panagathon

> but philosophically could it create a problem it couldn't solve? And what would be the value in writing that out?

Of course. It could hallucinate a nonsense, call it a problem, and fail to solve it.

The question is whether it could create an interesting problem, and be unable to solve it.

If the answer is yes, and yet we can solve it, it would have immense potential consequences for the ontology of mind and machine.

I'm quite excited to see how this intersection of the limited and the unlimited plays out.

For large language models, isn't everything nails?

Yeah, it's awful. I can't believe how much of a downgrade it is from the old Perseus Hopper.

Scientists and engineers with an interest in such things would have an easier time working on it, if the broader economic and civic context they work in wasn't being messed with by demagogues.

They shouldn't be drafted to resolve the rise of petty tyrants. It's a waste of their time.

> While this sounds generous (and in some ways it is), it does not address the general point that GP is making. That is, the systematic disadvantage which large parts of humanity have w.r.t. to access to the tools. You could say they can't drive a Lambhorgini either, but that also doesn't solve the problem.

This was also the case historically, when being at certain universities, with better professors, better scope of works available at the library, etc, would necessarily provide systematic advantage.

This is the reality of progress. It is always unevely distrubuted.

I do think the open source side of model development is a substantial counter to the pessimism here.

Oh yeah, that's clever

> I don't get why I would use Claude Code when OpenCode, Cursor, Zed, etc. all exist, are "free" and work with virtually any llm. Seems like a weird use case unless I'm missing something.

I'm with you on this. I've tried Gemma and Claude code and it's not good. Forgets it can use bash!

However, Gemma running locally with Pi as the harness is a beast.

It just comes off a bit victim-blamey, I think. If it's not intended, you should certainly be aware that it can.

You could do with building this isn't the system itself. Use an LLM to assess how well the user engaged, whether they got frustrated, etc. Give them a short (generated?) test on the discussed content to gain data regarding their mastery.

This looks unreal!

That's only potentially true for the Dionysian. Others were open to both. See 76e in Plato's Meno, for reference.

You can read Brian Muraresku's The Immortality Key for a detailed exploration of this topic.

It certainly isn't beyond criticism, but it's points are substantive and well referenced, giving the reader enough scope to tackle the controversial points themselves, not just take the authors presentation on face value.

Perhaps in industries you are familiar with. Many industries, however, would have a continued professional development disposition that benefits from continued access to fundamental research. Mine certainly does and the lack of access impoverishes that professional development.

This does sound like a lot of fun. Since this AI only reaches silver level, presumably such a contest could be quite competitive, with it not yet being clear cut whether a human or a tool-assisted human would come out on top, for any particular problem.