Associate professor @VUBrussel // Visiting scholar @Harvard.edu // Previously @jongeacademie.bsky.social // ♡ physics, math, AI, open science // Step 1: Find the pattern. Step 2: Follow the pattern. Step 3: Realize the pattern is following you.https://bsky.app/profile/vincentginis.bsky.socialhttps://bsky.app/profile/vincentginis.bsky.social/post/3mt2h6et2ms26Happy to see our work, published in Nature (rdcu.be/fzwzq), picked up by The Economist. We made a page where we answer the most pressing questions publicly. The Curious Case of the Declining Disruption's Disappearance (https://vincentholst.github.io/the_curious_case_of_the_declining_disruptions_disappearance/). Open Science meets Brandolini. [contains quote post or other embedded content]14 Aug 2026 15:14 +0000at://did:plc:ru62cybjy4vtskqlt73pfmjt/app.bsky.feed.post/3mt2h6et2ms26https://bsky.app/profile/vincentginis.bsky.social/post/3lihbnywnfk22Large Language Diffusion Models18 Feb 2025 12:37 +0000at://did:plc:ru62cybjy4vtskqlt73pfmjt/app.bsky.feed.post/3lihbnywnfk22https://bsky.app/profile/vincentginis.bsky.social/post/3licmdbhkdk2jToen Don’t look up uitkwam bedacht ik hoe de metafoor tot vervelens toe misbruikt zou worden in opiniestukken. Ik keek neer op mijn toekomstige zelf. Is er nog íémand bekommerd om de gevaren van AI? https://www.standaard.be/cnt/dmf20250214_9665528716 Feb 2025 16:05 +0000at://did:plc:ru62cybjy4vtskqlt73pfmjt/app.bsky.feed.post/3licmdbhkdk2jhttps://bsky.app/profile/vincentginis.bsky.social/post/3lhe4waajdc2iIf humanity goes down, let it at least be with poetry—not with 'OpenAI o5 (high),' but with a name the people chose. Something dignified. Something noble. Like Chatty McChatface.04 Feb 2025 13:09 +0000at://did:plc:ru62cybjy4vtskqlt73pfmjt/app.bsky.feed.post/3lhe4waajdc2ihttps://bsky.app/profile/vincentginis.bsky.social/post/3lh4zziezys2g🤲 [contains quote post or other embedded content]01 Feb 2025 17:29 +0000at://did:plc:ru62cybjy4vtskqlt73pfmjt/app.bsky.feed.post/3lh4zziezys2ghttps://bsky.app/profile/vincentginis.bsky.social/post/3lgtjilkkuk2uAcademics worldwide worked to create questions no state-of-the-art AI could answer in Humanity's Last Exam. But the real story isn’t just the questions that got in—it’s the ones that didn’t. Crafting problems that truly push AI is getting harder, and that says a lot about where we are. https://www.standaard.be/cnt/dmf20250127_9721515128 Jan 2025 22:39 +0000at://did:plc:ru62cybjy4vtskqlt73pfmjt/app.bsky.feed.post/3lgtjilkkuk2uhttps://bsky.app/profile/vincentginis.bsky.social/post/3lgld7jtmtk2oIt is 2025, and I vaguely remember the time when it was still possible to come up with questions that could baffle state-of-the-art LLMs. Of course, I’m joking. I don’t really remember. https://www.nytimes.com/2025/01/23/technology/ai-test-humanitys-last-exam.html?smid=bs-share25 Jan 2025 16:25 +0000at://did:plc:ru62cybjy4vtskqlt73pfmjt/app.bsky.feed.post/3lgld7jtmtk2o