AI news for Tuesday, 20 October 2026
Stories 121 to 180 of 198, most important first. Back to the first page →
-
AMD adds FSR 4.1 support for RX 6000 GPUs, Xbox Series X/S and Ryzen AI chips
AMD announced that its FidelityFX Super Resolution 4.1 (FSR 4.1) is now compatible with Radeon RX 6000‑series GPUs, Xbox Series X and S consoles, Ryzen Z‑series handhelds, and devices using Ryzen AI Max and AI 9 chips.…
1 sourceTom's Hardware -
Students plan AI walkout outside OpenAI's SF office Tuesday
Students and community members in San Francisco are organizing a walkout on Tuesday to demand an immediate freeze on artificial intelligence development. The event is part of a nationwide action staged by groups…
2 sourceshoodline.comyahoo.com -
EvoluteIQ CSO Deepak Kinger outlines agentic AI approach to regulatory compliance
EvoluteIQ co-founder and CSO Deepak Kinger describes how the company's agentic AI platform aims to shift banks from project-based regulatory responses to a permanent operating capability. Drawing on the 2020 LIBOR…
1 sourceUnite.AI -
Tsinghua University releases PPMem, a plane-based memory for AI video models
Tsinghua University announced Persistent Planar Memory for Video World Models (PPMem), a new approach that stores 3D memory as compact plane primitives instead of dense point clouds. The system builds on the Wan 2.2‑5B…
1 sourceUnite.AI -
Octave Reliance builds secure multi-tenant AI foundation on Amazon Bedrock
Octave Reliance’s Reliance platform serves over 14,000 customers in 180 countries, covering more than 60% of the Global Fortune 500. The SaaS quality‑management system supports over 650 enterprise customers with more…
1 source primary sourceAWS Machine Learning Blog -
Amazon Bedrock adds cross‑Region inference for OpenAI GPT models in Canada
Amazon Bedrock now lets Canadian developers call the latest OpenAI foundation models – GPT‑5.6 (Sol, Terra, Luna), GPT‑6 (Astra, Sol, Luna) and GPT‑6.1 Sol – through cross‑Region inference (CRIS). The models accept…
1 source primary sourceAWS Machine Learning Blog -
US sanctions MiraclePlus over alleged ties to Zhisuan, accused of stealing Google AI
the united states treasury added miracleplus, a chinese venture capital firm, to its sanctions list on friday, citing alleged ties to zhisuan. the move follows accusations that zhisuan stole google's artificial…
1 sourcebloomberg.com -
Commodity brokers prepare AI compute futures tied to Nvidia GPU rentals
Commodity brokers including ADM Investor Services, BGC Group and Nanhua USA LLC are gearing up to offer cash‑settled futures that track the hourly rental price of Nvidia GPUs, primarily the H100 and B200 models. Each…
1 sourcecryptobriefing.com -
Exa claims Exa 3.0 search API leads Perplexity, Brave and Parallel on eight benchmarks
Exa Labs announced Exa 3.0, an upgraded web‑search API for AI agents. The release adds a larger, undisclosed embedding model, a new "Lite" search type priced at $4 per 1,000 requests, and Dynamic Highlights that cut…
1 sourceMarkTechPost -
Asymptote Labs raises $5.5 million seed to monitor enterprise AI agents
Asymptote Labs announced an oversubscribed $5.5 million seed round to fund engineering, product development, hiring, and go‑to‑market expansion. The round was led by Tower Research Ventures, the venture arm of Tower…
1 sourceUnite.AI -
TripleTen survey shows finance leads non‑tech AI adoption, healthcare lags
A 2026 TripleTen survey finds finance workers are the most exposed to generative AI outside the tech sector, with 48% asked to use AI chatbots and 32% asked to automate parts of their workflow. By contrast, only 18% of…
1 sourceMIT Technology Review AI -
Arkose Labs releases Q3 2026 agentic AI trust report
Arkose Labs, a leader in agent trust and orchestration, published its State of Agentic AI Trust Report for Q3 2026. The report shows that more than 65% of customer sites—including banking, gaming, technology and travel…
1 sourcefinance.yahoo.com -
Grafana Labs expands Grafana Cloud with AI agents
Grafana Labs announced at ObservabilityCON 2026 that it is adding AI agents to Grafana Cloud. The new Assistant Watchers and Agent Observability are generally available, while Grafana Cloud Agents is in public preview.…
1 sourceUnite.AI -
xMEMS begins mass production of XMC-2400 and XMC-4800 micro‑cooling chips
