6 Comments

"En el lado negativo, esto significa que no tenemos idea de lo que el modelo ha aprendido"

Frase que me deja cuando menos pensativo. Parece que la realidad vuelve a superar a la ficción

Expand full comment

Es muy interesante lo que está pasando con el aprendizaje de estos modelos. La práctica va por delante de la teoría: no podemos explicarlo!

Expand full comment

Very interesting topic (again) and also, a very current one! It's kind of scary to think that this whole process is happening with humans knowing so little about it and not having too much control over it either 😥😥

Expand full comment

Yes, AI safety is a big concern nowadays. There is a whole field called AI alignment that deals with creating models whose objectives align with those of humanity. Pretty scary to even have to do this!

https://en.wikipedia.org/wiki/AI_alignment

Expand full comment

HAL?

Expand full comment

Me falta cultura 😂

Expand full comment