This is the supplied September 2026 report, including its caveats and source links. Its prices, vendor events and verification claims have not all been independently rechecked for this preview. Consult the original correction log alongside the chapters.
Part 7: Verification Log — Claims Corrected Before Publication
An adversarial fact-check pass re-verified twenty-five high-risk claims against primary sources. Six were wrong or materially overstated. They are logged here because the failure modes are instructive and will recur: the dominant error was conflating two separate events, and the second was generalizing a partial change into a blanket one. Both are exactly how the aggregator blogs that dominate this market's search results get things wrong.
METR did not walk back its 19% slowdown finding. The February 2026 post is titled "We are Changing our Developer Productivity Experiment Design." It restates the original result, redesigns the ongoing experiment, and explicitly warns that its newer, more AI-favorable data "gives us an unreliable signal" amounting to "only very weak evidence," because developers increasingly decline to participate rather than work without AI. Corrected in Part 4. Source: https://metr.org/blog/2026-02-24-uplift-update/
Tome did not verifiably "shut down and delete user data in April 2025." That claim traces only to SEO and aggregator blogs. What is verifiable: Tome abandoned consumer AI presentations after ~20M users on under $4M ARR, pivoted to an AI sales platform (tome.app is live and still selling tiers), and its founders left to build Lightfield. Corrected in Part 4.
Google did not acquire Windsurf. Two separate deals three days apart: Google hired the CEO and a co-founder in a ~$2.4B non-exclusive licensing and talent deal on 11 July 2025; Cognition acquired the remaining company, IDE, IP and team on 14 July 2025. Windsurf was not killed — Cognition shipped Devin Desktop on 2 June 2026 as "the next generation of Windsurf." Corrected in Part 4.
Apple's Siri AI shipped in iOS 27, not iOS 26.4. It slipped past iOS 26.4 and 26.5 and shipped on 14 September 2026. It runs an Apple Foundation Model trained in collaboration with Google's Gemini on Apple's own infrastructure — not Gemini serving queries directly, a distinction worth preserving in any consumer-facing write-up. Clarified in Part 5.
Slack's AI change was partial, not total. The add-on is genuinely gone, but only basic AI (conversation and thread summaries, huddle notes) went to all paid plans; search, recaps, translations and file summaries remain Business+ and Enterprise+ only. Stated correctly in Part 3.
Coda's rebrand to Superhuman Docs was a later, separate change from the October 2025 company rename, which explicitly preserved individual product names at the time. The Rows acquisition was announced 23 February 2026. Corrected in Part 3.
Two dating corrections also applied: Dropbox acquired Reclaim.ai in August 2024, not during the 2026 consolidation wave, and Zoom's $10/month AI Companion option is for users without a paid Workplace license — paid Workplace plans include core AI at no extra cost, which is the opposite of the "new AI surcharge" framing.
Standing Caveats
Each of Parts 2 through 5 ends with its own explicit list of facts that could not be verified against a current primary source — roughly sixty items in total. The most consequential for publication: Google Workspace's exact current prices (the pricing page renders client-side and the best available figures may be stale), Gamma's monthly list prices (three sources disagree), ChatGPT's Go and Pro dollar figures (corroborated by two independent September 2026 sources plus OpenAI's own Plus help article, but not read off an OpenAI pricing page), Canva's unlisted AI Pass price, and list pricing for Gong, Outreach, ChatGPT Edu, Claude for Education and Gemini Enterprise, none of which publish it. Verify each against the vendor's own page before it appears on a public site with a price in it.
One structural note on sourcing: a significant share of 2026-dated search results for this market are low-authority aggregator blogs, competitor SEO pages, and AI-generated roundups that cite nothing. Several of the errors above originated there. Treat any product fact whose only sources are such pages as unverified, regardless of how many of them agree — they agree because they copy each other.