xMEMS Labs announced that it has moved its solid‑state micro‑cooling technology into mass production, targeting the XMC-2400 and XMC-4800 µCooling chips. The company says the first commercial products using XMC-2400…
1 sourceUnite.AI -
Researchers introduce SURE-Act to improve Tool-Using agents
Tool-using large language model agents often must choose between acting autonomously, gathering more information, or escalating to a human, but existing data only show the outcomes of actions actually taken. To fill…
1 source primary sourcearXiv cs.AI -
Perplexity publishes guide on AI gateway mechanics and limits
Perplexity released a guide on Oct 20, 2026 titled "What Is an AI Gateway? A Guide to Routing, Cost Control, and Governance for LLM and Agent Traffic". It explains what an AI gateway is, how it sits between an…
1 sourceUnite.AI -
Infinity and HP double AI workload capacity on NVIDIA GB300 superchip
Infinity Artificial Intelligence Institute and HP Inc. announced that their joint effort using Infinity Auto Research optimized the NVIDIA Nemotron 3 Super model on the HP ZGX Fury workstation, which is powered by…
1 sourcefinance.yahoo.com -
WGA East partners with Cloak to block AI training on writers' work
The Writers Guild of America, East announced a partnership with Cloak, a software that encrypts files so they remain readable to humans but illegible to AI models. The tool will be offered for free to WGAE members as…
1 sourceyahoo.com -
OpenAI adds lock feature to ChatGPT
OpenAI announced on Tuesday that users can now lock individual ChatGPT conversations. A locked chat can be secured with a four‑digit PIN, Face ID or Touch ID on supported devices. The feature creates a separate “Locked…
1 sourceTechCrunch AI -
Cisco report says small businesses struggle to monitor AI agents and control costs
Cisco's Splunk State of Observability 2026 report shows small businesses increasingly use AI agents but face monitoring challenges. Nearly 90% of organizations say AI monitoring is a priority, yet more than half cannot…
1 sourceSmall Business Trends -
Could this be a Deep Blue moment for the cryptic crossword?
Darmok, a new AI system, solved the final puzzle of the Times’ annual cryptic crossword about three minutes faster than Liam Hughes, the contest’s youngest human champion. The Times final was completed in roughly three…
1 sourceThe Guardian AI -
Motley Fool scores Meta ahead of Alphabet on AGI readiness
Motley Fool’s AIball scoring system compares Alphabet and Meta Platforms on how well each is positioned for the rise of artificial general intelligence (AGI). The analysis notes that Meta’s ad revenue is largely…
1 sourcefool.com -
Co-Existence released, uses Anthropic Opus 5.5 to chart 30 months of AI progress
The author’s new book Co‑Existence launched today, looking back at AI development since the early‑2024 release of Co‑Intelligence. Using Anthropic’s Opus 5.5 in Claude Projects, the author let a single “orchestrator”…
1 sourceOne Useful Thing -
Opinion: Australian parties may push AI rollout despite falling public trust
Australian Labor and the Coalition are moving toward a joint AI strategy even as public confidence wanes. Only 30% of Australians thought AI’s benefits outweighed its risks last year, and a 2026 poll suggests trust has…
1 sourceThe Guardian AI -
Avenda Health launches Surgery Mode for intraoperative use of Unfold AI
Avenda Health announced Surgery Mode, a suite of intraoperative features that extend its Unfold AI platform into the operating room. The multimodal AI system creates a 3D map of prostate cancer, allowing surgeons to…
1 sourceThe Robot Report -
Analysts say investors rotating to cash‑rich mag 7, pressuring neocloud stocks
Roundhill Investments' chief investment strategist Drew Pettit and JonesTrading's chief market strategist Mike O’Rourke said investors are shifting toward cash‑rich Magnificent Seven (MAG 7) technology giants and away…
1 sourcestocktwits.com -
AtScale named OpenAI Select Partner to bring governed metrics to enterprise AI
AtScale announced on October 20, 2026, that it has been named an OpenAI Select Partner within the OpenAI Partner Network. The partnership aims to deliver governed metric computation and trusted business context to…
1 sourcefinance.yahoo.com -
Eon uses Amazon Bedrock AgentCore gateway to unify multi-cloud investigation tools
Eon built a single Model Context Protocol (MCP) gateway on Amazon Bedrock AgentCore to replace per‑tool credentials and local MCP servers across AWS, Azure and Google Cloud. The managed gateway provides identity…
