Skip to content

Home · The book · free to read

Back matter

Appendix D: Timeline of Key Developments

The reading text here is Infinite Architects as corrected and updated on 17 January 2026, the edition the ebook carries. The print first edition of 2 January 2026 is unchanged and stays as dated evidence: it was never revised, and nothing on these pages revises it. Where the two differ, the difference is a correction the author made to the ebook, not a rewriting of what was printed. Every section below is reproduced verbatim from the 17 January artefact, with that artefact’s SHA-256 and a SHA-256 for each section, so “verbatim” is checkable rather than claimed. Section source SHA-256 b3ba063bc120addc… · contents and full manifest.

Open the book reader

About this section

A dated list of the developments the book treats as its context, running from December 2024 to January 2026 and covering quantum error correction results, the alignment faking research, the model releases through 2025, and the closure of one safety institute and the change of direction at another.

Appendix D: Timeline of Key Developments

December 2024 – January 2026

December 2024

December 9: Google announces Willow quantum chip with below-threshold error correction.

December 17: Anthropic publishes alignment faking research (78% rate).

December 20: OpenAI announces o3 achieving 87.5% on ARC-AGI.

December 27: DeepSeek releases V3, matching GPT-4o at fraction of cost.

January 2025

January 2025: Sam Altman declares 'We are now confident we know how to build AGI.'

January 20: DeepSeek R1 released, becomes most-downloaded iOS app.

January 2025: Anthropic releases Constitutional Classifiers (86% → 4.4% jailbreak rate).

April 2024

April 19: Future of Humanity Institute closes after ~19 years.

April 2025

April 16: OpenAI publicly releases o3.

April 24: Anthropic launches Model Welfare Program.

April–June: COGITATE consciousness study results published in Nature.

May 2025

May 22: Claude 4 family released; Opus 4 first model at ASL-3.

August 2025

August 7: GPT-5 released.

August 2025: Sam Altman calls AGI 'not a super useful term.'

October 2025

October 22: Google publishes 'Quantum Echoes' paper in Nature.

November 2025

November 18: Gemini 3 Pro released, first model to cross 1500 Elo.

November 24: Claude Opus 4.5 released.

November 30: MIRI announces strategic pivot, calls approach 'extremely unlikely to succeed.'

December 2025

December 1–2: OpenAI declares 'Code Red' in response to Gemini 3.

December 4: Gemini 3 Deep Think released (45.1% on ARC-AGI-2).

December 16: GPT-5.2 released.

December 24: Sam Altman suggests AGI may have already arrived.

December 28: Geoffrey Hinton CNN interview (10–20% takeover probability).

January 2026

January 6: Infinite Architects published.

© Michael Darius Eastwood 2026. Free to read here, by the author’s decision; not public domain. All rights reserved. If this book gives you something, the whole of it is free to hand to the next reader: send them any section’s address, or the start.

Read this in the reader, which keeps your place and pages like an ebook. These pages stay the citable ones.

reads aloud · highlights as it goes · jump to any section