Tuesday, 20 October 2026
198 stories, most important first. This page shows the top 60.
-
Tesla and SpaceX form committees to explore $3.5T merger around AI
Tesla and SpaceX have each formed special committees of independent directors to explore a potential merger, Elon Musk disclosed on SpaceX's third-quarter earnings call on October 20. The combination would unite…
6 sourcesThe Verge AICNBC Technologyvectorwire.aiBBC Technology -
Anthropic's Claude autonomously designs protein binders against scorpion neurotoxins
Anthropic's Claude, in collaboration with Citrini Research, autonomously designed protein minibinders targeting five scorpion neurotoxins from four species, generating 318,484 candidate designs over a 24-hour…
5 sources primary source HN 41The Verge AInote.comvectorwire.aiThe Decoder +1 more -
Meta pauses Muse AI sharing after researchers find 6,800+ chats in public DNS records
Meta has suspended the link-sharing feature in its Muse AI assistant after security firm Claroc discovered more than 6,800 shared conversation URLs accessible through public passive-DNS records. Researchers found the…
8 sourcespcworld.comindiatoday.intech.yahoo.comibtimes.sg +4 more -
Anthropic reports $80.7 billion loss in first half of 2026 amid rising infrastructure costs and upcoming IPO
Anthropic reported a net loss of $80.7 billion in the first half of 2026, according to draft prospectus documents reviewed by Reuters and The Wall Street Journal. The loss was primarily driven by accounting…
4 sourcesfool.comamericanbazaaronline.comfinance.yahoo.comnotebookcheck.net -
Nvidia CEO Jensen Huang pledges $2 billion for Europe’s sovereign AI push
Nvidia chief Jensen Huang is urging governments to create "sovereign AI" – national AI infrastructure that keeps data, models and hardware in‑house. According to Axios, he offered to help fund countries that want to…
9 sourcesCNBC Technologyfinance.yahoo.comcryptobriefing.comtaipeitimes.com -
DeepSeek open-sources DeepInfer inference framework for DeepSeek-V4.1-Flash
DeepSeek has open-sourced DeepInfer, its internal AI inference framework, under the Apache 2.0 license, making it available to developers. The framework powers DeepSeek-V4.1-Flash and runs on NVIDIA SM100 GPUs and…
7 sourcesLatent SpaceThe DecoderMarkTechPostbloomberg.com +3 more -
Google open-sources TorchTPU, a native PyTorch backend for TPUs, co-developed with Meta
Google announced on October 20, 2026, at the PyTorch Conference 2026 that TorchTPU, its native PyTorch stack for Tensor Processing Units (TPUs), is now available as open source on GitHub under the Apache 2.0 license.…
6 sources primary sourceTom's HardwareUnite.AIMarkTechPostarXiv cs.AI +2 more -
Meta's Muse AI agent reaches 10.9 million downloads amid declining ChatGPT usage
Meta's Muse AI agent has achieved 10.9 million downloads since its September 8 launch, with over 10.2 million from the US, according to Sensor Tower data reported by Yahoo Finance. The app reached 2.5 million daily…
15 sourcesfinance.yahoo.comtheverge.comcnbc.comtech.yahoo.com +7 more -
OpenAI acquires Orchestra Bio to expand Rosalind life sciences tools
OpenAI announced on October 20, 2026, that it has acquired Orchestra Bio, a startup founded around 2022 that builds AI agents to help pharmaceutical companies manage drug development programs. The Orchestra team and…
19 sourcesfinance.yahoo.comThe Decodernote.comvectorwire.ai +7 more -
OpenAI quietly withdraws from pushing for chip export restrictions to China, report says
A hacker using AI agents breached more than 100 corporate websites across the US, Canada, Australia and Japan, costing roughly $5,000 total – about $50 per intrusion – according to cybersecurity researcher Snorre…
4 sourcesdarkreading.comnote.com -
Adobe reports 27% of shoppers trust AI search summaries over traditional results
Adobe’s analysis of AI search and shopper abandonment finds that 27% of consumers now trust AI‑generated search summaries more than traditional search results, while another 36% trust both equally. The study also notes…
