Why it matters: A major potential cybersecurity merger driven by the need to counter AI-powered attacks.
CrowdStrike is in talks with Zscaler as a wave of multibillion-dollar takeovers reshapes cybersecurity, with companies racing to defend against increasingly sophisticated AI-powered attacks.
Why it matters: A $7.5B AI software acquisition signals Amazon's push to counter Nvidia in compute.
Amazon is reportedly closing in on a $7.5 billion acquisition of AI startup Decart, whose chip software would help AWS compete with Nvidia in computing power.
Prabhkiran Singh, cofounder and former CEO of Bewakoof, has launched Save8, an AI shopping assistant that compares prices to help consumers find deals.
Why it matters: A first public case of autonomous AI agents finding and exploiting a real zero-day.
Amazon says a China-linked threat actor used autonomous Cairn AI agents to discover and exploit a critical zero-day in MONAI Label medical imaging software.
Why it matters: Cheaper idle pricing makes always-on voice agents more economical to run.
OpenAI released GPT-Live-1.1 for voice agents via its API, cutting idle cost to $0.02 per minute. The update targets developers building real-time speech applications.
Gladios Pte Ltd launched GladiosWAF, a web application firewall it describes as zero-rule and AI-driven, aimed at democratizing web security for sites without hand-written rules.
Why it matters: Congressional scrutiny of Nvidia-backed cloud debt could reshape AI infrastructure financing.
CoreWeave faces a $10B debt wall as Congress examines Nvidia's $6.3B backstop and role as key sole supplier to the debt-laden neocloud, raising questions about the AI boom's financing.
AlterContacts is a voice-first assistant for client-facing work: talk through a meeting and it updates the client's profile, tracks promises and renewals, and drafts follow-up messages for approval.
AI Context Flow keeps one shared memory over MCP so multiple AI tools and teammates work from the same version, with buckets by project or client, access controls and edit history.
Glassmail is a native Mac email client for Gmail, iCloud, Outlook and IMAP whose AI runs locally. Ask what needs attention today or dictate a reply in your own voice; no mail server, no subscription.
OpenVorld is an open-source harness that gives AI agents a visible pixel-art village, where each agent has a name, role and brain (Claude Code, Codex or any OpenAI-compatible API) and asks before running commands. Free and local.
Shipyard is a local CLI and MCP server for Claude Code, Cursor and Codex that connects accounts once, then automatically sets up domain, DNS, hosting, database, email and payments for each new app in about an hour.
Why it matters: A looped MoE design that runs locally on a laptop pushes open-weight efficiency forward.
Inkling 2 Mini is a multimodal model from Thinking Machines Lab that takes text and images and generates text. The 26B looped mixture-of-experts model can run on a 32GB Mac; transformers v5.24.0 adds support for it.
QuEra Computing and Nvidia introduced PPVM, an open-source digital-twin simulation framework for quantum processors. Separately, Quantum Motion integrated an Nvidia CUDA-Q backend into its QUICHE platform for end-to-end quantum chemistry workflows.
ClawDOS is a project that gets an AI coding agent running on vintage DOS hardware, demonstrated on a 1981 IBM PC XT. It shows agent tooling adapted to extremely constrained machines.
Why it matters: Brings desktop computer-use agents from a demo to a mainstream consumer assistant.
Google's Gemini app on Mac can now click and type in apps on the user's behalf, an agentic computer-control feature. A homepage redesign and prompt-to-GIF generation are also rolling out.
Observe.AI launched Agent Overwatch, which extends its CX platform to flag jailbreaks, failed workflows and compliance risks mid-conversation across AI agents it did not build, plus human agents, using a Companion Agent.
Kelt Forge is a $25 AI builder pitched by founder Ken Eynon for plain-language websites, web apps and mobile projects, returning reviewable Git commits for owners tired of packaged software.
Why it matters: The listing is a landmark test of public-market appetite for a leading AI lab.
Anthropic made its market debut as a public company, with its IPO failing to reassure investors as tech stocks fell worldwide. Commentary is now focused on who is to blame for the turbulent listing.
Why it matters: US export policy on AI chips shapes global access to Nvidia hardware.
The Treasury secretary said no further restrictions are needed on Nvidia chip exports to China, while reports say Commerce Secretary Lutnick was warned a year ago about China illegally obtaining AI chips.
Why it matters: AWS is making agentic purchasing a first-class path to cloud software while easing multi-region inference.
AWS Marketplace now lets AI agents buy software directly, even as Amazon blocks shopping bots on its retail site. AWS also launched a SageMaker HyperPod global inference gateway offering one endpoint across regions and clusters.
