@metahacker_@KarthikTabjul selling tokens is not a viable business model and their valuations are built on promises that are extremely hard to fulfill. 2027 will be in interesting year.
@TopG_NC93 yup. they need to find a better way to monetize. also important to note three very smart people are looking to agree to let china win the AI race, unless they know china will hit the same wall and can’t beat them. Aha!
@sparshselim@AnthropicAI@OpenAI i’d argue everything that you have mentioned needs time and a lot of effort. They have a lot more compute coming up and as he says, in plans to kick off a massive RL run. We should see meaningful advances in the coming years, if they’re not stopped that is.
@i_am_chandra_@greatbong task performance ≠ general intelligence. Human brain is a lot more than that. LLMs are just not there yet, from my daily experience with them. They’re incredible tools, heck i use them everyday. But thats everything which has lesser intelligence than us, even a cow.
@i_am_chandra_@greatbong thats just wrong. Specialization is enabled by the fact that i have general intelligence. The ability to learn, almost anything, and then even specialize in something, that is what makes my intelligence general. wake me up when an llm does that.
@greatbong Not to mention how efficient our brain is! People calling doomsday by looking at current AI systems sitting in large data centers have no idea how far away from actual intelligent systems we really are. this is the best time to enter this field.
@greatbong Even if we look at nature, anything which has lesser general intelligence than us, we have used such form of intelligence as tools. And that will be the fate of LLMs. We will ofcourse accelerate a lot of R&D, and may come up with better architechtires, but it has not happened yet
@REDBOXINDIA privatisation is a good step. government has no role managing airports of the whole country. although every airport being managed by adani is not the right spirit by any means