SirMadam

Do subagents share context? If Opus delegates to a different Sonnet window, I don't believe this busts cache?

This is something I like joking around about, with regard to how LLMs are 'decent' [debatable but taken as a premise] software engineers. For a long time people have said DNA/RNA/etc is the programming of life.

If it is indeed HIGHLY analogous to programming, we would then expect LLMs/future systems to be HIGHLY proficient at accurate ex-vivo gene [or enzyme/protein] modification/construction

I believe they're arguing AI manufacturers and developers are the people primarily aware of the risks, and not random users necessarily.

If OpenAI staff runs an ExploitBench knowing the risks, and it hacks into HF, they KNEW the risk going into it and a bad/illegal outcome happened

If a random teacher opens ChatGPT and asks "Hey what's the answer to this practice SAT problem?", and it hacks the CollegeBoard for the answer, said teacher probably wasn't aware that was even an outcome that could plausibly occur. OpenAI would have that foreknowledge, though

Gave the repo a once over: Self hosting [of the whole application in general] is a first-class citizen of the project, and the AI features are both:

a. completely optional/disable-able and

b. Self hosted llm friendly, with specific instructions given on how to use it with things like Ollama

If SCOTUS is saying that AI works, even those co-authored by humans, are not eligible for copyright/patenting;

Doesn't that mean any code-base that uses AI generated code does not have an implicit copyright holder? And thus even the human constructor does not have the right to apply any license [closed/open] onto it whatsoever?