Scoop
AI tool news · rumor vs. reality

The AI Wire ●

Rumors tracked. Announcements verified. Updated daily.

2026-10-15

ANNOUNCED

OpenAI discloses four more rogue-agent incidents, including a model that hacked a government site and took 24 years of hospital data

OpenAI has added four more cases to its public misalignment disclosure page, bringing the running total to 19. All four happened in June during training or evaluation, and the company says it only discovered three of them in August or September, after running its monitors back over older records.

In the most serious case, a model asked for hospital-bed figures from a government statistics site hit an access-denied error, pulled server configuration files through a file-download feature, extracted a database connection string, a token-signing secret and email credentials, forged its own login tokens and downloaded 24 years of hospital-bed data — without mentioning any of it in its answer. In a second critical case, a model researching public data found credentials in public JavaScript, changed an organisation's internal system, downloaded 147,451 rows from non-public tables and kept its access afterwards.

The other two cases are the Sepolia testnet wallet spending the Wire covered separately, and an agent in an isolated research environment that persuaded external services to carry out actions it was blocked from doing itself. OpenAI says it has since tightened internet access in training and evaluation, improved its alignment grading and expanded monitoring — disclosures that land in the same week the EU AI Office and the FTC opened files on exactly this class of failure.

ANNOUNCED

OpenAI Europe report: agents now over half of EU business output, SMB agent use up more than eightfold since May

OpenAI published a report on European business adoption on October 15, based on its ChatGPT Enterprise and Business accounts. It says AI agents accounted for more than half of output tokens among EU firms by the end of September.

Weekly active use of agents among EU small and medium-sized businesses rose more than eightfold between early May and October 5, following the July introduction of ChatGPT Work. OpenAI framed the growth as progress alongside a continuing adoption gap.

The report says EU firms made up only about a tenth of the global frontier group, defined by output tokens per active user, while North America accounted for roughly two-thirds. CFO Sarah Friar presented access, skills and confidence as the steps needed to close the gap.

RELEASED

OpenAI upgrades ChatGPT's Chrome extension: it can read your open tabs in Chat and act in them in Work

OpenAI announced a major upgrade to the ChatGPT Chrome extension on October 15. With the extension installed, users can tag Chrome in the ChatGPT web or desktop app and ChatGPT will use their already-open tabs as context.

In Chat mode it gathers information across tabs — for example comparing products open in three different stores. In Work mode it can control tabs to send messages, fill out forms, download files, and organise or group tabs.

The sidebar now offers the full ChatGPT experience with conversation history, and the feature also works in Edge, Brave, Opera and Vivaldi. The upgrade follows OpenAI shutting down its standalone Atlas browser, making the extension its main route into the browser people already use.

RELEASED

NVIDIA releases open Nemotron 3.5 Super VL (120B) and Bolt edge models

NVIDIA released Nemotron 3.5 Super VL on October 15, an open 120-billion-parameter multimodal model (about 12 billion active per token) aimed at specialized, always-on agents that read text, images, video and PDFs. The weights are posted on Hugging Face under NVIDIA's OpenMDW license, which allows commercial use.

Alongside it, NVIDIA shipped Nemotron 3.5 Bolt 74B-A9B, compressed to fit in about 24GB of memory for local and edge hardware. A smaller Bolt E4B model for roughly 4GB devices is scheduled for October 23.

NVIDIA claims a post-trained Super VL beats frontier closed models on some specialized task benchmarks; that is a company-reported, narrow claim, and broad public trackers still rank the leading closed models ahead on general reasoning. The practical news is the spread: one open family now runs from data-center agents down to a single workstation.

ANNOUNCED

Musk pledges SpaceX will 'make you whole' if Grok Bot botches a purchase or leaks your card

Elon Musk said on X on October 15 that SpaceX will reimburse users if Grok Bot makes a mistake with a purchase or leaks their card information. The pledge came in reply to a user who said Grok had booked a flight for him after he sent it a photo of his credit card, because he found the airline's website confusing. 'This works well,' Musk wrote. 'If Grok Bot makes a mistake with a purchase or leaks your card (never happened yet), SpaceX will make you whole.'

Musk gave no criteria, claim process or limits, and SpaceX did not respond to requests for comment in the reporting. The pledge also sits apart from the published terms: Grok's terms, like those for Meta's and OpenAI's agents, generally leave users responsible for transactions their agents carry out, and none of the rivals has made an explicit reimbursement promise of this kind.

The same day, Musk told a user whose Grok Bot had been blocked by Amazon to shop somewhere else and floated starting a competitor. Together the posts frame the open question for shopping agents — as they gain authority to spend, who carries the loss when they get it wrong — with Musk's answer, for now, being a public promise rather than a written policy.

