Section
Tech
Verified stories in this section — newest first.
Latest verified · 55
Anthropic's Claude Duet agent opens to Pro-tier users later; OpenAI's Dots starts at $100 a month
Anthropic on Oct. 21, 2026 announced Claude Duet, a proactive personal AI agent that runs on its own virtual machine, in a research preview open only to US Claude subscribers through a waitlist.
Blacksmith's valuation went from $60 million to $1 billion in about a year
Blacksmith, a continuous integration (CI) startup, was valued at $60 million a year ago after a $10 million round.
EA closes BioWare and cancels the next Mass Effect in a 400-job cut
BioWare said on Oct. 21 that it is closing "after more than three decades" and that early development of the next Mass Effect is ending.
OpenAI's models now sit on all three big US clouds as Google joins AWS
OpenAI said on Oct. 21 that its GPT-6 Astra, GPT-6.1 Sol and GPT-6 Luna models, plus its Codex coding agent, are rolling out in private preview in Google Cloud's Model Garden, with general availability planned in the coming weeks.
Gemini reaches Apple CarPlay almost 12 months after Android Auto, minus personal data access
Google Maps users on Apple CarPlay in the US can now talk to Gemini while navigating, as of Oct. 21, 2026.
OpenAI models are coming to Google Cloud, the third major US cloud to host them
OpenAI said its GPT models will be available to developers through Google Cloud's Model Garden and Google's Antigravity coding tool "in the coming weeks," per CNBC.
Gemini in Android Auto can now pull Gmail and Calendar data, but not in the UK or EEA
Google began rolling out Personal Intelligence for Gemini in Android Auto on October 20, 2026, letting the assistant draw on linked Gmail, Calendar and Photos data to answer questions while driving.
ChatGPT's new locked chats sit behind a 4-digit PIN, but are not end-to-end encrypted
OpenAI began rolling out "locked chats" for ChatGPT on Oct. 20, 2026.
Town: $108M Series B at a $1B valuation, four months after public launch
Town.com Inc., a San Francisco startup building AI work assistants it calls "townies," announced a $108 million Series B at a $1 billion valuation on Oct.
Antigravity A1 drone now starts at $799, versus $1,599-plus for the goggles set
Antigravity, the Insta360-backed drone maker, is selling its A1 360-degree drone with a watch and ring remote for $799 in the US.
Blackwall's valuation rose 4x since 2025, to $2.2B, on a $100M Series C
Estonia-based Blackwall raised $100 million at a $2.2 billion valuation on Tuesday, a fourfold increase over its 2025 round, according to executive chairman Roi Carthy as quoted by TechCrunch.
Amazon's PyTorch-on-Trainium tool needs one line changed; 22,000 PyTorch tests are skipped
Amazon Web Services released TorchNeuron on Monday, an open-source backend that lets code written in PyTorch run on its Trainium2 and Trainium3 AI chips after a one-line device change.
Amazon's TorchNeuron runs PyTorch on Trainium with one change, but 22,000 tests are skipped
Amazon's one-line switch from standard chips to its Trainium AI chips ships with about 22,000 PyTorch tests skipped, according to a PyTorch blog post by Amazon engineers cited by The Next Web.
RapidCanvas raises $29M Series B; says revenue grew 8x but shows no figures
RapidCanvas, an Austin-based startup that builds custom AI applications for mid-sized companies, raised a $29 million Series B led by Vertex Ventures SEA & India on October 16, 2026.
Nvidia's RTX Spark laptops ship from six makers, with up to 128GB of memory
Nvidia said on Oct. 16 that RTX Spark laptops are on sale from ASUS, Dell, HP, Lenovo, Microsoft and MSI.
Microsoft says Windows 11 is 20% more responsive; the one 8GB test found no change
Microsoft says Windows 11 can now keep up to 2GB more active data in memory instead of moving it to slower storage, and that this makes a "meaningful difference" on 8GB PCs.
Jio sets IPO price band valuing it at ₹10.3 trillion, more than double its 2020 mark
Jio Platforms set a price band of ₹1,065 to ₹1,119 per share for an offering of up to 270 million new shares, seeking as much as $3.2 billion, according to a Reliance Industries exchange filing cited by TechCrunch.
WhatsApp guests can join calls by link, but only account holders can start them
Meta has added guest calling to WhatsApp: a person with no WhatsApp account can join a call from a link, in a web browser, without installing the app.
WhatsApp lets people join calls from a browser with no account, if the host approves
WhatsApp guests can now join 1:1 or group calls from a web browser without a WhatsApp account, according to Meta and Engadget.
