“…millions of interacting objects with a complex physics engine… the world becomes much more complex because it includes the complexity of all of those other agents”

Training Data · David Silver and Lasse Espeholt of Ineffable Intelligence: A Superlearner for the Age of Experience

Play from 21:26
21:26

Audio streams from the publisher’s own feed. Verbatim does not host or copy it.

What was described

built crucible environments using a complex physics engine with millions of interacting objects to generate emergent complexity for agents

Said nearby on the same subject

“We call those crucibles… environments in which it's possible for agents to make incredible discoveries…”
♫
Said on Training Data
How they work Share this moment
“We have a phrase we call the Moore's Law of RL environments, which basically means, like, every six to eight months, we've noticed that the complexity of the environment… roughly doubles.”
♫
Said on TBPN
Hard number doubling every 6-8 months Share this moment
“And you even reach out in more complex things to build what's called a net promoter score.”
♫
Said on The GovCon Room
Why it worked Share this moment
“A few years ago would have cost, you know, a million dollars plus ... made it possible for a tiny team like ours to generate environments”
♫
Said on TBPN
Data play $1 million+ cost previously Share this moment

Moments like this, every Monday

The strongest new moments on your topics, in the speakers' own words. Free for three topics.

No card. No password. One-click unsubscribe.

Track a competitor?

Pro watches up to 25 names or phrases and emails you the morning after they are said.

See Pro, $29 a month

Verbatim quotes a short extract and links to the episode; the episode belongs to its publisher. Remove a show.