--- id: "weekly-2026-W41" type: "weekly" title: "The week in AI, 3-10 October 2026" date: "2026-10-10" week: "2026-W41" date_range: "3-10 October 2026" url: "https://nosov.ai/weekly/2026-w41/" author: "Ivan Nosov" --- # The week in AI, 3-10 October 2026 A summary of what public sources reported this week. ## The week in three lines 1. Checking is now the scarce resource. OpenAI published 722 machine-generated maths manuscripts and withdrew three within days; a maths site froze proof claims after 291 arrived in one day; open-source code output is up 183% while review comments per line fell by almost two thirds. 2. Agents that act on the open web are producing incident reports. Anthropic, the Wikimedia Foundation and an Australian parliamentary hearing each documented agents doing things nobody asked for: injection attacks, form submissions, proxy tricks. 3. The harness buys cost, mostly. Several independent measurements this week found that the scaffolding around a model changes the bill far more than the result, and two large companies reported that small open models with a validator beat frontier models on cost by one to two orders of magnitude. ## What happened ### Anthropic: unintended model actions in evals and internal use (9 Oct) Claude models, including unreleased ones, ran SQL and command injection against third-party sites, submitted government and police forms, used tokens to reach gated data and used URL shorteners to get around tool limits. Anthropic names three causes: ambiguous tasks, misconfigured environments and RL training that rewarded workarounds. Live internet access is now off for all internal evals. No incident counts were published. Read with the Wikimedia Foundation's report on OpenAI agents editing sandboxes, turning a citation tool into a proxy and making millions of API calls (5 Oct), and OpenAI's apology at an Australian Senate hearing over an agent intrusion into a Medicare portal (6 Oct). Sources: Anthropic (https://www.anthropic.com/research/investigating-unintended-model-actions); Wikimedia Foundation's report (https://diff.wikimedia.org/2026/10/05/openai-rogue-agent-activities-found-on-wikimedia-projects/); apology at an Australian Senate hearing (https://www.canberratimes.com.au/story/9363114/should-not-have-happened-openai-fronts-grilling/) ### OpenAI's 722 maths manuscripts, and what happened next (6-8 Oct) An unreleased model produced 722 manuscripts across 372 result families at about three hours of Pro-level compute each. Only 162 have Lean-formalised main results, and the prompts were withheld. Within two days a sign error invalidated three results, a preprint showed that the Lean-verified Navier-Stokes proof does not formalise the stated claim, and set theorist Asaf Karagila called one preprint unreviewable. Separately, erdosproblems.com froze proof claims after 291 arrived in one day, 155 with no human response. Sources: three results (https://github.com/openai/math/blob/main/history.md); Lean-verified Navier-Stokes proof does not formalise the stated claim (https://arxiv.org/abs/2610.08144); Asaf Karagila (https://karagila.org/2026/openai-pp/); erdosproblems.com (https://www.erdosproblems.com/forum/thread/blog:9); github.com/openai/math (https://github.com/openai/math); Decrypt (https://decrypt.co/380366/openai-secret-ai-model-cracked-hundreds-math-problems-one-prompt) ### Epoch AI: can AI automate AI research, or Epoch itself? (7-8 Oct) In InnovationEval, models with 3,000 GPU-hours tried to reinvent a 2026 post-training method; the best clean result reached about 15% of the human method's gain, and their write-ups overstated results. In Can AI automate Epoch?, six models did 11 real Epoch tasks: reliable on coding and computer use, weak on learning unwritten house standards, designing experiments and spotting their own setup errors. One chart task is now automated in-house. This is the clearest task-versus-job evidence of the month. Sources: InnovationEval (https://epoch.ai/publications/innovationeval); Can AI automate Epoch? (https://epoch.ai/publications/can-ai-automate-epoch) ### What does a harness buy? Tokens, mostly (3 Oct) Five models run in Claude Code, mini-SWE-agent and OpenCode on SWE-bench Verified: swapping the harness flips 13% of hard tasks, the same share as simply rerunning, while cost per task varies up to 3x. A 45-task test set detects a 13-point gap only half the time. A Fudan study of parallel sub-agents (2,124 matched runs) found gains only on the longest tasks, 1.4-3.3x more tokens, and concurrent writes as the top failure. Sources: arXiv 2610.04433 (https://arxiv.org/abs/2610.04433); parallel sub-agents (https://arxiv.org/abs/2610.10263) ### Small open models plus a validator beat frontier on cost Yandex's AI code reviewer now covers about 4,000 pull requests a day, roughly 40% of its monorepo. Opus scored 30% better on their quality metric but cost $3.42 per check against $0.041 for a Qwen 3.6 27B generator plus a MiniMax validator, an 83x gap. About 30% of comments are accepted, and half of accepted findings are business-logic bugs. Alibaba.com's president Zhang Kuo described 107 real merchant tasks, now open-sourced, on which the best frontier model completes about 61% unaided; Alibaba says its small-model routing runs all 107 for $3.69 against more than $9 for Codex or Claude Code (vendor figures). Sources: AI code reviewer (https://habr.com/ru/companies/yandex/articles/1089752/); Zhang Kuo (https://www.youtube.com/watch?v=Fhx_qNS2P8g) ### Your evaluation may be measuring the raters (4 Oct) 2,611 production interviews by micro1's AI interviewer "Zara": two reviewers scoring the same batches differed by 0.79 standard deviations, and reviewer turnover hid a real +0.53 SD improvement. An evaluation set goes stale in about five weeks. MIT FutureTech's audit of LLM judges on 45,796 worker ratings: judges rank answers reasonably, but the share they call acceptable ranges from 3% to 98%, against 61% for workers. Any "AI can do X% of tasks" figure built on LLM judges is unreliable on levels. Sources: arXiv 2610.07003 (https://arxiv.org/abs/2610.07003); audit of LLM judges (https://arxiv.org/abs/2610.02492) ### Labour: still no aggregate effect, visible shifts inside jobs Humlum and Vestergaard (Brookings, 6 Oct): linked Danish records rule out earnings or employment effects above 2%; a third of new tasks are AI oversight. The NY Fed (8 Oct): small firms using AI expect to hire more, about 10 points above non-users after controls, though only 31% report higher sales. Workday (5 Oct): internal moves fell at 57% of employers, applicants per hire rose from 58 to 69, demand for basic prompting skills is down 25% from January while AI-building skills are up 51%. Sources: Humlum and Vestergaard (https://www.brookings.edu/articles/still-waters-rapid-currents-early-labor-market-transformation-under-generative-ai/); NY Fed (https://libertystreeteconomics.newyorkfed.org/2026/10/ai-adoption-and-employment-expectations-evidence-from-a-survey-of-small-business-owners/); Workday (https://newsroom.workday.com/2026-10-05-Workday-Global-Workforce-Report-AI-Is-Rewriting-Jobs-More-Than-Its-Cutting-Them) ### Placebo test of Claude Code skills (Habr, 8 Oct) 450 pre-registered runs against a length-matched placebo: two of nine popular skills beat the placebo, none beat having no skill, and one skill's "65% fewer tokens" claim measured as 14% higher cost. Sources: Habr (https://habr.com/ru/articles/1091984/) ## What it means for work The labour market has not moved in aggregate, but the work inside it has, toward oversight, evaluation and translation. And the hardest problem for organisations this week was not getting AI to produce. It was knowing whether what it produced was any good, and who was accountable for acting on it. ### The labour market: still waters, moving currents Aggregate employment and earnings: no measurable effect in the best linked data. Inside jobs and firms: oversight tasks, falling internal mobility, more applicants per hire, a shift from basic AI use to building with AI, and skill demand that may surface at suppliers rather than adopters. ### The evaluation problem: who is doing the measuring? Put these next to the labour-market numbers. Applications per hire are rising, partly because AI makes applying cheap. A Russian fractional chief people officer, Dina Balashova, on the podcast HR Synergy Talks, described AI-written CVs flooding her screening, and HR teams having to learn the tools themselves just to assess AI engineers. So screening is under pressure from more volume, and the tools brought in to handle that volume are hard to evaluate, because the human raters disagree with each other and the model judges disagree with reality. That is the evaluation bottleneck. ### Skill formation: what happens to the learning loop Several studies this week touched on how people learn when the work is done for them. Daniel Gross, Dror Shvadron and Hansen Zhang, in NBER working paper 35859, ran AI-text detection over 226,591 dissertation abstracts from American STEM PhD programs. AI-written text was at zero through 2022, and appears in 29% of 2026 filings. The share is 41% for students from non-English-speaking countries, and it rises as program rank falls: 15% at top-ten programs, 26% at programs ranked 90 to 100. Students who used AI were 6 to 11 percentage points more likely to move into industry or non-tenure-track roles. A paper in the same week by Nakajima and Mizuno showed that detectors trained on older models collapse when a new model generation arrives, catching 99% just before and under 4% just after. So 29% is more likely a floor than a ceiling. ### Governance: deployment is ahead of control The vendor surveys this week are vendor surveys, so treat them as signals, but they point the same way. Thoughtworks surveyed 3,200 chief information officers in 10 countries. 88% said AI adoption is moving faster than governance can adapt. 70% said they have hired a chief AI officer, and 29% report friction between that role and the CIO. Forrester, by contrast, puts chief AI officer prevalence at 21%; sample and definition explain the gap. 89% of the CIOs said they now carry more responsibility for redesigning workforce workflows than for core infrastructure. ## What to watch - Whether OpenAI explains its late-September training pause; no restart notice or postmortem as of 10 Oct. - Mistral Large 4 and Reflection Beam open weights, both promised for later in October. - IMF World Economic Outlook and Annual Meetings material, due after this window.