News for hackers. Stuff that matters.
Re:DeepSeek gives Europe's tech firms a chance to catch up by rightbyte on Monday February 03, 2025 @05:16AM (#42916761)
You are hard to impress. Running a 1/20x sized version locally that would be sci fi level 10 years ago.

For such small models I would recommend specialized models only. Like Deep Seek Coder. But I think that one is lagging behind the state of art now.

Re:DeepSeek gives Europe's tech firms a chance to catch up by harry8 on Monday February 03, 2025 @05:56AM (#42916987)
Like an 8 year old factoring large numbers it's not amazing how well it is done it's that it is done at all that amazes. Sure. Amazing but not at all useful and not something one would expect the kind of fuss we've seen.

Seems the explanation is the deepseek-r1 models I was using are not, in fact, deepseek-r1. Thanks all for the heads up.

Re:DeepSeek gives Europe's tech firms a chance to catch up by rightbyte on Monday February 03, 2025 @06:20AM (#42917127)
Ye, to your defence it seems like the Ollama project did it confusing on purpose to mess with Deep Seek.