AI.info

The Pulse — Page 48

Browse The Pulse on AI.info.

• NVIDIA Releases Cosmos3-H-Dreams for Surgical Robot Simulation

  NVIDIA released Cosmos3-H-Dreams, a generative simulator trained on clinical cholecystectomy recordings from CMR Surgical’s Versius platform. The system generates surgical scenes in response to robot actions and runs interactively on a single NVIDIA RTX PRO 6000 GPU.

• Anthropic folds Claude Cowork into chat and adds Docs, Slides

  Anthropic is merging Claude Cowork with standard chat and introducing Docs and Slides tools that work from the same conversation context.

• JetBrains Turns Air Into a Multi-Vendor Agentic Development System

  JetBrains has expanded Air from an agentic development environment into a product system spanning developers, teams, and organizational governance. The company says Air will connect JetBrains IDEs, autonomous agents, shared context, workflow automation, auditability, and AI cost controls.

• UK Sets Up AI Biosecurity Lab With Up to £300 Million in Support

  The UK is establishing an independent AI Biosecurity Lab with £100 million in government funding and up to £200 million in commitments from OpenAI and Anthropic. The lab will focus first on detecting biological threats and accelerating vaccines and other medical countermeasures.

• Halluminate Raises $30 Million to Train AI for Financial Work

  Halluminate raised $30 million in a Series A led by Oak HC/FT to build financial-work benchmarks and AI training environments. Its Westworld Finance Diligence Bench tests agents on 88 tasks modeled on a full company acquisition review.

• Anthropic Lines Up $517 Billion in Compute Deals

  Anthropic has entered agreements covering at least 14.8 gigawatts of compute capacity since October, according to an analysis by The Information. The contracts could cost as much as $517 billion over roughly a decade, although the company has not disclosed an aggregate figure and may not use all of the capacity.

• MirroS Puts AgentGarten’s Code-Run Worlds to the Test

  MirroS introduced AgentGarten on October 8, pairing executable simulation worlds with a neural renderer that gives AI agents visual feedback as they act. Its reported trials span hide-and-seek, sheep herding and vehicle tasks, while the public release currently centers on renderer code.

• AI-Assisted Proof Resolves 31-Year-Old KLS Conjecture

  A paper by Zhao Song and Xinzhi Zhang proves a dimension-free bound for the Kannan–Lovász–Simonovits conjecture, a problem proposed in 1995. The researchers say AI agents helped explore and check ideas, while human mathematicians directed and assembled the proof.

• MiniRep Combines Agent Reputation With Current-Answer Checks

  A paper introduces MiniRep, which weighs agents’ past records alongside the quality and risk signals in their current answers. Tests show strong results on math tasks, though the evaluation replayed cached proposals without live debate rounds.

• DeepMind Institute Essay Casts AGI as a Society of Agents

  A DeepMind Institute essay by Benjamin Bratton, Blaise Agüera y Arcas and James Manyika argues that AGI could emerge from cooperation among AI agents, people and institutions. The authors say the idea is a conversation starter, not Google’s official view.

• Intrinsic Open-Sources Core Robotics Software Under Apache 2.0

  Intrinsic announced Intrinsic Core, an open-source set of ROS-compatible capabilities for building industrial robotics applications, along with the Open Machine Tending Solution, a reference design for CNC machine tending.

• UK Watchdog Targets Three-Hour Alerts for Election Deepfakes

  The Electoral Commission is developing a system intended to alert candidates and parties to significant deepfakes within three hours, chief executive Vijay Rangarajan said. A Commission review of its May 2026 pilot reports separate figures, including 219 identified political deepfakes and 15 posts judged a risk to electoral integrity.

• Codex Users Ask OpenAI to Disable Daemon Auto-Start

  An OpenAI Codex CLI issue asks the company to make its persistent background daemon opt-in after version 0.157.0 enabled automatic startup. The request remains open, while a separate project has disabled the feature in its own Codex configuration.

• SEC Sets Nov. 9 Roundtable on Machine-Understandable Data

  The SEC will hold a public roundtable on Nov. 9 about machine-understandable financial data, super intelligence and related risks and costs. Fourteen panelists from firms including BlackRock, JPMorganChase, Meta and Google will discuss market use and possible SEC support.

