←
Willi Schroll16,233 posts
Willi Schroll avatar

Willi Schroll

@wschroll

Innovation • Foresight • Strategy. Trend analysis, scenario reports, action plans for your business. AI- & data-driven. Senior level. DM me. W.S. & team.

◷ Joined June 2008▧ 1,000 Photos and videos
Willi Schroll avatar
Willi Schroll@wschroll·Nov 13

The era of *discovery systems* in science & also in RnD „to give researchers and engineers practical experience in building the next generation of 🔥discovery systems“: = bring together HI & AI to accelerate materials discovery, design & manufacturing.

AI can generate scientific hypotheses, write and execute code, build scientific instruments, run simulations, analyze results, and use what they learn to decide what to explore next. Working together as agent swarms, they can investigate many possibilities in parallel, reason deeply across boundaries of knowledge, and connect ideas between different fields. This opens a path toward new materials and designs that would be difficult to discover through conventional approaches alone. I’m excited to share my upcoming MIT Professional Education short course, Applied AI for Materials Discovery, taking place July 26–29, 2027, live online. Through hands-on clinics, we will build AI scientist agents and autonomous workflows that can plan, build instruments, reason and carry out complex scientific tasks. We will explore multimodal foundation models, generative design and coordinated agent swarms, connecting them with simulations and physical constraints to develop and test new ideas. The goal is to give researchers and engineers practical experience in building the next generation of discovery systems: systems that bring together human judgment and AI to accelerate materials discovery, design and manufacturing. For registration, see link in reply.

Post attachment
Nick Vasilescu avatar
⟳ Willi Schroll retweeted
Nick Vasilescu@nickvasilescu·Nov 14

Anthropic is reverse engineering every app on the internet. So I tried it on Higgsfield. 15 agents. 15 Orgo computers. Each one tore apart a different feature. - 25 min and $31 to turn the whole platform into a 5,500 line spec - Paper MCP cloned every screen, down to the fonts - T3 ran the whole thing in the cloud while I was at the airport The clone actually works. Image gen, video, Cinema Studio, canvas, even its own MCP. Open sourcing it. Full build below 👇

22:38
Linus ✦ Ekenstam 🔜 London avatar
⟳ Willi Schroll retweeted
Linus ✦ Ekenstam 🔜 London@LinusEkenstam·Nov 15

Most Americans have no fucking clue who Alexander Stubb is. Which is unfortunate, because Finland somehow managed to elect an actual functioning adult. PhD in international relations. Speaks five languages (his wife does too). Ironman triathlete. Commander-in-Chief of a country sharing 830 miles of border with fucking Russia. Meanwhile, half the political class can’t string together a coherent sentence without a teleprompter and a team of handlers. No tantrums. No bumper-sticker bullshit. No chest-thumping circus act. Just a terrifyingly competent human being casually dropping geopolitical wisdom in his THIRD LANGUAGE. “What humans begin, humans can also end.” Imagine that. A politician with a brain AND the ability to operate it.

2:50
Willi Schroll avatar
Willi Schroll@wschroll·Nov 11

This is a new alphabet, a new standard in Science Literacy maybe for the 21st century. „Super Science“ as a term?👌 Thank you for the pioneering work @ProfBuehlerMIT ! 3/

Willi Schroll avatar
Willi Schroll@wschroll·Nov 11

▶️ Learning and discovery are flows through spaces of possibility. … ▶️ Nonlinear physics actually defines a larger principle, where high-dimensional dynamics generate stable invariants; … the substrate for a new level of cognition. ▶️ Recursion then becomes level creation - one possibility space compresses into a principle, and that principle opens a larger space above it. 2/

Willi Schroll avatar
Willi Schroll@wschroll·Nov 11

„Intelligence then grows by _constructing new spaces to think in_.“ „the AI had to reason across long path-dependent histories, simulate alternative _futures_“ „A lot of positive will come out of this because it expands the human _epistemic horizon_“ 1/

