4 ms·
You mean like OpenAI, whose LLM tech is heavily based on original research done quite a while ago by Jürgen Schmidhuber in Germany? Does that now make OpenAI an
by Slartie 3y ago
You mean like OpenAI, whose LLM tech is heavily based on original research done quite a while ago by Jürgen Schmidhuber in Germany? Does that now make OpenAI and the LLM hype a paragon of European innovation, and its success "accidential"?
- hcfman 3y agoYeah, with that kind of argument nearly all successes could be called accidental.
- atemerev 3y agoBecause the origin of ideas is important. But yes, we have some original ideas in European academia. Our startup scene however is absolutely laughable (the only two countries that at least trying to do something are France and Estonia, the rest don’t even understand the problem). The “innovation grants” of 30 thousand euros (what anyone is supposed to do with these?), endless copycat companies working in “sustainability” or other European buzzword du jour which will never be scaled even in principle… etc etc.
- atemerev 3y agoIn Switzerland. And yes, deep learning scaling (by solving the vanishing gradient problem) was discovered by him, but this was a previous era of AI. LLMs required a lot of new ideas on top of that.