But you could train a large model to "understand" all the nanoassociations and nanopatterns among all the world's words, sights (images), sounds, smells, etc. That would be amazing. And that kind of ambitious project is now in reach.
You could train a large language model on every word that's been said since the dawn of time, and it still wouldn't know what a two-year-old does.
-
-
-
Nothing wrong with that. But it won't solve AI.
- Näytä vastaukset
Uusi keskustelu -
-
-
What about a neural net language (and world) model in a robot body with sensory inputs?
-
Yes, in principle, but we don't have the algorithms that could learn it yet.
- Näytä vastaukset
Uusi keskustelu -
-
-
Can a two-year-old do this?https://twitter.com/andrewwhite01/status/1432720845638672390?s=19 …
-
A two-year-old can't do addition, either.
Keskustelun loppu
Uusi keskustelu -
-
-
When we can explain how fish appear in water from nothing, we should be able to code that in a computer for reasoning. It is evolution written all over…
Kiitos. Käytämme tätä aikajanasi parantamiseen. KumoaKumoa
-
-
-
2 year old is a pretrained model which just needs residual training. !!
Kiitos. Käytämme tätä aikajanasi parantamiseen. KumoaKumoa
-
Lataaminen näyttää kestävän hetken.
Twitter saattaa olla ruuhkautunut tai ongelma on muuten hetkellinen. Yritä uudelleen tai käy Twitterin tilasivulla saadaksesi lisätietoja.