@walter4096@ChombaBupe The multi-generational airball of underestimating capabilities and timelines with no acknowledgement of the failure is disturbing.
I don’t disagree regarding intelligence, but the endless goal post moving is real.
Will one of them ever contribute something material?
GaryAI?
@ChombaBupe@walter4096 Looking forward to seeing your work.
No disagreement regarding intelligence.
Still my question is unanswered.
What are non-intelligences capable of?
With the inertia of infra, NN’s will now likely reach many capabilities first - if not ideally.
What can they do in 2028?
@ChombaBupe@walter4096 In five years, when we are still far from AGI, what curve-fit capabilities can we expect to see that are not already in their nascent stages?
Asking Gary M yields no reply.
Clearly many have been surprised so far.
Not you?
What will not surprise you in five years?
@jeffzwang Working with 15 years of hours-a-day personal dictation. Currently running simple batches on Zephry 7b and Llama 2 70b.
Looking to move towards RAG and agent driven approaches.
I’d love to to join.
@bengoertzel — DREAMERS, seeing the pattern of the WWW repeating. A brief imaginative age of invention followed by successive waves of opportunists whose curiosity is inversely proportional to the time it takes them to get involved. Eventual equilibrium in AGI skin care routines.
@filippie509 Similarly, if human intelligence is truly flexible, why are tens of thousands of human experts throwing themselves at the same ineffective approaches?
How much data do they need before they try something else?