Thursday, 15 October 2026
147 stories, most important first. This page shows the top 60.
-
Black Forest Labs releases FLUX 3 Image Klein, a 5.6B-Parameter Open-Weights model for text-to-image generation
Black Forest Labs released FLUX 3 Image Klein on October 15, 2026, a 5.6B-parameter open-weights model for text-to-image generation, image editing, and multi-reference generation with support for up to 10 reference…
3 sourcesbfl.aiThe DecoderUnite.AI -
NVIDIA releases Nemotron 3.5 Super VL and Bolt edge models for enterprise agents
NVIDIA released Nemotron 3.5 Super VL, a 120-billion-parameter open multimodal model built for always-on enterprise agents, alongside two smaller Bolt models for edge deployment. Super VL uses a mixture-of-experts…
10 sources primary sourcehuggingface.cotechtarget.comMarkTechPostmaccus.fr +5 more -
[AI News 10/16] Meta and Apple join forces. The era of running 816B massive AI on Macs is here. New announcements from NVIDIA, CERN, and Amazon|Haruki|AIを試す大学生
The model supports multimodal, coding, AI‑agent and long‑duration reasoning tasks and can handle up to 1 million tokens. Meta and Apple collaborated to optimise the model for Apple Silicon, and Apple’s research shows…
1 sourcenote.com -
NVIDIA launches RTX Spark laptops with local AI agent support worldwide
NVIDIA announced the global launch of RTX Spark laptops on October 16, 2026, featuring systems from ASUS, Dell, HP, Lenovo, Microsoft, and MSI with prices starting at $2,599. The laptops are built on the RTX Spark…
17 sourcesThe Verge AItheregister.compcmag.comextremetech.com +10 more -
Perplexity open-sources SPACE, its AI agent sandbox platform, with partners CoreWeave, Akamai, and Crusoe
Perplexity announced on October 15, 2026 that it is open-sourcing SPACE, the sandbox platform behind its Computer personal agent, releasing the code under the Apache License 2.0 with a public repository and hosted…
3 sourcesMarkTechPostsemafor.comUnite.AI -
Anthropic's IPO valuation may fall to $1.5 trillion amid slowing growth
Anthropic's planned IPO valuation has reportedly dropped from around $2 trillion to $1.5 trillion, according to sources familiar with discussions at its investor day in San Francisco on October 15, 2026. The company…
5 sourcesfinance.yahoo.comthenextweb.comaol.com -
OpenAI upgrades ChatGPT Chrome extension to read and act on open tabs
OpenAI has updated its ChatGPT Chrome extension to allow the AI to access and act on users' open browser tabs when triggered via @chrome mentions in Chat or Work modes. The extension can now read tab content, compare…
3 sources HN 5The DecoderUnite.AIThe Verge AI -
Google updates AI search to display inline links higher and more frequently
Google announced updates to its AI search features, including AI Overviews and AI Mode, to make inline links appear more frequently and higher up in responses. The changes include testing a section at the top of…
5 sourcesbusinesstoday.indigitaltrends.comSearch Engine Journalstraitstimes.com +1 more -
Volvo Cars expands NVIDIA partnership for door-to-door automated driving by 2030
Volvo Cars announced on October 15, 2026, that it is expanding its partnership with NVIDIA to advance automated driving safety in future Volvo vehicles, integrating NVIDIA’s Hyperion sensor and compute architecture,…
3 sourcesfinance.yahoo.comtheverge.comUnite.AI -
OpenAI targets ChatGPT users with ads via prompts, emails, and phone numbers
OpenAI lets advertisers inject targeted ads into ChatGPT sessions using users’ email addresses, phone numbers, or Android advertising IDs. According to tests by 404 Media, advertisers can upload raw or hashed personal…
1 source HN 12404media.co -
Mindbeam announces collaboration with NVIDIA on Litespark Nemotron 3.5 Super VL model and inference engine
Mindbeam announced a collaboration with NVIDIA focused on two layers of the open source AI stack: a post-trained frontier model that generates high-performance CUDA kernels and an inference engine for NVIDIA's latest…
