Notes on software, history, data, and whatever
The last few weeks saw another upheaval in the tech markets. The release of Kimi K3 seemed to shake confidence in the idea that OpenAI and Anthropic had a lock on the best models for the foreseeable future. Kimi K3, and GLM 5.2 before it, come close on some AI...
This morning it hit me why some of my AI software projects aren’t interesting to me. Basically it is the mirror of the sunk cost fallacy. There, we cast a desire to hang on to something – throwing good money after bad – as a flaw in our thinking. But...
Recently while building my iOS document reader app, I relearned a good lesson: You’re not done until you’ve successfully used your software on the least capable possible hardware, not the most. If you want to write fast software, use a slow computer Dominic Tarr Testing on a slow iPad saved...
Today I was trying to recall how C and C++ pack struct members. It’s important if you really need to use as little memory as possible. When you have 150 million records at 64 bytes each, it starts to matter that half of that is wasted space. It’s annoying to...
I was thinking that out of the three big upcoming IPOs: Anthropic, OpenAI and SpaceX that SpaceX would be the most promising one, since it does real things without a lot of competition. Their customers don’t have good alternatives, and their biggest customer is the U.S. government. Sounds reasonable. But...
From a recent New York times article Yanacek peered at the A.I.’s suggested fix, pondered for a moment, then hit “enter” to approve it. The A.I. took about eight minutes to figure things out, he told me. “By the time I’d opened my laptop, it’s ready.” One customer recently told...
The war against Iran is the most poorly planned, least legal and most chaotic large military operation perhaps ever done by the U.S. The President hasn’t offered consistent reasons or objectives either, so we can’t even argue with him. Was the war necessary? Was it necessary in the minuscule time...
Notes on unintuitive results from using C's log10() - What if you want to know the number of digits needed to represent an integer variable in base 10? You may want to know before printing if it will fit within a printed field. Furthermore, you’d like it to be done pretty efficiently. Why not simply convert to a string...
From Plain English straight to assembly - The other day I tried out a Z80 emulator Z80 Pack, and had the idea to make some software for the Z80, which I’ve never done. In the 80s I only ever programmed the 6502 CPU a little, and the 8088 a bit more. I was always interested in the...
Sometimes I like to saute sliced sweet potatos with some kale for lunch when working from home. Add a little vinegar and seasoning, maybe a bit of cheese and you’re done. This meal turns out great using typical orange sweet potatos, but fails badly using purple sweet potatos – they’re...
Much like our modern food and environment have made it hard to eat healthily without effort, and our modern transportation and farming and automated manufacturing have made exercise an optional activity, extreme wealth and power make difficult thinking an optional activity for the super rich. The billionaires can’t help it;...
The economics of the software business has been really good for developers since the 1980s at least. By developers I mean anyone in the business of making software, not just individual engineers – developers in the Apple App Store sense. Before the 1980s software was mostly a consulting style of...
I think this is a sub-genre of its own - Here are a few of the best (most disturbing) short stories I’ve read, ordered from newest to oldest. They’re squarely science fiction (maybe the Quickening is a slight exception here), which only adds to the horror. You don’t have to believe in the supernatural – there’s an explanation, however unlikely....
Like the surprising effectiveness of AI coding assistance itself, AI excels at types of coding tasks you might not expect - For fun, I’ve been trying out Claude Code on a variety of my personal projects, not expecting much. I am curious to find how hard is too hard for Claude. A fairly straightforward board game proved too much. And then, debugging broken features and adding new language features…
AI code assistants will cause great harm to business in the next ten years or so. Why? Because nobody will understand how their software works, what it’s doing exactly, if it’s actually secure, or in what ways it may fail to scale and so many more particulars. To some approximation...
I’ve tried out most of the AI LLM products at this point. For the most part I haven’t found a use for them beyond short term entertainment. Recent releases of Claude and Gemini have gotten good enough to perform some busy-work coding for me, but anything complicated needs too much...
This pizza uses almost none of the standard pizza toppings and it tastes great. Make it in a wood-fired pizza oven if you can, or use the highest temp available in your oven. Use convection if you have it. Ingredients Toppings 1 small ripe apple, thinly sliced (it should be...
With the still only partial return to offices, I expect lots of people are running into an unpleseant side of modern computing. Frequent software updates, logins and operating system patches will interrupt your work or at best, restart your computer between work sessions, potentially leaving you some wreckage with your...
What follows are a few sort of random observations on topics I’ve pondered while evaluating new languages and thinking about building my own language projects. They aren’t radical design choices or anything groundbreaking but they lend a language its feel for better or worse. Mainstream PL not quite big ideas...
“In a Hierarchy Every Employee Tends to Rise to His Level of Incompetence” Laurence J. Peter, The Peter Principle: Why things Always Go Wrong That’s the hypothesis anyway. More on the Peter Principle. Recently I learned about the study Promotions and the Peter Principle which supports the hypothesis with empirical...