.29 MBytes
[  5]   5.00-6.00   sec   728 MBytes  6.10 Gbits/sec    0   1.30 MBytes
[  5]   6.00-7.00   sec   727 MBytes  6.10 Gbits/sec    0   1.32 MBytes
[  5]   7.00-8.00   sec   729 MBytes  6.12 Gbits/sec    0   1.34 MBytes
[  5]   8.00-9.00   sec   729 MBytes  6.11 Gbits/sec   13   1.36 MBytes
[  5]   9.00-10.00  sec   729 MBytes  6.11 Gbits/sec    0   1.37 MBytes

The ITX motherboard doesn’t have enough SATA connectors for the 5x bay hard drive cage I purchased for this project. So I purchased an M.2 SATA controller, which had two SAS breakout cables that can connect up to eight devices. I had purchased a similar M.2 SATA controller in the past but never got it working in other boards. I assumed it was a compatibility issue, but after trying it on other motherboards, I think it was just defective.

M.2 SATA Controllers

The new SATA controller I purchased had a warning label instructing the user to install the SAS breakouts before installing the card. I assume this is to prevent stress to the card while in the M.2 slot, which could damage the controller’s PCB.

The Backup Pool

I created a zpool on the backup server without RAIDZ. This is effectively Just a Bunch of Disks (JBOD) array in ZFS. It doesn’t provide any redundancy, but I do gain the full storage capacity of each drive, which seems like a reasonable tradeoff for a backup.

zpool create -f -o ashift=12 \
                   -O compression=lz4 \
                   -O acltype=posixacl \
                   -O xattr=sa \
                   -O relatime=on \
                   -O encryption=aes-256-gcm \
                   -O keyformat=passphrase \
                   -O canmount=off \
                   -O mountpoint=none \
                   -f carmella /dev/sdf /dev/sdg /dev/sdh

I sent my first ZFS snapshot to the new pool. It was a full snapshot sent via SSH. With both the limitation of the NIC and the overhead of SSH encryption, it took a total of five days to complete this transfer! Luckily, most of this was over the weekend, so I only had to deal with the loud server in my home office for part of that transfer.

# zfs send arcadia/aerith@2026-03-04 | pv | ssh stacey zfs recv byleth/aerith
50.8TiB 5:02:08:10 [ 121MiB/s] [           <=>                                               ]

I decided to create a similar pool with my more recent set of backup hard drives, turning them into a JBOD pool. This time, I placed the backup hard drives in the existing NAS to see if it made a difference in transfer speed. The entire snapshot transferred in one day and twenty-two hours, a fraction of the network transfer time.

root@chloe ~ [SIGPIPE|0|1]# zfs send arcadia/aerith@2026-03-10 | pv | zfs recv carmella/aerith
50.8TiB 1:22:51:28 [ 316MiB/s] [                                            <=>              ]

These initial snapshots are a one-time cost, as subsequent backups only require incremental data between two previous snapshots.

Pool Name	Set	Type	Capacity	Current Available Space
arcadia	Master	RAIDZ	79.1 TB	28.4 TB
byleth	Backup	JBOD	52.7 TB	2.32 TB
carmella	Backup	JBOD	54.5 TB	3.78 TB

It may seem like my backup pools are far under the potential capacity of my primary storage. Fortunately, they’re JBODs and I have several smaller drives from when I lived minimally, ranging from 3TB to 8TB. My backup server supports five hot-swap drives, and both of these backup pools only use three. As my usage grows, I can add additional drives until my backup sets reach capacity.

8TB, 6TB and 4TB Capacity Hard Drives

Final Costs

Part	Price	Tax	Shipping	Total
10G Base-T PCI-e4.0 X1 Network Card	$113.99	$11.09	$5.99	$131.06
3x5.25” to 5x3.5” SATA Trayless Hot-Swap Cage	$130.99	$13.30	$12.69	$156.98
2U Rackmount Case	$105.00	$9.71	-	$114.71
M.2 to SATA 8 Port Controller	$55.99	$5.18	-	$61.17
Total for Backup Server				$463.92
4x 30TB Seagate Hard Drives	$2,639.96 (-$395.99 discount)	$207.57	-	$2,451.54
Total Including New Storage				$2915.46

The 10 Gb Ethernet card was double the cost of an equivalent SPF+ card with a fiber transceiver and ended up being somewhat of a waste compared to an equivalent 5 Gb card on this particular build. The hot-swap cage was also a little over twice as much as I’ve paid previously for similar drive bays. Other than those two items, most of the prices for this build are comparable to my previous server builds.

