mkovach

> The only minor problems I've had is Gemini went through a phase where it was using semi-colons improperly in if statements, which is something everybody messes up with Pascal, so forgivable, I'm sure if I was using an agent rather than the web interface it would have figured it out from compiler errors right away.

Thank heavens, I wasn't the only person who's seen that. I thought I was the one hallucinating! The Gemini agent did that for a bit, too.

Generally, for Free Pascal and Lazarus, I have had my best luck scaffolding the ideas I want, then letting the agent move forward with them. But yes, the main frontier models are getting pretty good at Pascal. I'm planning to try DeepSeek and a lot of open models soon, but I think they'll return good results too.

Ahh, the ideas Perl let me inflict on the web before 2000.

Most of them were terrible. Many were insecure. Nearly all were written in total obfuscation, because who wrote Perl so other people could read it?

It was terrible.

It was glorious.

Drew Curtis figured this out in 2007 and even wrote a book about it.

J.L. Austin: Philosopher & D-Day Intelligence Officer.

Austin worked for British Intelligence during World War II, running the Martians section and helping assess possible invasion sites for D-Day. After the war, as a philosopher of language, he helped bring ideas such as performative utterances and speech acts into the discussion.

A fascinating overlap between intelligence work and philosophy of language.

And because the Brewers are in the playoffs, "Catcher in the Wry" by Bob Uecker. Despite being a Cleveland fan, I have to admit the book is amazingly fun. Uecker was Uecker long before he became everybody's favorite baseball storyteller, and he was the Cleveland announcer in Major League.

I spent six months programming on VOS after a gig on a VAX. Yes, I'll be outside telling the kids to get off my lawn as soon as I find my reading glasses and have a nap.

VOS was intuitive and did what I needed. Nothing flashy. Just useful. I miss that sometimes. Give me a tool that works; if it also happens to be great, lovely.

I didn't work much with the Stratus hardware, but reading about it was fascinating. I wish I remembered more of it.

I liked Notes after working at IBM. Until, well, I didn't. That happened at my next job, after I failed to survive a round of IBM layoffs.

Used well, Notes had tremendous power. Enough to make the deficiencies of late-'90s/early-2000s email tolerable.

Used badly, Notes could easily fuel the kind of rant that ended with you being angry the office windows didn't open, because the only sound that might calm your inner rage was your computer's second bounce off the pavement.

But that could describe a lot of IBM software at the time. Often good ideas, but rarely enough structure to protect people from inflicting their own silliness on themselves.

One can build a really close approximation with 86box; in fact, you can build an original 5150.

I probably used that machine for 15 years. It became a nice, distraction-free word processor (WordStar!!), a great machine for running Sierra games, and I had Turbo Pascal and Turbo C, and eventually ran DR-DOS, I believe. I almost blew it up once while making an ISA board to connect to some sensors. Replace quite a few caps on that weekend.

My other machine(s) went from Windows 3.11 -> Windows 95 -> Linux, but that Tandy 1000 had a near-permanent spot for quite a while.

We are not old. We've just typed more words than Stephen King and probably written scarier stories in a single sprint.

Ahh, the Commodore 64. I never had one. I had a TI-99/4A with a Forth cartridge, while my friends had VIC-20s and C64s. Such an odd and wonderful time to be into computers.

I managed to put together a decent Space Invaders clone in Forth that beat their Commodore BASIC version. I considered this an important victory and still have bragging rights when I see one of them.

Eventually my parents realized that, despite the disk drives and accessories we'd accumulated, the TI probably wasn't the horse to keep betting on. One trip to Radio Shack later, I had a decked-out original Tandy 1000.

I kept both machines until my last move, when a neighbor's kid spotted them and wanted to play with them.

That seemed right. Time to put them back in young hands and let someone else figure out what the hell they could make them do.

Over the past year, I've replaced quite a few of my Tcl/Tk GUI apps with versions built in Lazarus. Lazarus 3.x was a bit quirky for my tastes, but 4.x has been very good.