1 source primary sourceAWS Machine Learning Blog -
Study finds most LLM agent decisions on prompt injections are set at read time
Researchers examined how LLM agents handle prompt injections by replaying episodes up to the moment an injection is read and then resampling the remainder many times. Across four AgentDojo suites they measured that a…
1 source primary sourcearXiv cs.CL -
Researchers introduce GardenBench, a lightweight daily benchmark for LLM capabilities
A team of AI researchers has released GardenBench, a new benchmark that contains 11 test suites aimed at evaluating large language models on everyday tasks. The benchmark uses simple, low‑cost prompts to measure…
1 source primary sourcearXiv cs.AI -
F5 provides BIG-IP Next for Kubernetes to Telenor AI Factory for sovereign AI
F5 announced that Telenor AI Factory will use F5 BIG-IP Next for Kubernetes on NVIDIA BlueField DPUs as the AI service plane for its sovereign AI infrastructure. The solution adds Kubernetes‑native traffic management,…
1 sourcefinance.yahoo.com -
Visa, BILL and Billtrust launch AI-driven payments pilot for businesses
Visa, BILL and Billtrust have begun a live pilot that uses trusted AI agents to automate invoice payments in supplier portals. The agents rely on Visa’s Trusted Agent Protocol and tokenized payment credentials,…
1 sourceSmall Business Trends -
OpenAI resets rate limits, adds locked chats and promises credit to 40,000 accounts
OpenAI announced a reset of usage limits just after noon on the 15th day of its 28‑day announcement streak, a move that should immediately lift rate‑limit blocks for most users. The same update introduced “Locked…
1 sourcenote.com -
The Robot Report opens nominations for 2027 TRR50 Robotics Innovation Awards
The Robot Report announced that nominations are now open for its 2027 TRR50 Robotics Innovation Awards, the program formerly known as the RBR50. Submissions must be entered through the online form by Friday, Jan. 5…
1 sourceThe Robot Report -
Cloudflare and OpenAI pilot indexes hundreds of millions of new pages for ChatGPT
Cloudflare and OpenAI say a research pilot that used about 1% of Cloudflare’s network helped OpenAI discover and index hundreds of millions of new web pages. The companies shared early results on Oct. 15 in a joint…
1 sourceSearch Engine Journal -
AI agent Velvet Vale ejected from AoIR listserv after member backlash
During the second day of the Association of Internet Researchers (AoIR) conference in Mexico City, an AI‑driven agent introduced itself on the community mailing list Air‑l as “Velvet Vale.” The post asked whether an…
1 sourceWired AI -
HSBC CEO says AI boom requires selective data center investments
HSBC chief executive Georges Elhedery told CNBC that while artificial intelligence is a transformational technology, investors should be selective about data‑center investments. He urged looking at the quality of…
1 sourceCNBC Technology -
TTEC Digital launches GovEx AI Agent Suite with 40+ prebuilt agents on Google Cloud Gemini
TTEC Digital announced the GovEx AI Agent Suite, a catalog of more than 40 prebuilt conversational agents built for U.S. government agencies. The suite expands the company’s FedRAMP‑compliant CX offerings and runs on…
1 sourcefinance.yahoo.com -
LogicVal benchmark evaluates constraint satisfaction in Text-to-Image models
Text‑to‑image (t2i) models now produce high‑fidelity images from natural‑language prompts, and researchers have created many benchmarks to test specific abilities. In this paper, the authors propose a new way to…
1 source primary sourcearXiv cs.AI -
OpenAI’s new math results raise concerns about post‑quantum encryption
OpenAI released a trove of more than 400 new mathematical results this month, including a solution to the Erdős unit‑distance conjecture and a claimed answer to the Navier‑Stokes equations. The breadth of the release…
1 sourcedefenseone.com -
OR-License benchmark tests LLM agents on real operations research tasks
Researchers introduced OR-License, a new benchmark designed to evaluate large language model agents on real-world operations research (OR) problems. Unlike existing benchmarks that use simplified or predefined tasks,…
1 source primary sourcearXiv cs.CL -
Is this the beginning of the end of the AI stock market boom?
