[larrysalibra]

I tried Deepseek R1 via Kagi assistant and it was much better than claude or gpt.

I asked for suggestions for rust libraries for a certain task and the suggestions from Deepseek were better.

Results here: https://x.com/larrysalibra/status/1883016984021090796

[progbits]

This is really poor test though, of course the most recently trained model knows the newest libraries or knows that a library was renamed.

Not disputing it's best at reasoning but you need a different test for that.

[gregoriol]

"recently trained" can't be an argument: those tools have to work with "current" data, otherwise they are useless.

[tomrod]

That's a different part of the implementation details. If you were to break the system into mocroservices, the model is a binary blob with a mocroservices wrapper and accessing web search is another microservice entirely. You really don't want the entire web to be constantly compressed and re-released as a new model iteration, it's super inefficient.

[nailer]

Technically you’re correct, but from a product point of view one should be able to get answers beyond the cut-off date. The current product fails to realise that some queries like “who is the current president of the USA” are time based and may need a search rather than an excuse.

[kemiller]

This only holds water if they are able to retrain frequently, which they haven't demonstrated yet. But if they are as efficient as they seem, then maybe.

[bobheadmaker]

That's interesting!