# AI Overwhelm Lab source ledger Retrieved: **2026-07-29** The lab uses dated local snapshots. It makes no runtime requests to data providers. ## Current-overwhelm opening The opening is intentionally anchored in the current experience rather than replaying 2022-2023 headlines. It combines local model records with recognizable AI product marks, feature notices, plan prices, limits, terminology, unusual AI product labels, and agent categories. The composition is randomized on every visit, but the local source set does not change at runtime. - The 2026-07-30 feed audit removed modulo-based repetition from the logo, glossary, product, connector, token, and interruption groups. - The curated language pool now contains 378 distinct labels before documented model names, glossary records, and 35 distinct logo marks are added. - Text fragments shorter than a meaningful label are not intentionally generated. Agent tiles use pictograms rather than unexplained letter abbreviations. - Product links and names are recorded in `app.js`. - Local SVG product and connector marks are stored in `assets/logos`. - Most monochrome marks were retrieved from the Simple Icons CDN: https://simpleicons.org/ - The OpenAI mark was retrieved from the OpenAI-sourced Wikimedia asset record and remains subject to OpenAI trademark guidance: https://openai.com/brand/ - The Midjourney emblem was retrieved from its Wikimedia asset record and remains subject to Midjourney trademark rights. - The marks identify products being discussed. Their use does not imply sponsorship or endorsement. The dated historical source records remain in `data/hero-events.json` for later timeline work, but they are no longer rendered in the opening swarm. ## Transition video - Title: `The Anatomy of Overwhelm: Deconstructing AI's Five Forces` - Supplied by Nathan Wildman on 2026-07-30. - Preserved source file: `assets/media/ai-overwhelm-five-forces.mp4` - Format: H.264 video with AAC audio, 1280 x 720, 30 fps, 4 minutes 51 seconds. - SHA-256: `89d10a5b3bfd71c5210bc86a5e620facb7acbf22198f2b754137c5d1e5c81934` - Web viewing copy: `assets/media/ai-overwhelm-five-forces-cropped.mp4` - Web viewing copy format: H.264 video with AAC audio, 1280 x 686, 30 fps, 4 minutes 51 seconds. - Web viewing copy SHA-256: `47bdd982cf52a1a095cf3e98330af97288aa38091e62778157fd1ca573f86121` - Edit: 34 pixels were removed from the bottom edge to omit the NotebookLM product footer. No horizontal crop, audio edit, time cut, or content rewrite was made. - Brand treatment: the webpage overlays the Built for Freedom logo on the viewing copy; it is not baked into or falsely attributed to the original master. - Use: user-controlled explainer inside the cinematic transition from the black overwhelm experience to the evidence timeline. - Research disclosure: the webpage links the principal data sources summarized by the film—Epoch AI, GitHub Octoverse, Stanford AI Index, Pew Research Center, and St. Louis Fed/FRED—plus this full ledger. - Boundary: the video is bundled locally and does not autoplay or make a runtime request. The unmodified 1280 x 720 source file remains preserved locally for future editorial and attribution review. ## Quantitative sources - Epoch AI, all-models dataset: https://epoch.ai/data/ai-models - Documentation: https://epoch.ai/data/ai-models-documentation - Inclusion and coverage notes: https://epoch.ai/data/ai-models-documentation/inclusion - Local use: 3,547 dated release marks from 3,572 source records, including 93 dated 2026 records through July 24. - Boundary: non-exhaustive and uneven across private, undocumented, niche, and historical models. Epoch says prominent new-model coverage can lag by days and less-prominent releases can lag by months. The graph stops at the latest observed date and never plots unobserved future quarters as zero. - GitHub Octoverse 2024: https://github.blog/news-insights/octoverse/octoverse-2024/ - Local use: 70,000+ new public generative-AI projects in 2024 and nearly 150,000 total. - Boundary: public GitHub projects are not a census of AI products or companies. - Stanford AI Index 2025, public AI-project series: https://hai.stanford.edu/assets/files/hai_ai-index-report-2025_chapter1_final.pdf - Local use: broader GitHub AI-project growth from 1,549 in 2011 to 4.32 million in 2024. The ≈3.08 million 2023 point is explicitly derived from Stanford’s reported 40.3% 2024 growth. - Boundary: GitHub expanded its AI-topic classification methodology; counts from older report editions are not silently spliced into this series. - Epoch AI trends: https://epoch.ai/trends - Local use: a clearly labeled illustrative index using the reported approximately 5× annual training-compute trend for frontier language models since 2020 and Epoch’s rounded statement of approximately 10,000× growth since 2020. - Boundary: 2020 = 1×; the index is not a measured series for named models and is not extended as a forecast. The 2026 endpoint is rounded rather than a literal 5-to-the-sixth calculation. - Stanford AI Index 2026: https://hai.stanford.edu/ai-index/2026-ai-index-report/ - Local use: 5.6 million public AI projects, Hugging Face upload growth, 17.1 million H100-equivalents of global AI compute capacity, 29.6 GW of