How leaders win over their team’s biggest AI skeptics | Superhuman’s Loic Houssier

**Andrew Zigler** (0:06)
Welcome to Dev Interrupted. I'm your host, Andrew Zigler.

**Ben Lloyd Pearson** (0:10)
And I'm your host, Ben Lloyd Pearson. This week, I'm sitting down with superhuman CTO, Loic Houssier, who joins the pod to discuss how he overcame skepticism towards AI by empowering his most cynical developers instead of issuing top-down mandates. He'll detail his team's biggest and unexpected productivity gains along the way, which were achieved by a balanced mix of high-trust qualitative feedback and quantitative adoption data. But before we dive into this interview, let's cover this week's news. We have an accusation of brain drain at Amazon, Clod skills are awesome, how to say no and a field guide to AI Slop. Andrew, where are we starting this week?

**Andrew Zigler** (0:49)
Well, you know I love a little bit of gossip. What's the accusation? Oh, you must be talking about the article you and I read in the register by Corey Quinn about the Amazon brain drain phenomenon and how that might be at play behind that AWS outage just this past week. It was quite interesting. It gave us a little bit of context on the outage and where it sits for AWS and its history of outages and what it means for its downstream customers. But it was also interesting to compare this with, in the Pragmatic Engineer, there was a deep dive on why this failure happened at AWS, ultimately caused by DynamoDB's DNS. And so once again, it's always DNS that seems to knock down the world, whether it's on AWS's systems or Cloudflare's. But for me, this took down my beloved Asana. Ben, what did it take down for you?

**Ben Lloyd Pearson** (1:39)
Yeah, well, for me, it was really just all the things that our podcast depends on that went down. I was like, thankfully, I still had most of my AI tools, I guess. But I know our producer Adam was panicking a little bit because some of the tools we used weren't available to get our episode ready for last weeks.

**Andrew Zigler** (1:58)
Yeah, nothing really worked.

**Ben Lloyd Pearson** (1:59)
That was a nice last-minute fire drill.

**Andrew Zigler** (2:02)
Yeah, I know. I think everyone felt a little bit of the pressure downstream from the outage, and it really kind of shines a light on how much of the world and the internet that we use every day runs on AWS systems.

**Ben Lloyd Pearson** (2:14)
I love that you mentioned to Sonic because they also had the saddest and most brilliant error message, I think, of any tool that we used calling this some difficulties. I was like, yes, this is somewhat difficult, I think, at least. But, you know, I mean, first of all, I think, you know, the thing to really take note of is that an AWS outage makes national news. Like, that is actually generally, like, a positive thing for AWS when you think about it, because they are a company that's known for being pretty reliable. I mean, a lot of the internet runs on them at this point. So, like, the fact that this is a huge deal is actually something that is kind of a testament to what it's become. I kind of went down the rabbit hole a little bit, and I read about some former AWS people, including one, Justin Garrison, who wrote an article back in either late 23 or 24, where he made this claim that he suspected there would be a major AWS outage. He called out the year 2024, but then he said, no amount of multi-region redundancy will protect you. And that was actually the case for this outage, which is really interesting. I've read a lot of articles that sort of summarize what happened here, and from my understanding is that there was this outage in US East 1, but there's a lot of this critical infrastructure for all of the AWS regions that only exist within US East 1 And when that went down, like DynamoDB, sometimes the impact of outages can actually resonate out to people who either don't even host in US East 1 or have multi-tenant redundancy or multi-region redundancy. So there's a lot of opinionated takes out there. Like what happened in like, holy cow. So, I mean, fortunately, we all survived, it seems like. And we got through our period of some difficulties.
And now I think life is back to normal. So hopefully, there's not another one for a while.

**Andrew Zigler** (4:08)
Yeah, no more outages for a little while, please. What about these Claude skills? You know, what came across our desk on these?

**Ben Lloyd Pearson** (4:14)
Yeah, so we found this article from Simon Willison titled Claude skills are awesome and maybe even a bigger deal than MCP.