I’m thankful I purchased my storage drives last year. I’m also thankful this build utilized existing parts and didn’t require RAM or NVMe storage, which have skyrocketed in just the past three months.

Future Work

The next steps for this project involve automating ZFS snapshots and backup transfers. I’d like to build a job that will make regular snapshots, wake the backup server, perform snapshot transfers and server updates, and maintain the required incremental snapshots needed to update both backup pools. For now, I have more than enough primary storage capacity, along with two sets of full backups and a plan for long-term continual rotations. With the current state of the consumer PC hardware market, this might be my last build or upgrade project for a considerable amount of time.

nice gamer rack. 28 August 2022. @duponin@udongein.xyz. mirror ↩
3U Intelligent Green Saving LCD Display & Fan Control. Retrieved 18 March 2026. Archive ↩
Western Digital Has No More HDD Capacity Left, as CEO Reveals Massive AI Deals; Brace Yourself For Price Surges Ahead!. 14 Feburary 2026. Zuhair. Wccftech. ↩
Western Digital is Sold Out of Hard Drives for 2026. 17 Feburary 2026. msmash. Slashdot. ↩
ZFS / RAIDZ Capacity Calculator (beta). WintelGuy.com. Retrieved 20 March 2026 ↩

Search the Past

Sun, 15 Mar 2026 00:00:00 +0000

In today’s world of content generated by the Weighted Random Word Generators (also known as Large Language Models (LLMs); too often incorrectly referred to as Artificial Intelligence), the vast majority of top search results are a mix of mainstream news sites and automatically generated articles, which are difficult to read and verify for correctness. Too many people simply read and trust the generated answer at the top of the search results. For those who want to look deeper, the first step is to set your Internet time machines to the past, before the era of the modern chatbot.

Time Machine

Here is a simple search form that supports date ranges for several search engines. I’ve made the source code available under the search the past project. You can easily modify it or embed it into your website or personal dashboard.

Technical challenges

Adding query parameters to Bing was a challenge. Their web user interface is broken, and you cannot manually type in dates on all web browsers I’ve tried. You have to click to cycle dates.

You cannot enter a date in Bing's filters. The UI is broken.

The query parameters Bing adds are complex, and even repeated interrogations to LLMs only produced queries that were wrong. Thankfully, someone on the Fediverse¹ dug up a five-year old-post explaining Bing’s weird date system, which encodes dates as a filter in days since a specific epoch².

Yandex's Limited Time Period Selection

Yandex’s user interface limits searches to time periods from the previous day, the last two weeks or last month. Their official documentation claim dates can be added directly to the query³, but I couldn’t get such searches to work. I suspect their documentation is no longer up to date. I’ve found Yandex to be an incredible resource for many things that are either not indexed or intentionally removed from Google, Bing and DuckDuckGo. However, I’ve removed it from this project for now, until I can figure out how to get date-specific queries working.

Fake Time Travelers

I found the following search result for the RTX 3080-Ti video card from 2006.

The trouble is, this video card didn’t exist in 2006. I purchased a 3080-Ti in 2021, a year after its release. This is a modern post that has its metadata set to be far in the past. This could just be an honest mistake, some type of weird search engine optimization (SEO) or an LLM-generated article.

It doesn’t really matter why this particular article exists, but it is important to realize the dates on web pages may not indicate their actual publication date. Another limitation of using major search engines is that we can only search for old websites that are still up. Websites that have left our world are only available via archival sites, which may have their own integrity problems.

Even with these limitations and caveats, using time-based queries for past search results can be incredibly useful for research and avoiding documents generated by modern LLMs.

Other Search Engines

There are several engines that specialize in indexing smaller websites. They can be useful in finding blogs, fringe websites and other interesting content.

Wiby - “Search engine for the classic web.”
Shodan - Search Engine for Internet-connected devices.
Grep.app - Fast GitHub code search.
ooh.directory - “A place to find good blogs that interest you.”
Filmot - YouTube subtitle and caption search.

Closing Thoughts

The vast majority of people don’t go past the first page of search engine results⁴, not that it matters since most search engines never allow people to view more than a few pages anyway⁵⁶. Google has been formed into a verb, with far too many people leaning on a sole search engine. In a 2005 interview with Eric Schmidt, who was the CEO of Google at the time, he stated that Google returning a list of results was a bug and that he desired for it to return one perfect result⁷. Two decades later, it seems as if the LLM era has gotten search engines closer to Schmidt’s original ~~goal~~ dystopian-level of thought control.

