Evidence receipt / evaluation
Published · transcript-backedLenny Rachitsky: evaluation
14 Dec 2025 Lenny's Podcast Why humans are AI’s biggest bottleneck (and what’s coming in 2026) | Alexander Embiricos (OpenAI Codex Product Lead)
“Probably looking at it and just making sure this is the right thing to do, yeah.”
Source trail
Everything needed to verify it.
- Speaker
- Lenny Rachitsky
- Attribution
- Verified speaker
- Claim type
- evaluation
- Recorded
- 14 Dec 2025
- Publisher
- Lenny's Podcast
Transcript context
…Yeah, that's super interesting. And I bet you, so if we went and asked them what the bottleneck to that productivity is, did they share what it is? Probably looking at it and just making sure this is the right thing to do, yeah. Yeah. So we see this now. We have a Slack integration for Codex. People love if there's something that you need to do quickly, people will just @ mention Codex, "Why do you think this bug is happening?" It doesn't have to be an engineer. Even maybe data scientists often here are using Codex a ton to just answer questions like, "Why do you think this metric moved? What happened?" So questions, you get the answer right back in Slack. It's amazing, super useful. But as for when it's writing code, then you have to go back and look at the code. So the real, I think, bottleneck right now is validating that the code worked and writing code review. So in my mind, if we wanted to get to something like the friend you were talking about's world, I think we really need to figure out how to get people to configure their coding agents to be much more autonomous on those later stages of the work.…
Stored transcript either side of the excerpt. The highlighted words are the published quote; the surrounding text is unedited source, never generated.