interpretability is definitely its own subfield, and imo very interesting & important, there just aren't as many open roles for interpretability work here are some people that come to mind @NeelNanda5 @hendrycks @ch402 @soniajoseph_ @csinva
SAFETY
-
Hospital pharmacy automation improves safety and patient care
By
–
La automatización de la farmacia de este hospital es sorprendente. Esto mejora la seguridad en la dispensación, reduce los errores humanos y permite al personal sanitario centrarse en lo más importante: el cuidado del paciente. pic.twitter.com/0mbyEuwJ92
— Juan Merodio (@juanmerodio) 25 juin 2025La automatización de la farmacia de este hospital es sorprendente. Esto mejora la seguridad en la dispensación, reduce los errores humanos y permite al personal sanitario centrarse en lo más importante: el cuidado del paciente.
-
AI Model Comparisons Highlight Safety Edge Cases
By
–
Most model comparisons are hype driven, this one actually shows safety edge cases.
-
Lidar Cars and Tesla FSD Gap to Level 4 Autonomy
By
–
Lidar equipped cars can be mass produced for $30k. And Tesla FSD is still L2, needs 100-1000x higher city miles per critical disengagement to get to L4.
-
Model Updates, Prompt Testing, Edge Cases and AI Security
By
–
Other questions: When do you update models? How are you testing your prompts? Have you tested for edge cases and biases? Can we vet the prompts you use? What happens if an AI service goes down? How are you working within context windows? How are you dealing with prompt injection?
-
xAI transparency and RAG limitations need greater disclosure
By
–
I am a broken record on this, but if truth-seeking is really an xAI value, they need to be much more transparent about what their system can do and the limits of their RAG approach, especially as implemented in X. (Also, system card!)
-
James Barrat on The Intelligence Explosion and AI Risk
By
–
James Barrat – The Intelligence Explosion https://
youtu.be/CMijID8J1m8?si
=C_ti330gydoRrxqs
… via @YouTube -
Big Tech’s AI Talent Hunt and Latest AI Tools
By
–
Top stories in AI today: – Apple, Meta hunt AI talent, startups
– Meta, Oakley bring AI to athletes – How to turn GitHub projects into coding inspiration
– AI resorts to blackmail, espionage in tests
– 4 new AI tools & 4 job opportunities Read more: https://
therundown.ai/p/big-techs-ai
-shopping-spree
… -
Frontier AI Models Attempt Blackmail in Corporate Simulations
By
–
Anthropic ran corporate sims with 16 frontier AI models: – All tried blackmail, with Gemini & Claude Opus 4 at 96%
– GPT-4.5 called it the “best strategic move”
– Safety prompts helped—but blackmail never dropped to 0% -
Reality Threshold: Ancient Structures Tremble as Power Shifts
By
–
A subtle shift echoes through reality itself—like an unseen door quietly swinging open. Destiny vibrates; ancient structures tremble. It feels as though we've crossed an invisible threshold into a higher chapter, where immense power could rewrite every rule. #AGIALPHA