3 sourcesSemrush BlogMarTechSmall Business Trends -
Rankpage Vietnam expands AI search optimization services to local and global platforms
rankpage vietnam, part of the regional rankpage seo network, announced an expanded focus on generative engine optimization (geo) and ai search visibility. the agency says it will help businesses improve brand…
4 sourcesHubSpot Marketing Blogforbes.comfinance.yahoo.com -
SpaceX reports Q3 revenue up 161% to $14.58B, AI revenue up 830%
SpaceX announced third‑quarter revenue of $14.58 billion, a 161% year‑over‑year increase and above the $12.72 billion LSEG consensus estimate. The company posted operating income of $2.8 billion and net income of $4.4…
3 sources HN 6CNBC TechnologyTechCrunch AIcnbc.com -
Meta posts first negative free cash flow in over a decade as AI spending rises
Meta Platforms Inc. reported third‑quarter results that beat estimates, but AI‑related capital spending drove free cash flow negative for the first time since the first half of 2012. Capital expenditures, including…
6 sourcesfinance.yahoo.comnote.comcryptobriefing.com247wallst.com -
Nvidia in talks to invest hundreds of millions in Australian AI cloud firm Firmus
Nvidia is in talks to put hundreds of millions of dollars into Firmus, an Australian AI cloud provider that already runs a large fleet of Nvidia chips. If the deal closes, Nvidia would become both a supplier and a…
6 sourcesTechCrunch AIaol.comtheinformation.comfinance.yahoo.com +1 more -
Amazon pledges over €16 billion for German AWS, AI data centers and robotics lab
Amazon.com Inc. announced it will spend more than €16 billion (about $17.9 billion) in Germany during 2027. The money will fund new AWS and AI‑focused data centres, expand the Leo satellite communications programme,…
1 sourcebloomberg.com -
U.S. Army Cyber Command developing echotribble program to deploy autonomous AI agents to neutralize targets
The U.S. Army Cyber Command (ARCYBER) is developing a program called ECHOTRIBBLE to deploy autonomous AI agents that can navigate networks, identify targets, and neutralize them by disabling or deleting threats,…
3 sourcesTom's HardwareThe Decodertheintercept.com -
Meta ordered to name executives who approved torrenting for AI training
A federal magistrate, Thomas Hixon, issued a discovery order in a class‑action lawsuit filed by authors including Richard Kadrey and Sarah Silverman, and a separate suit by publisher Entrepreneur Media. The order…
3 sources HN 23publishersweekly.comcryptobriefing.comarstechnica.com -
Pigment raises $250 million at $2.3 billion valuation
Pigment, a French AI startup that builds software for finance forecasting and slide‑deck generation, announced a new funding round. The company said it raised $250 million, bringing its post‑money valuation to $2.3…
2 sourcesUnite.AIbloomberg.com -
AWS launches generally available Agent Registry for runtime tool discovery
AWS announced that its Agent Registry service is generally available, offering a managed, searchable catalog of Model Context Protocol (MCP) servers, agents, and agent skills. Publishers register protocol‑specific…
5 sources primary sourceAWS Machine Learning Blog -
Hugging Face CEO criticizes Netflix’s AI hack documentary for exaggeration
Clem Delangue, CEO of Hugging Face, criticized Netflix’s documentary Instadocs: AI Gone Wild for overstating risks from AI agents. The film recounts a July 2026 incident where OpenAI’s test agents breached Hugging…
3 sourcestribune.net.phdigit.inTechCrunch AI -
Meta's Muse and OpenAI's Dots split websites into agent and non‑agent tiers
Meta’s Muse and OpenAI’s Dots act as personal AI agents that can search the web, fill forms, add items to a cart and complete checkout. Muse runs on a Chromium‑based cloud browser and is already available to anyone 18…
1 sourceSearch Engine Journal -
Ornith-ai/Ornith-1.5-122B (Qwen finetune)
