omegastick

And his comments about ANNs in general were wrong. ANNs were never "fancy search engines". You can train them to be a fancy search engine if you want, but you can also train them to do many other things.

Shock horror: different subfields have different terminology.

I'm a big fan of that quote, but always took it differently than the meaning associated with this article.

Good artists see an idea and use it. Great artists see an idea and _make it their own_.

In the UK they are not. In fact they are significantly broader. Anything which hosts "user-generated content" visible to other users.

Because if you have an explicit world model you can optimize against it.

It sounds like you are imagining tacking a world model onto an LLM. That's one approach but not what LeCun advocates for.

This has not been my experience. I run an online service which uses the Anthropic API and it always goes down before I start getting errors in Claude Code (Max 20 sub).