Outcome engineering, AI hit pieces, and the end of the backlog

**Ben Lloyd-Pearson** (0:05)
Welcome to the Friday Deploy. I'm your host, Ben Lloyd-Pearson.

**Andrew Ziegler** (0:09)
And I'm your host, Andrew Ziegler.

**Ben Lloyd-Pearson** (0:11)
Yeah, this week, we're covering OpenAI joining forces with OpenClaw, AI writing a hit piece about a person, outcome engineers declaring that the backlog is dead, and Gemini agents going full tragedy. And you know, Andrew, this has been, this has been a remarkable week for me, like fully adopting agentic AI for the first time in my career, I suppose. And it's, it's really changed my perspective just on so much. But how are you feeling right now?

**Andrew Ziegler** (0:39)
I mean, that's actually the reality of this is, I usually have several things spinning when I'm in meetings or doing other stuff. And then I come back and I see what they did and I realign, give feedback. It's almost like this, just this constant feeling, there's only a burden, this need to be productive all the time. So it really makes you think of like Yage's recent article on the AI vampire and like the idea of constantly feeling like you need to be working. Even Zach, when he was here last week, mentioned that just like the living with the burden of how fast you can go is really exciting. But it's also an opportunity to really check in with yourself and be like, what's important? How do I want to use this new power to build the future?

**Ben Lloyd-Pearson** (1:18)
I have this tendency right now to triple task. So I'm doing something passive, like watching something while I have Claude building something while I'm writing the prompt for the next thing that it's going to build. It's like doing three things at once.
I mean, I can't always do that, but it's not as difficult of a workflow to manage once you get into it. It's pretty wild.

**Andrew Ziegler** (1:41)
Yeah, it becomes a world where you're building systems, systems that move knowledge and information around, and ultimately, you have to build the perfect glove fit for yourself. As we're in a new challenging new world where a lot of the best productivity comes from really, really finely understanding the problems that you're facing, and then creating these custom tailored solutions to solve it.

**Ben Lloyd-Pearson** (2:04)
Yeah. Well, anyways, let's get into the news because I think we're going to relate a lot of this to what we're experiencing. Let's start with this OpenAI acquiring OpenClaw, or at least the developer who built OpenClaw joining OpenAI. What's going on here?

**Andrew Ziegler** (2:18)
Yeah. So this is more like an aqua hire, I would say, behind for the very brilliant Peter Steinberg. He's the creator of OpenClaw, which originally hit the markets as Claude Bot. He's joining OpenAI to work on bringing agents to everybody. I think that's what OpenClaw taught the world is that agentic assistants are closer than we thought, and again, rapid adoption. As soon as it hit the scene, lots of engineers and non-engineers alike are buying new devices, old devices like Mac servers, putting things out of business, basically trying to get all of these things hosted in online. There's like a mad dash of people using these assistants. It really speaks to the hunger, I think, for this kind of technology. So it's interesting to see him join OpenAI. It makes me wonder what's next over there.

**Ben Lloyd-Pearson** (3:05)
I feel like with OpenAI, they benefited immensely from how viral the chat interface that they innovated on a few years ago. That was what made like GPTs and LLMs go viral for the first time. But I really think that's just the beginning of the changes that AI will introduce and that when you really think about it, a chat interface is not going to be nearly the best interface for AI. There's going to be a lot better ways of interacting with it.
OpenAI is really to me seems like they're trying the strategy of just trying a bunch of different things that are sort of around AI, and basically building as many of those different things as they possibly can, like workflow engines and integrations and stuff like that. But I feel like the result is that it's turned into a lot of really cool experiments rather than something that actually gets worked into my workflow and changes things. It kind of makes sense that they would want someone like Steinberger because he's even admitted, like this is just really just a really cool experiment that he built in OpenClaw. And I think some of the issues we've seen and that we've covered here on Dev Interrupted kind of proves that OpenClaw is not a production grade type thing. Like there's still significant changes and improvements that had to be made before something like that could actually go mainstream. And I can't help but like contrast that against a company like Anthropic, who I feel like they just seem way more focused on what they're doing. You know, like they really have a lot of intention around the things that they're building. So it's just, you know, it's two very different takes.