Video · 2026-08-05 · 1h 3m · 6 moments

The Rise of the Design Engineer

✦ AI generated

GUEST Scott GUEST Wes GUEST Courtney Tolchin

timeline · colored by role

Claim◆

The design engineer—who blends aesthetic sensibility with technical skill to practice design built and shipped autonomously—is poised to become the new full stack developer, and this skill set will only grow more important as vibe-coded UIs proliferate.

Wes argues the design engineer merges the aesthetic and technical disciplines that were once fought over (designers who code, full stack vs. specialization). He predicts the skill set becomes the new full stack developer, mattering even more in the next six months as everyone vibe-codes mediocre UIs.

transcript

Wes: this idea of a design engineer, which is somebody who very well understands how products work, how user interfaces work, how they what happens when you click it, how what happens when you have multiple items together. ... these apps look exactly the same, but they are not the same in how they function. ... this skill is going to be the new full stack developer, right? ... like that was the huge thing that everybody fought over as well is like should you be a full stack engineer or should you be backend and front end? That's no longer a question. In most cases, people are are doing both in the web world.

Claim

Developers are biased against designers because of jealousy, and the bias is compounded by gender; but someone who can read and write code and also has great design taste is unstoppable, so the asker is in a strong position.

Scott responds to Birthday Girl's question about credibility bias: engineers undervalue design because of jealousy toward a discipline they haven't learned, and the bias is likely compounded by her being a woman. He affirms the hybrid role's value.

transcript

Scott: on the regard of are people biased against design because it's not seen as technical. Uh I think that this has always been one a bias that I've noticed and I would like to pose that it's out of jealousy. developers look at somebody who's good at design and think they possibly could not be as good of an engineer as me because well I'm not good at design and I would uh was you know that's a whole discipline you've had to learn and I've spent that time in the engineering side of things. So I I've personally seen this to be the case. ... I would also say it is probably also because you're a woman. I I that is also a bias that we see in this industry. So you you have a multiple factors here and I think that's working against you. But I do think that this role in particular is such a huge role in the future now uh of the world that we exist in with AI and all these things and somebody who has design taste and design style is going to be able to create some really amazing stuff.

supports · 1

Example◆

Good design is not about what an app looks like—visually similar apps can justify radically different usability—and current AI-generated 'slop' interfaces repeat the same mistake as the Bootstrap era and McMansion subdivision housing: prioritizing appearance and templates over character and usability.

Scott and Wes extend the Homeowner's analogy: like McMansion homes that all look identical and have terrible usability, AI-generated UIs look samey on the surface. The real problem is how they function—Threads desktop looks like Twitter but is infuriating to use, and apps like TikTok and Instagram are visually indistinguishable.

transcript

Scott: The the biggest point is not what it looks like. It's it's how usable it actually is. Because if you take a look at like all of the apps, like if I open Threads, Twitter, Blue Sky, and and all of these other like short form blogging apps, sometimes I cannot even tell which one I'm in. ... Even though it looks very similar to Twitter, there is so much more thought and care put into the design of Twitter than there is on the threads desktop website. And that there is is a perfect example of it's not what it looks like. It's how the you interact with it, how you use it. ... If somebody spent the time to say, hm, when you click on this button where it says a notification, does that notification clear? or all of these little tiny examples as to how the interface is to to get where you want.

Fact

Tests are more important than ever as AI accelerates code changes, and the most effective strategy is to convert as much as possible into deterministic, automated checks—hand-written unit tests for complex logic and especially end-to-end tests—so outcomes do not depend on the AI's discretion and token spend stays lower than letting agents fumble.

In response to TLC's question about how to test AI-generated code, Wes emphasizes hand-written unit tests for complex logic (e.g., URL parsing) and Scott stresses end-to-end tests as the way to make verification deterministic, noting deterministic tests use fewer tokens than letting agents open a browser to poke around.

transcript

Wes: your tests are more important than ever as as people are just prompting and adding features to absolutely everything. You need to feel extremely confident. There's the amount of pull requests that are coming in, the amount of code changes that are coming coming in have skyrocketed. ... Unit tests first for for complex logic is something that I will do quite a bit. ... the end-to-end integration tests are super important for these types of things because you want to make sure that you can still use it. ... Tools like Playright are so good for this because you can one see it happen. ... I really lean hard on end to-end tests, especially in this and moving those possible things out of the AI's discretion, right? Oh, fire up a Chrome MCP or fire up an agent browser and make sure this thing works. Uh, how about instead write the playright test, click through it, and then you can run those tests. ... you you see them doing that all the time and and that's that's way more expensive than just running a deterministic test.

supports · 1

Prediction

A speculative bubble in AI company valuations and GPU ownership is real, but the underlying technology and its use are not going away; however, the pricing bubble—AI providers charging every engineer $5,000 a month—is collapsing under cheaper open-weight models that are 98% as good.

