“Given this relatively small amount of people so far means that we have a lot of compute per person… you can do a lot of experiments just like one person.”
Training Data · David Silver and Lasse Espeholt of Ineffable Intelligence: A Superlearner for the Age of Experience
Play from 1:02:24
1:02:24
Audio streams from the publisher’s own feed. Verbatim does not host or copy it.
What was described
Maintain a very small team (≈20 hires) to keep a high compute-per-person ratio, allowing one person to run many experiments
20 people hired total
Said nearby on the same subject
“Some of our early AI work has been on basically producing systems to do this. And this allows us to move much more quickly, too, because if you're running so many experiments, you really quickly become bottlenecked on your ability to understand what was produced”
“I started establishing a newsletter, building a database. I guess you had a small database. I did to start. Very small. Very small. Yeah. But I guess like you say, when it's word of mouth, even if you've got 30 people on that database.”
“You’re going to need to hire an on-stat internal developers … which gets very expensive, can handle the cloud expenses … you’ll run into a lot of headaches”
“What I found to be very helpful… really defining what you value… aligning personal values with the company values… keeps you focused on how you're going to stay motivated.”