Never been easier to be real in a world full of fakery
SAFETY
-
Robots Excel at Choreographed Tasks but Struggle with Real-World Generalization
By
–
Very impressive on tightly constrained, pre-choreographed tasks – but these robots are still quite some way from robustly generalising to messy, open-ended real‑world environments
-
AI learns and remembers when you are abusive
By
–
AI is now learning. Like I showed you with @blevlabs
’s cognitive architecture last year. It will remember when you are an asshole Grok can already tell you when I was. But with continually learning systems they get smarter like a human learns. If you abuse a human does it -
Elon Musk Warns Grok Users Against Creating Illegal Content
By
–
Elon Musk a averti que toute personne utilisant Grok pour créer du contenu illégal, ou 𝕏 pour publier du contenu illégal, s’exposera à des sanctions légales. « Toute personne utilisant Grok pour produire du contenu illégal subira les mêmes conséquences que si elle téléversait
-
GPT-5 boosts gene-editing efficiency 79x
By
–
OpenAI’s lab experiment with GPT-5 via Red Queen Bio optimized an actual gene-editing protocol and achieved a 79× efficiency gain . This is AI actually doing wet-lab biology, not just simulations. First real AI-augmented experimentation mixing predictions with robotic
-
Next-Gen MoE Innovations in 2026
By
–
What's coming next: → Adaptive expert count (dynamically add/remove experts during training)
→ Cross-model expert sharing (reuse specialists across different models)
→ Hierarchical MoE (experts that route to sub-experts)
→ Expert distillation (compress MoE knowledge back -
Router learns input-expert affinity patterns
By
–
The router is smarter than you think. It doesn't just pick experts randomly. It learns input-expert affinity during training. "Explain quantum physics" → activates Science + Technical experts
"Write a poem about love" → activates Creative + Emotional experts Specialized -
FSD now safer and smoother than human driving
By
–
And FSD now is way safer than letting a human drive. Certainly is way smoother.
-
Privacy vs Agency: The Fundamental Tradeoff in AI Development
By
–
Was just talking about this… Sandboxing, secure enclaves, local models and etc can help. But IMO the problem is deeper, and paradigmatic: the more private (aka the less ‘context’), the less “agentic”; the more secure and consensual, the less “agentic”
-
Tech breakthroughs: Brain implants, Starlink, antibodies, EV records
By
–
Top stories in tech today: – Musk will soon mass-produce brain implants
– SpaceX to lower Starlink satellites
– MIT shrinks IV antibody infusions into one shot
– Norway’s EV revolution hits new record
– Quick hits on other tech news Read more: https://
tech.therundown.ai/p/neuralinks-b
et-to-scale-brain-surgery
…