valleyer

Not if the harness applies an OS-level sandbox to the process, as Codex does.

https://learn.chatgpt.com/docs/sandboxing?surface=cli

No, merely analyzing or discussing the law generally does not constitute the practice of law. If it did, plenty of newspaper articles, law review journals (often written by law students), non-attorney legal aid, and legal pamphlets would be prohibited.

If you want something more concrete, I googled it, and in New York, there appears to have been a case New York County Lawyers’ Association v. Dacey, in which Dacey wrote a book "How to Avoid Probate!", and the NYCA accepted a dissent in a lower court, stating in part:

    Does the writing, publication, advertising, sale and distribution of "How To Avoid Probate!" constitute the unauthorized practice of law within the meaning of subdivision B of section 750? It cannot be claimed that the publication of a legal text which purports to say what the law is amounts to legal practice. And the mere fact that the principles or rules stated in the text may be accepted by a particular reader as a solution to his problem does not affect this.

    "[I]t is a prized American privilege to speak one's mind, although not always with perfect good taste, on all public institutions" (Bridges v. California, 314 U. S. 252, 270)

The law applies to everyone, so it's reasonable for everyone to try to understand it, not just attorneys.

Similarly, it's fine for people to have opinions on food, dental hygiene, and the tax code without being a chef, a dentist, and an accountant.

By contrast, you can, right now, try believing that what’s happening is what’s supposed to be happening, in exactly the right order.

My government is currently killing people for looking Mexican. Any advice for how I should apply this belief there?

This lawsuit alleges violation of federal, not state, law.

Arabic or Hebrew would be better tests for that.

A proper fix is not to shell out to a command at all; use getpwnam(3) or similar.

This doesn’t tell me a whole lot, because my intuition for QALYs is weak. How strongly should I prefer an intervention with a net benefit of 0.025 QALYs over other things I might do with my time? No idea!

However! When marketing the effect of global health interventions, a count of 27 qalys is typically considered “a life saved”. A life also happens to be a million micromorts, and I have a much better intuition for micromorts!

This came across as unintentionally funny to me. It goes from making the joke that one obscure unit of measure is inscrutable to saying, don't worry, because we can put it in another equally obscure unit of measure!

Right, that's the fix. So saying that it's not fixable is incorrect.

Your second paragraph directly contradicts the first.

Yeah, god forbid we explore new things as a species.

I think this is what you meant, but just to clarify: it doesn't allow it to write to .git. Read access is allowed.

Codex (at least) already imposes the macOS sandbox on the shell commands it runs. If it wants to run something without sandbox imposition, the harness makes me approve it manually.

Is the difference with your script mostly that you choose to impose a stricter sandbox profile (and not allow any user-approved exceptions at runtime)?

Oh, I was also quite happy with Fable. I was just answering the question asked.

Well, in part because the phenomenon has been discussed on Web forums that (a) have at this point made their way back into training data and (b) are accessible in Web searches that the model can invoke. And in part because the model can "know" what its initial instinct is and "decide" to go against it.

Has the US government limited access to any non-LLM ML models?

Unfortunately nowadays even the built-in apps on the major desktop OSes are inconsistent, so the temptation for third-party apps not to care is somewhat understandable.

Ah, I hadn't noticed _SimpleSetNode.

You're right, and I guess I've been breaking that rule for a while. What's the purpose there? The double-underbar and underbar-capital rules seem to be allowing for non-conflicting introduction of keywords. Is the single-underbar rule to protect standard library headers or something?

Names beginning with double underbar (or single underbar + capital letter) are reserved. Single underbar + lowercase is not. C23 §6.4.2.1.

You might not have used one, but there have long been parking meters / payment kiosks that take charge cards and even cash. Neither an app nor a human attendant is required. It bugs me that these are slowly being replaced by smartphone app systems.

What (U.S., I assume) state do you live in that has a state law requiring green grass? I've never heard of that.

Why is this? Do labs reinforce the model name during training? I was under the impression that this sort of "self-knowledge" always came from the system prompt, but I guess not...

Refund for the subscription I started after the announcement of Fable.

I have one that still works fine today (except for removing the notorious RIFA film capacitors, but those didn't fail after "a couple years").

https://news.ycombinator.com/item?id=47911524

It has ads already — for its own sports broadcasts on Apple TV+.

Yes, civil legal proceedings (and/or hiring a collections agency) are generally how debts are pursued in the United States.

There seems to be an 8 meg image here: https://github.com/luke8086/gentleos32/releases/download/202...

The root comment of this thread does purport to know when that will happen (well, pending public announcement of those two IPO dates).