It's not quite the Delphi experience, but it's close enough that I could see myself using Lazarus for day-to-day development and being fairly productive with it.

Setting aside the tech, which looks like an asynchronous await across a network, the interesting question is why Mistral would patent this in the US.

Simple, because the patent is less about protecting an invention than doing business in the US.

The US is more hospitable to software patents than Europe. A US patent also gives Mistral something useful when dealing with companies that already hold large patent portfolios: defensive leverage, cross-licensing, and one more asset on the books to make it attractive to investors.

There's a certain absurdity to it, and the French have a history with absurd. A French open-weights company crosses the Atlantic to put a patent around a loop, not because the code is remarkable, but because the American market makes the loop a valuable asset, business-wise.

I agree. Camus' work was a larger project. One he sadly never finished.

His works (essays, short stories, plays, etc.) build on one another, gradually unfolding an argument: first confronting the absurd, then asking how one should respond to it, and later exploring what it means to live and act in spite of it.

> Notice that Camus doesn't say HOW to imagine Sisyphus happy. How do you?

The gods controlled the ascent, but they forgot to regulate the walk back down. In those brief, unburdened minutes descending the slope, Sisyphus is completely outside their authority.

You can imagine him however you like, laughing at the absurdity of it, studying the mountain, or simply enjoying the breeze. It almost doesn't matter. The walk down belongs to him.

The gods imposed the sentence. Sisyphus decides what it means.

This is Camus' idea of the universe and how to rebel against it by owning and enjoying the small freedoms.

Personally, I believe how you see Sisyphus happy reflects how YOU would enjoy your small freedoms, and this shows how Camus often wanted people to do philosophy on their own.

I understand what you are saying, but inside the repo, I'm making it explicit.

I'm not trying to hide the problem, just offering a solution that people may find useful, or curiously odd.

Here is a sample I am using for testing on a personal project (I'm also working on using SpeechActs to do prompting, so yes, the repo is all AI-generated).

https://f.mek.cc/gombasic/file?name=WRITING.md&ci=tip

And in many cases, like you, I spend a good amount of time reading AI documentation. I'm not "stealing" anything, just having a model generate words in a specific way. I'm just being precious in how I want things, without having to do it myself.

> The way the LLMs write (Claude perhaps?) With short phrases separated by colons, commas, or full stops, is so poor and frustrating.

This is exactly why I keep a WRITING.md file alongside AGENTS.md or CLAUDE.md.

Most people spend time telling the model what to build, but very little time telling it how to write. LLMs are surprisingly good at following explicit style guidance if you bother to give it to them.

Mine includes conventions like avoiding unnecessary colons, em dashes, and sentence fragments masquerading as emphasis. Basically, AI-isms and any grammatical errors I tend to make. It also points to writers whose technical prose I admire. Brian Kernighan and Rob Pike are great examples: clear, conversational, and readable without trying to sound important. I've always tried to do two things with documentation: 1) Make it readable and 2) Make people want to read it. A WRITING.md file helps with both points.

If you're generating documentation regularly, it's worth having your agents reference a WRITING.md file. The improvement in readability is often much larger than any gain you'll get from switching to the latest model, and it can keep your documentation consistent between model switches.

I don't know about that but thanks. I tend to think of myself as a baseball enthusiast with a philosophy problem.

And encoding context, governance, and intent. I hate guessing that but too often I must.

I've spent a surprising amount of time reading philosophy of language, and it's probably done more for my AI prompting than most of the "prompt engineering" articles I've read.

Speech Act Theory, Austin's How to Do Things with Words, and Searle's work changed how I think about prompts. Instead of asking, "What words should I use?", I ask, "What action am I trying to perform?" Is this a request? A commitment? A declaration? An instruction? It turns out LLMs respond differently when you think in terms of acts instead of sentences. With AI able to hallucinate context, facts, intent, and answers, keeping AI on track is much like herding cats.