DreamCoder is a workspace that puts coding agents, terminals and code reviews in resizable tiled browser panes; free for personal use and businesses of up to five people.
RYPTASK released a managed AI agent offering aimed at managed service providers and their clients, letting MSPs deploy and run AI agents on customers' behalf.
Why it matters: Extends AI agent security monitoring beyond Google Cloud to multi-cloud deployments.
Agent Anomaly Detection now monitors ADK agents on AWS, Azure or on-prem by ingesting OpenTelemetry traces into Google Cloud and evaluating full conversations for behavioral drift, without needing access to other clouds' systems.
Amazon Textract Insights is a generative AI feature of Textract that lets users describe fields in natural language, then returns structured results with confidence scores and page-level grounding.
Why it matters: Signal that AI datacenter siting is becoming national policy and a midterm election issue.
President Trump said he would block or impose a "tremendous tax" on U.S. companies that build AI data centers overseas, framing the push as keeping AI infrastructure in the country.
Why it matters: Time-horizon doubling is a key benchmark for how fast agentic capabilities advance.
Nonprofit METR reported that recent models including OpenAI's GPT-6 Astra can complete tasks a human would need about a full work week for, with capability time horizons doubling every 3.6 months, even as research judgment lags.
Why it matters: Extends collaborative, self-updating document hubs beyond top tiers to ChatGPT's mainstream paid users.
ChatGPT Spaces, collaborative hubs where users upload documents and files to share, is now rolling out to Plus subscribers worldwide, with self-updating documents.
Why it matters: A broad unwind in AI-linked equities tests how much of the boom rests on stock momentum.
European stocks logged their worst week in more than a year as the AI trade unraveled, prompting US investors to rotate into value stocks and Dan Ives to call the moment scary but not the end.
Houston-based Bot Auto launched Capacity Grid, a freight network matching shipper demand with autonomous trucking capacity, backed by agreements covering 200 million miles of driverless freight.
The Knoxville Police Department launched a Drone as First Responder program that dispatches semi-autonomous drones to 911 calls, reaching nearly all of Knoxville in 90 seconds or less.
Turkish defense firm Aselsan developed Ejderha ("Dragon"), a defensive EMP weapon designed to counter drone swarms; a test fried 51 jam-proof drones simultaneously.
Why it matters: First formal path toward an international treaty governing autonomous weapons.
At UN talks on lethal autonomous weapons, countries agreed to open negotiations aiming for a regulatory pact by 2028, but stopped short of firm commitments to rein in the weapons.
Nvidia released GeForce Hotfix Driver 621.35 to address a bug causing excessive and unnecessary writes to SSDs. The patch targets GPU driver behavior that wore down storage.
Waddle-1 is a cloud reasoning agent for robots from a two-person YC startup. Waddle Labs says it scored 81 on Robocurve, beating GPT-6.2 Astra and Dream-1 on unseen bimanual tasks, and opened a preview API.
Why it matters: Clarifies financing behind one of the largest AI datacenter buildouts.
Oracle said its infrastructure contract with OpenAI is fully funded after completing a final capital raise, addressing concerns about its ability to finance the buildout.
Why it matters: Shows Huawei pushing on-device AI services into a new phone sub-brand.
Huawei launched the OSTA smartphone sub-brand and its first handset, the O1 Pro, priced RMB 5,499–8,799. It runs on the Kirin 9030S chip and an AI-focused HarmonyOS 7 that accepts natural-language service requests.
Phison showed a computational-storage SSD with an NPU and 64GB of memory that runs AI frameworks on the drive, plus Mixed Media pSLC and QLC drives, at SC26.
AAEON launched its first box PC built on Nvidia's Jetson Thor platform, aimed at edge AI in humanoid robots. The device brings Thor-class compute to robot onboard processing.
ARCS.ai launched the ARPI Center in Yerevan on November 20, 2026, a shared robotics and physical AI center aiming to graduate 50 specialized engineers annually by combining faculty, students and equipment.
Nvidia's DAQIRI GPU data-acquisition software update roughly doubles small-packet throughput, though not at 8 KB packet sizes, according to Tech Times.
Why it matters: A colossal debt-financed GPU buildout tied to space data centers tests how far AI infrastructure leverage can stretch.
Apollo and Goldman Sachs are leading a $44 billion GPU-backed debt financing for SpaceX, which closed a $32 billion debt deal at steep yields as lenders reportedly balk at orbital AI chip plans.