AI data-center power capacity, selected capability gaps, and documented AI-incident growth. - Pew Research Center, Americans and AI 2026: - https://www.pewresearch.org/internet/2026/06/17/americans-and-ai-2026-chatbots-smart-devices-and-views-on-impact/ - Local use: pace, privacy, ChatGPT reach, chatbot frequency, search summaries, chatbot brands, and smart-device adoption. - Boundary: U.S. adult survey estimates. The 2026 ChatGPT question wording changed, so that line segment is dotted. - Federal Reserve Bank of St. Louis / FRED, Generative AI Adoption Tracker: - https://fred.stlouisfed.org/series/RPSGENAIUSAGESHAREALL - https://www.stlouisfed.org/on-the-economy/2025/nov/state-generative-ai-adoption-2025 - Local use: quarterly U.S. working-age adult adoption from 47.98% in Q1 2025 through 57.87% in Q1 2026, plus the three-year mass-market adoption comparison. - Boundary: survey estimates and technology-launch comparisons depend on the cited definitions and time windows. - Federal Reserve Board, Monitoring AI Adoption in the U.S. Economy: - https://www.federalreserve.gov/econres/notes/feds-notes/monitoring-ai-adoption-in-the-u-s-economy-20260403.html - Local use: 18% firm adoption, 41% worker use, 54% employment-weighted LLM-firm exposure, and 78% employment-weighted AI-firm exposure. - Boundary: these are different survey units and must remain separately labeled. - Google Trends U.S. dated snapshot: https://trends.google.com/trends/explore?geo=US&q=Artificial%20intelligence,ChatGPT,Generative%20AI - Local use: relative U.S. search interest for the compared phrases, shown separately from survey adoption. - Boundary: 0–100 relative interest within this comparison, not search counts or percentage of people. The older Accenture, Deloitte, and control-paradox snapshots remain in `data/` for provenance and later research. They are not used as timeline lines because their populations and definitions do not match the current views. ## Content studies - The Impact of AI-Generated Text on the Internet: https://doi.org/10.48550/arXiv.2604.26965 - Detecting the emergence of AI-generated newspaper articles: https://aclanthology.org/2026.acl-long.663/ - Delving into ChatGPT usage in academic writing through excess vocabulary: https://doi.org/10.1126/sciadv.adt3813 - Prevalence and Propagation of AI-Generated Content in the 2024 U.S. Presidential Election: https://arxiv.org/abs/2502.11248 - AI Forensics, AI-Generated Algorithmic Virality: https://aiforensics.org/work/gen-ai-slop - Deezer, 13.4 million AI tracks detected and tagged in 2025 and 44% of April 2026 uploads: https://newsroom-deezer.com/2026/04/ai-generated-tracks-represent-44-of-new-uploaded-music/ - Deezer, 90,000 AI tracks per day and more than 50% of uploads at the June 2026 peak: https://newsroom-deezer.com/2026/07/deezer-h1-2026-results-direct-growth-profitability/ - TikTok, 1.3 billion videos labeled AIGC by November 2025: https://newsroom.tiktok.com/more-ways-to-spot-shape-and-understand-ai-generated-content?lang=en-GB - TikTok, more than 3 billion videos labeled AIGC by July 2026: https://newsroom.tiktok.com/helping-people-spot-and-understand-aigc-on-tiktok?lang=en-150 - YouTube official platform statistics: https://blog.youtube/press/ - YouTube CEO 2026 letter: https://blog.youtube/inside-youtube/the-future-of-youtube-2026/ - Adobe Firefly generation milestones: - 1 billion, July 2023: https://news.adobe.com/news/news-details/2023/adobe-firefly-expands-globally-supports-prompts-in-over-100-languages - 3 billion, October 2023: https://news.adobe.com/news/news-details/2023/adobe-releases-next-generation-of-firefly-models - 7 billion, April 2024: https://news.adobe.com/news/news-details/2024/adobe-introduces-firefly-image-3-foundation-model-to-take-creative-exploration-and-ideation-to-new-heights - 8 billion, May 2024: https://news.adobe.com/news/news-details/2024/adobe-unveils-firefly-powered-generative-remove-in-lightroom-for-fast-and-easy-ai-editing-across-surfaces - 12 billion, September 2024: https://news.adobe.com/news/2024/09/091124-adobe-previews-firefly-video-model - 13 billion, October 2024: https://news.adobe.com/news/2024/10/101424-adobe-launches-firefly-video-model - 18 billion, February 2025: https://news.adobe.com/news/2025/02/firefly-web-app-commercially-safe - 22 billion, April 2025: https://news.adobe.com/news/2025/04/adobe-revolutionizes-ai-assisted-creativity-firefly - 24 billion, June 2025: https://news.adobe.com/news/2025/06/adobe-firefly-revolutionizes-creative-ideation-new-mobile-app The live content evidence uses different samples, stages, and methods. The interface never sums Firefly assets, TikTok labels, and Deezer detections into a fictional universal total. YouTube's total daily upload volume is shown as a denominator with the AI share explicitly unknown; Firefly's generation count is not treated as published internet content; TikTok's label count is not treated as a platform-wide percentage. ## Tool and glossary records Tool records link to official product pages and carry a shared `2026-07-28` verification date. They are curated examples, not rankings or an exhaustive directory. Glossary entries link to primary papers, official documentation, or established reference sources.