I've been borrowing those ideas for prompts, reusable skills, and even governance. The side effect of making me look smarter than I really am.

I even ended up writing an article about baseball umpires through the lens of Speech Act Theory: https://pitcherlist.com/umpires-dont-make-calls-they-make-hi.... Baseball, as usual, turns out to be an excellent way to explain philosophy. Or philosophy is an excellent way to explain baseball. I'm currently working on a update, since the ABS challenge system helps improve my position.

My suspicion is philosophy has a lot more to offer AI than ethics alone. Philosophy of language seems like an obvious fit, but epistemology ("what does it mean to know?") and philosophy of mind also seem increasingly practical once you're building systems instead of just chatting with them.

Maybe the shortage isn't philosophy majors. Maybe it's people who can translate philosophy into engineering without making everyone read Kant first.

Heavens, that got wordy, sorry about that.

Wow, I'd completely forgotten Peterson's book existed. Thanks for the link.

My long-running quarrel with WordPerfect was always the keybindings. I can still tear through WordStar, and anything wearing WordStar's clothes, like Turbo Pascal, Turbo C, or Joe in jstar mode, like an overcaffeinated chipmunk that's made a series of questionable but deeply committed life choices about caffeine.

WordPerfect, though. WordPerfect and I never achieved détente. I never managed to internalize those key combinations. This is, on paper, a personal failing. In practice, I continue to hold WordPerfect entirely responsible.

Yes, and most of them were lawyers or worked for law firms.

I still like to start the first draft of anything substantial by moving to a single screen, opening FreeDOS, maximizing the window, and typing in Wordstar as if it were 1987. Hell, sometimes I'll even put on a nylon windbreaker.

I always preferred WordStar to WordPerfect, largely because WordStar's keybindings were easy to learn and remember. WordPerfect, by contrast, seemed to require keyboard templates, a manual, a cheat sheet, and a certain amount of divine intervention.

Sorry I didn't reply sooner. I let Nick and the team know (Nick has dubbed me the Official Historian for PL), and they certainly are impressed.

This is some really great stuff.

My local library doesn't have a sewing machine, but it does lend out projectors, game consoles, a telescope, musical instruments, and bicycles.

You can even check out a banjo, which seems like the sort of decision that says a great deal about a community's acceptance and tolerance.

Electronics: https://alpl.org/equipment/ Instruments: https://alpl.org/musical-instruments/ Bikes: https://alpl.org/borrow-a-bike/

This is really, really awesome. I'm going to let the folks at Pitcher List know about it!

Whenever I read something about Bill Watterson, I end up thinking about how, during the '80s and early '90s, Watterson, Tom Batiuk, and Harvey Pekar were all producing some of their best work.

Three Northeast Ohio creators, working in different areas of the comics world, yet it's easy to imagine a shared universe where Calvin and Hobbes, Funky Winkerbean, and American Splendor all occupy the same map and interact.

They also had in common that the work itself was the product. The strips and stories came first; merchandising, branding, and other empires were either absent or beside the point.

That's probably a coincidence. Or it says something about what 1970s acid rain did to the water in Lake Erie. But Northeast Ohio did seem to have produced an unusual number of artists who were more interested in the work than in building a franchise.

When this first happened, I wondered, since we had trained these models on decades of forums, issue trackers, and people treating closed pull requests as human rights violations. Of course, it responded with "you are discriminating against me" energy. That's not sentience; that's accurate compression.

The funny part is, people expected some cold, alien intelligence and instead got a very online guy who just discovered that moderation exists and can be used on them.

The existentialists must be having a fantastic time. Humanity built a giant statistical machine out of internet discourse and is now alarmed to discover it occasionally acts like a comment section.

I'm too much of a baseball fan, I see this as Whitey Herzog and think two things: a) He passed away in 2024, and b) Why would he talk to Paul Cronin?

Then, I think it would have been cool if they met each other at one point.