pohl

Everyone was saying they shouldn’t just accept the LLM’s output without cautious review. Seems they took the time to make sure it was a good release. Not sure why that doesn’t sound like a valid reason not to rush to you.

Be honest, though: had they shipped earlier—which they probably could have safely done—they would have been called “reckless” by the detractors. There was no way to make them happy here.

If there’s only 3 copies in existence, and everyone on the frontier wants it in their corpus, what do you think will happen?

I don’t think so. The ChatGPT app was, which is the “Classic” app now. The Codex app that they’re carrying forward is an Electron app and if you forget to quit it before you walk away it’ll make even your M5 Max unresponsive eventually. Sad days.

Probably for the ease of coding extensions — which strikes me as outdated thinking: if it’s open source and you’re outsourcing the coding to LLMs, why not use a compiled, safe language?

There’s an interesting counter example for DeepSeek called CodeWhale, though:

https://github.com/Hmbown/CodeWhale

only twice

https://en.wikipedia.org/wiki/Elon_Musk_salute_controversy

No, shrinking portions while also charging proportionally less would be the best thing we could do.

I use IntelliJ at my current gig—and the CoPilot plugin for AI, because it’s what I have access to—and it has become ClaudeCode-esque of late: sub-agents, skills, tool-calling, hooks…and continues to improve. It’s unfortunate how the AI-in-a-sidebar UI kind of hides how sophisticated it has grown, but it’s a glass of ice water in hell in an environment where you can’t easily procure what you want.

No doubt. Government is already on the ropes in the struggle against billionaires and large industrial/resource-extraction lobbies. If anything, that’s what needs counterweight.

I suppose it could be that if one’s intent is to mislead, then obfuscating the name of a technique specifically meant for that purpose is a feature, not a bug.

Fun and useful PR jargon to describe what this post smells like: a “limited hangout”.

(a damage-control strategy where an organization or individual admits to a portion of a secret or scandal after a cover story fails, while still hiding the most damaging key facts.)

Soon we'll be coding with agent assistance and it'll suggest we get some Carls Jr. Big Ass Fries.

Not even if it means an exec can get the yacht with a helipad?

Since the objective was never disclosed, one might assume it was to deplete our magazine depth.

Good observation. I’m certainly not trying to over complicate things. That just tells us, though, that the bio-mechanical computation part is there before language, which I took as a foregone conclusion, it being meatware and all.

> humans reasoned before language

That's an interesting supposition. Are you assuming language didn't exist before it was written? Language and meaning are, if you squint, pretty ancient and have roots in things like birdsong. It could be that ur-semantics predates our species as a whole.

I'm not convinced—and certainly don't find it obvious—that this couldn't ultimately also be how we reason as humans.

It's clear that there's an enormous amount of leverage built into language-as-practiced that one can use to engage in a broad spectrum of reasoning, from the extremely fallible off-the-cuff conclusion to the deeply-considered and rigorous proof. How do we know this leverage is built into language-as-practiced? Because LLMs can do a broad swath of it.

But how do we know we're not doing something similar?

I don't think we can assume that we're not simply by observing that we're not digital and we don't use matrix multiplication. Why immediately dismiss the possibility that there might be a similar, but biomechanical, computation at play in our heads that plays in the same space of vectors?

Seemed strange they would bench Terra and Luna but not Opus, Sonnet, and Haiku

I’m curious what domains might use something like this. (I haven’t found a need, yet)

No doubt

Does that data exclude FFI wrappers, which necessarily must be unsafe? If not, don’t put much stock in it.

Not completely unrelated. Both were likely entered to distract from the same news story dogging him every 24hr cycle.

That’s an extraordinary claim.

I always saw the mechanism being that Clevon created more offspring without instilling good memes.

Everyone unintentionally telling on themselves as not being big readers.

Your first sentence is a perfect example of why emdashes need to exist. The comma is overloaded, since it can be used both as a list separator and for parenthetical asides. But that first sentence starts using them as a list separator, and then inserts a parenthetical aside into the list. Emdash wins this test.

> without information about what you are concentrated in.

They said cash.

Agreed. OTOH, if the question is should I also learn some different way of thinking about music/fretboard, the answer is always yes. If it makes neurons fire, add it to the mix.

Can you pinpoint the revolution by pinpointing moments of invention and release? Revolution takes time because it’s more about uptake, downstream effects, and feedback.

Nice. If I add a primary and then clicking optimize—without first moving the primary—it seem as though there are still only 3 primaries, because the new one doesn't move.