“We don't want to be the new waterfall.” Cassie Shum · Thoughtworks Technology Podcast
Hands-On Engineering Podcasts · September 2026
Shum is describing what she and Tim Cochran keep hearing from enterprises bolting agents onto their development process: an intense new focus on getting the spec right so it can be handed off. Her objection is that heavy up-front specification recreates the thing agile was meant to end. She asks teams when an idea ever reached the customer as the same idea it started as, and says almost none of them say it did.
Transcript
I think loops is our next generation word for, you know, what, how do we close the feedback gap, right? Like, how do we learn from what we're actually doing? And so when we talk about product and execution loops, they are very much the STLC process or the PDLC process that we saw from years ago, all before AI. So I think that's number one. Like that's we're bringing that back, right? As a first-class concept. I do think, and I'm excited to hear Tim's opinion here as well, because we've been kind of focused on both of these loops is that. As engineers, we are seeing in the enterprises and companies right now that there is a major focus on the execution loop, right? So there's a lot of software factories and harnesses because there's the agentic technology that has come about in the last couple years has really accelerated that. And so I think the major focus of that has been accelerating, right? But one of the things that Tim and I have discussed at great length and we're trying to work with our customers on what that actually means is it doesn't just stop from the spec to the PR, right? So how do you close that loop from a value sense? How does that actually come back to the product? And that's why we talk about the two closed loops and they're both equally important and they need to feed off of each other.
Definitely agree. And I think what's interesting is that they're optimizing completely different ways, right? I think from the beginning of software development, we've always wanted to sort of industrialize software development, but it's kind of been difficult because most of the time we're actually creating something that's new and novel. But I think now, especially with AI and some of the insights that we get from agent traces and things like that, we can look at, we can be explicit about these two loops where actually, you know what? We do actually want to create a software factory or a dark factory. And we do actually want to industrialize things that are concrete and deterministic. But there's a lot of software development that is not concrete and deterministic. And those things we actually need to sort of like acknowledge and accept that they are fuzzy and they have to be discovered and evolved and iterated on. And sort of embracing that asymmetry between those two loops is sort of where we see the team working on. So obviously right now there's a ton of focus on improving agency efficiency and creating harnesses and that's all great, all in the execution loop. But we also want to use the sort of impetus of AI to actually close the gap between product and engineering. Because, you know, right now developers don't have to spend two weeks to build a feature, right? They can actually come up for air, you know, after like half a day or a day and they can maybe collaborate a bit more. So that's that's more about like why we're sort of thinking about these two loops.
And one last thing to add on this one as well, and Tim and I discussed this a lot because we've been hearing this from actually some of our previous places that we've been is that because now the agentic lifecycle is up and front center and everyone says, let's bolt on the AI, there's such a focus on spec writing. Like how do we actually hone in that spec writing so we can hand it over to the agents? And as I do that, I feel like I'm saying, how do we make things more waterfall? Like I'm going to upfront design specs and then the agents will take care of all of that. And so I always ask people when we're talking to them, I said, when have you ever had an idea that was that same idea at the very end when it's in front of your customer? And half the time, and not more than time, they said, oh, no, it's not. Because going back to some of the fundamentals that we've talked about for years is like you have to iterate and learn on that idea. And so the question now begs is, how do you do that with the agents? How do you do that with builders and agents and iterate that and create a system to do that? And so that's why those two loops are incredibly important as well. We don't want to be the new waterfall. We want to actually embrace that collaborative approach of actually creating an idea and finding value in that using the agents that we have now.
Is it the same problem frame differently? Because like if we think about the DevOps movement, what the last 16, 17 years, that was all supposedly, I say supposedly because it got hijacked a little bit about feedback loops and making sure that product teams were seeing how things run in production, they could react faster, you know, et cetera. Is this the same problem with different clothing or is it really that different because it's AI involved?
I would say it's absolutely the same problem. I think what AI, and it's, you know, what old is new and all that kind of thing. And I think it's really what AI is doing is accelerating, right? And, you know, the promise of agile or the dual track software development, we've always wanted to have this close collaboration and this feed these feedback. And yes, it's just kind of like embracing that and just because of the acceleration of AI. I think.
I think that acceleration is going to expose all of the different bottlenecks in the software development lifecycle process. It's not just engineering anymore. And I'm going to hubble bragg Tim a little bit. So five years ago, he wrote about the developer effectiveness. You know, he wrote on Martin's blog about how to actually hone that in and make that more effective, right? And so he talks about the bottlenecks that developers faced back then. And I think now we're in a new phase where new bottlenecks will come out because agents and AI is accelerating, if you do it well, like accelerating the engineering area, right? This, this, the developers, the friction that was held and felt for many decades. And now we're going to find the friction and the bottlenecks in other areas as well. And so I think the premise of the feedback is the same, but the bottlenecks will appear in different areas, Tim.
More from Thoughtworks Technology Podcast
