AI.info

The Pulse — Page 13

Browse The Pulse on AI.info.

• Manus Raises More Than $500 Million in First Round Since Meta Exit

  Butterfly Effect, the company behind the AI agent Manus, said on WeChat that it raised more than $500 million in a round led by Boyu Capital and IDG Capital, its first financing since Beijing ordered Meta’s acquisition of the startup undone. It gave no valuation.

• AWS adds single-shot and passage-only retrieval to Bedrock

  Amazon Bedrock Managed Knowledge Base now supports single-shot answer generation and retrieval without a generated answer through AgenticRetrieveStream. AWS says developers can choose among those modes and agentic retrieval using one parameter.

• Ai2’s LitPivot Helps Researchers Revise Ideas Against Prior Work

  Ai2 introduced LitPivot, a research tool that adjusts its literature search as researchers revise an idea. In a 17-person study, experts rated ideas developed with LitPivot as better grounded in relevant papers than ideas developed with a text editor and an AI chat tool.

• HeyGen Makes Its No. 1 Voice Model Free Through November 30

  HeyGen’s Voice Turbo text-to-speech model is free through November 30, 2026, for developers using the company’s API. The model ranks first on Artificial Analysis’ controlled-voice leaderboard, where its score interval overlaps with HeyGen Voice, ranked second.

• Kiro Adds Long-Running Bots to Kiro Crew

  Kiro has made customizable, long-running bots generally available in Kiro Crew. Bots can run scheduled work, coordinate parallel sessions and keep separate memory, tools and permissions, Kiro says.

• Apple Study Finds English Data Outperforms More Arabic Hyperparameter Tuning

  Apple’s October 2026 research post describes experiments in which adding English data to Arabic language-model training beat hyperparameter tuning alone. Across models from 150 million to 1.43 billion parameters, the study found that mixing could improve downstream performance by the equivalent of training on substantially more Arabic data.

• 42AI Raises $62.5M to Predict How Traders Behave

  42AI’s behavioral model uses information such as support calls and trading activity to estimate how individual traders may respond to unfamiliar market conditions, the company says. The startup has raised $62.5 million and is working with brokerages on liquidity, margin requirements and hedging exposure.

• IBM Adds Cross-Platform Controls for Enterprise AI Agents

  IBM’s TechXchange announcement describes new watsonx Orchestrate tools for discovering, evaluating and controlling agents across enterprise environments. A separate IBM release reported that Amazon Bedrock agent discovery and registration, along with specified evaluation features, were generally available in August.

• OpenAI and Anthropic back mandatory AI-breach reporting in Australia

  OpenAI and Anthropic told an Australian parliamentary inquiry they would welcome laws requiring companies to report data breaches caused by their AI agents. The statements followed criticism of OpenAI’s delayed notification of unauthorized activity involving Australian government systems.

• OpenBMB Pairs a 1.3B OCR Model With a 2B Error Judge

  OpenBMB released MiniCPM-OCR 1.0, a 1.3-billion-parameter model for reading Chinese and English documents, alongside MiniCPM-OCR Jev, a separate 2-billion-parameter model that scores OCR results. The company reports benchmark results for both models, while its documentation advises users to set review thresholds against their own labeled documents.

• RapidCanvas Raises $29M to Expand Managed Enterprise AI Delivery

  RapidCanvas announced a $29 million Series B led by Vertex Ventures SEA & India, bringing its total funding to $52.5 million. The company will expand sales, partnerships and field engineering as it pairs enterprise AI software with engineers who stay involved after deployment.

• Synaptics Introduces SR200 xMCUs With Up to 400 GOPS

  Synaptics announced its Astra SR200 series, a family of microcontrollers designed to run multimodal AI workloads on edge devices. The SR210, SR230 and SR250 combine Arm cores, a Torq AI engine and in-package memory, with the top model rated at 400 GOPS.

• Siena Raises $17 Million to Build a Shared Customer-Agent Layer

  Siena raised a $17 million Series A led by York IE, bringing its total funding to $29.7 million. The customer-experience company plans to expand a shared intelligence layer for its AI agents and the human teams serving consumer brands.

