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.
Jun 21, 2026
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 public. Instead, they provided a limited release of this model to select organizations as part of Project Glasswing. Then, last week, Antrhopic released a version of this model called Fabel. 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subscribers were allowed an entire three days to work with this model before the US federal government issued an unpublished export control order for Anthropic, citing concerns of model jail breaking and weaponization, to restrict access to the model from any foreign nationals living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 abroad. Anthropic denied the government’s finding, but complied with the order by revoking access to the model for all customers. 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.
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 2025. This was after Micron received massive amounts of taxpayer money from the CHIPS Act. 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.
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.
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.
I’ve owned a lot of mechanical keyboards. Over two years ago, I started using an Ergodox layout, starting with the SliceMK. It was difficult to get used to the ortholinear layout of the Ergodox, but I’m glad I stuck with it, as I really enjoy the experience of typing on split keyboards. I later added the W-Ergo to my keyboard lineup, the first keyboard I’ve used with SA profile keycaps. A few weeks ago I built out another Ergodox profile keyboard: the IF-Ergo. It uses the Gateron Baby Kangaroo V2 switches, as well as the Domikey Crisis and Cyberpunk SA keycaps. This keyboard replaces my Slice MK as the primary keyboard for my work laptop.
In 2019, I placed a pre-order for a Purism Librem 5. It was touted as a Linux mobile phone, providing an alternative to the two big players in the cellphone operating system space. Two years later I attempted to cancel this pre-order, as the promised device had failed to materialize. Three years after my original order, I got an e-mail indicating the device I no longer wanted would finally ship. It took almost six years from my original order to get a refund. I was met with tons of excuses from support staff and eventually had to file a complaint with my credit card company to finally get the money restored. I’m honestly shocked Purism is even still in business. They’re a horrible company filled with liars. No one should give them any money for what must be garbage products.
I’ve been using Sonatype Nexus for the past few months as a repository for build artifacts (Python and RPM packages) for project I’ve been developing. Recently several of my build pipelines broke with RPM checksum errors. Doing some searches, I only found old, unanswered forum questions and absolutely useless AI-generated answers. Upgrading to the latest version of Nexus didn’t solve the issue either. After looking through some documentation, I did discover how to rebuild the package indexes and fix my issue. However, I didn’t realize the Nexus upgrade shifted me from a truly open-source artifact repository to closed proprietary commercial software with no way to downgrade back to the open-source version. This was infuriating. Now, even though my problem was fixed, I had to extract all my artifacts, rebuild an older Nexus installation and then reupload all my package. I’ll also have to create my own build pipeline if I want to continue getting updates for the open source-version of Nexus.
Nearly a decade ago, I was purchasing beer for a friend’s party at a drug store and the cashier wanted to scan my drivers license. I refused and her manager said my ID had to be scanned to make my purchase. The cashier was surprised no one had refused a scan before. Sometimes people have scanned my IDs before I could stop them, so I put a sticker on the back of my drivers license. Increasingly, more businesses and restaurants are demanding this invasive tracing of customers for entry or alcohol purchase. No matter how much a bouncer may claim “we don’t store your information,” they have literally no way of being sure of that. If you value your privacy, do not let venues or stores scan your drivers license.
When Elon Musk originally bought Twitter, a number of people moved over to Mastodon. There were a lot of misunderstandings around how Mastodon works. People are used to mega-websites, now known as platforms. Mastodon is composed of tens of thousands of independently run servers that communicate with each other using a protocol known as ActivityPub (AP). There are other services that can communicate over AP, including Pleroma, PeerTube, Misskey, PixelFed and more. This interconnected network is colloquially known as The Fediverse. People use to the mega-sites and large-scale algorithmic content moderation were completely unprepared for a return to distributed, chronological, non-algorithmic social networking.
