Are you using AI to help people, especially those who can't interact with the world the way you do?

Human-empathetic design in an AI world? I'm giving a talk tonight about how AI is about to change design. Alan Cooper, @MrAlanCooper, who built what became Visual Basic, and I spent a lot of time back in the 1990s talking about design and the process of designing software (I worked at a Visual Basic programming magazine so got to know him well). He authored influential books like "About Face: The Essentials of Interaction Design" and "The Inmates Are Running the Asylum," which helped establish modern UX principles. He hated prototypes. Why? Because others would think they were real and then disappointed when they didn't actually work. He recommended that all designers work with pencil and pen to make sure that everyone knew they were in a brainstorming phase, which led to better designs. For instance, I have in my head a new kind of camera interface. Here, I got @lovart_ai to build it with just a little bit of prompting: Design is about solving problems for human beings. That means having deep empathy for human beings. AI can "pretend" to do that, sometimes pretty well, but it can come across as "souless." Sometimes design is about "hiding" functionality. Snapchat did that very well. It added a little bit of a tax to figure out how to slide interfaces in and out. But by doing so it made something beautiful and addicting, particularly to young people who wanted a very different interface than their parents had with Facebook and Twitter. In playing around with Lovart it is OK at some of that, but it often can't come up with a really novel way of doing something like that. Yeah, you can try to prompt that, and it'll try, but it can't really come up with something with deep empathy and if you just try to get it to be unique (I prompted it to "get weirder") it'll go in a direction that doesn't lead to a good place. I once interviewed a famous architect who built sports stadiums. He told me he had the fun job, which is to come up with the concept. The interns, he told me, had to do the grungy work, like figure out what the door handles on the bathrooms would be. AI can do the grungy work so well, which is why Lovart is so good at doing brochures, stickers, outdoor advertising, among other things. But it is edging into the "fun work" too. The thing is it really doesn't understand user goals very well. Look at the camera interface it designed. It didn't get the goal of a camera is to capture an emotion, a moment, an aesthetic. Its shutter speed button is way too big and doesn't fit what's in my head. Shutter speed is used by photographers to either blur the scene, by using a slow one, or to capture fast-moving subjects, like a runner, sharply. A human designer might get the user goals a lot better than my simple prompt communicated by asking for manual overrides. The nice thing is that with AI you can keep iterating until you get something usable. But now you have to sell it to other people, and you are back to Alan Cooper's point: that to get other people to add their ideas, or to buy into yours, if you show them real looking interfaces they stop thinking. Of course you can just prompt that, which gave me this image. I added all the designers I can find to this list, if you have others, please let me know: Oh, and had @Grok make me an interview with a "fake" Alan Cooper about the future of design that properly laid out his thinking, and mine: How do you think design will change thanks to AI? AI might be souless, but it sure is fun!

MCP is here! You can now give Devin access to your favorite servers via the MCP Marketplace. Think Datadog, Linear, Sentry, Figma, and thousands more. Demos from our team + getting started 👇

Introducing Moby. A hyper-intelligent agent suite that can add $100,000+ of revenue to any 7-figure brand. Bold claim, I know. Let's see it in action:

We’re thrilled to see our advanced ML models and EMG hardware — that transform neural signals controlling muscles at the wrist into commands that seamlessly drive computer interactions — appearing in the latest edition of @Nature. Read the story: Find more details on this work and the models on @github:

It’s time for the American AI community to wake up, drop the "open is not safe" bullshit, and return to its roots: open science and open-source AI, powered by an unmatched community of frontier labs, big tech, startups, universities, and non‑profits. If we don’t, we’ll be forced to build on foreign foundations and risk losing the race in AI altogether by lack of local innovation and competition. Let’s go!

Very few people have been inside a datacenter. 19 days to get 230,000 GPUs zero and working is absolutely incredible. I helped Zoho’s founders setup a single rack in its datacenter’s early days and it took all Saturday for something like 20 computers. Congrats @elonmusk and the team that did it. We live in incredible times.

This is a game changer for me. The @perplexity_ai Comet browser can automatically go through the YouTube upload wizard for me It's always so annoying to do that for every video. Now I can just tell it what to write.

Coming in an hour. A discussion with @samuelekpe who launched @grupa_ai today. With @IrenaCronin and me. An AI startup that helps founders.

Even the 13-year-olds are building agents. Meanwhile I can’t get my 15-year-old to do same. Parenting failure. Hey @grok how do I get my son to stop playing video games and talk to you instead? :-)