AGI 1-1: we're almost there! Baby versions of concept formation and
internal dialogue are here. Common sense distillation is pretty mature.
What's next? "GOAL Formation". Agents like OpenAI Operator, aka "Embodied
Intelligence", are one leap away from REAL autonomy.

Real meaning: coming up with goals and ideas of their own, based on the
read of sensory data, then engaging actuators.

Prompt-free driving. Autopilot!

https://x.com/KeyvanMSadeghi/status/1884314139117621465

Riding the hype with the lingo there, positive reinforcement play. My
non-marketing take is, the path along LLMs, as they stand, leads to
proto-AGI. Think dumb robot tin cans. But those can be iterated upon to
arrive at above average human intelligence, and beyond.

Key point is, I think, and I may be wrong, that goal formation compatible
with current multimodal models, is gonna give us the first proto-AGI.
Autopilot, not copilot.

"Compatible" is the keyword. Our field is filled with genius egotistical
dumbasses. My way or highway type. That won't fly, AGI is a consensus
problem.

Peace ❤️

------------------------------------------
Artificial General Intelligence List: AGI
Permalink: 
https://agi.topicbox.com/groups/agi/T8db887b0af928f03-M1c4b5e51a4fb8411e3976f49
Delivery options: https://agi.topicbox.com/groups/agi/subscription

Reply via email to