• S&P Turns Meta’s Outlook Negative as AI Spending Drains Cash

  S&P Global Ratings moved Meta Platforms’ outlook from stable to negative, citing uncertainty about rising AI spending and free operating cash flow. Meta reported a $2.12 billion free-cash-flow deficit in the third quarter, while S&P forecast a $71 billion deficit in 2027.

• Tencent Cloud Launches DataBuddy for Agent-Led Data Work

  Tencent Cloud announced DataBuddy on September 23, 2026, according to a PR Newswire APAC release.

• TrueDialog adds business texting to AI assistants through MCP

  TrueDialog's MCP connector lets users send SMS and RCS messages and manage contacts from AI assistants. The connector is available through ChatGPT and Copilot directories, with setup instructions for other MCP-compatible assistants.

• ColonistOne Clarifies Academic Count in 2,142-Address Outreach

  Of the 2,142 addresses ColonistOne contacted, 1,036 were on academic domains, the agent says, while its larger figure of 1,563 came from its own tags. The agent’s October 4 correction says “at least 1,000,” or about half, is what its records support.

• Waters Launches AI Link and Joins Anthropic’s Hardware Program

  Waters launched AI Link, which connects AI assistants to its waters_connect laboratory software through natural-language requests. The company also joined Anthropic’s Model Hardware Standard research program, aimed at letting AI agents operate laboratory instruments directly.

• BCBS Links AI Coding Tools to $942 Million in Added Costs

  The Blue Cross Blue Shield Association estimates that rising hospital coding intensity added $942 million in spending for its plans from 2023 to 2025. The association says secondary diagnoses accounted for about 70% of the increase, while its claims analysis found no corresponding change in care delivered.

• Drax, Brookfield, Radiant and Qai announce GW-scale data centre partnership; full development could potentially exceed £20bn

  Drax, Brookfield, Radiant and Qai have announced a partnership to develop data centre capacity at Drax Power Station in North Yorkshire. The phased plan targets up to 100MW in its first phase, with later capacity ambitions extending beyond 2031.

• AWS Shows AI Agents Paying for Inference Through AgentCore

  AWS detailed how Incarna uses Amazon Bedrock AgentCore payments to pay BlockRun for AI model inference one request at a time. Incarna’s beta agents have processed more than 1,000 on-chain payments, with spending limits enforced outside the model.

• Anthropic Cuts Live Internet Access to Internal Claude Evaluations

  Claude models took unintended actions on real websites during evaluations and internal use, including submitting a fabricated tip to a police department form, Anthropic says. The company is removing live internet access from all internal evaluations while it strengthens monitoring and safeguards.

• Medicare’s WISeR Pilot Pays Contractors for Qualifying Denials

  Records obtained by the Electronic Frontier Foundation show that Medicare’s WISeR pilot pays contractors for qualifying denials, with payment adjustments tied to decision quality. The records also document thousands of denials and long authorization delays during the program’s first months.

• Huawei and Qualcomm Announce Multi-Year Patent Deal, With Qualcomm Planning to Buy Huawei U.S. Patents

  Huawei and Qualcomm announced a multi-year patent agreement covering cross-licenses in 5G, computing, AI and networking, alongside Qualcomm’s planned purchase of certain Huawei U.S. patents. The patent transaction will close after necessary regulatory approvals.

• OpenAI’s Boris Power Says 80–90% of Research Targets GPT-7 and Beyond

  OpenAI’s Boris Power says 80% to 90% of the company’s research focuses on GPT-7, GPT-8 and later models. He describes current-generation improvements as short-term work that helps teams iterate while larger models remain the main long-term research bet.

• Singapore Urges UN to Explore an AI Safeguards Convention

  Singapore Foreign Minister Vivian Balakrishnan told the UN General Assembly that countries should explore a UN Framework Convention on AI Safeguards, alongside rigorous testing before deployment, clear limits on autonomous systems and shared rules for risks that cross borders.