Tesla software 2026.44.1 adds a Grok button to the navigation search bar and, per a hacker's teardown, rebrands the in-car AI computer as a 'Super Intelligence' computer with Grok sync.
Tencent Cloud Enterprise launched a full-stack intelligent private cloud offering aimed at sovereign agentic AI deployments, letting organizations run AI agents on dedicated infrastructure.
Why it matters: Nvidia's $390B supply lock-in signals the scale of AI memory and compute demand through 2027.
Nvidia reported a $115.8B quarter and $118B in buildout support while raising supply commitments to $390B, largely for memory, raising questions about where the chipmaker's cash flows.
Google's Gemini app for Android is rolling out a redesigned camera interface, making it easier to point the lens and ask questions about what the phone sees.
A Hugging Face dataset of 250 reasoning traces, pairing prompts from the MiMo-2.6-Pro 250x set with full reasoning and final answers from qwen/qwen4-max, roughly 6,691 output tokens per row.
Open-artifact is an open-source TypeScript project for creating and editing artifacts such as slides, UI designs, whiteboards, sheets, dashboards and docs. Everything is saved locally as plain files, works with any agent or model, and can be commented on and exported.
tiny-voca is a synthetic American English speech corpus with a female and male voice reading the same 17,371 sentences, about 19.1 hours of 24 kHz audio, meant to train the tiny-voca TTS model. Released under CC BY 4.0.
Apple says from 2027 software releases Siri AI and Apple Intelligence are free on supported devices, but server-backed features have usage limits that rise with eligible iCloud+ tiers.
PiBot is an open-source, local-first AI agent team for your desktop, with multi-agent group chats, computer use, browser control, MCP, routines and approvals, running on macOS, Windows, Linux and Android. It drew 107 stars in its first week.
AI4XE/ai4ai is a TypeScript toolkit that lets AI agents manage themselves, covering runtime status, context renewal and self-duplication across agent harnesses. It drew 101 stars in its first week.
The New Stack reports Nvidia used 20,000 training examples to fix mistakes in its internal AI and saw its coding score jump, detailing the data and training approach behind the improvement.
FastFrames-S is an open text-to-video model from Apple aimed at interactive-latency generation. It is built on Wan2.1-T2V-1.3B and published with code and a paper.
A Google guide for AI coding agents on adding Google Analytics to iOS and Android apps, covering setup steps so agents can instrument screens, events and conversions correctly.
Anthropic's Claude app for iPhone and iPad reached version 1.261119.19, a minor release described only as squashing some bugs and improving the overall experience.
AlterContacts is a voice-first assistant for client-facing work: talk through a meeting and it updates the client's profile, tracks promises and renewals, and drafts follow-up messages for approval.
AI Context Flow keeps one shared memory over MCP so multiple AI tools and teammates work from the same version, with buckets by project or client, access controls and edit history.
Glassmail is a native Mac email client for Gmail, iCloud, Outlook and IMAP whose AI runs locally. Ask what needs attention today or dictate a reply in your own voice; no mail server, no subscription.
OpenVorld is an open-source harness that gives AI agents a visible pixel-art village, where each agent has a name, role and brain (Claude Code, Codex or any OpenAI-compatible API) and asks before running commands. Free and local.
Shipyard is a local CLI and MCP server for Claude Code, Cursor and Codex that connects accounts once, then automatically sets up domain, DNS, hosting, database, email and payments for each new app in about an hour.
Why it matters: A looped MoE design that runs locally on a laptop pushes open-weight efficiency forward.
Inkling 2 Mini is a multimodal model from Thinking Machines Lab that takes text and images and generates text. The 26B looped mixture-of-experts model can run on a 32GB Mac; transformers v5.24.0 adds support for it.
BHASHINI, India's multilingual AI platform, demonstrated AI applications for hospitals that cut doctors' documentation time and bridge language barriers, at the Global Digital Health Summit 2026.
ClawDOS is a project that gets an AI coding agent running on vintage DOS hardware, demonstrated on a 1981 IBM PC XT. It shows agent tooling adapted to extremely constrained machines.
Cline is an autonomous coding agent that can create and edit files, run commands and use a browser. The CLI version 3.0.75 installs via npm; the VS Code extension hits 4.1.27 and the desktop app 0.0.50.
New open-source JavaScript repo collecting agent skills for the NestJS Node.js framework, letting AI coding agents work with Nest projects; gained 100 stars in its first week.
Why it matters: Brings desktop computer-use agents from a demo to a mainstream consumer assistant.
