Not exactly a benchmark, but I always tell people the best way to test a model is to have a conversation with it about something you know a ton about. Gardening, running, music, tv, whatever it is. You get a feel fast for where its strengths/weaknesses are, then go from there
@mustafasuleyman
-
Social Sciences Hackers Drive Next Wave Innovation
By
–
It's the year of the social sciences hacker.
— Mustafa Suleyman (@mustafasuleyman) 15 juillet 2025
We're about to see leaps in innovation that don't come from engineers. Instead, they'll come from people who've never gotten to build before.
I couldn't be more excited about it. pic.twitter.com/vgJ7wxdOdRIt's the year of the social sciences hacker. We're about to see leaps in innovation that don't come from engineers. Instead, they'll come from people who've never gotten to build before. I couldn't be more excited about it.
-
Can Computers Be Conscious? Domain-Specific Superintelligence Debate
By
–
CatGPT really put me in the hot seat. Can computers be conscious? How close are we to domain-specific superintelligence? And most importantly WHY did I wear that outfit in my TED Talk? Check out the full convo for more hot takes + style highs and lows:
-
Mobile Uploads: From Hard Problem to Trivial Task
By
–
Interviewed a candidate today whose big hard job 10 years ago was "mobile uploads". At the time people thought it was a really hard problem. Absolutely crazy to think how far we've come.
-
Implicit Feedback Mechanisms in AI Model Training Systems
By
–
TikTok fyp is a textbook example here. How fast you swipe, how long you watch…feedback cues you can get (and learn from) way more often than explicit thumbs up/down or those 'what did you think of this' post-video surveys
-
Great PMs Design UIs to Collect AI Feedback
By
–
Good PMs gather feedback. Great PMs design UIs to collect it. In an AI-driven world, relying on feedback forms or even thumbs up/down isn't enough. Feedback often comes dressed in great UI design. You have to bake it in from the outset.
-
Human Impact Over Technology Advancement in AI Building
By
–
Tech for tech’s sake is pointless. It’s all about what tech that makes you feel. How it makes our users feel. Calmer? More confident? Capable?
— Mustafa Suleyman (@mustafasuleyman) 9 juillet 2025
Stories like Noel's are why I'm so passionate about building AI. It's not the advances or the benchmarks, it's the human impact. pic.twitter.com/MSSr1h3u5BTech for tech’s sake is pointless. It’s all about what tech that makes you feel. How it makes our users feel. Calmer? More confident? Capable? Stories like Noel's are why I'm so passionate about building AI. It's not the advances or the benchmarks, it's the human impact.
-
AI Tools Simplifying Everyday Tasks: Cooking, Fitness, Shopping
By
–
lately been doing:
in Edge – parsing product reviews, going hands-free while cooking
Windows – learning Photoshop, anything settings related
on mobile – estimate macros in breakfast, fix things around the house, figure out if packing label is prepaid or not. Bonus: outfit roast -
Copilot Voice and Vision Features Now Available Free
By
–
Try it out for yourself! You get unlimited free voice usage in Copilot + Vision usage in Edge. And in the U.S., Vision in Windows and on mobile are out now.
-
Voice and Vision AI Interfaces Replace Text Prompting
By
–
Noticed I'm prompting a lot less these days. Not because I'm using AI less – it's actually the opposite. I'm just defaulting to using voice + vision more and more. It's so natural, it's like the UI just melts away. Less explaining, more engaging