Most search engine and social media algorithms seem to favor posts that are current or recent. Their user interfaces often make it difficult to search for more niche content. Just as LLMs can be “hacked” using prompt injection to give unintended results, using date ranges on the major search engines is one technique that can be used to find obscure content from real human beings in the time before the modern LLM era.

He’s memeing, anyway here’s the only thing I could find on their retarded search query system. 13 Feburary 2026. @caronnov@dill.burggit.moe mirror ↩
So…how do you sort search results by time range?. 21 August 2025. cutcss. /r/bing. Reddit. ↩
Search by date, language, and file types. Retrieved 10 March 2026. Yandex. Archive ↩
Using the Internet: Skill related problems in users’ online behavior. Deursen. Dijk. June 2009. Interacting with Computers. Science Direct ↩
Where Did the Rest of the Internet Go?. 11 April 2022. TruthStream Media. ↩
YES, They Really Are Deleting the Internet And it’s WAY Worse Than You Think. 15 August 2022. Bright Insight. ↩
Eric Schmidt - Charlie Rose. 3 June 2005. Charlie Rose. (Interview) Specific Clip ↩

2026 Will be the Year of Wars

Fri, 16 Jan 2026 00:00:00 +0000

Last year, I thought 2025 would be the year of psychological operations. It’s not a risky prediction to make since false flags and propaganda infiltrate everything these days, from Wikipedia to film and cinema. Still, it was a year filled with economic hardships for most Americans, combined with questionable narratives surrounding political assassinations. Now that we’ve entered 2026, the year has already started off with military operations and uprisings. We will undoubtedly be in for a year filled with more economic hardships, but now with a side order of armed conflicts and war propaganda reminiscent of George W. Bush-era neoconservatism.

Venezuela and the New Era of Regime Change

In 2003, George W. Bush took America to war in Iraq over contrived narratives about “weapons of mass destruction.” Saddam Hussein, a man previously photographed shaking hands with Donald Rumsfeld in 1983, turned from an American ally to an enemy. The resulting attack and “Mission Accomplished” banner on the USS Abraham Lincoln led to years of civil unrest¹, and the dissolving of every Iraqi institution in a disastrously executed reconstruction effort².

I previously made the argument that Donald Trump would be the better option for leading America away from war, and it turns out that prediction was entirely wrong. On January 3rd, on the breaking of a brand new year, Trump announced the U.S. had carried out large scale strikes against Venezuela and captured their leader, Nicolas Maduro³. People in Caracas posted videos of explosions in the capitol⁴. Attorney General Pam Bondi stated Maduro and his wife had been indicted for charges of “Narco-Terrorism Conspiracy, Cocaine Importation Conspiracy, Possession of Machineguns [sic]” and a list of other charges⁵.

The extraction of Maduro was a grandiose narrative surrounded by a fog of war-style operation, similar to when Obama allegedly had a team assassinate Osama bin Laden and dump his body in the ocean. Immediately after Obama’s operation in 2011, a witness interviewed by Al Jazeera, who frequently walked by the house Bin Laden allegedly was hiding in, said he didn’t believe Bin Laden ever lived there⁶. French intelligence agencies suspected Bin Laden was already dead in 2006⁷. An Obama-appointed judge refused to release the photos of Bin Laden’s death⁸, and my own personal group of Australian friends believed the entire operation had been faked, when I lived in Melbourne in 2012.

We’re told the military employed more than 150 aircraft, including bombers, helicopters and drones, in the operation to capture Maduro⁹. Mainstream news sources released photos of Maduro in a funny hat, surrounded by DEA agents, giving two thumbs up¹⁰. Immediately people questioned if such photos were real or fake, and Snopes claimed they couldn’t independently verify the authenticity of one such photograph¹¹. Some have questioned why Venezuela’s military didn’t fight back, as their military did possess air defenses such as the S-300 and portable MANPADS¹².

If the smiling expressions and thumbs-ups photos are in fact real, it might be the case that Maduro made some kind of deal with the U.S. government to put on a dramatic military stage show with minimal casualties. Only time will tell what the true nature and purpose of this January operation was, but it is certainly not the sensational, headline-grabbing story it’s being reported to be¹³.

