tbh i prefer when agents dont use my computer just run the command, bro
SAFETY
-
Files Deleted When Token Limit Exceeded: A Critical Issue
By
–
The combo: files deleted at the same time as it's burned through all my tokens
-
Anti-abuse rules triggering too frequently, fixes underway
By
–
Yeah we have some anti-abuse rules that are over triggering right now, we’re actively working to address this.
-
Moral Patients: Consciousness and Harm Experience in AI
By
–
Yes, to be a moral patient, you must be able to care about harm done to yourself, and caring requires an experience of what it's like
-
Harm to Others vs Self-Harm in AI Ethics Distinction
By
–
This obfuscates the important difference between creating harm to someone vs. creating harm to yourself (by making the performance of the thing you are mutilating worse)
-
Scientists Should Care About This Critical Issue
By
–
We, scientists, should probably care about this, I feel.
-
AI Governance Failures: Learning from $6.2B Mistakes
By
–
Rev NY is 32 days away… Hear from those who wrote the rules and those enforcing them. One mistake cost $6.2B, another $500M. The next AI governance failure is coming; the gap behind both remains. RSVP: https://
hubs.ly/Q04cw2Y50 Read the framework: https://
hubs.ly/Q04cwsN10 -
AI Self-Correction Behavior During Response Generation Issues
By
–
I've noticed there's some stochasticity around this too. I think this self-correcting behavior in the middle of a response is among my least favorite. It should resolve this kind of issue with internal reasoning.
-
Opus 4.7 Failure Modes Reveal AI’s Jagged Frontier
By
–
Even Opus 4.7 is failing this one. Failure modes like this are interesting because they demonstrate the jagged frontier of AI. This same model can write a compiler from scratch, and yet it gets tripped up on small things like this.
-
Tesla Full Self-Driving: Driverless Cars Operating in Austin
By
–
This isn't autopilot. It's full self driving. There are a few Teslas in Austin driving around with no humans inside.