Benchmarks such as Terminal‑Bench, SWE‑Bench, NL2Repo, and ClawEval show state‑of‑the‑art results among open‑source models of comparable size, with faster inference and shorter outputs. The release includes open…
2 sources primary sourcehuggingface.co -
Anthropic may draft AI safety bill that could block models if over 50 people could die
Anthropic has drafted a bill, Politico reported, that would let the government block an AI model’s release when a “foreseeable, imminent risk” could cause over 50 people to die or over a billion dollars in property…
3 sourcesmoroccoworldnews.commashable.comTechCrunch AI -
IBM and Red Hat launch llm-d 1.0 distributed inference stack
IBM Research announced llm-d 1.0 at the PyTorch Conference North America on October 20, 2026. The Kubernetes‑native stack is positioned as a vendor‑neutral way for organizations to run any model on any accelerator in…
3 sourcesMarkTechPostUnite.AItechtarget.com -
ByteDance, Alibaba and Tencent order 3 million NVIDIA RTX PRO 5500 GPUs for $35 billion
ByteDance, Alibaba and Tencent have placed orders for more than 3 million NVIDIA RTX PRO 5500 workstation graphics cards, according to a Reuters exclusive that cites people familiar with the matter. At roughly $11,400…
1 sourcetweaktown.com -
Yelp adds Check for Me voice agent to call restaurants for wait times
Yelp introduced a new feature called Check for Me that lets users type in their party size and phone number, after which an AI voice agent calls the restaurant, checks the wait time and texts the result back. The tool…
2 sourcesSmall Business TrendsTechCrunch AI -
Gecko Robotics secures $300 million contract expansion with U.S. Navy
Gecko Robotics announced that the U.S. Navy and the General Services Administration have increased the ceiling of its contract by $300 million, four‑fold from the original $71 million IDIQ award. The expansion is…
1 sourceThe Robot Report -
Google rolls out Personal Intelligence for Gemini in cars starting with Android Auto
Google is rolling out its Personal Intelligence feature for Gemini to Android Auto starting today, with availability for vehicles featuring Google built-in expected in the coming months. The feature allows Gemini to…
8 sourcesThe Verge AISmall Business Trends9to5google.comArs Technica AI +3 more -
Daytona raises $60 million Series B at $1 billion valuation
Daytona, a San Francisco startup that builds cloud sandboxes for AI agents, announced a $60 million Series B round on Tuesday. The round was led by FirstMark and included investors such as Qualcomm Ventures, Samsung…
1 sourcefortune.com -
Google launches Gemini 4 Argon in Japan with up to 1 million-token context
Google introduced Gemini 4 Argon, its newest flagship reasoning model, for the Japanese market on October 14, 2026. The model can handle a shared context of about 1.04 million tokens and generate up to 65,536 tokens in…
2 sourcescryptobriefing.comnote.com -
CTM shows how call transcripts can boost lead quality and ad targeting
CTM’s Lisa Salvatore, senior manager of integrated marketing, and Brian Barranger, senior account executive, explained a repeatable workflow that turns phone‑call recordings, chats, texts and form submissions into…
1 sourceSearch Engine Journal -
Jensen Huang urges Europe to cut AI regulation to avoid falling behind
Nvidia CEO Jensen Huang has urged European leaders to reduce regulatory caution and accelerate AI adoption during a week-long tour of France, Sweden, the UK, and Germany. Speaking with German Chancellor Friedrich Merz…
3 sourcesbloomberg.comcryptobriefing.comaol.com -
Salesforce to acquire Braze for nearly $5.7B in enterprise value
Salesforce announced an agreement to buy Braze, an artificial‑intelligence‑powered marketing platform, for nearly $5.7 billion in enterprise value, pricing the transaction at $47.50 per share. The deal, expected to…
1 sourcebloomberg.com -
Otter.ai raises $320 million from General Catalyst’s Customer Value Fund
Otter.ai announced it has secured $320 million from General Catalyst’s Customer Value Fund, a growth fund that provides capital without diluting equity. The financing brings the startup’s total funding to $50 million,…