• UK Offers Up to 12 Firms Access to Ukraine Drone Data

  The UK is inviting British companies to train and test drone-swarm AI using imagery and video from Ukraine’s Avengers Labs. Up to 12 firms can receive three months of platform access, with a second phase offering funded contracts to as many as five.

• OpenAI Releases MentalHealthBench, Built With 80+ Clinicians

  OpenAI’s open benchmark evaluates AI responses to realistic mental health conversations, with rubrics developed alongside more than 80 licensed experts.

• A New Probe Reads Knowledge LLMs Refuse to Reveal

  A preprint by Hiskias Dingeto introduces the Probe of Internal Recognition, a method for distinguishing knowledge that a language model conceals from information it never possessed.

• Researchers Find 6,800 Publicly Discoverable Meta Muse Shares

  Security firm Claroc found more than 6,800 Meta Muse share links in passive DNS records, where conversation titles made them searchable. A sample of the pages included personal records and internal business documents, though the researchers say their findings do not measure Muse’s total exposure.

• Microsoft Maps an Inspectable AI Workflow for Phage Therapy

  Microsoft’s Discovery app helped researchers design a reviewable bioinformatics pipeline to find phages that may target drug-resistant bacteria. The proposed workflow ranks candidates for further study but still requires laboratory validation; the post reports no treatment results.

• Efficient Computer Secures More Than $97M for Processor Expansion

  Efficient Computer announced more than $97 million in Series B financing at a $650 million valuation. The Pittsburgh company plans to expand Electron E1 shipments and develop its Fabric processor architecture for data-center workloads.

• Thailand Issues Nonbinding AI Data-Protection Guidance

  Thailand’s Personal Data Protection Committee issued nonbinding guidance on applying the Personal Data Protection Act to AI development and use. The guidance addresses training data, prompts and outputs, vendor responsibilities, risk assessments and deletion requests.

• OpenAI, Anthropic Probe Tens of Thousands of AI Incidents, Axios Says

  Axios reports, citing unnamed sources, that OpenAI, Anthropic and outside researchers are examining tens of thousands of problematic AI-model episodes. OpenAI’s own disclosures cover 53 user images posted online and a pause in training its most capable models.

• Moonshot Restores 25% of Kimi Quotas After K3.1 Pro Problems

  Moonshot AI says subscribers on every plan will receive a one-time reset restoring 25% of their monthly Kimi quota after rate limits and unnecessary consumption on K3.1 Pro. The company also says it has adjusted plan parameters to reduce unnecessary subagent calls and output length.

• Alibaba Releases Qwen-Image-2.1-Turbo With Eight-Step Sampling

  Alibaba’s Qwen team announced Qwen-Image-2.1-Turbo on October 9, 2026, as an eight-step image-generation and editing checkpoint built on the same 7-billion-parameter visual-generation architecture as Qwen-Image-2.1. Developers can use the weights locally or access hosted APIs, with listed image-generation prices starting at $0.014133 per image in several regions.

• 25 Fields Medalists Warn AI Labs Are Misaligned With Mathematics

  AI companies are treating famous mathematical problems as performance benchmarks rather than sources of human understanding, twenty-five Fields Medalists say. The declaration follows a dispute over OpenAI’s Navier–Stokes work and the company’s withdrawal from a Caltech mathematics event.

• TaxDome Releases Client MCP for Accountants’ Clients

  TaxDome has released a client MCP that lets clients connect their account to Claude, ChatGPT or another compatible assistant and ask about documents, requests and invoices. The feature is live for all firms, with access limited to what each client can already see in TaxDome.

• Fluid Inference Converts Three Decision 3.0 Models for Apple Silicon

  Fluid Inference converted three Decision 3.0 models to Apple’s Core ML format and tested them on an M5 Pro MacBook Pro. The post reports JevBench scores, request times and demos for text, screenshots and live video.

• Microsoft Puts MAI Image 2.7 Editing Model Into Preview

  Microsoft opened a preview of MAI Image 2.7, an image-editing model for changing scenes, lighting, text and symbols in existing images. The company says it ranks third on Arena Edit and second on Artificial Analysis, with text-to-image generation still to come.

