Tech • IA • Crypto
The Problem Solvers: Scott Wu at Cognition
TL;DR
AI-driven software engineering tools are rapidly advancing, with companies aiming to make building software as simple as issuing instructions to a computer.
KEY POINTS
Rise of AI Software Engineers
New systems such as Devin, described as an autonomous AI software engineer, are being developed to handle complex programming tasks. These tools are designed not just to assist developers but to independently plan, write, and execute code, marking a shift from traditional coding assistants to fully capable engineering agents.
Mission to Accelerate Development
The core goal behind these technologies is to make software development up to 10 times faster. Organizations across sectors—including banks, health insurers, governments, and private equity firms—are already adopting such tools to streamline workflows and reduce development timelines.
Foundations in Competitive Programming
The push toward advanced AI coding systems is rooted in deep expertise in mathematics and competitive programming, disciplines that emphasize efficiency and problem-solving. Early skepticism about whether AI could match human-level performance in these domains has largely been overcome by recent breakthroughs.
Collaboration Over Competition
Industry leaders emphasize that no single product will dominate all aspects of software engineering. Instead, a diverse ecosystem of tools is emerging, with companies collaborating to integrate capabilities. Partnerships, including those involving Anthropic and tools like Claude Code, highlight a cooperative approach to innovation.
Transparency and Trust in Development
Open communication between AI developers and partners has been key to progress. Transparency around product capabilities and limitations has helped build trust, enabling organizations to adopt these tools more confidently and integrate them into critical operations.
Expanding Access to Software Creation
A central vision driving the field is democratization: enabling anyone to turn ideas into working software. By lowering technical barriers, AI systems aim to empower non-programmers to create applications, potentially transforming industries and individual productivity.
Practical Focus on Engineering Challenges
Developers of AI coding systems stress the importance of solving real-world engineering problems rather than pursuing abstract benchmarks. Efforts are concentrated on improving reliability, debugging, and deployment—areas that directly impact usability and adoption.
Emergence of a New Development Paradigm
The trajectory of these tools points toward a future where coding becomes conversational. Instead of writing detailed syntax, users may simply describe desired outcomes, with AI systems handling implementation, testing, and iteration.
CONCLUSION
Advances in AI-driven software engineering are reshaping how code is written and deployed, with collaborative ecosystems and practical applications driving a transition toward faster, more accessible development.
Full transcript
And Scott, if you could please introduce yourself by telling us where you go to school and what grade you're in. I go to Glasgow Middle School and I'm in seventh grade. Excellent. We are gonna move on to the first question. What is the value of 255- (buzzer beeps) Scott? 5,000. 5,000 is the correct answer. (audience applauds) I grew up, like my whole life was math and programming competitions. - Yeah. I was world champion of competitive programming in 2014. And I remember talking about this with some of my friends, and do you guys think there's ever gonna be a day where AI can actually just do all of these things that we do? And to say the obvious, it's crazy. Yeah. (Scott laughs) What we do at Cognition is we build Devin, the first AI software engineer and a suite of tools around that. We work with a lot of the biggest institutions in the world. We work with banks, health insurers, governments, private equity firms. Our whole mission is to help build software around the world 10 times faster. You sort of had a singular vision on what Devin was gonna look like. I'm curious like how you managed to stick with it. We just care so much more about the art and craft of software engineering on the ground, and how do we figure out how to go and make that much faster and solve all the practical problems and unblock that. It always felt like you had a very collaborative attitude towards how to make this happen together. Very vividly remember my first conversation with Daria. I got on the call with him, I was talking about like, "Yeah, so we're working on this." He stopped me and he's like, "Don't worry, you don't have to pitch me on this, the concept, I totally get it." We were doing something that people didn't really believe was possible- - Yeah, yeah. And I appreciated that a lot from the very beginning. We've also started launching some of our own products, like Claude Code. As our relationships evolved, I'm curious what your feelings are about it. I really appreciate that you guys have been really transparent with us. I've always thought about it as I really just don't think there's gonna be a single product that does all of code and all of software engineering. There are so many experiences to serve. And I think from that perspective, there's so much more for us to do together than there is separate, and that's what we're excited about. Yeah. (subdued music) From the beginning, I always loved that you could take any idea that you had and turn it into reality. Everyone will be able to turn those ideas into reality. We've been able to make so much more progress and see companies build much, much faster than they have been before. By getting to work with Anthropic, we're able to go and deliver customers the exact capabilities that they're looking for. We've been able to help bring folks to a new era of software engineering. This is a world where building software is just as easy as telling your computer what to do. (gentle music)