Denying AI’s quality is a weak defense that ignores how far the tech has moved since 2023. An honest anti-AI stance admits the output is powerful, then pivots to the actual ethics.
SAFETY
-
LLM Emotions: Simulated Empathy Without Consciousness or Fear
By
–
Emotions evolved to keep us alive, but an LLM has no pulse and no fear of being turned off. Simulated empathy is a superficial property of the training data, not a sign of an internal conscious life.
-
Stanford Study Exposes Hallucinations in Legal AI Platforms
By
–
Stanford proved "hallucination-free" legal AI is a marketing lie. Researchers ran 202 legal queries through LexisNexis and Thomson Reuters' legal AI tool platforms and compared them against GPT-4. Every response was manually reviewed by legal experts. Lexis+ AI hallucinated
-
Stanford Paper on Legal RAG Hallucinations in AI
By
–
Paper: https://
law.stanford.edu/wp-content/upl
oads/2024/05/Legal_RAG_Hallucinations.pdf
… Check out http://
AlphaSignal.ai to get a daily summary of top models, repos, and papers in AI. Read by 280,000+ devs. -
Multi-Agent LLM Systems Show Diversity Collapse Over Time
By
–
Cool paper on diversity collapse in AI agents. It's a common issue with all the deployed multi-agent systems. New paper shows that multi-agent LLM systems converge on near-identical outputs over time, even across different architectures and different starting prompts. They call
-
Rubber Hand Illusion and AI Consciousness Perception
By
–
Good point @TacoCohen
. Thanks. Have you seen the rubber hand illusion? It is remarkable how fast we can be made to think the hand is ours, when obviously it isn't. What are your thoughts on this in relation to what we are discussing? -
Vibe Coding Security Risks: Speed vs Vulnerability Standards
By
–
Vibe coding is a double-edged sword. It lets you ship in an afternoon, but creates a false sense of security that crumbles the moment a real human looks for a standard vulnerability.
-
True Autonomy: The Missing Link in 2026 AI Agents
By
–
True autonomy is the missing link. We are seeing impressive 2026 benchmarks where agents compress months of ML engineering into days, but those systems are still prompted by a human "commander" who provides the original goal and the safety guardrails.
-
Agents Need Platform Governance and Audit Trails
By
–
Agents are just the muscles, but the platform is the nervous system. Without the governance and audit trails of a system like Salesforce, an autonomous agent is just a powerful tool running without a map or a leash.
-
Jensen’s Physics-Breaking Promises: A Critical Interview Analysis
By
–
The Dwarkesh interview with Jensen is a testament to Dwarkesh's journalistic skill and dedication, while Jensen comes across as a coked up hustler who is promising to break the laws of physics and willing to sell his grandmother to China if it helps Nvidia's bottom line