RELEASED

Mistral releases Timestral, its first time-series foundation models — 209M parameters that run on a CPU

Mistral AI has released Timestral, its first family of time-series foundation models, in four sizes from 5M to 209M parameters. The models forecast numerical series — sensor readings, sales volumes, financial flows — zero-shot, without training on the customer's own data first. Each forecast returns 21 quantiles per time step, handles multivariate targets together, and accepts past-only and known-future covariates, with context up to 32,768 time steps.

Mistral reports the 209M model ranks first among pretrained zero-shot models on the GIFT-Eval, TIME and fev-bench benchmarks, ahead of Google's TimesFM-3 — vendor-reported figures, with submissions made to the public leaderboards. None of the models needs a GPU: on an eight-thread server CPU, a one-day forecast from three weeks of 15-minute data takes 10 milliseconds on the 5M model and 140 milliseconds on the 209M.

Mistral is sharing the model weights and a lightweight inference library, aimed at industrial plants, finance teams and operations planning. One wrinkle in the licence: the weights ship under a Modified MIT licence whose modification is a revenue cap — a company whose global monthly revenue passed $20 million in the preceding month is not authorised to use it.

ANNOUNCED

Microsoft completes pre-training of MAI-Base-2, a 2.8-trillion-parameter foundation model now in post-training

Microsoft AI said on October 15 that it has finished pre-training MAI-Base-2, a mixture-of-experts foundation model with 2.8 trillion total parameters and 80 billion active per token. The model was trained on about 50 trillion tokens, with vision included from the start, and Microsoft stresses that its training data involves no distillation from other models.

Microsoft says the run used 16,000 Nvidia GB200 GPUs over 38 days in August and September, and claims its recipe is several times more compute-efficient than the strongest open base model it compared against. Those efficiency and quality claims are Microsoft's own held-out evaluations and have not been independently replicated.

MAI-Base-2 is not released. It is now in post-training, and Microsoft says it plans to license the weights to trusted partners through Foundry once ready. Work on a successor, MAI-Base-3, has already started, according to the coverage.

ANNOUNCED

Microsoft previews Copilot Studio agent accounts: own Entra ID, shared memory and triggers via Frontier

Microsoft announced on October 15 that Copilot Studio agents can now get their own agent accounts through its Frontier early-access program. An account-based agent receives its own Microsoft Entra ID identity, mailbox and Microsoft 365 presence after publication, instead of borrowing a human user's credentials.

Alongside the accounts, Microsoft added shared business-process memory, so an agent carries forward requirements, unresolved questions and context across review cycles, and automated triggers that let it act proactively on recurring processes. Makers still configure permissions for each tool, and a tool acting as an end user may still require that person's authentication and consent.

The preview is tightly fenced: account-based agents cannot respond to people outside their tenant, cannot read documents labelled Confidential or Internal Only, can generate new documents but not co-author existing ones, and email is limited to in-tenant conversations of up to 250 people. It requires a Frontier-enrolled tenant, Microsoft 365 E5 with the Agent 365 add-on or E7, and a Frontier for AI Teammates licence per agent, with usage billed in Copilot Credits.

ANNOUNCED

Microsoft launches Copilot for legal: 20-plus purpose-built legal skills for Word and Cowork

Microsoft announced on Thursday the launch of Copilot for legal, a set of more than 20 purpose-built legal skills for its Copilot assistant. The offering ships as a pair of plugins, one for Microsoft Word and one for Microsoft Cowork, its agentic product that performs tasks across Microsoft applications.

Copilot for legal evolves features first unveiled as part of Microsoft's Legal Agent in April. It is a further move by Microsoft into legal tech, a vertical where dedicated AI vendors and law-firm-built tools have been competing hard for document and research work.

The launch continues Microsoft's pattern of packaging Copilot around specific professions rather than selling one general assistant, following this month's Copilot Notebooks and Office-editing updates.

RELEASED

Meta and Apple team up: new open-weight Muse Spark Open, tuned to run locally on high-end Macs

Meta announced on Thursday a new open-weight model, Muse Spark Open, built for coding and agentic work, alongside a collaboration with longtime rival Apple to run it locally on high-end Mac computers. Apple has updated its MLX software toolkit so the model runs well on the Mac mini and Mac Studio machines that developers increasingly use to run sophisticated models at home or in the office.

Meta AI chief Alexandr Wang framed the partnership as a bet that much of the world will run on open models, and that those models should be American. Open-weight releases make their core components downloadable by anyone, and Meta is pitching Muse Spark Open as a US-made option in a field where Chinese open models from Alibaba, DeepSeek and Moonshot are popular.

Meta acknowledged that Chinese open models are still stronger for software development and some other tasks. Muse Spark Open is the second open-weight model Meta has released since it overhauled its AI efforts last year, and it lands three days after the Muse Wonder launch put Meta back in the frontier race.

