GPTs are fundamentally the same since release. Only dataset sizes increased. They are still suffering from the same problems(eg hallucinations) as their first iterations, and these problems are the direct result of their inherent design and cannot be solved. GPTs will not lead to AGI and are a dead end.

>>105756194 (OP)
>GPTs will not lead to AGI and are a dead end.
wow youre only 2 years late
how retarded

>expecting one token at a time text autocomplete to be smart
It's gotten farther than I ever would have expected but it's not AI and there is no easy solution to it spitting out wrong information because of the way LLMs work probabilistically instead of logically.

>>105756327
Then explain why all the youtubers who are smarter than you say its the future????

>>105756194 (OP)
>GPTs are fundamentally the same since release. Only dataset sizes increased.
There have been improvements. Newer models are capable of working with larger contexts that would've made old models shit themselves, improved system prompts have helped to reduce hallucinations and get better answers, and reasoning training has helped models to work through complex problems that would have been impossible for older models.
But all of those are improvements at the edges. They don't fix the fundamental flaws. It's like upgrading your PC by putting in better fans. Sure it will probably improve performance but not by much, you'll still be limited by the speed of the core parts, like the CPU, GPU, and RAM.
>GPTs will not lead to AGI and are a dead end.
Yep.

>>105756704
Youtubers will say anything for views (money).