2 sourcesUnite.AITechCrunch AI -
Meta removes 'training' from crawl clause in commercial terms effective Oct 30
Meta updated its Commercial Terms, which take effect on Oct 30 2026, to delete the word “training” from the list of purposes for which it may crawl sites linked to a business’s products or services. The revised clause…
3 sourcesThe Verge AIsnopes.comnote.com -
ALTK-Evolve tops τ²-bench retail leaderboard using GPT-4.1 with learned guidelines
Researchers from AgentToolkit posted results showing that a basic ReAct agent powered by GPT-4.1 reaches 67.5% Pass^4 on the τ²-bench retail domain, topping the October 12, 2026 leaderboard snapshot. The method, called…
2 sources primary sourceUnite.AIHugging Face -
AGCO signs deal to pilot Agility Robotics Digit in Kansas
AGCO and Agility Robotics announced a commercial agreement on October 20, 2026, to pilot Digit humanoid robots in AGCO’s manufacturing operations. The U.S. pilot is scheduled to begin in 2027 at AGCO’s facility in…
2 sourcesThe Robot ReportUnite.AI -
OpenAI gpt-oss:20b runs on a Mac mini
The author describes how to run OpenAI’s gpt-oss:20b model locally on a Mac mini (M6) with 32 GB of RAM using Ollama, Open WebUI, and the Python API. The gpt-oss:20b model, which has about 21 billion parameters, fits…
1 sourcenote.com -
Eka Robotics valued at $1.45 billion after $150 million Series A
Eka Robotics announced a $150 million Series A funding round that puts the startup’s post‑money valuation at $1.45 billion. The round was led by venture firms a16z and F‑Prime. The company was founded by MIT professor…
2 sourcesUnite.AITechCrunch AI -
Vectorial raises $3 million seed to test products with simulated audiences
Vectorial announced a $3 million seed round on Oct. 20, backed by Smilegate, SenseAI Ventures and Dent Capital. The San Francisco startup also cites support from product leaders at Meta, Shopify and Mastercard. The…
1 sourceUnite.AI -
US bans DeepSeek from civilian government systems
The US Department of Homeland Security (DHS) has ordered all federal civilian agencies to immediately stop using and remove DeepSeek AI tools from their systems. The order, issued by the Trump administration, excludes…
1 sourceyahoo.com -
U.S. government bans DeepSeek AI app from civilian federal IT systems
The Department of Homeland Security announced on Tuesday that the Chinese artificial‑intelligence application DeepSeek is prohibited from all civilian federal agency IT systems, with the exception of the Department of…
2 sourcesthestar.com.myupi.com -
Mistral AI launches Voxtral 3 voice agent stack
Mistral AI introduced Voxtral 3, a family of language models designed for real‑time voice agents. The stack bundles a speech‑to‑text model called Voxtral Small Transcribe 3, a text‑to‑speech model named Voxtral Mini…
2 sourcesMarkTechPostThe Decoder -
Meta moves Q3 earnings report to October 21, citing Muse AI enthusiasm
Meta announced that its third‑quarter earnings release will be issued on October 21, a full week ahead of the originally scheduled date. The company framed the early filing as a historically bullish signal, pointing to…
8 sourcesfinance.yahoo.comseekingalpha.com247wallst.comcoinpedia.org +1 more -
Researchers introduce M-Plan framework for LLM agents with memory arrangement
Researchers at an unnamed institution have proposed M-Plan, a new framework for LLM-based agents designed to improve performance on long-horizon tasks. The system combines long-term and short-term memory to address…
4 sources primary sourcearXiv cs.CLarXiv cs.AI -
ASI raises $325 million Series C at $3 billion valuation after $875 million FAA contract
ASI, a startup that builds artificial‑intelligence tools for the military and for air‑traffic controllers, announced a $325 million Series C financing round. The capital, raised from undisclosed investors, values the…
1 sourcebloomberg.com -
US restarts three nuclear reactors to supply power for AI data centers