Wes and Scott argue the 'AI bubble' discourse is correct about unsustainable valuations and GPU/capital costs, but wrong that the technology disappears—like the dot-com crash, the internet survived. They see the real correction coming in AI pricing, as companies realize open-weight models are 10x cheaper and nearly as good.

transcript

Wes: I think the information around AI being a bubble is very real. I think the the house of cards there in terms of uh CPU or GPU ownership and things like that. I I think all of that very much makes sense. Do I think that the bubble will pop and that AI will go away? Absolutely not. ... when the dot bubble burst, did the internet go away? ... It's just that uh some people got shut down. ... the bubble of Enthropic being able to charge every single engineer $5,000 a month, I think, is is going away. Um, and we're seeing that by people realizing, oh, some of these new open weight models are 10 times cheaper and 98% as good. ... you got to get good at working with this stuff if you want to continue to be effective. That's just where I'm at on this.

Fact◆

You should teach children to be critical thinkers about AI—understanding that it is a fallible, highly agreeable predictive model, not an error-proof, authoritative source—and expose them to it in a guided, monitored way, while avoiding things like AI image transformation that can damage self-image.

Dr. Courtney Tolchin (Tinsky) explains how to handle children's exposure to AI: talk about it at home, teach that AI hallucinates and is not a definitive answer, model critical thinking (even in the car before asking Grok), and keep exposure guided and monitored—while avoiding AI image transformations that can make kids dislike their own appearance.

transcript

Courtney Tolchin: having that understanding really built into your kids of it can make mistakes and it is not errorproof is really important and again to be critical thinkers. So to be able to like think about is this answer correct? Is this you know aligned with what I know? ... I I feel like I've taught my kids have have a high level of skepticism towards a lot of the stuff. ... introducing kids to AI is important because obviously we know it's not going away, but you have to do it in a way where you're like guiding their usage and monitoring it. ... Having them understand that it is, you know, a predictive model. It does not care about you. It does not have feelings for you like a human being would. ... we specifically don't let our kids have the um AI image transformation. Um because ... that could be very damaging to children where ... What would I look like if I was prettier? What would I look like if I had a different color hair? And then all of a sudden you just don't start to not like yourself because the photo that the AI comes up with is looks so much better.

Highlight slides

Design engineers: the new full stack developers✦ from: The design engineer—who blends aesthetic sensibility with technical skill to practice design built and shipped autonomously—is poised to become the new full stack developer, and this skill set will only grow more important as vibe-coded UIs proliferate.A convergence replacing an old fight✦ from: The design engineer—who blends aesthetic sensibility with technical skill to practice design built and shipped autonomously—is poised to become the new full stack developer, and this skill set will only grow more important as vibe-coded UIs proliferate.Good design = usability, not appearance✦ from: Good design is not about what an app looks like—visually similar apps can justify radically different usability—and current AI-generated 'slop' interfaces repeat the same mistake as the Bootstrap era and McMansion subdivision housing: prioritizing appearance and templates over character and usability.The McMansion analogy for AI-generated UIs✦ from: Good design is not about what an app looks like—visually similar apps can justify radically different usability—and current AI-generated 'slop' interfaces repeat the same mistake as the Bootstrap era and McMansion subdivision housing: prioritizing appearance and templates over character and usability.AI is not an authority — it's a fallible predictive model✦ from: You should teach children to be critical thinkers about AI—understanding that it is a fallible, highly agreeable predictive model, not an error-proof, authoritative source—and expose them to it in a guided, monitored way, while avoiding things like AI image transformation that can damage self-image.Guided, monitored exposure — with firm boundaries✦ from: You should teach children to be critical thinkers about AI—understanding that it is a fallible, highly agreeable predictive model, not an error-proof, authoritative source—and expose them to it in a guided, monitored way, while avoiding things like AI image transformation that can damage self-image.

Related episodes

Video · Lenny's Podcast

Chatbots are not the final interface: OpenAI’s Head of Design on what’s next | Ian Silber

→

Video · Lenny's Podcast

OpenAI’s Head of Design: This is the best time in history to be a designer | Ian Silber

→

Video · 20VC

Jensen's Open-Weights Letter | Google Cloud Grows 82% But The Market Tanks

→