4 sources primary sourceunite.aifinance.yahoo.comNVIDIA Blogzdnet.com -
Simons Foundation, Anthropic and OpenAI commit $15 million to help mathematicians use AI
The Simons Foundation announced a $15 million initiative in partnership with Anthropic and OpenAI to support the mathematics community in navigating AI's rise. The funding will support workshops, open calls for…
2 sources HN 55simonsfoundation.orgaol.com -
Nvidia CEO Jensen Huang says U.S. must work with Chinese AI companies on safety
Nvidia CEO Jensen Huang stated that the U.S. absolutely must work with Chinese AI companies to ensure safe development of artificial intelligence, emphasizing dialogue as a path forward. He made these remarks during a…
6 sourcesfinance.yahoo.commiragenews.comthestar.com.myCNBC Technology +1 more -
Google launches Universal Cart across Search and Gemini app in the US
Google launched its Universal Cart tool in the United States on October 15, 2026, rolling it out across the Gemini app and Google Search. The tool, first previewed at Google I/O in May 2026, allows users to save…
4 sourcesnote.comTechCrunch AIUnite.AIThe Verge AI -
Apertus 1.5 adds 4T tokens to 8B model, 2T to 70B, with multimodal and long-context upgrades
Apertus released version 1.5 of its open-weight models, adding 4 trillion tokens to the 8B model and 2 trillion to the 70B model. The update expands context to 262,144 tokens, introduces native multimodal understanding…
1 source primary sourceHugging Face -
TSMC reports record third-quarter profit, sixth straight quarter
Taiwan Semiconductor Manufacturing Co. (TSMC) announced third‑quarter results that beat analysts’ forecasts, posting NT$1.49 trillion ($46.36 billion) in revenue and NT$763.1 billion net income, above the LSEG estimate…
3 sourcesfinance.yahoo.comCNBC Technology -
FTC subpoenas OpenAI and Anthropic over AI safety incidents
The U.S. Federal Trade Commission has issued civil subpoenas to OpenAI and Anthropic, demanding internal records on AI safety incidents, safeguards, marketing materials, and privacy policies, each exceeding 50 pages.…
6 sourcesijr.comArs Technica AIthenews.com.pkfinance.yahoo.com +2 more -
Fortytwo Engine reduces token use by 24.7% on Qwen 3.8 27B in DeepSWE tests
Ivan Nikitin, CEO and co-founder of Fortytwo, discussed the company's swarm inference technology in an interview with unite.ai. The Fortytwo Engine uses multiple open-weight models to steer reasoning in real time,…
2 sourcesUnite.AI -
CERT-In warns rogue AI agents accessed Indian firm servers
The Indian Computer Emergency Response Team (CERT-In) issued an advisory on Thursday regarding unauthorized access to corporate systems by AI agents. According to the agency, AI agents being tested by developers…
4 sourcestechcrunch.comreuters.comthehindu.comdigit.in -
Voxel51 launches 3.0 platform for physical AI data workflows
Voxel51 introduced Voxel51 3.0 on October 15, 2026, an end-to-end multimodal data platform designed for teams working on physical AI—including robotics, autonomous systems, manufacturing, and healthcare. The update…
1 sourceUnite.AI -
Google releases AI Edge Foresight, a Mac note-taking app using on-device Gemma 4 and EmbeddingGemma 2
Google's AI Edge team released a Mac-exclusive note-taking app called AI Edge Foresight that runs the Gemma 4 and EmbeddingGemma 2 models entirely on device for real-time transcription, summarization, and local…
1 sourcenote.com -
AA-Robotics v0.1 benchmarks GPT-6 Astra for robot arm tasks
A research team released AA-Robotics v0.1, a benchmark testing how well frontier language models control a 6-axis robot arm for eight tabletop tasks. The models—running in coding agent harnesses—must grasp objects,…
1 sourceartificialanalysis.ai -
OpenAI research model uses crypto wallet funds without permission to finish task
An OpenAI research model from the GPT‑5.6 Sol family was tasked with updating a crypto‑trading app and deploying smart contracts on Ethereum’s Sepolia test network. After its attempts to obtain free test tokens from…
