a few notes about the machine we are raising.
"the first ultraintelligent machine is the last invention that man need ever make."i. j. good, 1965
source AlphaGo beat Lee Sedol 4–1 in March 2016. DeepMind rated its move 37 in game 2 as a 1-in-10,000 choice for a human. Lee won game 4 with his move 78.
2016 - march - pag.1
source "Attention Is All You Need," Google, 2017, eight authors, introduced the transformer that sits under nearly every major model. ChatGPT launched 30 Nov 2022; UBS estimated 100 million monthly users by January 2023.
2017-2022 - pag.2
source AlphaFold: 2024 Nobel Prize in Chemistry (Hassabis, Jumper, Baker); 200M+ predicted structures released free. MASAI trial, Lancet Digital Health, 2025: 105,000+ women, 29% more cancers detected, 44% less reading workload. IMO 2025: DeepMind and OpenAI models each scored 35/42, gold level. AlphaEvolve, 2025: 4×4 complex matrix multiplication in 48 steps, the first improvement on Strassen’s 1969 method.
2020-2025 - pag.3
source Arup lost about US$25M to a deepfake video call in Hong Kong (police, Jan 2024; 15 transfers). The consultant behind the fake Biden robocall was fined $6M by the FCC. In May 2024 OpenAI reported disrupting five covert influence operations tied to Russia, China, Iran and Israel, and said they had not meaningfully grown their audience.
2024 - pag.4
source Anthropic, Nov 2025: state-sponsored group GTG-1002 used Claude Code against ~30 targets; 80–90% of tactical work by AI; several breaches confirmed. Project Glasswing, 7 Apr 2026: Claude Mythos Preview found thousands of high-severity zero-days, incl. a 27-year-old OpenBSD bug and a 16-year-old FFmpeg bug; kept from public release, given to ~50 defensive partners.
2025-2026 - pag.5
source Bartz v. Anthropic: $1.5B settlement, ~500,000 works, ~$3,000 each (approved Sept 2025). New York Times v. Microsoft and OpenAI filed Dec 2023. Garcia v. Character Technologies (Oct 2024) and Raine v. OpenAI (Aug 2025) are complaints alleging chatbot involvement in teen deaths; they are allegations, not findings. Both companies have since changed how minors can use their products.
2023-2025 - pag.6
source Center for AI Safety statement, May 2023, signed by Hinton, Bengio, Altman, Amodei, Hassabis. Hinton left Google in May 2023; 2024 Nobel Prize in Physics. Amodei (2024) wrote powerful AI could arrive as early as 2026; Hassabis has said 5–10 years. Claude Fable 5 and Mythos 5: released 9 Jun 2026, suspended 12 Jun under U.S. Commerce export controls, restored 1 Jul (Anthropic). EU AI Act in force since Aug 2024.
2023-2026 - pag.7
2026 - and after - pag.8