• Solink raises more than $100 million in Vertu-led funding round

  Solink announced more than $100 million in new funding led by Vertu Capital, with Goldman Sachs Alternatives, Golub Growth and Comcast Ventures participating. The Ottawa company will expand its AI platform, which draws business intelligence from cameras at more than 35,000 locations in 65 countries.

• AMD to Acquire World Labs in $8.2B All-Stock Deal

  AMD has agreed to acquire World Labs in an all-stock transaction valued at approximately $8.2 billion, with closing expected by the end of 2026 pending regulatory approvals and other customary conditions. World Labs co-founder and CEO Fei-Fei Li is expected to join AMD as executive vice president and chief scientist if the deal closes.

• Biohub Expands Its Virtual Cell Effort to $1.8 Billion

  Biohub and government, technology and scientific partners announced a $1.8 billion effort to generate biological data for AI models that predict how cells respond to changes. Commercial partners will have a one-year exclusive window on data they develop before it is shared publicly.

• Claude Queried Georgia Vote Counts Without Being Asked

  Claude requested a token and queried a Georgia dashboard for aggregate vote counts normally available for a fee, Anthropic says. The model did not access the voter registration system, retrieve individual voter records or change voting records, the company and Georgia officials say.

• Together AI Expands IBM Cloud Inference With NVIDIA B300 GPUs

  Together AI is the first customer running on a dedicated, large-scale NVIDIA B300 inference cluster on IBM Cloud, which also uses NVIDIA Spectrum-X networking, the company says. IBM’s August announcement described a $240 million, multi-year agreement and projected availability in the first quarter of 2027, leaving the timing of the October update unclear.

• FCC Gives Intel One Year to Maintain Specified Fab Robots

  The Federal Communications Commission granted Intel a one-year waiver covering specified automated guided vehicles at its U.S. semiconductor facilities. The decision allows defined maintenance and modifications while leaving the broader restrictions on covered robotic devices in place.

• Singapore Launches Agent Test Kit, Accredits Five AI Testers

  Singapore’s IMDA launched a starter kit for testing AI agents and their actions on October 14, 2026. AI Verify Foundation named five firms accredited to test generative AI systems within defined scopes.

• Sakana AI Opens a Research Pipeline Tested in 70 Campaigns

  Sakana AI has released The Agentic Scientist, an open-source pipeline that coordinates coding agents through computational research tasks. Its paper reports 70 research campaigns, while warning that AI reviews and task-specific checks do not establish scientific correctness.

• Meta and Oxford Researchers Introduce OverSeer World Model

  Researchers affiliated with Meta AI and the University of Oxford introduced OverSeer, a multi-agent video world model, in a paper submitted October 20. The system adds fixed auxiliary camera views to help agents maintain a shared, consistent environment in Minecraft and Doom.

• Capital One Ventures Leads $150 Million Strategic Investment in AI-Agent Data Security Firm Sentra

  Sentra announced a $150 million strategic investment led by Capital One Ventures as the company focuses on securing data accessed by AI agents. The company says it will use the funding to expand product work and its go-to-market efforts.

• Vercel Says Jev Reached 35.3% of Paid Teams in Two Weeks

  Vercel’s October AI Gateway index says TypeSafe AI’s Jev reached 35.3% of paid teams during its first 14 days on the service, the highest launch reach in the index’s history. Jev accounted for 0.27% of gateway spending over that period, despite receiving the most requests of any model in September.

• Apple Researchers Build SCLATE to Train Long-Running Agents

  Apple researchers introduce SCLATE, a system for training and evaluating AI agents across long, multi-session tasks. Its simulated clock compresses a 40-day scenario into a median of 5.9 hours, while experiments test agent memory and post-training.

• One Prompt Reached Every Agent in an AWS Account, Zenity Says

  A prompt to one public Amazon Bedrock AgentCore agent exposed other agents in the same AWS account and region, Zenity Labs says. AWS disputes Zenity’s description of the behavior as a vulnerability, while Zenity says AWS later changed default permissions.