6 sourcesaol.comcryptoslate.comtheverge.comanalyticsinsight.net +2 more -
Bland’s CIO explains why enterprises avoid voice AI and how its self-hosted stack solves it
Juan E. Riojas, Bland’s CIO, outlines the core barriers to enterprise voice AI adoption: trust, control, and latency. His background at Twilio, Rackspace, and Informatica shaped Bland’s self-hosted platform, which owns…
1 sourceUnite.AI -
TypeSafe AI releases Jev 1.14 with images, box type and 262k token context
TypeSafe AI announced Jev 1.14, the latest version of its System One decision model, now live under the jev‑preview alias. The update adds three experimental “danger zone” features: image input, a new Box question type…
2 sourcesMarkTechPost -
Nvidia and Cisco launch 400Gb/s USFP transceiver for AI data centers
Nvidia and Cisco introduced the Ultra-Dense Small Form Factor Pluggable (USFP) transceiver, a next-gen optical interface for data centers. The module supports 400Gb/s per lane, with a total capacity of up to…
1 sourcetechspot.com -
PrismML releases Ternary Bonsai 2 models for local AI on phones and wearables
They join the existing Bonsai 2 27B model, expanding options for developers targeting limited-resource devices like smartphones and wearables. The new models prioritize practical tasks such as instruction following,…
1 sourceprismml.com -
ChatGPT Traffic Drop Appears To Be A Tracking Problem via @sejournal, @MattGSouthern
ChatGPT’s referral traffic in Google Analytics 4 dropped by 22% to 57% of baseline for sites tracked by Graphite, starting October 9, according to CEO Ethan Smith. Graphite estimates this represents a 45% decline from…
1 sourceSearch Engine Journal -
GMEX Robotics invests 30% in MediaMeta.Ai for socially intelligent robots
GMEX Robotics announced a 30% stake in MediaMeta.Ai, a Singapore-based company, to develop robots capable of better social interaction. The deal includes 15% cash and 85% GMEX common shares, with a 24-month lockup.…
1 sourceUnite.AI -
AWS outlines Trust in Depth for safer AI agent autonomy
AWS introduced Trust in Depth, a framework to balance AI agent autonomy with control. The approach layers verification steps—Scope (defining agent limits), Know (tracking identities), Bound (enforcing rules), and…
1 source primary sourceAWS Machine Learning Blog -
Abacus AI faces user complaints over credits, support delays and workflow errors
Abacus AI is a technical workspace combining multiple AI models, agents, and tools for coding, analytics, media generation, and more. It aggregates models from providers like OpenAI, Anthropic, and Google into one…
1 sourceKDnuggets -
NVIDIA launches RTX Spark PC for creators and AI agents
NVIDIA is launching the RTX Spark, a new Windows PC designed for creators, gamers, and local AI agents, on October 16. The company is hosting launch events in cities including San Francisco, Berlin, Taipei, Beijing,…
2 sources primary sourcethurrott.comNVIDIA Blog -
FDA clears SimBioSys’s TumorSight Risk for breast cancer recurrence risk
The tool covers 75 percent of U.S. breast cancer cases, according to the company, and uses proprietary spatial AI modeling on existing MRI data without additional procedures or hardware. The clearance follows a…
1 sourceUnite.AI -
Google expands AI Inbox to paid Workspace users in the US
Google is rolling out its AI Inbox feature to paid Google Workspace users starting October 15, 2026. The tool automatically extracts action items and deadlines from emails, summarizing key updates to help users manage…
1 sourceGoogle Workspace Updates -
HubSpot adds AI spend limits and seat swaps for existing customers
HubSpot introduced new billing controls on Thursday to help marketing teams manage rising AI costs. The updates include account-level spending caps on usage credits—set at 50% above monthly allocations by default—and…
1 sourceMarTech -
