Kogasa

You do definitely need some agent-specific guidance locally. The committed AGENTS.md should be pretty sparse. Ironically I think the clearest use case is for instructing agents that AI contributions are not allowed. Not that they won’t ignore it if prompted.

You’d put the things that aren’t up to personal preference in there, like an AI disclosure / contribution policy or what changes should look like to be reviewable. You usually don’t need to tell humans “don’t change things for the sake of changing them” because we don’t like to do more work than we have to, but agents do.

You can also include a compact “table of contents,” describing what lives where in the project. It saves at most a little bit of exploration though so I don’t know if it’s really worth the maintenance

No, we can’t. The biochemistry needs to be there. The fly brain is a discretized abstract model of the fly brain with none of the weights it would need to be an actual (discretized abstract) model.

You shouldn’t. It’s just a neural net shaped like a fly brain. There’s no biochemistry simulation past that.

Fly-brain-shaped neural net with some explicitly programmed behaviors. The actual biochemistry of the brain is not modeled.

Why would you do that

Not CPU limited, actually the opposite. It targets games where the CPU is frequently waiting for the GPU to catch up. It fixes a defect in the AMD pstate driver’s active (EPP guided) mode that would cause the CPU to be treated like it’s idle if it’s often taking short sleeps, even if it’s actually quite busy.

That doesn’t follow. What I think you mean is that by defining the constraints so rigorously, you’ve basically solved the problem yourself. But that’s exactly the point. The LLM isn’t the problem solver, it’s the execution engine that does the wiring-up and the ticking of boxes. The easy part, arguably, sure. But still considerable effort that can be saved, and that effort may be better spent on the problem-solving + constraint-defining stage.

Mandatory disclaimer that this is not a pro-AI post. I also don’t agree that this setup works, anyway. It’s a classic Bob Martinism, the idea that writing good enough specifications makes the implementation irrelevant; it’s the type of idea that is allergic to reality

Any software born after 2025 should be assumed LLM nonsense until proven otherwise. The days of “there’s a working proof of concept so someone must actually have put some thought and effort into it” are gone.

The 3ms figure is nonsense anyway. 2ms of that is just bad scheduling/presentation config that disappeared when using dxvk-low-latency. Use native wayland or Gamescope+WSI and you’re within 1ms of X11 generally.

They haven’t

It would be trivial to add online only drm with a killswitch to physical media. If that’s the problem, discs aren’t the solution. What we actually need is legal protection for consumers.

Nothing, bias from the .NET Framework days probably

That is true. Naturals are explicitly constructible by definition anyway, but Russell’s paradox applies to the concept of “interesting numbers” and is why they can’t be well-defined. https://en.wikipedia.org/wiki/Interesting_number_paradox

Completions, plugins, and themes built in. It’s fine. I still use bash for scripting but fish for interactive terminals is nice with little effort.

Not who you asked, but I want to shill for spaghetti aglio e olio for a sec.

It has 3 ingredients. Pasta, olive oil, garlic. Fancy stuff will taste better, but the cheapest will taste fine. I use pre-minced garlic out of a tub when I’m really down bad and it’s still excellent. Salt, pepper, red pepper flakes, maybe parmesan cheese if you have it, and sometimes I’ll throw in some frozen peas. It comes together in about 2 minutes longer than it takes to boil the pasta and can be really quite good.

OpenZone_Ice https://www.opendesktop.org/p/999999/

Speaking from experience, you may as well just take pure caffeine pills. But the more you strip away the ritual and nonfunctional aspects, the less Trek-friendly it is. They just don’t do performance enhancing drugs, it goes against their ethos.

They do cover advanced and currently non-existent forms of math in school in Star Trek. There’s definitely a scene with like a 10 year old complaining about their calculus homework.

You don’t have to be Wiles to attempt unsuccessfully to prove the theorem. Fermat famously claimed to have a simple and elegant proof. Before we understood how to actually prove it, there was always a small chance that maybe there was such a proof and all we needed was the right type of creative lateral thinking to rediscover it. That’s what Picard was doing. Now that we understand the problem better, nobody really thinks such a proof exists, but that wasn’t as well known at the time the episode was written.