Every material factual change to a published page is recorded here: what the page said, what it says now, and why. A correction means the page was wrong when it was published. An update means the facts changed afterwards and the page was re-checked. The rules behind this log are in the editorial standards.
17 entries: 12 corrections, 5 updates. The log starts on October 4, 2026.
October 15, 2026 — Update
Updated /answers-meta-muse-vs-claude-cowork-vs-gemini-spark-founders-2026/
- Was: Meta Muse is US-only
- Now: Meta Muse is available in the US and, since September 18, 2026, Canada
- Source: https://www.aiagentslibrary.com/blog/meta-muse-availability/
- Page: /answers-meta-muse-vs-claude-cowork-vs-gemini-spark-founders-2026/
October 15, 2026 — Update
Updated /answers-meta-muse-vs-claude-cowork-vs-gemini-spark-founders-2026/
- Was: Gemini Spark is gated behind Google AI Ultra ($99.99/$199.99), select countries only
- Now: Gemini Spark requires Google AI Pro ($19.99) or Ultra; unavailable in the EEA, UK, Switzerland and Nigeria
- Source: https://support.google.com/gemini/answer/17094507?hl=en
- Page: /answers-meta-muse-vs-claude-cowork-vs-gemini-spark-founders-2026/
October 15, 2026 — Update
Updated /answers-meta-muse-vs-claude-cowork-vs-gemini-spark-founders-2026/
- Was: Claude Cowork is Anthropic's separate agent workspace, included in paid plans from Pro
- Now: Anthropic merged Cowork and chat into one Claude on September 16, 2026; plan prices unchanged (Pro annual / monthly)
- Source: https://claude.com/blog/cowork-is-now-claude
- Page: /answers-meta-muse-vs-claude-cowork-vs-gemini-spark-founders-2026/
October 15, 2026 — Update
Updated /answers-flow-state-when-coding-with-ai/
- Was: Forge took an 8B model from 53% to 99% reliability on agentic tasks
- Now: Forge's launch post reported 53% to 99%; its current README reports single digits to 84% on the harder 26-scenario v0.7.0 suite
- Source: https://github.com/antoinezambelli/forge
- Page: /answers-flow-state-when-coding-with-ai/
October 15, 2026 — Correction
Corrected /answers-flow-state-when-coding-with-ai/
- Was: Newport's research shows a five-second context switch leaves residual attention for up to twenty minutes
- Now: Sophie Leroy's 2009 experiments show residue from unfinished work hurts performance on the next task; no universal number of minutes is attached
- Source: https://www.brnsft.com/blog/attention-residue-sophie-leroy-context-switching
- Page: /answers-flow-state-when-coding-with-ai/
October 15, 2026 — Correction
Corrected /answers-flow-state-when-coding-with-ai/
- Was: Stanford CS336 AI agent guidelines tell students to write a plan, run the agent in long autonomous tasks and intervene only at checkpoint reviews
- Now: The CS336 guidelines tell AI assistants to act as tutors that explain, ask guiding questions and review, and not to write, run or edit students' code
- Source: https://github.com/stanford-cs336/assignment1-basics/blob/main/AGENTS.md
- Page: /answers-flow-state-when-coding-with-ai/
October 15, 2026 — Correction
Corrected /answers-adviserry-vs-readwise-reader-vs-claude-project-newsletters-into-actions-2026/
- Was: There is a seven-day free trial
- Now: There is a 14-day free trial (card required, charged on day 15 unless you cancel)
- Source: https://adviserry.com/
- Page: /answers-adviserry-vs-readwise-reader-vs-claude-project-newsletters-into-actions-2026/
October 13, 2026 — Correction
Corrected /answers-avoid-ai-brain-fry-daily-ai-users-survey-founders-october-2026/
- Was: 41% need a full evening of rest to recover after an AI-heavy workday
- Now: 40% need a full evening of rest to recover after an AI-heavy workday (41% in the survey's own recovery breakdown)
- Source: https://allwork.space/2026/10/ai-brain-fry-is-hitting-young-workers-hardest-what-can-employers-do
- Page: /answers-avoid-ai-brain-fry-daily-ai-users-survey-founders-october-2026/
October 10, 2026 — Correction
Corrected /answers-chatgpt-plus-vs-claude-pro-vs-google-ai-pro-vs-perplexity-pro-decisions-october-2026/