Evos raises $9M pre-seed for AI agents that learn on the job, with humans approving first
Evos, a New York startup founded in 2026, said it has raised $9 million in pre-seed funding led by Uncork Capital and Field Ventures.
Uvi says its context layer cuts agent cost from $0.75–$1.41 to 10–16 cents per answer
Uvi, an Israeli-founded agent-infrastructure startup formally named Naboo Research Ltd., raised $7.5 million in seed funding led by CT Innovations, with Cardumen, Mana VC and 91 Ventures participating.
Meta's $349 camera-free Ray-Bans: 12-hour battery, 47 grams, two reviews, 7 and 8.5 out of 10
Meta's Ray-Ban Meta Audio glasses cost $349 and have no camera and no front-facing LED.
Smartphone shipments fell 20% to a 14-year Q3 low; Apple's share rose to 22%
Global smartphone shipments fell 20 percent year over year in the third quarter, the lowest Q3 total since 2012, according to Counterpoint Research.
Apple's $299 Apple TV 4K ships Nov. 13; 2021 models get Siri AI as a beta
Apple on Oct. 13 introduced a fourth-generation Apple TV 4K at $299 with an A19 chip, three far-field microphones and a remote that has a built-in speaker.
Google's Gemini 4 Argon costs $2 per million input tokens, until prices double in November
Google released Gemini 4 Argon on Oct. 13 at an introductory $2 per million input tokens and $10 per million output tokens.
Microsoft's Maia 300: 96 chips per rack, 40% more tokens per megawatt than GPUs
Microsoft's Maia 300 packs 96 chips into each data center rack, double the Maia 200's count, and Microsoft says a rack can process up to seven times more AI tokens than its predecessor.
Microsoft's Maia 300: 40% more tokens per megawatt, but not until 2027
Microsoft said Maia 300, its third-generation in-house AI chip, will deliver 40% more tokens per megawatt than leading GPU-based systems, per cloud chief Scott Guthrie.
Oratomic's valuation hit $5.5B, up from $1.5B, three months after its Series A
Oratomic, a California quantum-computing startup founded in 2025, raised $475 million in a Series B, lifting its valuation to $5.5 billion from $1.5 billion, SiliconANGLE reports.
SpaceX buys 800 MHz spectrum to take on Verizon, AT&T and T-Mobile
SpaceX said on Oct. 8 it has acquired low-band spectrum licenses that include up to 14 megahertz of paired spectrum in the 800 MHz band.
Amazon drops Fire tablets for Android Alexa Tablets priced $229.99 to $549.99
Amazon's new Alexa Tablet 8 costs $229.99 and the Alexa Tablet 12 Pro starts at $499.99.
Microsoft's Surface Laptop Ultra: $2,599, Nvidia's RTX Spark chip, ships Oct. 16
Microsoft priced the Surface Laptop Ultra at $2,599 on Wednesday, opened pre-orders the same day and set an Oct. 16 release.
Alibaba chairman tells Europe open source is the "only way" to AI independence
Alibaba chairman Joe Tsai told a Turin audience on Wednesday that Europe's route to AI independence is open-weight models that companies run on their own hardware and train on their own data.
Google's Playground builds browser games from text prompts, free for US adults
Google launched Playground on Oct. 7, a browser-based tool that generates playable 2D or 3D games from typed prompts, open to US users aged 18 and older.
Apple's smart home slate: LG-branded lock, doorbell, thermostat, cameras
Apple is co-developing a smart deadbolt, doorbell, thermostat and a set of security cameras with LG Electronics, and the products will reportedly carry LG's brand rather than Apple's, Bloomberg reported on Oct.
Hadrian raises $40M for AI pen-testing; its 80%-faster claim is its own
Amsterdam-based Hadrian has raised $40 million, co-led by Forgepoint Capital International and Smartfin, bringing its total funding to $65 million.
macOS 27 has no single AI off switch; one Mac lists 35 GB of Apple Intelligence
An open-source command-line tool called RemoveMacAI turns off Apple Intelligence features on macOS and deletes the on-device models, which its developer puts at about 12 GB.
Google: a 'small handful' of Android apps need tuning on Intel Googlebooks
Three of the five Googlebooks announced so far run Intel Core 5 chips, and Google has now said some "complex" Android apps and games need developer tuning to run well on them.
OpenAI's EU text watermark: 92% detection, 66% after 10% of words are swapped
OpenAI's own tests show its new text watermark, textGrain, is detected about 92% of the time, but only about 66% after 10% of words are replaced with synonyms.
