Stop being skeptical about AI for development with Charity Majors

**Gergely Orosz** (0:00)
Why are there firmly two camps within software engineering when it comes to AI? Those hating its effects and those who are AI-pilled. Charity Majors emphasizes both camps and thinks they are talking alongside one another.
Charity is a co-founder and CEO of Honeycomb, previously worked at Facebook and Parse, and is one of my favorite voices in engineering. Today, we discuss what it would take for us engineers to ship code we have never read, and why this is more of a when question, not an if question. Why are reliability is quietly getting worse across the industry, and why it will take some time to recover? Career advice in this age of AI. Why middle managers should consider going back to being IC, and why junior engineers will be okay? If you want to hear from someone who was skeptical about AI in 2025, but has changed her mind based on the evidence, this episode is for you. In today's episode, Charity will say, spoiler alert, that the question is not if we will stop reading code written by AI, but when, and we should take lessons from ops and QA on how they prove that software that others wrote works in prod.
And she's got a very good point. As any ops engineer or SRE will tell you, that's how software has always been written by unreliable agents from their point of view, that is software engineers like me, your colleagues or you.
And let's face it, you probably haven't read all the code in your code base either.
This is where I need to mention our presenting sponsor, Antisys. Antisys verifies software written by unreliable agents. It runs your whole system in a hostile simulation and roots out the bugs for you. It does this by using an approach called deterministic simulation testing or DST. Antisys is turbochargers testing by running your whole system under aggressive fault injection. Imagine Antisys is hundreds or thousands of versions of the Mario game running, each instance aggressively trying to break the game with increasingly weird input combinations. If it finds a breakage, this is where the determinism comes in. Instead of you having to try to reproduce a tricky bug you saw in production, Antisys can provide you with a perfect deterministic replay of anything it finds every time. With Antisys, you can specify properties at the whole system level, and Antisys will actively try to disprove them, so you can be confident that if your system holds up in Antisys, it will hold up in production. Head over to antisys.com/pragmatic to learn more. Charity, it's so nice to do this in person.

**Charity Majors** (2:08)
You're in my city. This is amazing.

**Gergely Orosz** (2:11)
So today, I wanted to get up with AI, but before we kick off with AI, I just want to make it clear for people who don't know you that you're not an AI hater or an AI lover, you actually built a lot of cool stuff pre-AI, right? Starting at, we just saw Linden Labs. Was that your first job?

**Charity Majors** (2:27)
My first job at Linden Lab right across the street.

**Gergely Orosz** (2:29)
Right across the street. We were just talking about that. So you were building Second Life?

**Charity Majors** (2:32)
Yeah, we were building Second Life, yeah.

**Gergely Orosz** (2:35)
Then from there on, one of the big hits was Parse, the developer tool which was beloved by developers, the best backend for mobile services. I used to use it.
Then what happened? Facebook bought you.

**Charity Majors** (2:47)
Facebook bought it. Yeah, it was my first great lesson in most acquisitions fail. Most were terrible. This one failed. They shut it down.
Ultimately, I'm very grateful to have had the experience because if it wasn't for that, I've always been a startup kid and so nobody knew my name. It wasn't until I was leaving Facebook that investors were like, would you like some money? That's how we started Honeycomb.

**Gergely Orosz** (3:11)
Then you saw stuff at Facebook, right? It inspired you.

**Charity Majors** (3:15)
Yeah. Facebook, there is a tool called Scuba. We were in a weird position. We were building on AWS, Ruby on Rails, all this stuff. Then we got to use the internal Facebook tools.
Facebook had this tool called Scuba and we were experiencing hockey stick growth. It was just like we had over a million mobile apps hosted on Parse by the time I left.

**Gergely Orosz** (3:36)
Yeah.

**Charity Majors** (3:38)
Every single week, a new one would break. It would hit the top 10 on iTunes or something out of nowhere. I would just be like, oh, wait.
These apps need a little bit of a stack. It went from, it would take hours or weeks. We have to get lucky. We finally find, because it might be one app that's spamming the logs, but that might not be the reason. They might all be backed up behind the reason. We started getting our datasets into Scuba and finding them. It just went from being a really hard engineering problem with a little bit of luck to just being like a port problem, click, click, click. There it is. It was just mind blown.