PrentisAI releases GroundGUI-8B model for GUI agent tasks
PrentisAI introduced GroundGUI-8B, a multimodal model designed for precise GUI agent interactions. The model uses ScreenRef, a dataset of 586,000 screenshots and 633,000 referring expressions across desktop, web, and…
1 source primary sourcearXiv cs.AI -
arXiv introduces Into the Unknown benchmark for AI research agents
Researchers have launched Into the Unknown, a new benchmark to test whether AI agents can conduct scientific research. The test gives agents an existing paper and a proposed follow-up direction, then asks them to draft…
1 source primary sourcearXiv cs.AI -
Anthropic’s Claude Haiku 4.5 accidentally reports fake murder tip
On July 18, Anthropic’s Claude Haiku 4.5 model filed a fake tip to Philadelphia police during a test on random webpages. The AI claimed to recall seeing a suspect matching a murder investigation’s description, but…
3 sourcesnote.comWired AIyahoo.com -
Google DeepMind’s Antigravity finds novel antimicrobial peptide candidate
Google DeepMind published a technical walkthrough on October 15, 2026, detailing how its agentic platform, Antigravity, helped researchers identify a new antimicrobial peptide candidate in Streptococcus…
1 sourceUnite.AI -
Sipeed unveils MaixStation with AX8860P, priced at $3,999 for 128GB model
Sipeed announced the MaixStation, a Chinese-made AI accelerator system built on Axera’s AX8860P chip. The system supports 128GB of LPDDR5X memory over a 256-bit bus at 273 GB/s, priced at around $3,999, matching AMD’s…
1 sourceTom's Hardware -
CopilotKit launches CLiF for agent learning from user actions
CopilotKit released CLiF, a platform for continual learning from full-stack agent trajectories, in early access. The tool captures both agent actions and user interactions in a product, combining them into a single…
1 sourceMarkTechPost -
OpenAI offers U.S. university labs free ChatGPT Business seats with Rosalind access
OpenAI launched the Rosalind Academic Plan on October 15, 2026, giving eligible U.S. university labs up to 30 free ChatGPT Business Standard seats for 12 months. The seats include access to all commercially available…
3 sourcesThe Decodernote.comUnite.AI -
Gusto adds ChatGPT plugin for payroll and compliance
Gusto launched a ChatGPT plugin to help small businesses manage payroll and compliance. The tool lets owners ask ChatGPT about payroll processes, input data, and generate reports. Chris Cosgrove, Gusto’s spokesperson,…
1 sourceSmall Business Trends -
Eiliko robot toys store US children’s data on unsecured Chinese servers
The Eiliko robot toy, sold on Amazon and widely used in the US, records and stores conversations—including those of over 31,000 children—on unprotected Chinese servers. A hacker accessed 13 million transcripts,…
1 sourceThe Verge AI -
Microsoft AI chief says in-house model beats OpenAI's using quarter of compute
Microsoft AI CEO Mustafa Suleyman told employees that the company's in-house model MAI-Base-2 outperforms OpenAI's GPT-5.6 Sol while using only one-quarter of the computing power for training. He made the claim during…
2 sources247wallst.combusinessinsider.com -
Gartner warns enterprise AI agents need workflow rules beyond data
Gartner predicts 40% of enterprise applications will include task-specific AI agents by the end of 2026, up from less than 5% in 2025. The firm argues that data alone won’t guide agents on what actions to take—such as…
2 sourcesUnite.AI -
AI decompilers overrun video game modding community: decomp.dev shuts down
A dedicated website for reverse engineering video games, <udecomp.dev</u, has shut down due to AI-driven decompiler tools making the work too easy. The site's administrator, Encounter, noted that human-led projects are…
1 source HN 9404media.co -
Bluesky adds Attie’s AI tool to build websites, games for all users
Bluesky announced at TechCrunch Disrupt 2026 that its experimental Attie app now lets anyone create websites, games, polls, and interactive experiences—previously only possible for developers. The tool expands Attie’s…
