[dead]
Grok3 Launch [video]
111–120 of 1001 posts
Re: Grok3 Launch [video]
#112[flagged]
Re: Grok3 Launch [video]
#113Companies have hijacked the open source concept to mean downloadable blob and we follow them as I see in the comments. It’s a real shame.
it's closed source and open outcome
Re: Grok3 Launch [video]
#114Earlier quoted context omitted.
[flagged]
Doesn’t OpenAI claim their work will lead to artificial general intelligence? That seems like a much steeper gradient to climb. Hype is the fuel that bootstraps fortunes in techtopia.
Yes, but we have a long track record of one overpromising and underdelivering, charging money for what doesn't yet exist; while the other released stuff on a minimal website with an associated blog post for fanfare, and actually (metaphorically) turned the world upsidedown, and only charges for what they actually deliver. Yes there's hype now, but that's how it began.
So I think people are less distrustful of Altman when he says "thousands of days".
Re: Grok3 Launch [video]
#115Elon just said they are launching an AI game studio. Does this mean they will be building games that are mostly built with AI, or will they make AI tooling available for anyone to build games easily? Probably the former, but it would be nice if they would make it fully available to everyone.
They will be building games played for you, like the fascist's Path of Exile account but without exploiting humans for the task.
Re: Grok3 Launch [video]
#116Companies have hijacked the open source concept to mean downloadable blob and we follow them as I see in the comments. It’s a real shame.
this is correct. "open source" means everything required to recreate from scratch and improve. not "here's a massive binary, an interpreter script and permission."
How can you even "open source" an AI model without all of the, presumably copyrighted and extremely voluminous, training data?
Re: Grok3 Launch [video]
#117[flagged]
Re: Grok3 Launch [video]
#118Earlier quoted context omitted.
Regardless of which it is, we can assume it'll be here after Tesla's full autopilot that he promised as well as the Mars colonies.
Full self driving exists, as well as starship. So I take it you mean “imminently”. Like him or loathe him, he executes , which is more than can be said for most.
"Full Self Driving" does not exist, only "Full Self-Driving (Supervised)"
Re: Grok3 Launch [video]
#119If what they say is true, then you have to give them credit for catching up incredibly fast. And slightly pulling ahead. Not only with the models, but also products.
To me, it seemed like they spent there money to get there. They talked about the massive datacenter they built, but will it pay off is the question.
They may not need direct subscription revenue to recoup their investment.
A variant of multi-modal LLMs may be the solution to self-driving cars, home robotics, and more.
I keep saying that to be a really effective driver, an AI model will need a theory of mind, which the larger LLMs appear to have. Similarly, any such model will need to be able to do OCR and read arbitrary street signs, and understand what the sign meant. Most modern LLMs can already do this.
Re: Grok3 Launch [video]
#120We know RLHF and alignment degrades model quality. could it be that Grok, due to its less restrictive training guidelines (and the fact that its creators aren't afraid of getting sued), can achieve higher performance partly due to this simple factor?
> We know RLHF and alignment degrades model quality. I feel you can't make statements like this without giving some sources. IIUC, without RLHF/alignment, the model won't even be able to chat with you, it would just be a document completion engine.
You're both right because RLHF and fine-tuning are just techniques.
It's dependent on the training data and not as much the method.
So, if you make the RLHF/finetune data such that it avoids certain topics, then you reduce model quality in practice since your training data might accidentally cast a net wide enough that you make it avoid certain legitimate questions.
On benchmarks these things don't typically show up though.
But yes. Those techniques are required for making it chat. Otherwise it just autocompletes from the internet. It is also used in a couple of other places (reasoning/search(hallucination mitigation))