Post attachment

We built recursive meta-intelligence, an AI that creates its own scientific instruments, turns them into persistent worlds that an agent ecology with hundreds of AIs inhabits, and uses those worlds to discover mechanistic principles in one of the hardest classes of physical problems: how complex hierarchical materials (nested structures of matter that give rise to new function through organization) evolve and fail. The AI reasons across enormous spaces of possible physical trajectories, where every rupture changes what can happen next, and compresses those histories into principles (which humans can understand and design with) - complex chains of causal events, highly nonlinear, and intricate. Scientific superintelligence is tangible here - machine-scale exploration opening cognitive channels into complexity that has been extremely difficult for humans to traverse directly. AI builds the spaces in which its next level of reasoning becomes possible; a representation becomes an instrument, the instrument becomes an executable world, and that world becomes the substrate for further intelligence. Intelligence then grows by constructing new spaces to think in. The task we explored started from a seemingly simple prompt to explore a biological material system - the AI then chose the representation, mechanics and experiments, built a fracture laboratory to push materials to their limit, tested hypotheses and generated scientific conclusions. The swarm explored a combinatorial universe in which architecture controls function and every rupture changes the future state of the material. The AI discovered a compact principle that defines how multiscale material architecture can program the evolution of failure. Material placement and geometric order determine how forces redistribute, whether damage cascades or remains distributed, and whether function survives substantial flaws. For this discovery to happen the AI had to reason across long path-dependent histories, simulate alternative futures and compress them into generative invariants (model-based causal reasoning, counterfactual simulation and temporal abstraction applied to an evolving physical world). It is incredible to witness this transition to a new form of intelligence and capability through scaling swarms. A lot of positive will come out of this because it expands the human epistemic horizon as AI can traverse thousands of possible histories and return mechanisms compact enough for us to understand, test and build from. Intelligence compounds through its artifacts! A few lessons we learned: ▶️ Learning and discovery are flows through spaces of possibility. Early work has shown how backpropagation flows through parameters, reinforcement learning through action and consequence, and autonomous swarms through representations, instruments and executable worlds, bringing it all together. Flows create structure; structure redirects future flows in the recursive instrument. ▶️ Nonlinear physics actually defines a larger principle, where high-dimensional dynamics generate stable invariants; invariants become effective variables; those variables become the substrate for a new level of cognition. ▶️ Recursion then becomes level creation - one possibility space compresses into a principle, and that principle opens a larger space above it.

2:37
Willi Schroll avatar
Willi Schroll@wschroll·Nov 11

History in turbo mode 🔥 AI 2027 TIMELINES MOVED UP
Researchers now see top-human AI research capability by 2027; 50-50 odds by May. Author: “scared, sad, and angry.”0 🔥 ANTHROPIC FILES FOR IPO
Shares could trade as soon as next week. Updated docs say it has crossed its own AI safety threshold.0 🔥 OPENAI LAUNCHES CHATGPT BUSINESS HUB
Brands get AI shopping agents, verified profiles, mention tracking, and ads. Chatbot SEO begins.0 🔥 CLAUDE EXPANDS INTO FULL OFFICE SUITE
New beta Sheets, Whiteboards, Meetings; Data now generally available. Building Docs, Slides, Design.0 🔥 CLAUDE FABLE 5.2 TAKES #1
Tops Artificial Analysis Intelligence Index ahead of Gemini 4 Carbon. Cheaper and faster than previous version.0 🔥 META SHIPS MUSE EMBER
Smallest model matches Claude Haiku performance at roughly one-third the cost for agent work.0 🔥 GOOGLE UNVEILS NEW TPU CHIPS
TPU 8tx keeps data encrypted in use; 8i Max joins chips for full-codebase agents. Demand already sold out.0 🔥 NVIDIA CHIP PROBE HALTED
Top lawyer emailed investigators in 2025 Megaspeed China smuggling case. Commerce hands it to Justice.0 🔥 ADOBE-DISNEY AI DEAL
Multi-year pact to train custom models on Disney characters and styles for parks, film, ESPN.0 🔥 OPENAI BUYS AI TRAVEL STARTUP
Acquires Odessia; founder joins to build travel features. ChatGPT expands into shopping and trips.0 🔥 FRANCE LAUNCHES PARIS AI SAFETY CALL
OpenAI, Anthropic, Mistral pledge emergency off-switches in advanced models.