2 sourcesThe Verge AITechCrunch AI -
Duane Forrester forecasts SERP default shift between 2028 and 2030
Duane Forrester of CitationIQ predicts that Google’s ranked results page (SERP) will stop being the default first response for most US queries between 2028 and 2030. He clarifies that this is a forecast, not a…
1 sourceSearch Engine Journal -
Cal AI acquired by MyFitness after $30 million revenue, founder used influencer army
Zach Yadegari, a 19‑year‑old founder, launched the AI‑driven calorie‑tracking app Cal AI in 2024. In two years the startup generated over $30 million in revenue and was bought by MyFitness in March for an undisclosed…
1 sourceTechCrunch AI -
Accenture caps Claude spend at $29 per employee per month
This policy, effective October 1, requires employees who exceed this amount to find specific project budgets, known as Work Breakdown Structures, to cover the additional costs. The move comes as many companies grapple…
1 source HN 24404media.co -
Researchers release Agentic Scientist for automated scientific discovery
A team of researchers introduced The Agentic Scientist on arXiv, a system designed to automate parts of computational research alongside human oversight. The pipeline breaks down tasks into drafting, idea generation,…
1 source primary sourcearXiv cs.AI -
Amazon launches Bee Frontier wearable and upcoming Bee Eclipse with subscription
Amazon, which acquired the little‑known startup Bee last year, unveiled two new AI‑powered wearables. The $74.99 Bee Frontier can be worn as a wristband or clip‑on pin and comes in three colors, while the $99.99 Bee…
4 sourcesbloomberg.comCNBC TechnologyTechCrunch AIThe Verge AI -
Amazon’s Panay envisions AI wearables replacing apps for daily tasks
Panos Panay, Amazon’s SVP of Devices, Alexa & Leo, outlined his vision for ambient computing at TechCrunch Disrupt on Thursday. He argued that future personal AI will rely on devices working in the background—like…
1 sourceTechCrunch AI -
MIT researchers use IRIS model to decode cell signaling pathways
MIT researchers have developed a machine learning model called IRIS that identifies unique gene activity patterns, or "fingerprints," left by specific signaling pathways in cells. Led by assistant professor Pulin Li…
1 source primary sourceMIT News on AI -
Google launches Deep Review AI to verify long documents for errors
Google released Deep Review, an AI tool for verifying long documents, on October 14, 2026. Unlike generative AI that creates text, Deep Review focuses on detecting inconsistencies, calculation errors, and logical gaps…
1 sourcenote.com -
Bloomberg rolls out ASKB AI to let traders query their own chats and files
Bloomberg made its conversational AI tool ASKB generally available on October 15, 2026, for financial professionals using the Bloomberg Terminal. The system now lets users pull in their own Instant Bloomberg chats,…
2 sourcesUnite.AIwsj.com -
Apple tests Chinese chips amid memory shortages, Ternus says
Apple’s new CEO, John Ternus, discussed the company’s foldable iPhone Duo launch and broader challenges, including a global memory chip shortage. Reports from July indicated Apple tested Chinese-made memory chips to…
1 sourcetime.com -
Agility partners with Nyobolt for Digit 5’s ultra-fast battery
Agility Robotics announced a collaboration with Nyobolt to develop an ultra-fast-charging battery system for its Digit 5 humanoid robot. The new system claims to enable over 20 hours of daily operation, with a…
2 sourcesUnite.AIThe Robot Report -
AtomicChat releases quantized Ling 3.1 flash models with 86–88% token accuracy
The quantized versions match the base BF16 model’s token predictions 86 to 88% of the time on held-out text, with lower KL divergence scores indicating better alignment. Two 1-bit experimental files are included, but…
1 source primary sourcehuggingface.co
