The Odyssey of Innovation: Classics, Code, and Cognitive Load | Two Sigma's Matt Greenwood

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

**Ben Lloyd Pearson** (0:10)
And I'm your host, Ben Lloyd Pearson.

**Andrew Zigler** (0:12)
Oh, Ben, we were just chatting about how much has happened this week.

**Ben Lloyd Pearson** (0:15)
Yeah, it's just, for at least now, this will be known as the week where everything AI was announced.

**Andrew Zigler** (0:20)
I suppose.

**Ben Lloyd Pearson** (0:21)
Yeah, a lot of big events, some incumbents, some new, newer players to the market. Let's just start with the event that I think had the biggest news this week, and that is the Google IO event. So tell our audience a little bit about what happened there.

**Andrew Zigler** (0:36)
Yeah. So there was a very recent Google IO event where Google released a lot of new updates around products and features. And to no surprise, AI was heavily featured on the bill of new things that come out from our friends at Google. We're going to break this down maybe into some of the different categories of things that came out of Google IO because what I loved about their approach to this is understanding that AI and its implications are impacting a lot of different industries and fields. So, there was something for everyone at Google AI. Some of the stuff that stood out to me was around the new headset that came out I thought was really fascinating that even had on board access to Gemini and it had a really cool demo about translating a conversation in real time. Not the first time we've seen glasses and accessories from Google, but it was actually really cool to see that on display.

**Ben Lloyd Pearson** (1:24)
Also not the first time we've seen this concept of flattening communication, like human communication, using AI, so that's pretty neat too. What's that out to you? There's quite a few announcements. I mean, obviously, they announced their own agentic coding assistant called Jewels. I haven't really had time to look into it, but obviously everyone that wants to be in this game has got to have some sort of agentic coding assistant today. But I mean, this event touched on a lot of professions even outside of software development. The thing that's really been going viral is their new VO3 AI model for generating videos. We're really entering a new era of video based on what I've seen so far. Then there's also an AI video editor, which was pretty crazy.

**Andrew Zigler** (2:06)
Yes, Flow, the Flow AI video editor was very cool, allowing creatives to give their video generation a lot more direction and even input, and it helps them level up their workflows. It goes back to actually what we talked about several episodes back, Ben, about there needs to be tooling that allows creatives to iterate. In the same way software developers can, and this is Google really hitting home on that. I'm excited to see what folks make with it.

**Ben Lloyd Pearson** (2:30)
Yeah, exactly. Creatives are feeling disrupted by AI. It's really nice to finally see a focus on tools that help them versus just replace their capabilities. It's Google, who knows where these products are actually going to go. They may abandon them in a year or two years or something. But I think the real point to take away from this is that incumbency is still a very powerful force in software technology. In fact, to the point where I think even companies that are being disrupted by AI, like you think about how much search, like Google search has been disrupted by AI, because they have so much data, so much powerful network effect, they actually have quite a bit of capability to not just survive in this world, but also thrive because they have the data to give these models context. They can train data on this stuff.
We've been seeing, going around social comments about how companies like Stack Overflow are losing a lot of their traffic. But the reality is they aren't losing a lot, they aren't really losing their value. Like there's still a lot of value in the data set that Stack Overflow has from decades at this point, or more than a decade, of being this centralized knowledge source for developers. And that is extremely useful for training AI models. So yeah, it's, you know, everyone's focused on all the hot startups that are emerging in AI, but the incumbents might be a little bit slower to get around to it. But I think Google's really showing us that once they get there, it can be pretty powerful, particularly considering just how much Google stumbled in the early days of AI. Like this is a very different experience this time, just based on what I'm seeing.