Gwern: evaluation

Transcript context

…Was there a time that you felt you had updated too far? Yeah, there were a few times I thought I had overshot. I thought people over-updated on AlphaGo. They went too far on AI hype with AlphaGo. Afterwards, when pushes into big reinforcement learning efforts kind of all fizzled out—like post-Dota, as the reinforcement learning wasn't working out for solving those hard problems outside of the simulated game universes—then I started thinking, “Okay, maybe we kinda overshot there…” But then GPT came out of nowhere and basically erased all that. It was like, "Oh, shit. Here's how RL is going to work. It's going to be the cherry on the cake. We're just going to focus on the cake for a while.” Now we have actually figured out a good recipe for baking a cake, which was not true before. Before, it seemed like you were going to have to brute-force it end-to-end from the rewards. But now you can do the LeCun thing, of learning fast on generative models and then just doing a little bit of RL on top to make it do something specific. Now that you know that AGI is a thing that's coming, what’s your thinking around how you see your role in this timeline? How are you thinking about how to spend these next few years?…

Stored transcript either side of the excerpt. The highlighted words are the published quote; the surrounding text is unedited source, never generated.