But What About The History?

The argument in favor of Trump’s operations in Venezuela takes us decades back to President Hugo Chavez and how he nationalized the nation’s oil fields in 2006 and 2007, arguably committing theft against international oil companies British Petroleum, Exxon Mobil, Chevron, ConocoPhillips, Total and Statoil, whose investments into the nation’s oil infrastructure were estimated at $17 billion¹⁴. This was in the wake of a failed coup against Chavez, and a strike of 18,000 workers, which brought oil production to a standstill¹⁵.

When Chavez nationalized the nation’s oil and fired experienced workers, the nation’s oil production plummeted. While redistribution of wealth to the nation’s poor might seem like a great idealistic virtue, in reality it can lead to a loss of industrial skill. It’s similar how, when Mugabe took farms away from white farmers and gave them to his supporters who knew nothing about farming, it plunged the nation into years of food shortages¹⁶.

Still, does that justify the Trump administration’s military intervention? Many of Trump’s supporters believed he would lead America away from an era of war in the Middle East, and yet last summer he ordered airstrikes on Iran, coordinated with Israel.

Trump has said the U.S. will run Venezuela for now, even as the Venezuelan supreme court granted Vice President Delcy Rodríguez full presidential power and the military recognized her as the country’s acting leader¹⁷. The original claims of fighting “Narco-Terrorism⁵” were quickly superseded with plans to subsidize U.S. oil firms to get Venezuelan oil production “up and running” with increased operations within 18 months¹⁸.

American oil firms are less than enthusiastic about developing Venezuelan oil again. Fracking has increased American oil and natural gas production since Chavez’s nationalized oil production 20 years ago. Rebuilding the oil industry in Venezuela could cost up to a billion dollars with uncertain payoffs¹⁹. Their oil is a heavy grade crude and requires thermal cracking and solvents for extraction and processing²⁰.

Trump’s operation was not about any type of liberation for the Venezuelan people. Although there are many Venezuelans who are celebrating the removal of Maduro²¹, does anyone think whoever becomes their next leader will honestly be any better? Will America, using their intelligence operatives, allow anyone to run as a serious candidate in Venezuelan elections who does not side with the interests of the U.S. government? This was not a story of liberation, stopping the flow of drugs, or even a story about oil. This is a story about regime change and an obvious continuation of neoconservative imperialism, grasping for the tail end of a dying American empire.

Shooting of Federal Officers

I’ve tried to avoid all coverage of the recent federal officer-involved shootings. Any commentary seems to bring out the absolute worst in all sides. I’ve personally gone through immigration processes to work in other countries and am the descendant of legal immigrants. I do not like the defense of illegal immigration, nor do I like defending any type of constable just because they’re a police officer, for incidents that are questionable and result in serious injury or death.

When it comes to the recent Immigration and Customs Enforcement (ICE) involved shootings, things immediately don’t seem to be adding up. In Minneapolis, the shooting of an ICE agent occurred only a few blocks from the George Floyd incident years earlier on Portland Avenue²². Meanwhile, another shooting involving Customs and Border Patrol (CBP), just happened to occur in Portland, Oregon²³. The protestors in Minneapolis, who are presumably against borders, have constructed a barricaded border²⁴, reminiscent of the autonomous zone in Seattle years ago.

ICE officers refused to let someone who claimed they were a doctor administer aid to the woman who was shot²⁵. While this seems understandable in some respects, as this was a protest that seemed to turn violent and an officer cannot simply trust anyone on the street who claims to be a doctor, it could have also been because the woman was not really shot or injured. The entire incident, broadcast immediately into the national spotlight, might have been entirely staged²⁶.

JD Vance has already stated publicly that the ICE officers have been given immunity²⁷, and local police have been denied their right to investigate the incident²⁸²⁹. Barriers were erected in Minneapolis and stayed up for three days around the area of the incident³⁰. These events only further the argument this could very well be some type of staged event or cover-up.

What’s amazing is that the conservative stance perpetuated by the outrage media is causing people to side with federal authority over local authority and in favor of federal agents without investigation or due process. The news cycle is also conditioning those on the left to side against federal officers and with the protestor, regardless of the details of the incident or if the officer’s life was threatened by a deranged individual who used a car like a weapon.

