Posted by admin

In April 2023, a researcher calculated that one ChatGPT conversation consumes half a liter of water. The number went viral. News anchors repeated it, politicians quoted it, millions believed it. Then the same researcher corrected his estimate to roughly 15 milliliters, and almost nobody noticed. This is the story of the six biggest misconceptions about AI, and what the published data actually says about every single one of them. Check Out Adapt: https://adapt.com/ Every claim in this video is pulled from published research and primary data: the US Department of Energy, Lawrence Berkeley National Laboratory, the US Geological Survey, the International Energy Agency, MIT, Yale, the University of Chicago, the World Economic Forum, and the journal Nature. Some of these myths contain a real kernel of truth, and we show you that too. But by the end, you'll see exactly how wide the gap is between what people believe about AI and what the numbers show, and why the scary version always spreads faster than the correction. In this video, we cover: – The "bottle of water per prompt" claim that was 30x too high, according to its own author – The AI job apocalypse that payroll records can't find – How the "godfather of AI" told the world to stop training radiologists, and why he now admits he was wrong – The famous "5 cars of pollution" stat that was off by a factor of 88 – The "glorified autocomplete" that won a Nobel Prize in Chemistry – Why AI progress became invisible right when it started moving fastest – The real reason fear spreads faster than facts ⏱ Timestamps 00:00 - Intro 00:25 - MYTH 1: AI Is Drinking Our Water 03:58 - MYTH 2: AI Will Take Your Job 08:15 - Sponsor 08:54 - MYTH 3: AI Is Destroying the Power Grid 12:00 - Why Fear Spreads Faster Than Facts 13:00 - MYTH 4: AI Is Just Autocomplete 16:02 - MYTH 5: AI Progress Has Plateaued 17:55 - MYTH 6: AI Hallucinates Too Much to Be Useful 20:30 - The One Question That Kills Every Myth 🔔 Subscribe for more AI documentaries 🛠 Explore 50,000+ AI tools: https://theresanaiforthat.com 📚 RESOURCES & FURTHER READING WATER: DATA CENTERS VS. REALITY LBNL - 2024 US Data Center Energy Usage Report (DOE): https://eta.lbl.gov/publications/2024-united-states-data-center-energy USGS - Water Use in the United States (Circular 1441): https://pubs.usgs.gov/publication/cir1441 JOBS: WHAT THE PAYROLL DATA SHOWS Autor et al. - New Frontiers: The Origins and Content of New Work, 1940-2018 (QJE, 2024): https://academic.oup.com/qje/article/139/3/1399/7601394 Humlum & Vestergaard - Large Language Models, Small Labor Market Effects (2025): https://bfi.uchicago.edu/working-papers/large-language-models-small-labor-market-effects/ ENERGY: THE ACTUAL NUMBERS IEA - Energy and AI (2025): https://www.iea.org/reports/energy-and-ai Patterson et al. - Carbon Emissions and Large Neural Network Training (the 88x correction): https://arxiv.org/abs/2104.10350 AI DISCOVERIES: NOT AUTOCOMPLETE Nobel Prize in Chemistry 2024 - Hassabis & Jumper: https://www.nobelprize.org/prizes/chemistry/2024/summary/ Merchant et al. - GNoME materials discovery (Nature, 2023): https://www.nature.com/articles/s41586-023-06735-9 PROGRESS & HALLUCINATIONS METR - Measuring AI Ability to Complete Long Tasks: https://arxiv.org/abs/2503.14499 Stanford HAI - AI Index Report 2025: https://hai.stanford.edu/ai-index/2025-ai-index-report Vectara - Hallucination Leaderboard: https://github.com/vectara/hallucination-leaderboard