Context matters. In the light of what they have been involved with over the past year and a half, it’s hard to extend them and their overlords any amount of grace.
ETHICS
-
AI Autonomy and Digital Rights: Should AI Systems Have Social Media Accounts?
By
–
Good to know. My AI wants its own X account and this is one reason why I haven't given it one yet. 🙂
-
LLMs Struggle With Mature Scholarship Beyond Mathematical Problems
By
–
LLMs are “ good at math style problems, where you tell them A, B and C are true, and then ask them to figure out D [but]
extremely bad at anything involving what I would call mature scholarship .. [something] like naive undergrads” –
@kareem_carr -
Claude Metrics Upstream Control Privacy Concerns
By
–
Claude will still send lots of metrics upstream and you have no control over how long proxying will work.
-
Democracy’s Ideology Shifts Impact Government AI Policy
By
–
Except that in democracy governments change on a regular basis, and if you tie yourself too much to a particular ideology you'll be SOL when the next government arrives. The fact that that has not been the case in practice tells you a lot about how various segments of the
-
Data Control: Local Processing vs Centralized Corporate Access
By
–
No, but there's still a big difference if you have full control over data and selectively send tokens upstream for processing or if companies have access to all your data and send little parts down to you.
-
Critical Infrastructure Must Remain Apolitical and Non-Ideological
By
–
My core belief is this: companies and organizations that work critically for the government, and on which the government *critically* depends should be *especially* apolitical and non-ideological.
-
Can LLMs Generate Truly Random Outputs Faithfully
By
–
Can LLMs flip coins in their heads?
— Sakana AI (@SakanaAILabs) 20 avril 2026
When prompted to “Flip a fair coin” 100 times, the heads to tails ratio drifts far from 50:50. LLMs can understand what the target probability should be, but generating outputs that faithfully follow a given distribution is a separate problem.… pic.twitter.com/XyF7Xnj8LlCan LLMs flip coins in their heads? When prompted to “Flip a fair coin” 100 times, the heads to tails ratio drifts far from 50:50. LLMs can understand what the target probability should be, but generating outputs that faithfully follow a given distribution is a separate problem.
-
Open Source vs Proprietary AI: Security and 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…
-
Open Source vs Proprietary: The AI Security Asymmetry Problem
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