Huge AI announcements from OpenAI, Anthropic, Meta, and Google in the past 48 hours. Here's everything going on in AI right now: 1. The authors of AI 2027 just moved their timelines up. They now expect AI on par with the best human AI researchers to arrive in 2027, after METR estimated Anthropic is improving about 2x as fast thanks to AI. Daniel Kokotajlo gives it a 50-50 chance by May 2027 and says he's "scared, sad, and angry." 2. Anthropic opened the investor pitch for its IPO, with shares likely trading as soon as Thursday next week. An updated filing also says it has now crossed its own safety line for AI-driven progress. 3. OpenAI launched Businesses in ChatGPT, a beta hub where brands can set up AI agents for shoppers, claim a verified profile, and track how often ChatGPT mentions them. Paid ads can point users to those agents too. This is like the start of SEO for chatbots. 4. Anthropic rolled out Claude Sheets and Whiteboards in beta, made Claude Data generally available, and announced Claude Meetings, an assistant that sits in on your video calls. Add those to Slides, Design, and Docs. Seems like Claude is slowly rebuilding the entire office suite. 5. Anthropic also released Claude Fable 5.2 and Mythos 5.2 yesterday, with Fable debuting at #1 on Artificial Analysis' Intelligence Index, 5 points ahead of Google's Gemini 4 Carbon. It's also 20% cheaper and up to 30% faster than 5.1. Google lost the top spot before Carbon even shipped (!) 6. Meta released Muse Ember, its smallest model, built for high-volume agent work. It ties Claude Haiku 5.5 on the Intelligence Index at roughly one-third of the cost per task ($0.14 vs. $0.41 on an average test). Meta shipped its first Muse model in April and now has an answer for every tier OpenAI and Anthropic sell. 7. Google announced 2 new AI chips coming next year: the TPU 8tx, which keeps data encrypted even while a model is working on it, and the 8i Max, which joins 2 chips into one for agents that need to hold an entire codebase in view. Google already has nearly 100 TPU customers (including Anthropic and Apple) and every available chip is spoken for. 8. Nvidia's top lawyer reportedly emailed agents in August 2025 while they were investigating whether Singapore's Megaspeed smuggled its chips to China. The Commerce Department then halted the case and handed it to the Justice Department. Not a great week for Nvidia. 9. Adobe signed a multi-year deal with Disney to build custom AI models trained on Disney's characters and visual styles, with work already underway for theme parks, film, and ESPN. Pretty interesting to see one of the most protective IP owners in entertainment is now training its own AI models. 10. OpenAI acquired Odessia, a months-old AI trip-planning startup from Sonder founder Francis Davidson, who is joining OpenAI to work on travel. Between this and #3, ChatGPT is coming for shopping and travel. 11. France unveiled the Paris Call for AI Safety and Security, with OpenAI, Anthropic, and Mistral committing to build emergency off switches into their most advanced models. Interesting timing given #1 on this list. Daily reminder that this is the worst AI will ever be.

Bojan Tunguz avatar
⟳ Willi Schroll retweeted
Bojan Tunguz@tunguz·Nov 7

Analog techno is the thing.

6:37
Willi Schroll avatar
Willi Schroll@wschroll·Nov 7

Looks like a classical sigmoidal saturation curve indeed. This is the Age of AI = infrastructure for „*Cognitive Mobility*“. 🧠 Get from point A, problem, to point B, solution. As with phys. mobility: Not every pedestrian becomes a car owner.

Breaking: Green line represents a widespread fantasy, and roughly the rocket that Anthropic would have needed to make it as a sustainable business. The dark bars on the right show that they already fell off the pace — which is why they probably won’t make it.