The reality is that there is a reasonable middle. If an officer tells you to get out of the car, do so and stay silent until you can speak to a lawyer. Know and assert your rights. Don’t attempt to drive off or into an officer. At the same time, any officer-involved shooting should be carefully investigated, including by local authorities even in a federal incident, to eliminate even any hint of bias or corruption. People should be held accountable for their actions.

There is significant evidence to suggest the George Floyd incident was a staged event³¹. The recent set of ICE escalations has many of the same characteristics. They are designed to psychologically condition conservatives to continue supporting federal raids, which have little to no impact on immigration. Instead, they normalize federal police in municipalities while also conditioning resentment of liberals as people with no respect for borders, legal jurisdictions or immigration law. These are events that turn both sides of the political spectrum against their own interests.

Iran

Protests have broken out in Iran, with US-based Human Rights Activist News Agency claiming more than 2,500 people have been killed so far³². How this information is getting out is questionable, as the nation has shut down the Internet, including jamming Starlink satellite transmissions, for well over a week³³. Former CIA director Mike Pompeo has already chimed in on how the collapse of the Venezuelan regime was a direct blow to the Ayatollah, claiming they rely on Venezuelan oil³⁴. This is an absolutely laughable and absurd statement, as Iran imports less than 2% of their oil³⁵.

In 1951, Dr. Mohammed Mossadegh was the democratically elected president of Iran. In an effort to bring more of the nation’s wealth to his own people, he nationalized oil reserves that had been previously held by the British. Amid fears of communism, President Dwight D. Eisenhower, in conjunction with the British, started Operation Ajax, which succeeded in a coup against the democratically elected president and led to the installation of the Shah, who would come to be known as a tyrannical dictator³⁶.

It’s impossible to know at this point what is really happening in Iran. Reports vary wildly about the number of people killed, and America’s intelligence agencies have a long history of operations to unseat the nation. I have serious doubts if any of the recent protests are in any way organic. The nation of Syria collapsed in December of 2024, and Iran is another nation that has been in the crosshairs of regime change rhetoric by American presidents for decades. This could simply be a continuation of the escalations from the previous summer; yet another attempt at destabilization and regime change.

The Year of War and Mirrors

The opening events of January seem to be used by the media to set the tone for the entire year. In the past, we’ve seen January as a stage for political prisoners, terrorism and assassinations. The rhetoric leading up to 2026, including how Trump restored the tiles Department of War and Secretary of War³⁷, was preparing the way for events that would be psychologically burned into the minds of all Americans at the very start of the year.

It sets the stage for 2026: a year that pushes America into a regressive policy of imperial expansionism. It’s hardly a new concept. The technocratic class is embracing computing power they could only dream about a century prior in order to advance the surveillance state and consolidate power into a small set of kingdoms.

New Year’s Day is often seen for many as a new beginning and a time to make resolute the goals one has for the coming year. It’s a period of rebirth, pushing through the decay of winter, past the solstice with hopes of longer, warmer and brighter days. That is also why it is the perfect time to induce trauma, like a baby being poked and prodded and subjected to invasive procedures in a hospital, tainting and conditioning it before entering our modern world³⁸. We are slowly being conditioned to the annual media cycle, bringing us propaganda and trauma during the cold winter months, preparing us to turn further against one another in what will surely be a year of endless wars.