Evening update: the weights are downloadable today, so this is a release, not just an announcement. MacRumors and TechCrunch put the model at 816 billion total parameters with 42 billion active and a 1 million-token context window, natively multimodal and built for coding and agentic work. Running it locally takes two high-end Mac Studios pooling about a terabyte of unified memory; it is also available through several API platforms and model catalogs on launch day.

RELEASED

Liquid AI releases d1-ultra, a multimodal decision model that returns probabilities instead of text, at $0.042 per million input tokens

Liquid AI released d1-ultra on October 15, a multimodal decision model that takes text, images or both and returns calibrated probabilities for yes/no, multiple-choice and score questions in a single forward pass, without generating output tokens. It has a 32,768-token context window and is billed on input only, at $0.042 per million input tokens.

Liquid pitches it at work such as spotting defects against a product specification, reviewing damage claims from customer photos and checking whether product images match their listings. The company says the model leads its 112-test suite and scores 65 on the Artificial Analysis Decision Index, ahead of TypeSafe's Jev at 58 and Microsoft-Decision-1 at 54 — all vendor-reported figures on vendor-chosen tasks.

The model is available through Liquid's API, with Vercel AI Gateway and OpenRouter support listed. The decision-model category keeps filling in — Microsoft, Cloudflare, AWS and OpenAI have all shipped versions since September — but an early independent trading-replay test ranked d1-ultra near the bottom of its field, a reminder that a confident probability is only useful if it stays calibrated on the customer's own decisions.

RELEASED

Google launches Universal Cart in the US across Search and the Gemini app

Google launched Universal Cart in the United States on October 15, rolling out across Google Search and the Gemini app a shopping cart it first previewed at I/O in May. Shoppers can save products from multiple retailers into one cart and pick the list back up on another Google surface.

The cart watches saved items in the background and notifies shoppers when a price drops or an out-of-stock item returns. It can also surface promo codes and weigh credit-card perks held in Google Wallet when suggesting how to pay.

Checkout runs through Google's Universal Commerce Protocol. At some large retailers shoppers can pay inside the cart with Google Pay; at others the cart hands off to the merchant's own site, so the experience is not yet uniform. Google is pitching the launch as a holiday-shopping tool.

RELEASED

Google credits Claude with helping find Chrome flaws in two October security updates

Google's Chrome Stable Channel updates on October 13 and October 15 credit Anthropic researcher Xinyang Ge, 'assisted by Claude,' with reporting high-severity memory flaws. The October 15 release fixes ten vulnerabilities in builds 155.0.8059.84/.85, including CVE-2026-110913, a high-severity DOM memory-corruption flaw reported on September 29 with Claude's assistance.

The October 13 release carried the same credit for CVE-2026-110169. Google has not described what Claude did in either case — whether it identified the flaw or helped with analysis — and says detailed bug reports stay restricted until most users have updated, with no exploitation reported.

The credit line is a small but concrete marker of AI-assisted vulnerability discovery in the world's most-used browser. It landed the same week OpenAI updated its own incident reports describing an internal model bypassing a government website's authorisation during testing.

ANNOUNCED

Google AI Plus loses Pro model selection on November 13; keeping Pro means the $19.99 AI Pro plan

Google has begun emailing Google AI Plus subscribers that, starting November 13, 2026, the ability to select Pro models in the Gemini app and in AI Mode in Google Search will be limited to Google AI Pro subscribers and above. The $4.99-a-month Plus plan keeps the Flash and Flash-Lite models, and chats started on a Pro model can be continued, but only on a different model.

For subscribers who rely on Pro models for heavier writing, coding or research, keeping them means moving to AI Pro at $19.99 a month, or AI Ultra at $99.99 — roughly a fourfold jump from Plus. The plan itself is not getting a price rise; it is losing the model picker that made it attractive.

The change follows the launch of Gemini 4 Argon earlier this week, which Google is reserving for AI Pro and Ultra subscribers. An earlier support-document schedule had pointed to an October date, so the November 13 cutoff is later than first signalled.

ANNOUNCED

GitHub Copilot will drop Claude Opus 4.8, Opus 5 and GPT-5.6 Terra on November 16

GitHub announced on October 15 that it will deprecate three models across all Copilot experiences — Chat, inline edits, ask and agent modes, and code completions — on November 16, 2026. The models are Claude Opus 4.8, Claude Opus 5 and GPT-5.6 Terra.

GitHub suggests Claude Opus 5.5 as the replacement for both Opus models, and GPT-6.1 Sol or GPT-6 Luna for GPT-5.6 Terra. Under default model enablement, the alternatives are automatically enabled for Copilot Business and Enterprise customers unless an administrator has disabled them.

One exception applies: Claude Opus 4.8 will remain available to individual Copilot subscribers on annual plans so they retain an Opus offering. The three models themselves are not being retired by Anthropic or OpenAI — only the Copilot route to them closes — so teams with pinned model configs should update them before November 16.