851–860 of 1001 posts
Earlier quoted context omitted.
This is grok 3, so not a debut
It is a debut of their thinking mode iirc. Unfortunately LLMs are shifting compute time to test time instead of train time. I don't really like this and frankly it shows a stalling of the architectures, data sets, etc...
To put it this way: after seeing examples of how a LLM with similar capabilities to state-of-the-art ones can be built with 20 times less money, we now have proof that the same can be done with 20 times more money as well!
Imagine what they'll achieve if they'll apply deepseek methods here with this insane compute
The politics in the comments here are really toxic. What’s happening to HN? This is the largest computer cluster the world has ever seen. Can someone please post interesting comments about things I can learn?
Bottom line: a technology that has the ability to shape human thought perhaps more than any other in history is owned by a man with some truly vile ideas. (Remember, his primary stated goal is eliminating the “woke mind virus,” i.e. reshaping global politics and culture in the image of the far-right.) We can make happy marketing noises all we like, but at the end of the day, that’s the thing that’s actually going to have a meaningful impact on the world. Once his audience is captured, the model will say what Musk needs it to say and people will believe it.
If we can’t discuss the potentially catastrophic consequences of new technology, then none of us deserve to call ourselves “engineers.” We are just docile consumers latched onto Silicon Valley’s teat.
Grok has gotten to the top of one benchmark: https://x.com/lmarena_ai/status/1891706264800936307 It's been said before but it is great news for consumers that there's so much competition in the LLM space. If it's hard for any one player to get daylight between them & the 2nd best alternative, hopefully that means one monopolistic firm isn't going to be sucking up all the value created by these things
It's not good news when this competition comes at cost of a gigantic over inflated bubble, in which all the big players keep on sucking billions from investors without even having a business model. This hype will burst sooner than later and will trigger yet another global recession. This is untenable.
Earlier quoted context omitted.
Elon's strength is a massive wallet.
I guess you didn't watch the videon in OP, because if you had, you'd know that they tried to buy the buildout and got quotes for 12-18 months, then decided to do it themselves instead.
Earlier quoted context omitted.
You'd still need a fairly large amount of compute power to be able to run DeepSeek R1 locally, no?
Well yes, but not so large that it's completely prohibitive. People have been running the full models on computers going as low as $6000: https://x.com/carrigmat/status/1884244369907278106 Of course this is for a personal instance, you'd need a much more expensive setup to handle concurrent users. And that's to run it, not train it.
This takes advantage of the sparsity of MOE and the efficient KV-cache of MLA.
https://garymarcus.substack.com/p/elon-musks-terrifying-visi... I'm not sure if this was a very bad joke by Elon, or if Grok 3 is really biased like that.
Karpathy, which is IMHO a serious and balanced person, lamented that it looks too censored (see recent tweets). Elon Musk is (for me) a very scary person, and it is important to evaluate AI safety (but I believe that the safety that matters in AI is of a different kind), yet to listen to Gary Marcus does not make any sense: it's just an extremely biased person that is riding the anti AI wave.