hooloovoo_zoo

In my experience students will just keep using the old name (sometimes even the name before the old name) in perpetuity. With regard to the new campus, satellite campuses never carry the prestige of the mothership, for schools or companies.

I feel the fact that these models always modify the body design of a pelican to fit the bike rather than the other way around represents a fundamental issue with AI.

I’d be much more interested in particulate pollution limits for tires.

I suspect LotR is a singularly unrepresentative choice here considering how much info exists about it.

Anything morality-related is going to be subjective, but I think there are just too many practical problems with LLMs as coauthors. For me, in a scientific context 'chatgpt' is too vague, and I think it would be logically inconsistent to have LLMs as coauthors and not other forms of Monte Carlo. I also think LLMs are just too mechanical to be ascribed 'people words' (in the same way I don't consider my automated coffee machine a barista).

Tool-assisted proofs typically already get a description of the tool usage. Even if one wanted to start giving authorship credit to llms, they’re too non-atomic (did it have web search, which mcp?, etc.)

Doesn’t seem like much value-add beyond pointing Claude code at org mode.

Except the Waymo can do 150 mph bumper to bumper with other Waymos if you let them.

Mm, the code is constrained to run inside a game 'tick'?

They also have to replace 20%+ of their satellite network every year.

They are training the model to 1. Produce code (as opposed to answer a question, write a poem, etc.) 2. Produce long enough output to be a valid solution. So they are doing exactly what I said. Cheers.

One sentence summary: We fine-tuned a general-purpose model to produce valid benchmark code results and it got better at producing benchmark code results; we didn't bother to evaluate it on anything the model used to be good at.

Proposition: Every power user feature added lowers Anthropic’s market cap $1B and OpenAI’s $10B.

If you ask the agent a product question you can trigger it into web search mode where it provides web links and autocomplete suggestions. You can follow the a/c suggestions down some pretty elaborate chains of questions but eventually it hits a rate limit and reverts to the default dumb bot.

Well, they seem to be using 3rd party LLMs for websearch.