Dwarkesh Patel says we haven't yet achieved AGI for one simple reason: the models can't learn on the job. They can rewrite a transcript or suggest a clip to tweet out, but they won't remember what you liked or why it worked. Without learning loops, they can't evolve from... tool to employee. “Why can't we just delegate all language work?”show more

Dwarkesh Patel on an all-too-real moment (for me) on TBPN.

Scan any documents, convert images into text, PDF files, etc. 👍

for now™ ... I think agentic AI may make this a short lived truth by the end of the year.

I will let you know when I can hire an agent to do what I do!

Has he missed the whole self improving paradigm that is trending

this is the part that was built into ADIN from the start. if LPs vote, pass, or leave comments on funding a startup, that data loops back in. over time the agents start to pick up patterns: what gets funded, what gets ignored, what people hesitate on.

we need the Pensieve 🐝

We need Continual self improving model. (A model that can edit its own weights after each response / action )

Very true!

I really would have expected someone to release a dumb, experimental model that can learn on the job

Hmmm How to say this? Please advise… … .. ….. … “Language or Job… you think this is LLM or do your read what am laying down?” Matters not what we think… job is Job to that language that never leaves