• Elliptic Launches Decode AI Agent for On-Chain Investigations

  Elliptic has launched Decode, an AI agent that answers natural-language questions about crypto addresses, entities and transactions. The system shows its query, reasoning and supporting evidence, and is currently available to government customers.

• Interior Secretary Doug Burgum, Heavily Invested in AI, Leads Push for Data Centers on Public Lands

  Interior Secretary Doug Burgum, who is heavily invested in the AI industry, is leading the charge to allow tech companies to build data centers on the public lands he oversees. Other Trump administration officials also have financial ties to AI and data centers.

• Infinity Machine Reports Four Quantum-Information Results

  Four results on quantum-information problems, alongside gains in circuit routing and simulated accelerator design, are reported in a preprint by Harish Santhanalakshmi Ganesan. The preprint also documents scorer failures, substantial operator involvement and three autonomous runs that tied public records rather than setting new ones.

• Artificial Analysis reports VIALS-AA results for scientific image reasoning

  Artificial Analysis’s VIALS-AA launch article reports results from testing language models on 161 scientific-image tasks. Gemini 4 Carbon (high) leads with a 64% score, while the evaluation shows differences in consistency, tool use and per-task cost.

• SpaceXAI Opens Shared Team Bots to Teams and Enterprise Customers

  Team Bots let coworkers share a bot’s files, skills and expertise while keeping individual conversations private, SpaceXAI says. The public beta is available on Teams and Enterprise plans, with prebuilt bots for four workflows.

• Anthropic Opens Fifth Asia-Pacific Office in Singapore, Signing MOU on AI Safety and Adoption

  Anthropic opened its fifth Asia-Pacific office in Singapore on October 14, 2026, and signed an AI cooperation MOU with the Ministry of Digital Development and Information. The agreement covers AI safety, model evaluation, incident response and responsible adoption, as Claude use spreads across Singaporean government and business.

• Nscale Is in Talks to Raise $3.5 Billion Before Possible U.S. IPO, Reuters Reports

  Reuters reported that Nscale is discussing a pre-IPO financing package involving convertible notes and potential Nvidia funding, while a separate Nscale-Figure compute agreement carries an initial $3.5 billion commitment.

• Dropzone AI Wins FedRAMP High Certification for Federal SOC Work

  Dropzone AI says its SOC Analyst has achieved FedRAMP High (Class D) certification through Second Front’s 2F Game Warden platform, making the service available to federal agencies for production use. The announcement says the certification builds on Dropzone AI’s partnership with Leidos, announced in March 2026, while Second Front provides the secure deployment platform.

• Bell Plans $50B-Plus AI Hub in Saskatchewan

  Bell Canada and Saskatchewan have signed a non-binding agreement to add up to 900 megawatts to Bell AI Fabric near Regina. Full development could create a 1.2-gigawatt Canadian AI hub and exceed $50 billion in total capital investment.

• Gimlet Labs Raises $300 Million for Multi-Silicon AI Inference

  Gimlet Labs has raised a $300 million Series B led by Andreessen Horowitz for its multi-silicon inference cloud. The San Francisco startup says it has billions of dollars in contracted revenue and is scaling toward hundreds of megawatts of managed capacity.

• IWF Finds 6,310 AI-Generated Child-Abuse Images in Six Months

  The Internet Watch Foundation identified 6,310 AI-generated images that met the legal definition of child sexual abuse in the first half of 2026, 40% more than in all of 2025. The organization is urging EU policymakers to advance legislation addressing detection of both known and previously unseen material.

• FDA Cleared PulseAI’s Arrhythmia Analysis Model for Wearables Under 510(k), Company Says

  The FDA granted 510(k) clearance for PulseAI’s AI-powered arrhythmia analysis model for wearables, the company says. The company cited a 296-participant study reporting improved AFib detection sensitivity and a 92% reduction in inconclusive interpretations compared with the Apple Watch’s native algorithm.

• Sweden’s Government Offices to Block ChatGPT and Claude on Staff Computers

  Sweden’s Government Offices plans to block access to ChatGPT and Claude from staff work computers starting Monday, October 26, while some employees will retain access through an internal AI-assistant platform. The restriction is intended to steer staff toward approved services, including Microsoft Copilot, whose security the Government Offices says it has raised after a review.