Profile avatar
framant.bsky.social
Manager, speculative fiction lover, intrigued by foresight and future studies (and Belgian beer). I am liberal (if you know Herbert Spencer and Mill, that kind of liberal).
33 posts 20 followers 51 following
Prolific Poster

Supreme court justice calls out lawyer for citing AI-generated 'phantom rulings' flip.it/LutB9c

Supreme court justice calls out lawyer for citing AI-generated 'phantom rulings' flip.it/LutB9c

Microsoft Unveils Majorana 1, the World’s First Quantum Chip Powered by a Topological Core flip.it/1Ovee8

We are starting to see what "AI will accelerate science" actually looks like. This Google paper describes novel discoveries being made by AI working with human co-scientists (something I think we have all been waiting to see), along with an early version of an AI scientist for wider use.

Changing conspiracy theory beliefs is very hard, but a replicated finding shows a short chat with GPT-4 changes people’s belief in conspiracy theories for the long term. Why? It isn’t rhetorical tricks, it's that AI provides relevant facts and evidence tailored to each person's specific beliefs.

Two scientists (Vopson and Lepadatu) at the Jeremiah Horrocks Institute in the UK have formulated what they call the second law of infodynamics, a principle that could offer the first tangible evidence supporting the #simulationhypothesis. flip.it/Uc_8.s

DeepSeek R1 Replicated for $30 By Researchers at UC Berkeley flip.it/T0G7Hh

Decoding the Knowledge: How Taxi Drivers Think Differently From AI - Neuroscience News flip.it/xeJl2I

Google search ends in frustration after AI images dominate results: 'The internet is dead' flip.it/JuC7EP

www.livescience.com/technology/a...

AI Agents Are Becoming More Humanlike — and OpenAI Is Launching a New One in January. This could be the next big thing: flip.it/NeZ4tS

€ 22.700 NETTI donati alla #CasaDelleDonne di Bologna. Grazie a chi ha partecipato a questa meravigliosa raccolta fondi per #Bologna5 per ripristinare la grafica sul taxi #Bologna5 dopo il vile atto vandalico. A metà gennaio verrà presentata la macchina allestita e sarà ancora più bella ❤️🌺🏅

Physicists Discover 'Quantum Embezzlement' Could Offer Infinite Source of Entanglement flip.it/PDTxfZ

Reasoning AI models require training on human reasoning. One of the real gaps in pushing forward these models is going to be the old problem of how to figure out how to get experts to explain what they do. AI keeps bumping up against our limited knowledge of how expertise works

With previous technologies, younger people launching startups had big advantages as they were outside firms & early users In AI, they may not. Why? Because AIs produce good-enough work that novices can’t tell how useful they are or (as we found in our paper) risks & how AI fits into organizations.

Quantum teleportation is coming… how will this change the world… flip.it/_DN-eZ

Google Street View camera captures highly suspicious act, leading to arrests…

APpaREnTLy THiS iS hoW yoU JaIlBreAk AI flip.it/xA2WfY

1/20

"Allowing a certain name or phrase to always break ChatGPT outputs could cause a lot of trouble down the line for certain ChatGPT users, opening them up for adversarial attacks and limiting the usefulness of the system" arstechnica.com/information-...

ChatGPT solving problems like a manager would do: vm.tiktok.com/ZNe37gFob/

Getting started with prompting AI in three (free) posts: 1) Easy prompting for most: www.oneusefulthing.org/p/getting-st... 2) Approaches to prompt "engineering": www.oneusefulthing.org/p/innovation... 3) A bit on how LLMs work that is relevant for prompting: www.oneusefulthing.org/p/thinking-l...

Large language models (#LLMs) consistently outperformed neuroscience experts in predicting experimental outcomes, with an average accuracy of 81.4% compared to human experts’ 63.4%. The study suggests a future where LLMs assist researchers in designing studies and making discoveries faster.

A post on the mysterious viral spread of “Black Friday” across the globe since 2005 The craziest cases are Venezuela & Iran!! The post frames the puzzle & the beginning of a theory inspired by @emollick.bsky.social’s classic take on Moore’s Law Comments please! open.substack.com/pub/ezrazuck...

Europe is late on #AI: Gallup study involved leaders from 40 large European enterprises representing $1.3 trillion in revenues and 3 million workers... only 13% of European workforce is engaged (-10% vs global average). Our competitiveness as a Region is in danger. www.gallup.com/workplace/65...

Fascinating: In 2-hour sprints, AI agents outperform human experts at ML engineering tasks like optimizing GPU kernel. But humans pull ahead over longer periods - scoring 2x better at 32 hours. AI is faster but struggles with creative, long-term problem solving (for now?). metr.org/blog/2024-11...

AI can help learning... when it isn't a crutch. There are now multiple controlled experiments showing that students who use AI to get answers to problems hurts learning (even though they think they are learning), but that students who use well-promoted LLMs as a tutor perform better on tests.

The Future of AI Revealed by Microsoft and OpenAI : AI Scaling Laws

Retrieval, Augmentation and Generation made easy.

Crazy interesting paper in many ways: 1) Voice-enabled GPT-4o conducted 2 hour interviews of 1,052 people 2) GPT-4o agents were given the transcripts & prompted to simulate the people 3) The agents were given surveys & tasks. They achieved 85% accuracy in simulating interviewees real answers!

A really intriguing set of findings of AI researchers that needs more exploration, replication, and expansion: AI agents given personalities and backgrounds, and placed into a virtual formal organization (with CEOs, VPs, etc.) outperform normal AI in doing complex tasks.

Robot trained on surgery videos performs as well as human docs www.futurity.org/robot-surger...

Former Google X employees come out of stealth with #TwinMind, an AI app that hears and remembers everything about you. Digital twin replicas of yourself are finally here? www.businessinsider.nl/former-googl...

Articolo molto interessante. Ho dei dubbi con la definizione di “liberals”, tuttavia l’analisi di come le dittature stiano vincendo nel cuore degli elettori è corretta secondo me: i dittatori promettono una realtà migliore a scapito di qualcun altro, i liberal ti informano che dovrai fare sacrifici.

Ettore Majorana, fisico teorico, si diplomò al liceo classico (al Tasso a Roma, ma ho visto alcune fonti secondo cui si diplomò a Catania).

Your digital twin will coach you, answer questions and email of lower importance, draft documents, videos and posts for you. Should we start training our own digital twin, a proprietary one?

After all, flying cars could be a #future: “China just debuted a 'flying' car and it could release as soon as 2026.”