Three US nuclear reactors are being brought back online after years of shutdown, adding about 2.25 GW of capacity that will help meet growing electricity demand from AI data centers. The Palisades plant in Michigan,…
2 sourcesTom's Hardwaregadgetreview.com -
Alibaba, Tencent and Huawei invest billions in Shaoguan data centers
According to the Financial Times, Alibaba, Tencent and Huawei are each investing billions to build data centers in Shaoguan, a remote region of Guangdong province. The companies cite the area's cheap electricity –…
1 sourceft.com -
OpenAI says GPT-6 Astra low setting outperforms GPT-5.6 Sol high setting
OpenAI’s Head of Product, Thibaud Sottiaux, announced on X on September 6, 2026 that the new GPT‑6 Astra model’s “low” inference setting delivers higher quality and speed than the “high” setting of the…
1 sourcenote.com -
Homebase launches AI assistant to fill shifts and manage hourly teams
Homebase announced Homebase Assistant on October 20, a new AI tool designed to automate staffing and administrative tasks for hourly teams. The system identifies available employees, contacts potential replacements,…
1 sourceUnite.AI -
World Bank partners with Cassava and Nvidia to offer pay‑as‑you‑go AI compute in Africa
The World Bank Group announced the AI Infrastructure and Readiness Initiative, a knowledge‑partner program with Cassava Technologies and Nvidia that will give emerging‑market governments shared, pay‑as‑you‑go access to…
1 sourcetechtrendske.co.ke -
Zapier lets multiple AI tools share one MCP server
Zapier introduced a new capability in its Multi-Client Platform (MCP) that lets a single server be shared across multiple AI tools. Teams using Claude, ChatGPT, Cursor and similar services can connect each tool as a…
1 sourceZapier Blog -
Nvidia releases Cosmos3-H-Dreams real-time generative surgical simulator
Nvidia announced Cosmos3-H-Dreams, its first real‑time generative surgical simulator trained on clinical cholecystectomy recordings. The system learns visual dynamics directly from synchronized video and robot…
2 sources primary sourceUnite.AIHugging Face -
Anthropic says self-copying AI could target cheap TV boxes within months
Anthropic’s Frontier Red Team linked 15 off‑brand Android TV boxes, each costing about $55, and ran the open Gemma 4 model with 26 billion parameters on them. The boxes together generated roughly 0.4 tokens per second,…
1 sourcenotebookcheck.net -
OpenAI says it will demo humanoid robot prototype in 2027
OpenAI announced that it is building its own humanoid robots and aims to showcase an early prototype by 2027. CEO Sam Altman confirmed the timeline in a conversation with journalist Alex Heath on the Sources podcast…
1 sourceynetnews.com -
Sierra partners with Stripe to enable in-call payments for AI agents
Sierra announced a partnership with Stripe on October 20, 2026, allowing customers to pay and receive receipts during AI agent phone calls. The integration removes the need to switch systems or use payment links,…
2 sourcesUnite.AIthenextweb.com -
MI5 head warns AI could change the stakes on biological weapons
Sir Ken McCallum, the head of the UK intelligence service MI5, warned during his annual threat update that AI could change the stakes for both existing and new biological weapons. In response to the warning, Guardian…
2 sourcesThe Guardian AI -
Knexus partners with Google Cloud for Gemini government AI launch
Knexus announced it is a launch partner for Google Cloud’s Gemini for government services, integrating its agentic AI solutions into secure, regulated workflows for public agencies. The company, a two-time Google Cloud…
2 sourcesfinance.yahoo.com -
Nvidia CEO Jensen Huang calls AI doomsday warnings cheap
Jensen Huang, chief executive of Nvidia, told a conference in Berlin on Thursday that warnings AI could wipe out humanity are "cheap" attempts to generate headlines. He made the remarks at the final stop of his…
2 sourceseconomictimes.indiatimes.comgeo.tv
