yes – was more meant as figure of speech obvious genuinely – this is the most exciting thing i have seen a long time in that space true 3d worldmodels – not 2d hacks
RESEARCH
-
AI Reasoning and Truth: When Better Thinking Doesn’t Guarantee Honesty
By
–
Think, Then Lie: When AI Reasoning Doesn’t Guarantee Truth
— Satya Mallick (@LearnOpenCV) 30 mars 2026
In this episode of Artificial Intelligence: Papers and Concepts, we explore “Think, Then Lie,” a concept that challenges a key assumption in modern AI—that better reasoning always leads to more truthful outputs. As… pic.twitter.com/6LPmHiinPBThink, Then Lie: When AI Reasoning Doesn’t Guarantee Truth In this episode of Artificial Intelligence: Papers and Concepts, we explore “Think, Then Lie,” a concept that challenges a key assumption in modern AI—that better reasoning always leads to more truthful outputs. As
-
Yoav Shoham Named AAAS Fellow for Agentic AI Contributions
By
–
Congratulations to our Co-founder and Co-CEO @yshoham, named an @aaas Fellow for pioneering contributions to agentic AI. 28 years as a professor at @Stanford, now building the future of enterprise AI at @AI21Labs. Read the full story: news.stanford.edu/stories/20…
-
Vector Databases Explained in Three Levels Difficulty
By
–
Vector Databases Explained in 3 Levels of Difficulty https://
machinelearningmastery.com/vector-databas
es-explained-in-3-levels-of-difficulty/?utm_source=dlvr.it&utm_medium=twitter
… -
Scene Generation Technology: Applications in Robotics, Gaming, VR
By
–
this is really interesting! does this mean it builds a scene "meter by meter" (tokens as splats done over distance)? if yes there is tons of usecases for this – eg worldmodels for robotics, interactive games, VR, or simply generation of scenes, etc etc all combined with
-
AI Agents Accelerating GLP-1 Drug Development at Novo Nordisk
By
–
Do you want to understand how AI is already changing medicine? Novo Nordisk is using AI agents to accelerate its GLP-1 drug pipeline, shaving *weeks to months* off clinical trials, potentially worth hundreds of millions in faster time-to-market. The Ozempic maker uses agents
-
Fix RAG hallucinations by protecting tables and structured content
By
–
Your RAG pipeline answers everything correctly. Except anything from a table. Pricing data. Comparison charts. Structured specs. Ask about any of these and the answer is either wrong or completely made up. The model isn't hallucinating because it's bad. It's hallucinating because it never saw the full table. When you chunk documents, you split them by a fixed token count. The splitter doesn't understand what it's cutting through. It just counts and splits. So your pricing table gets sliced in the middle. Half the rows in one chunk, half in another. The model receives an incomplete table and fills in the blanks on its own. Same thing happens with code blocks and any structured content. The moment you start treating tables and code as protected blocks and never let the chunker split through them, the accuracy on structured questions jumps. Same documents. Same model. Same prompt. Just keep structured content whole. I wrote a free playbook (its on git, no email wall or anything) that covers this decision framework (and 6 others like model selection, evaluation, and production optimization) as simple find-your-situation, follow-the-row tables. Link in the first comment.
-
New Clinical Practice Guidelines on AI in Medicine
By
–
New clinical practice guidelines on this axis @CMAJ
-
All-Terrain Wheeled Quadruped: Hybrid Mobility Innovation
By
–
Hybrid Mobility Unleashed: The All-Terrain Wheeled Quadruped
— Ronald van Loon (@Ronald_vanLoon) 30 mars 2026
via @ZappyZappy7
#Robotics #RPA #ArtificialIntelligence #Innovation #Technology pic.twitter.com/xZXDAoiHuTHybrid Mobility Unleashed: The All-Terrain Wheeled Quadruped
via @ZappyZappy7 #Robotics #RPA #ArtificialIntelligence #Innovation #Technology -
Google’s AI debates itself without training
By
–
HOLY SHIT… Google just proved that reasoning models aren't thinking, they're arguing. > DeepSeek-R1 spontaneously generates internal multi-agent debates inside its own chain of thought. Nobody trained it to do this. Reinforcement learning for accuracy produced it