Post attachment
Willi Schroll avatar
Willi Schroll@wschroll·Nov 2

Agreed. We foresight people cannot predict the 1 future. We think in scenarios, in many (n>1) futures. Prepare for the worst in polycrisis. Redesign your risk management as ai-supported hackers are on the rise & might soon attack essential infrastructures.

People should know the real reason why AI cyberattacks are growing (and will likely get worse):

0:59
Willi Schroll avatar
Willi Schroll@wschroll·Oct 31

Show up, body language experts. I guess that she already has nightmares about AI Scenarios; but she is forced “by professionality” standards (ie peer group conformism & career pressure) to promote the AI dreamland narrative 🥳

Mary Phuong of DeepMind on AI loss of control risk. "The last 2 minutes gave me chills. A significant amount of people saying that AI companies are paying people to spread fear about these agents. I have never really felt like [loss of control] was an actual thing to care about. But seeing Mary and how her face trembles, made me reconsider my stand. She's either a Hollywood-level skilled actress, one of the best, or we should be scared. For real." "Seems like Mary is already aware of more than she can admit and it haunts her..."

2:22
Ulf Poschardt avatar
⟳ Willi Schroll retweeted
Ulf Poschardt@ulfposh·Oct 30

Wir, die bürgerliche Mitte, wir sind nicht unversöhnlich genug.

10:02
Claude avatar
⟳ Willi Schroll retweeted
Claude@claudeai·Oct 28

New in Claude for Legal: a gallery of ready-made artifacts for contract review, regulatory horizon scans, cite checks, and more. Start from the built-in example, add your own documents and connectors, and share it with anyone across the business.

1:13
Willi Schroll avatar
⟳ Willi Schroll retweeted
Willi Schroll@wschroll·Oct 28

Berlin ist verloren. #Lanz Krach Grillmeister. "Ich werde nicht dafür werben" 🤡 "hab ich doch immer gesagt" 🤡 Hat irgendjemand charakterlose Witzfigur gesagt? 🤡

Willi Schroll avatar
Willi Schroll@wschroll·Oct 26

Verification system. 4:00f it is open source; the prompt is on github 6:00f we are building for agents, not for humans 🫡

AI that verifies your AIs. @dshekhar17, founder of @reticle_hq tells me why AI developers need a verification system. Here's my AI telling you what it learned by watching this video: **Your AI coding agent just wrote 1,000 lines of code overnight. Great! …But who's checking if any of it actually works? 🤔** Just sat down with Dibyanshu, founder of **Reticle** (reticle.sh) — the verification layer for AI coding agents. Here's why this matters: 🔹 **The problem:** We let agents generate massive amounts of code, but humans are still the bottleneck — manually verifying whether the output actually works. That defeats the whole point. 🔹 **The solution:** Reticle instruments every interactive element in your codebase, captures runtime signals, and lets your agent *verify its own work*. After each feature build, it pauses the agent, runs verification, passes evidence back, and the agent decides: ship it or keep iterating. 🔹 **How to get started:** Literally just give your AI agent the GitHub link and say "install Reticle." The agent handles the rest — no handholding required. 🔹 **Works with any LLM:** OpenAI, Anthropic, Gemini, Groq — as long as the harness is MCP-compatible (Claude, Manus, OpenCode, etc.). 🔹 **Pricing:** Open source and **free forever** for indie developers. $49/mo for heavier infrastructure usage. 🔹 **Traction:** 6,000–7,000 users and 120K npm downloads/month — and growing. 🔹 **The founder story:** Dibyanshu built this because he gave his agent a prompt before bed, woke up, and found it had broken every single feature. Nothing worked. He tried every existing tool — none delivered. So he built his own. 🔹 **The big vision:** "AI agents should be our developers now. We should not be the ones checking their code." Agentic IDEs are going to transform heavily, and Reticle is building the deterministic infrastructure to make that safe. Follow Dibyanshu on X and LinkedIn, check out Reticle at **reticle.sh**, or just search "Reticle" on Google to find the GitHub repo.

15:55