- Was: ChatGPT Plus context: 27,000 to 400,000 tokens, by plan and mode
- Now: ChatGPT Plus context: 54,000 tokens for instant answers, 256,000 with Thinking
- Source: https://www.ai-toolbox.co/chatgpt-models/chatgpt-context-window-token-limits-2026
- Page: /answers-chatgpt-plus-vs-claude-pro-vs-google-ai-pro-vs-perplexity-pro-decisions-october-2026/
October 10, 2026 — Correction
Corrected /answers-ai-adoption-senior-up-junior-down-41-country-study-founders-october-2026/
- Was: 80% of the executives Davenport and Mittal surveyed leaned toward replacing employees rather than retraining them
- Now: in one of their AI surveys only 10% of US respondents stated a clear preference for retraining and keeping current employees
- Source: https://akos.ma/blog/notes-about-all-in-on-ai-by-thomas-davenport-and-nitin-mittal/
- Page: /answers-ai-adoption-senior-up-junior-down-41-country-study-founders-october-2026/
October 10, 2026 — Correction
Corrected /answers-ai-adoption-senior-up-junior-down-41-country-study-founders-october-2026/
- Was: King's announced the study on October 7, 2026
- Now: King's announced the study on October 6, 2026
- Source: https://www.kcl.ac.uk/news/ai-adoption-linked-to-increase-in-employment-globally
- Page: /answers-ai-adoption-senior-up-junior-down-41-country-study-founders-october-2026/
October 8, 2026 — Correction
Corrected /answers-best-ai-agents-founders-long-game-july-2026/
- Was: Manus broke out in early 2026
- Now: Manus launched in March 2025; after Meta's deal was unwound it announced a return to independence on 11 Aug 2026
- Source: https://en.wikipedia.org/wiki/Manus_(AI_agent)
- Page: /answers-best-ai-agents-founders-long-game-july-2026/
October 8, 2026 — Update
Updated /answers-best-ai-agents-founders-long-game-july-2026/
- Was: Lindy pricing starts around $49.99/month for personal use; 1,600+ integrations
- Now: Lindy is $29.99 per user/month for 3,000 pooled credits; 1,500+ integrations
- Source: https://www.lindy.ai/pricing
- Page: /answers-best-ai-agents-founders-long-game-july-2026/
October 8, 2026 — Correction
Corrected /answers-best-book-about-alexander-the-great/
- Was: Alexander wept when there were no more lands to conquer
- Now: In Plutarch, Alexander wept because, with infinite worlds, he had not yet conquered one; the 'no more worlds' version is a later misquote
- Source: https://www.perseus.tufts.edu/hopper/text?doc=Perseus%3Atext%3A2008.01.0267%3Asection%3D4
- Page: /answers-best-book-about-alexander-the-great/
October 8, 2026 — Correction
Corrected /answers-best-book-about-alexander-the-great/
- Was: Green's Alexander of Macedon was first published in 1974 and revised in 1991
- Now: First issued in 1970, revised and expanded edition 1974, first US publication 1991
- Source: https://en.wikipedia.org/wiki/Peter_Green_(historian)
- Page: /answers-best-book-about-alexander-the-great/
October 8, 2026 — Correction
Corrected /answers-how-to-stop-being-addicted-to-ai/
- Was: Psychology Today's research confirms reducing ease of access is the single most effective intervention
- Now: Psychology Today (Jan 2026) lists deleting the app as one of eight steps; it does not rank it as the most effective
- Source: https://www.psychologytoday.com/us/blog/living-your-best-life/202601/8-tips-for-managing-ai-dependence
- Page: /answers-how-to-stop-being-addicted-to-ai/
October 8, 2026 — Correction
Corrected /answers-is-prompting-still-valuable-skill-workday-report-october-2026/
- Was: Marshall Goldsmith's mentor told him he had become "too successful" and was no longer thinking or investing in his future.
- Now: Removed: no primary source for the quote could be found; the Career Waves point stands on Clark's book.
- Source: https://store.hbr.org/product/the-long-game-how-to-be-a-long-term-thinker-in-a-short-term-world/10452
- Page: /answers-is-prompting-still-valuable-skill-workday-report-october-2026/
Found an error? hello@andreihirvi.com