Category intelligence

Social Media Briefing — June 3, 2026

516 current items analyzed and ranked.

Executive synthesis

Social Media Summary

AI economics skepticism dominated discussions today. Gary Marcus went viral arguing AI will eventually 'fall apart' financially, citing commodity tech, no moats, and unsustainable capex. Timnit Gebru echoed criticism of consulting-firm hype cycles.

Policy and institutional moves rounded out the day. Anthropic welcomed a US AI Executive Order and expanded Project Glasswing / Claude Mythos Preview to ~150 more organizations. NeurIPS 2026 announced its Position Paper Track will require substantially human-written submissions, sparking debate on AI authorship norms, while a notable open-model researcher announced their departure from Ai2.

Key Themes

AI Economics and Bubble Skepticism · 18OpenAI Codex Expansion · 10AI Coding Productivity · 9Anthropic Programs · 2AI for Scientific Discovery · 5AI Policy and Regulation · 8Model Evaluation and Access · 6NVIDIA GTC Taipei and AI Infrastructure · 11AI Governance and Policy · 2Agentic AI and Skills · 8

Primary evidence

Top Ranked Signals

80 score
AI Analysis

Gary Marcus lays out a five-point thesis that AI will eventually fall apart financially because everyone builds the same commodity tech with no moat, preventing monopoly pricing and forcing price wars that make returns modest relative to spending.

Why things will eventually fall apart: 1. Everybody, even Google, seems to be treating AI as if it were some kind of winner take all competition like web search was, in which Google taking over 95% 2. But everybody is building essentially the same technical solution with essentially the same data, so there is no moat. 3. If there is no moat, nobody is going to take 90% of the market. 4. With no clear winners, nobody can charge monopoly prices; instead, you get price wars and commodity prici
AI economicsno moatinvestment bubblecompetition
80 score
AI Analysis

OpenAI launches Codex Sites, letting Codex turn work, ideas, and plans into shareable interactive websites or apps, rolling out to Business and Enterprise plans.

Building apps has never been easier. With Sites, Codex can turn your work, ideas, and plans into an interactive website or app your team can explore, use, and share with a URL. Rolling out to Business and Enterprise plans, before expanding more broadly. t.co/fF17Y2EzCP
OpenAI Codexno-codeagentsproduct launch
78 score
AI Analysis

Ethan Mollick describes a study where Gemini 2.5 answered law professors' office-hours questions and beat human professors with a 75 percent win rate while being rated less harmful, with newer models doing even better.

Law professors wrote questions they were asked during office hours. Gemini 2.5 & humans answered them then other law professors blindly judged the results: -Gemini had a 75% win rate vs. professors -Gemini's answers were rated LESS harmful than humans -Newer models do even better
AI evaluationlawhuman vs AIresearch
78 score
AI Analysis

Anthropic publicly welcomes a US Executive Order on AI and signals intent to collaborate with the White House on implementation.

This Executive Order is an important step in strengthening America’s leadership in AI. We look forward to collaborating with the White House to support its implementation. t.co/ZwDimPrp3t
AI policyregulationAnthropic
75 score
AI Analysis

Ethan Mollick summarizes a big paper using GitHub data showing autocomplete tools led to 2.2x more code, local agents 7.4x, and remote coding agents 17.3x, but human bottlenecks meant releases only rose 30 percent.

Big paper on AI coding agents using Github & other data The auto-complete tools (Copilot) led to 2.2x more code, local agents like original Claude Code led to 7.4x, & current remote coding agents 17.3x(!) But human bottlenecks in coding means actual releases "only" went up 30% t.co/GiXEr94s4i
AI codingproductivityresearchcoding agents
72 score
AI Analysis

DeepMind introduces Co-Scientist, a Gemini-based multi-agent system that generates, debates, and evolves novel scientific hypotheses.

We believe AI can be a dedicated research partner to help discover the next breakthrough. Enter Co-Scientist: our latest Gemini-based multi-agent system that can generate, debate and evolve novel hypotheses for complex scientific problems 🧵
AI for sciencemulti-agent systemsGoogle DeepMindCo-Scientist
72 score
AI Analysis

Summarizes a GitHub-data study: auto-complete tools yielded 2.2x code, local agents 7.4x, remote coding agents 17.3x, yet human bottlenecks meant releases rose only about 30%.

Big paper on AI coding agents using Github data. The early auto-complete tools (like Copilot) led to 2.2x as much code, local agents like original Claude Code led to 7.4x, & current remote coding agents 17.3x(!) But human bottlenecks in the coding process means actual releases "only" went up 30%.
AI codingproductivityhuman bottlenecks
70 score
AI Analysis

Anthropic expands Project Glasswing, extending access to Claude Mythos Preview to roughly 150 additional organizations across more than fifteen countries.

We’re expanding Project Glasswing. We’ve extended access to Claude Mythos Preview to approximately 150 additional organizations, based in more than fifteen countries. Read more about this expansion and our future plans for Project Glasswing: t.co/QrtHSBdRbh
AnthropicClaude Mythosaccess programsProject Glasswing
70 score
AI Analysis

Announces his departure from Ai2 after 2.5+ years building Olmo, Tulu and other open projects, planning to stay in open model and open science work, with personal life updates and a team note.

My time at Ai2 / @allen_ai has come to an end. Ai2 is a wonderful place. The last 2.5+ years building Olmo, Tulu, and other projects will be one of the peaks of my entire career. I'm extremely thankful for my teammates and the open community who made this work possible. For me, it's time to try something different. I will still be working in the open model & open science spaces (more news on that soon). In the meantime I'll be spending a few months learning, chatting with a broader network,
career newsopen modelsAI researchAi2
70 score
AI Analysis

NeurIPS announces that its 2026 Position Paper Track will require papers to be substantially human-written, with AI limited to copy-editing.

This year, the NeurIPS 2026 Position Paper Track made the decision to require that all papers be substantially human-written, with AI used for only copy-editing or similar peripheral changes to the main text! For more details, please check our blogpost: t.co/wrWuMQJwrx
AI policyacademiaAI authorshipNeurIPS
70 score
AI Analysis

Reports a study where Gemini 2.5 answered contract-law questions with a 75% win rate versus law professors and rated less harmful, with newer models doing better.

Apparently superhuman performance from the AI in answering questions about contract law in this paper, as judged by law professors compared to the answers of their peers -Gemini 2.5 had a 75% win rate vs. professors -Gemini's answers were rated LESS harmful than humans -Newer models do even better
AI in lawAI benchmarksprofessional automation
68 score
AI Analysis

OpenAI announces an expansion of Codex plugins that turn it into a role specialist with one install, giving access to 62 apps and 110 skills across sales, analytics, creative, design, and investing.

We’re making Codex more useful for your work by expanding plugins beyond individual tools. These plugins turn Codex into a specialist for a specific role with a single install, no coding required. Codex can access 62 popular apps and 110 skills for work across sales, data analytics, creative production, product design, and public equity investing. t.co/nunrYP2uMI
OpenAI Codexpluginsagentsenterprise AI