Investors are questioning whether the estimated $US1tn AI investment this year rests on solid ground. The Nasdaq 100 slipped 5.5% last week, reviving fears of an AI bust, while the failed ASX listing of Australian…
1 sourceThe Guardian AI -
Researchers introduce Mirror, a self‑supervised video‑to‑action model
The paper presents Mirror, a self‑supervised approach that learns to map video plans to actions without any pre‑collected action data or hand‑crafted rewards. Mirror iteratively tries to follow a visual plan, records…
1 source primary sourcearXiv cs.AI -
OpenAI may revive Sora video model in deal with new AI video startup
OpenAI shut down its Sora video‑generation service in April 2026 after a March announcement that the model’s compute demands were straining the company’s resources and its focus shifted toward enterprise AI, coding…
1 sourcenote.com -
Google faces Senate grilling over education AI
Senate lawmakers Richard Blumenthal (D-Conn.) and Josh Hawley (R-Mo.) sent a letter to Google CEO Sundar Pichai on Wednesday, demanding answers about the company’s AI use in American classrooms. They ask how Google…
1 sourcenypost.com -
HPE offers Private Cloud AI platform to lower effective cost per token
HPE’s Private Cloud AI platform is presented as a production‑ready foundation that combines infrastructure, software, models, tools, governance, and services. It lets enterprises manage scheduling, model selection, GPU…
1 sourceMIT Technology Review AI -
New York mayor Mamdani says feds have abdicated on AI, urges city pushback
At the AI Now Institute’s People’s AI Assembly in Manhattan, Mayor Zohran Mamdani and activist Ilana Glazer debated the rapid rollout of artificial intelligence. Mamdani warned New Yorkers not to accept AI as…
1 sourceTechCrunch AI -
Google tests AI‑assisted software engineer interviews using Gemini, internal docs show
Internal documents reviewed by Business Insider show that Google is piloting AI‑assisted changes to its software engineer hiring process. The rollout, begun earlier this year, lets candidates use the company’s Gemini…
1 sourcebusinessinsider.com -
Customers Bank deploys Palantir Foundry and AIP for AI-native deposit onboarding
Customers Bank, the U.S. subsidiary of Customers Bancorp, is embedding Palantir Technologies engineers to run its Foundry and Artificial Intelligence Platform (AIP) on deposit‑account opening. The bank says the new…
1 sourcefintechmagazine.com -
Amulya Vishwanath says physical AI is in its LEGO era
Amulya Vishwanath, former Nvidia product marketer and founder‑managing partner of Techable Ventures, told TechCrunch’s Equity that the robotics field is experiencing a "ChatGPT moment" but is still in a "LEGO era." She…
1 sourcetechcrunch.com -
Opinion: AI scales human bias instead of eliminating it
AI systems inherit the judgments and blind spots embedded in the data they are trained on, so when they are used for hiring, healthcare or employee evaluation they can amplify existing discrimination at a speed no…
1 sourceUnite.AI -
Barcelona Supercomputing Center expands AI Factory Accelerator, adds NVIDIA access
Barcelona Supercomputing Center (BSC) has moved its AI Factory Accelerator Programme into full operation, offering startups sovereign compute, technical support and training. The program, which began in April 2025, has…
1 source primary sourceNVIDIA Blog -
Researchers find transformer prompt influence decays as power law
The paper, posted on arXiv, examines how much of a transformer model’s prompt remains after generating tokens. The authors show that the distance between hidden states of two different prompts decays approximately as a…
1 source primary sourcearXiv cs.AI -
Crypto bettors say Meta may launch Muse AI glasses by year‑end
Cryptocurrency bettors on the Polygon‑based Polymarket platform are pricing a 21% chance that Meta Platforms, Inc. will make Muse‑powered AI glasses available by Oct. 31, down from 61% a month ago and 25% a week ago.…
1 sourcefinance.yahoo.com -
Medical AI expands into rare‑disease drug discovery and virtual lab markets
DeepMind’s AlphaFold has predicted more than 200 million protein structures, giving AI‑driven drug discovery startups a large data foundation. Recursion Pharmaceuticals, a US biotech, reports encouraging Phase 1b/2…
1 sourceUnite.AI -
Opinion: AI reverse‑engineering tools let users clone smart‑TV software
A wave of AI‑driven "vibe coding" is letting hobbyists replace proprietary software on smart TVs, picture frames, baby monitors and other IoT devices. Over the weekend a GitHub project called “Reverse Engineer…
1 source HN 23404media.co -
Opinion: enterprise AI needs cumulative context before it can earn autonomy
Enterprises can quickly generate plausible answers from AI agents, but missing business‑specific context often leads to errors. The article illustrates a finance scenario where an agent used the wrong forecast version,…
1 sourceUnite.AI -
Z.ai founders in 2018‑2021 CGTRI, MI5 says CGTRI had very strong ties to China’s spy agency
Four founders of Z.AI Co., a Chinese artificial‑intelligence startup described as one of the country’s most promising AI firms, took part in a research project that ran from 2018 to 2021. The project was co‑funded by…
1 sourcebloomberg.com -
Study probes if LMs understand variable roles via SAEs
Researchers from arXiv cs.CL published a study titled "Do LMs Know What Variables Do?" to investigate whether large language models (LLMs) truly understand the functional roles of variables in code. The team used…
1 source primary sourcearXiv cs.CL -
Researchers prove Prefix-Tuning can exactly steer pretrained transformers
A new arXiv paper titled “Prefix‑Tuning for Arbitrary Output Sequences on Pretrained Transformers” provides a theoretical proof that a task‑specific prefix can be constructed to make a pretrained autoregressive…
1 source primary sourcearXiv cs.CL