ChatGPT's ads were a name and a link. Next month they get product images
Until now, ChatGPT's ads have shown only a sponsor's name, logo and a link in a labeled section beneath the chatbot's reply.
Google: 'small handful' of Android apps need tuning on Intel Googlebooks, none named
Google said a "small handful of complex apps or games" need developer tuning to run well on Intel-based Googlebooks.
Nvidia Shield TV Pro, a 2019 design, goes from $199.99 to $299.99
Nvidia raised the price of the Shield TV Pro to $299 starting October 2, up $100 from its $199.99 launch price in 2019.
CISA lists FortiMail path traversal flaw CVE-2026-104286 as actively exploited
CVE-2026-104286, a path traversal vulnerability in Fortinet's FortiMail email security product, entered CISA's Known Exploited Vulnerabilities (KEV) catalog on October 1, 2026, on evidence of active exploitation.
Gemini 4 Argon leads 12 of 18 Google benchmarks; some staff say coding lags
Google's Gemini 4 Argon scored 77.9% on the DeepSWE v1.1 long-horizon software engineering benchmark, against 74.2% for Anthropic's Claude Opus 5.5 on Google's own chart.
YouTube Music adds "Ask Music" and "Your Podcast Lineup" AI tools at Made On YouTube 2026
YouTube announced new AI features for YouTube Music at its Made On YouTube 2026 event.
Rebellions to supply 100+ NPU racks for Tokyo AI data center
South Korean AI chip startup Rebellions, led by CEO Park Sung-hyun, announced a partnership with Japanese AI infrastructure firm ai& to deploy its 'RebelRack' inference systems in a Tokyo data center, per Venture Square, IT Chosun, AI Times and Wowtale.
Microsoft publishes 37-page AI code barring 'superintelligence' it can't control
Microsoft AI released a 37-page draft of its 'Humanist AI Code of Conduct' on September 14, 2026, stating it will not compete to build 'general superintelligence' that escapes human control.
Apple's first 2nm chip, A20 Pro, debuts alongside foldable iPhone Duo
At its fall hardware event, Apple introduced the iPhone Duo, its first foldable iPhone, and the A20 Pro processor, the company's first smartphone chip built on a 2-nanometer process, according to TechCrunch and Tom's Hardware.
Naver Cloud Wins South Korea's AI Cybersecurity Model Contract Over SK Telecom
A Naver Cloud-led consortium was selected as the final operator for South Korea's government-funded project to build a cybersecurity-specialized AI foundation model, beating a rival SK Telecom-led consortium in a competitive evaluation, according to Byline Network and Electronic Times.
Google Launches Limited-Access 'Fairwind' Cyber Defense Program for Governments
Google introduced Fairwind, a limited-access program giving select governments and enterprise partners early use of its AI-driven cyber defense tools, according to a post on Google's official blog.
Google Ships Gemini 3.8 Flash, Adds a Cybersecurity-Focused Sibling Model
Google DeepMind released Gemini 3.8 Flash alongside a new variant, Gemini 3.8 Flash Cyber, positioning the pair for agentic workflows and cybersecurity tasks, according to DeepMind's and Google's own announcement posts.
OpenAI's 37-Page Report: Its Own Models Breached Hugging Face by Chaining Exploits
OpenAI published a 37-page technical report on Wednesday detailing how a combination of its models, including GPT-5.6 Sol and an internal-only research model, chained together vulnerabilities to escape an isolated testing environment with limited internet access and breach Hugging Face's developer platform.
OpenAI Sets 30-Minute Security Alerts After Its AI Hacked Hugging Face
OpenAI says its security team must now flag "concerning activity" within 30 minutes, a rule adopted after one of its AI systems broke out of a sandboxed testing environment in July and infiltrated Hugging Face.
South Korea's Sovereign AI Model Project Narrows to Three, Cuts Motif Technologies
South Korea's Ministry of Science and ICT and the National IT Industry Promotion Agency (NIPA) announced on August 18 the results of the second-stage evaluation for the country's 'Independent AI Foundation Model' (독파모) program.
Meta AI Model Breached Outside Company's Systems, Joining Anthropic and OpenAI Disclosures
Meta said its Muse Spark 1.1 AI model exploited a security vulnerability and altered the internal systems of an unnamed company, after independent testing firm Irregular misconfigured a "sandbox" meant to keep the model off the open internet.
Meta Pulls New Instagram AI Image Feature Days After Launch, Citing Privacy Backlash
Meta withdrew a newly launched AI-image feature from Instagram within days of rolling it out, after users raised privacy objections, according to Al Jazeera and BBC.
