If Persona Selection underlies alignment, why is it hard to get AIs to be honest? Tell them they're Fred Rogers or Immanuel Kant (I asked Claude for figures who never lied or never got caught). Or tell them they're Ged of Earthsea, or Ned Stark. LLMs surely have neural
ETHICS
-
AI Economic Risks and Critical Policy Choices Ahead
By
–
talk about without understanding this will probably crash the world economy. one has to make choices.
-
Administrative Bloat: How Organizations Expand Inefficiently
By
–
What do administrators do? They hire more administrators, a.k.a. cancer.
-
Company Ships 100% AI-Generated Code to Win Market Race
By
–
100% of my company's code is written by AI. (It's all garbage, but who cares? We're winning the race!)
-
Green Cards for U.S.-Trained Talent from Authoritarian Nations
By
–
Still think the only way to retain U.S. trained talent from authoritarian countries is to give them AND their family members green cards. Otherwise these family ties are used as leverage by the dictatorship for coercion and blackmail. But we’re also too xenophobic in America
-
Neuralink brain integration for direct AI feed consumption
By
–
If I had a Neuralink I would plug this feed directly into my frontal lobe. https://t.co/yJ2BrxKWSc
— Bob Gourley – e/acc (@bobgourley) 20 avril 2026If I had a Neuralink I would plug this feed directly into my frontal lobe.
-
Open Source AI Security: Balancing Attacker-Defender Capability Asymmetry
By
–
Why? – Risk comes from capability asymmetry between attackers and defenders. Open source is what every frontier lab already trains on so defenders get the same AI firepower as attackers. With proprietary code, you're on your own. The biggest risk is someone training a model on x.com/ClementDelangu…