I think most ML experts (that weren't being paid out the wazoo for saying otherwise) have been saying we're on the tail end of the LLM technology sigma curve.
-
I think most ML experts (that weren't being paid out the wazoo for saying otherwise) have been saying we're on the tail end of the LLM technology sigma curve. (Basically treating an LLM as a stochastic index, the actual measure of training algorithm quality is query accuracy per training datum)
Even with deepseek's methodology, you see smaller and smaller returns on training input.
-
System shared this topic on
-
I think most ML experts (that weren't being paid out the wazoo for saying otherwise) have been saying we're on the tail end of the LLM technology sigma curve. (Basically treating an LLM as a stochastic index, the actual measure of training algorithm quality is query accuracy per training datum)
Even with deepseek's methodology, you see smaller and smaller returns on training input.
At this point, it is useful for doing some specific things so the way to make it great is making it cheap and accessible. Being able to run it locally would be way more useful.
-
At this point, it is useful for doing some specific things so the way to make it great is making it cheap and accessible. Being able to run it locally would be way more useful.
Yeah it is useful, but it is not an industry worth trillion of dollars in valuation. The only use cases LLMs have is to make shitty summarizations of text, use it as shitty google search alternative or to write shitty code
-
At this point, it is useful for doing some specific things so the way to make it great is making it cheap and accessible. Being able to run it locally would be way more useful.
100% this. Wouldn't it be something if they weren't overtly running their companies to replace all of us? If feel like focusing instead on creating great personal assistants that make our lives easier in various ways would get a lot of support from the public.
And don't get me wrong, these LLMs are great at helping people already but that's definitely not the obvious end goal of OpenAI or any of the others.
-
At this point, it is useful for doing some specific things so the way to make it great is making it cheap and accessible. Being able to run it locally would be way more useful.
Sure, but then what would they do with their billions of dollars data center plugged into a nuclear power plant?
-
Sure, but then what would they do with their billions of dollars data center plugged into a nuclear power plant?
Can we skip the dog and pony show, and get straight to paying directly to crush the orphans?
-
Yeah it is useful, but it is not an industry worth trillion of dollars in valuation. The only use cases LLMs have is to make shitty summarizations of text, use it as shitty google search alternative or to write shitty code
The really useful part is just natural language recognition, which is much better than previous things. Using it in games or software would be a big improvement.
-
System shared this topic on