Inside Iraq: The Untold Stories (2004 Documentary) ↩
No End in Sight (2007 Documentary) ↩
The United States of America has successfully carried out a large scale strike against Venezuela and its leader, President Nicolas Maduro, who has been, along with his wife, captured and flown out of the Country…. 3 January 2026. Trump. Truthsocial. Archive ↩
NOW - Explosions detonating in Venezuela’s capital of Caracas, reports.. 3 January 2026. @Disclosetv. X. ↩
Nicolas Maduro and his wife, Cilia Flores, have been indicted in the Southern District of New York. Nicolas Maduro has been charged with Narco-Terrorism Conspiracy, Cocaine Importation Conspiracy, Possession of Machineguns and Destructive Devices, and Conspiracy to Possess Machineguns and Destructive Devices against the United States…. 3 January 2026. @AGPamBondi. X ↩ ↩²
Bin Laden Death - Eyewitness Neighbors Say It’s Bullshit. 3 May 2011. Al Jazeera news clip posted by TheFlexEffect. ↩
French Intelligence Also Suspects Osama bin Laden Is Dead. 23 September 2006. White. About.com US Liberal Politics. (Archived Captured 27 November 2012) ↩
Judge Who Blocked Release of Osama bin Laden Death Photos Now Blocks Release of Senate Torture Report. 26 May 2015. Brinkerhoff. Straehley. AllGov. ↩
U.S. had 150 aircraft in the sky during Maduro raid. 3 January 2026. Demarest. Axios. ↩
Nicolas Maduro smiling and giving thumbs up in posed pictures with DEA agents. 3 January 2026. Smith. MSN. ↩
Inspecting image of Maduro giving thumbs up next to DEA agent. 8 January 2026. Christenssen. Snopes. ↩
Why Venezuela’s Military Did Not Fight. 4 January 2026. García. GlobalResearch. ↩
The Gathering Storm: Snatch-and-Grab Regime Change is the New New Normal. 11 January. Corbett Report. Substack ↩
Venezuela nationalizes Orinoco oil fields. 27 February 2007. The New York Times. ↩
Hugo Chavez’s Sad, Oil-Soaked Economic Legacy After a decade drunk on petroleum profits, Venezuela’s economy is looking at a dicey future.. 6 March 2013. Weissmann. The Atlantic. ↩
Mugabe and the White African. (2009 Documentary) ↩
Trump says the U.S. will “run” Venezuela for now. What do we know about the plans?. 4 January 2026. Quinn. CBS News. ↩
Trump claims US oil firms could be ‘up and running’ in Venezuela within 18 months. 6 January 2026. Epstein. MSN. ↩
Making Imperialism Great Again?. 13 January 2026. Paul. The Ron Paul Institute (via Lew Rockwell). ↩
Venezuelan Oil and U.S. Refining Capacity Constraints in 2026. 14 January 2026. Hidayat. Discovery Alert. ↩
Venezuelans are PISSED OFF... 5 January 2026. Asmongold Clips. ↩
Minneapolis on edge: ICE agent shot Renee Good ‘in the face,’ just blocks from where police killed George Floyd. Garg. 8 January 2026. WION. ↩
Portland shooting latest: Police chief alleges gang ‘nexus;’ 2 shot remain in hospital. 9 January 2026. Ault, Stone, Deliso, et. al. ABC News. ↩
“Wake up babe, the open borders people are constructing borders again”. 9 January 2026. @lance_corporate. X. ↩
ICE ‘blocked doctor for 15 mins’ and ambulance from trying to help dead shooting victim. 7 January 2026. Orrico. MSN. ↩
33’S COMPANY! WAKE UP AMERICA THEY ARE REVEALING THEMSELVES THROUGH MAJOR CODED PSYOPS RIGHT NOW!. 9 January 2026. Official Rumble Channel for A Call For An Uprising. ↩
JD Vance Declares ICE Agent Involved in Fatal Shooting Has ‘Absolute Immunity’. 8 January 2026. Bahney. Mediaita. Yahoo! news. ↩
What’s behind the highly unusual move to block Minnesota officials from investigating ICE shooting. 9 January 2026. Rabinowitz, Perez, Holmes. CNN. ↩
Minneapolis Cops Blocked From Investigating The ICE Shooting. 12 January 2026. Timcast IRL ↩
Minneapolis deadly ICE shooting: City removes barricade near scene. 9 January 2026. Turtinen. FOX 9. ↩
The Timeline Of George Floyd With Maryam Henein and Sean Hibbeler. 24 September 2022. Tin Foil Hat with Sam Tripoli. ↩
‘The world needs to know what’s happening’: families of protesters killed in Iran tell of heartbreak. 15 January 2026. Parent. The Guardian. ↩
Iran’s internet shutdown is now one of its longest ever, as protests continue. 15 January 2026. Franceschi-Bicchierai. TechCrunch. ↩
The collapse of the Venezuelan regime is a blow to the Ayatollah, whose regime relies on Venezuelan oil to stay afloat. We must ensure that Maduro isn’t succeeded by yet another corrupt, Iran-aligned Chavista, but a government that represents the Venezuelan people.. 5 January 2026. Mike Pompeo. Facebook. ↩
Iran Oil Reserves, Production and Consumption Statistics. World Meter. Retrieved 15 January 2026. Archive ↩
A Very British Coup. ↩
Trump Renames DOD to Department of War. 5 September 2025. Olay. Pentagon News. ↩
Ep. 215 - Entre Dos Auguas (pt.1). 10 October 2025. WAR MODE. (podcast). YouTube. Apple. Rumble. ↩