Google's Gemini app on Mac can now click and type in apps on the user's behalf, an agentic computer-control feature. A homepage redesign and prompt-to-GIF generation are also rolling out.
Kelt Forge is a $25 AI builder pitched by founder Ken Eynon for plain-language websites, web apps and mobile projects, returning reviewable Git commits for owners tired of packaged software.
Qwen Code is an AI-powered coding assistant for the terminal, installed with npm. Version 0.28.3 is a routine update to the open-source CLI, released via npm.
A Hugging Face dataset of 250 reasoning traces, pairing prompts from the MiMo-2.6-Pro 250x set with full reasoning and final answers from qwen/qwen4-max, roughly 6,691 output tokens per row.
tiny-voca is a synthetic American English speech corpus with a female and male voice reading the same 17,371 sentences, about 19.1 hours of 24 kHz audio, meant to train the tiny-voca TTS model. Released under CC BY 4.0.
FastFrames-S is an open text-to-video model from Apple aimed at interactive-latency generation. It is built on Wan2.1-T2V-1.3B and published with code and a paper.
Inkling 2 Mini is a multimodal model from Thinking Machines Lab that takes text and images and generates text. The 26B looped mixture-of-experts model can run on a 32GB Mac; transformers v5.24.0 adds support for it.
EVIDENT is a Hugging Face dataset of 220 GUI/web agent task records (150 Korean, 70 English) pairing user requests with intended outcomes and evidence checks, for evaluating agent trajectories.
Aion Instruct is Microsoft's 2.3B-parameter small language model built for on-device AI and OS use cases in Windows, with MIT-licensed weights on Hugging Face.
Model · text-generation · mit · ♥ 102 · 1.6k downloads
Inkling 2 Mini is a compact open-weight model from Thinking Machines designed to run locally on personal devices; weights are on Hugging Face, including a GGUF build for local inference.
Open-artifact is an open-source TypeScript project for creating and editing artifacts such as slides, UI designs, whiteboards, sheets, dashboards and docs. Everything is saved locally as plain files, works with any agent or model, and can be commented on and exported.
PiBot is an open-source, local-first AI agent team for your desktop, with multi-agent group chats, computer use, browser control, MCP, routines and approvals, running on macOS, Windows, Linux and Android. It drew 107 stars in its first week.
AI4XE/ai4ai is a TypeScript toolkit that lets AI agents manage themselves, covering runtime status, context renewal and self-duplication across agent harnesses. It drew 101 stars in its first week.
Inkling 2 Mini is a multimodal model from Thinking Machines Lab that takes text and images and generates text. The 26B looped mixture-of-experts model can run on a 32GB Mac; transformers v5.24.0 adds support for it.
New open-source JavaScript repo collecting agent skills for the NestJS Node.js framework, letting AI coding agents work with Nest projects; gained 100 stars in its first week.
TraxAgent is a read-only MCP server and Claude Code plugin that lets Claude, ChatGPT, Claude Code, Cursor or VS Code show who spent your AI budget and what to fix; the repo drew 109 stars in its first week.
A Python GitHub project integrating AnythingLLM with a workspace called Verona to chat over your own documents, so answers are grounded in their content. The new repo drew about 124 stars in its first week.
A community-built JavaScript workspace project described as a friendly, feature-rich environment for AI workflows. It is a new GitHub repo with about 107 stars in its first week.
A TypeScript project that provides a streamlined workspace for organizing and running AI workflows in one place, built on Open WebUI. The new GitHub repo drew about 102 stars in its first week.
Antigravity, an agent orchestration platform, released version 2.34.0 following 2.33.1, for macOS, Windows and Linux. A separate plugin, agy-fast, accelerates MCP workflows.
TensorRT-LLM is Nvidia's Python API and runtime for running LLMs efficiently on Nvidia GPUs. Version 1.3.0 adds new optimizations and runtime components.
An open-source tool that lets Mac users run several Claude Desktop accounts at once with separate sign-ins and shared Claude Code setup, installable via Homebrew.
Google's Gemini app for Android is rolling out a redesigned camera interface, making it easier to point the lens and ask questions about what the phone sees.
Apple says from 2027 software releases Siri AI and Apple Intelligence are free on supported devices, but server-backed features have usage limits that rise with eligible iCloud+ tiers.
Anthropic's Claude app for iPhone and iPad reached version 1.261119.19, a minor release described only as squashing some bugs and improving the overall experience.
AlterContacts is a voice-first assistant for client-facing work: talk through a meeting and it updates the client's profile, tracks promises and renewals, and drafts follow-up messages for approval.
