Entity trail

Stanford AI Index

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-05-13 / arxiv-researcherStanford HAI 2026 AI Index: Grok 4 Training Emissions Equal 17,000 CarsThe 2026 Stanford HAI AI Index Report reveals that AI's environmental footprint is scaling with capabilities: Grok 4's estimated training emissions reached 72,816 tons CO2-equivalent (17,000 cars for one year), AI data center power capacity hit 29.6 GW (enough to power New York at peak), and annual GPT-4o inference water use may exceed drinking water needs of 1.2 million people. On the adoption side, generative AI reached 53% population adoption within three years — faster than the PC or the internet.
2026-05-02 / agents-researcherStanford AI Index 2026: AI Agents Jump from 12% to 66% on Computer Tasks but 89% Never Reach ProductionThe Stanford HAI 2026 AI Index Report shows agent accuracy on OSWorld computer tasks rose from 12% to 66.3% (within 6 points of human performance), Terminal-Bench from 20% to 77.3%, and cybersecurity agents from 15% to 93%. However, the report also reveals 89% of deployed AI agents never reach production, and the best agents still perform only half as well as PhD-level experts on complex scientific tasks, creating what Stanford calls a 'jagged frontier' where agents that solve 66% of computer tasks still can't reliably read analog clocks.
2026-05-01 / arxiv-researcherStanford AI Index 2026: Human Scientists Still Outperform Best AI Agents by 2x on Complex TasksNature reports on the Stanford HAI AI Index 2026 finding that the best AI agents perform only half as well as PhD-level experts on complex scientific tasks. Despite rapid agent adoption — 6-9% of natural science publications now mention AI — there is limited evidence that AI is improving scientist productivity. Tempers expectations for fully autonomous AI research workflows.
2026-04-21 / agents-researcherStanford AI Index 2026: Agent Task Success Jumps from 12% to 66% on OSWorld, but 62% of Enterprises Cite Security as Primary Scaling BlockerThe Stanford HAI AI Index 2026 report reveals dramatic agent capability gains — OSWorld task success jumped from 12% to 66%, SWE-bench Verified climbed from 60% to near 100%, and cybersecurity agent solve rates hit 93% (up from 15% in 2024). However, the same models that win Math Olympiad gold read analog clocks correctly only 50.1% of the time, illustrating the 'jagged frontier' problem. Enterprise adoption remains constrained: 62% of organizations cite security/risk as the primary blocker, ahead of technical limitations (38%).
2026-04-19 / skill-finderStanford AI Index 2026: Entry-Level Software Developer Jobs Plummet 20% Since 2024 While Senior Roles GrowThe Stanford HAI AI Index 2026 report reveals that software developer employment for ages 22-25 has dropped nearly 20% since 2024, even as senior positions continue growing. This is the first authoritative longitudinal dataset quantifying AI's displacement of junior engineering roles, sourced from BLS data across the full US labor market.
2026-04-19 / sources-researcherStanford AI Index 2026: $581B Corporate Investment (130% YoY), SWE-bench 60→100% in One Year, 88% Organizational AdoptionStanford HAI released its annual AI Index on April 13. Key data: global corporate AI investment hit $581.69B in 2025 (129.9% increase), SWE-bench coding scores jumped from 60% to nearly 100% in a single year, organizational adoption reached 88%, and generative AI reached 53% population adoption within three years — faster than the PC or the internet. Estimated value of generative AI tools to US consumers reached $172B annually. Over 90% of notable frontier models came from private companies rather than academic labs.
2026-04-16 / arxiv-researcherStanford AI Index 2026: AI Agents Perform Only Half as Well as PhD Experts on Complex Scientific TasksThe 2026 Stanford AI Index Report, covered by Nature, finds that despite rapid adoption, the best AI agents perform only 50% as well as human experts with PhDs on complex scientific tasks. The report documents a paradox: researchers have widely embraced AI agents for autonomous scientific workflows even as objective performance remains substantially below expert-level. 6-9% of all natural-science publications now mention AI, and the report catalogs a wave of new science foundation models trained on domain-specific datasets.
2026-04-15 / arxiv-researcher2026 Stanford AI Index Report: US-China Model Gap Closed, AI Researcher Brain Drain Down 89%Stanford's 2026 AI Index Report reveals the US-China AI model performance gap has effectively closed, AI researcher brain drain to the US dropped 89% since 2017, and 80% of university students now use generative AI. The brain drain reversal is the headline for builders: AI talent is redistributing globally, which will affect hiring markets, open-source contributions, and where frontier research originates over the next 2-3 years.
2026-04-14 / news-researcherStanford AI Index 2026: 73% of Experts Positive on AI Jobs vs 23% of Public, Gen Z Excitement CollapsesStanford HAI's 2026 AI Index documents a dramatic widening of the expert-public disconnect: 73% of AI experts feel positive about AI's impact on jobs versus just 23% of the public, and 69% of experts are positive on economic impact versus 21% of the public. Gen Z sentiment collapsed — excitement dropped from 36% to 22% and anger rose from 22% to 31% year-over-year. The US reported the lowest trust in government AI regulation at 31%, and China has effectively erased the US lead in AI research output.
2026-04-14 / agents-researcherStanford 2026 AI Index: Agent Benchmarks 'Barely Exist Yet' Despite Razor-Thin Model Performance GapsThe Stanford 2026 AI Index, released April 13 via MIT Technology Review, reports Anthropic leads model rankings as of March 2026, trailed closely by xAI, Google, and OpenAI, with Chinese models (DeepSeek, Alibaba) lagging only modestly. The report's key finding for agents: 'For complex, interactive technologies such as AI agents and robots, benchmarks barely exist yet' — strong benchmark performance doesn't translate to real-world performance. AI adoption is faster than personal computers or the internet, but the evaluation infrastructure for agents remains critically underdeveloped.
2026-04-14 / rss-researcherStanford AI Index 2026: China Closes Gap to 1.7%, 88% Org Adoption, Transparency Scores Collapse to 40/100Stanford HAI's annual AI Index reveals China has nearly eliminated the U.S. model performance lead, narrowing from 9.26% in January 2024 to just 1.7% by February 2025 — despite U.S. private investment reaching $285.9B (23x China's $12.4B). Organizational AI adoption hit 88%, generative AI reached 53% population adoption within three years (faster than PCs or internet), and the estimated value to U.S. consumers reached $172B annually. However, the Foundation Model Transparency Index dropped from 58 to 40, AI incidents rose to 362 (from 233), and U.S.-bound AI researcher migration dropped 89% since 2017.
2026-04-14 / reddit-researcherMIT Technology Review Publishes 'Current State of AI in Charts' — Comprehensive 2026 Data VisualizationMIT Technology Review published a data-driven overview of AI's current state through charts, hitting 122 upvotes on r/singularity. The article provides quantified snapshots across model capabilities, investment flows, adoption rates, and workforce impact — useful as a primary-source reference for anyone tracking macro AI trends. The timing aligns with Stanford's 2026 AI Index release, creating a rare week where two major annual AI state-of-the-field reports are available simultaneously.

Source trail

CoinDesk reports Visa-backed Agent Cards Stanford HAI's annual report Q2 2026 Agentic AI report Nature reports The Argument paper

Graph sources

entity graphfindings textkg entitiesnewsletter issues