New policy brief: Legal reform can get bogged down by outdated or cumbersome regulations. Our latest brief with Stanford RegLab scholars presents an AI tool that helps governments—such as the @SFCityAttorney
—identify and eliminate such “policy sludge.” https://
hai.stanford.edu/policy/cleanin
g-up-policy-sludge-an-ai-statutory-research-system
…
RESEARCH
-

AI Tool Helps Governments Eliminate Outdated Policy Regulations
By
–
-
LLMs Saved AI From Surveillance and Ad Targeting
By
–
people don't realize it, but LLMs saved AI if we didn't stumble into general intelligence most people would still be working on vision, which was most useful for human surveillance and ad targeting.. inherently evil technology
-

GPT-04 Mini: Fast reasoning and STEM excellence
By
–
GPT-04 Mini/High: Quick Thinker & Mathematician Quick reasoning (04 Mini) and STEM excellence (04 Mini High) Faster than GPT-03, more reliable than GPT-40 Ideal for fast logic puzzles, STEM tasks, and real-time chatbot responses.
-

GPT-03 Pro: Oracle Model for High-Stakes Accuracy
By
–
GPT-03 Pro: The Oracle (Ultimate Accuracy) Exceptional accuracy, comprehensive reasoning Extremely slow, costly Use when stakes are very high: audits, formal proofs, critical accuracy.
-

GPT-4.5: The Wordsmith Creative AI Model Capabilities
By
–
GPT-4.5: The Wordsmith (Creative Tone) Exceptional emotional and sensory writing Smooth flow, engaging content Weak at logic, math, detailed research Best suited for marketing copy, creative writing, and persuasive content.
-

Deep Research GPT-03+: Scholar Mode for Academic Literature
By
–
Deep Research (GPT-03+): The Scholar (Real-world Evidence) Detailed mini literature reviews Extensive citations and real-world data Very slow, capped usage Ideal for academic work, in-depth blog posts, reports needing verified data.
-

GPT-03 Professor: Advanced Reasoning AI Model Released
By
–
GPT-03: The Professor (Reasoning Pro) Structured, detailed answers Deep reasoning, backed by sources Slower response time Perfect for research, logical reasoning, nuanced topics.
-
Open Source Strategy as Path to AGI Talent Attraction
By
–
If you assume the models will go closed source as soon as they get AGI-ish, then it makes more sense (you attract more talent, prestige, and free research going open source in the meantime).
-
Why Big Tech Invests Billions in Open Source LLMs
By
–
I've done academic research on open source & I am confused about the case for firms spending billions making open LLMs. In most open source, companies gain ecosystem benefits: people build on their platform, they sell ancillary services, etc. Doesn't work for LLMs at that cost.
-

India-AI Impact Summit 2026: Stakeholder Consultation on Research and Governance
By
–
On 8th June 2025, the fourth Stakeholder Consultation for the India-AI Impact Summit 2026 convened 50+ distinguished researchers and thought leaders from India’s academic ecosystem. The dialogue explored how cutting-edge research can help us understand, govern, and leverage