Glassmail is a native Mac email client for Gmail, iCloud, Outlook and IMAP whose AI runs locally. Ask what needs attention today or dictate a reply in your own voice; no mail server, no subscription.
BHASHINI, India's multilingual AI platform, demonstrated AI applications for hospitals that cut doctors' documentation time and bridge language barriers, at the Global Digital Health Summit 2026.
Google's Gemini app on Mac can now click and type in apps on the user's behalf, an agentic computer-control feature. A homepage redesign and prompt-to-GIF generation are also rolling out.
Meta's personal AI assistant app for iPhone and iPad adds scheduled tasks and reminders, email and calendar connection, and in-depth research reports in version 300.0.0.
Apple is moving into AI agents that can browse, control apps and run multi-step tasks, making operating-system access the new competitive front. Meanwhile a Siri hands-on finds the assistant private but not very useful.
Gemini Notebook (NotebookLM) for iPhone and iPad hit version 1.68.7, letting users view any interactive site directly in the app plus bug fixes. Separately, Gemini can now make GIFs from a prompt with a Google AI plan.
ChatGPT Spaces, collaborative hubs where users upload documents and files to share, is now rolling out to Plus subscribers worldwide, with self-updating documents.
Brizy released version 1.9.5 of its AI website builder, which now generates imagery instead of using Unsplash stock photos and can generate individual page sections from a prompt.
Android Authority highlights an Android player that uses AI to cut disruptions from podcasts, plus Pocket Relay and Buddy in its weekly AI apps roundup.
Google is offering Nigerian students 12 months of free Gemini worth ₦92,400, and added a Gemini feature that records where items are kept, integrated into Android's Find Hub; Gemini also appears in Search and Gemini grocery carts for Target and Albertsons.
The New Stack reports Nvidia used 20,000 training examples to fix mistakes in its internal AI and saw its coding score jump, detailing the data and training approach behind the improvement.
A Google guide for AI coding agents on adding Google Analytics to iOS and Android apps, covering setup steps so agents can instrument screens, events and conversions correctly.
AI Context Flow keeps one shared memory over MCP so multiple AI tools and teammates work from the same version, with buckets by project or client, access controls and edit history.
OpenVorld is an open-source harness that gives AI agents a visible pixel-art village, where each agent has a name, role and brain (Claude Code, Codex or any OpenAI-compatible API) and asks before running commands. Free and local.
Shipyard is a local CLI and MCP server for Claude Code, Cursor and Codex that connects accounts once, then automatically sets up domain, DNS, hosting, database, email and payments for each new app in about an hour.
ClawDOS is a project that gets an AI coding agent running on vintage DOS hardware, demonstrated on a 1981 IBM PC XT. It shows agent tooling adapted to extremely constrained machines.
Cline is an autonomous coding agent that can create and edit files, run commands and use a browser. The CLI version 3.0.75 installs via npm; the VS Code extension hits 4.1.27 and the desktop app 0.0.50.
Kelt Forge is a $25 AI builder pitched by founder Ken Eynon for plain-language websites, web apps and mobile projects, returning reviewable Git commits for owners tired of packaged software.
Qwen Code is an AI-powered coding assistant for the terminal, installed with npm. Version 0.28.3 is a routine update to the open-source CLI, released via npm.
DreamCoder is a workspace that puts coding agents, terminals and code reviews in resizable tiled browser panes; free for personal use and businesses of up to five people.
Amazon Textract Insights is a generative AI feature of Textract that lets users describe fields in natural language, then returns structured results with confidence scores and page-level grounding.
Claude Code, Anthropic's terminal AI assistant that reads codebases, edits files and runs commands, reaches version 2.1.330 for the CLI (npm) and the VS Code extension.
Zed is a multiplayer code editor built in Rust for collaborating on code with teammates and AI agents. Version 1.29.3 is a maintenance release for macOS, Windows and Linux.
Anthropic says Claude writes and reviews 90% of its team's code, with the catch being how the work is checked; an engineer warns Claude Code issues could worsen tech outages, and Musk and Jassy praise Anthropic models.
TechTarget reports IT leaders advising firms to constrain AI agents and re-run tests whenever the underlying models are updated, to keep agent behavior reliable in production.
Antigravity, an agent orchestration platform, released version 2.34.0 following 2.33.1, for macOS, Windows and Linux. A separate plugin, agy-fast, accelerates MCP workflows.
Kilo Code CLI is a terminal-based AI coding agent that generates code from natural language, automates tasks and runs terminal commands using 500+ AI models. Version 7.8.32 is available via npm.