Why it matters: Escalating scrutiny of Nvidia's China dealings could affect chip export policy and industry.
The FBI is reportedly investigating Nvidia CEO Jensen Huang as scrutiny grows over China chip access, with a short seller alleging more companies helped China obtain chips.
Why it matters: An AI-generated proof on a Millennium Prize problem would be a landmark for AI in mathematics.
A Nobel laureate-led committee validated an OpenAI-generated proof establishing the existence of quantum Yang-Mills theory. The full proof will be released publicly on November 10.
Red Hat released OpenShift 5 at KubeCon + CloudNativeCon North America on November 9, 2026, adding agentic troubleshooting and a single platform for AI workloads, VMs and containers.
Why it matters: Pushes open training stack to multi-million-token contexts, easing long-context model work on TPUs.
Google Cloud's open-source MaxText framework now supports long-context training up to 14 million tokens on Cloud TPUs via four context-parallel strategies (All-gather, A2A, P2P, USP), powered by the Tokamax Splash Attention kernel.
Kyndryl Autonomous Operations for ServiceNow, co-developed and sold in the ServiceNow Store, uses Bridge operational data and AI agents for incident analysis, risk assessment and more.
Why it matters: Extends OpenAI's persistent agents to multiple local machines, a step toward distributed agent workflows.
OpenAI's persistent agents can now route a task to a second Mac or PC, letting developers test Xcode builds and older SDKs while continuing to work.
Airbnb is building an AI agent that can arrange itineraries and manage holiday reservations, even as CEO Brian Chesky dismisses the AI booking threat as negligible.
Maybelline New York launched a real-time makeup virtual try-on experience inside ChatGPT, letting shoppers see shades on themselves while chatting with the assistant.
AWS Machine Learning · 6h
#
Why it matters: It brings a large open multimodal model to a mainstream cloud deployment channel.
Nemotron 3.5 Super VL is a 120B-parameter multimodal vision-language model for text, image and video inference. It is now deployable in Amazon SageMaker JumpStart in NVFP4 or BF16 variants.
Cisco introduced frontier AI red teaming and agent observability tools aimed at securing the agentic enterprise.
Why it matters: One of the largest AI IPOs could reshape funding and valuation benchmarks for the sector.
Anthropic reportedly plans to raise up to $69.3 billion in an IPO, though it sent and then retracted details of the filing. It is moving forward with IPO plans.
Coinbase Wallet plans to add AI agent capabilities in 2026, letting agents act inside the self-custody wallet. Details of what agents will be allowed to do were not given.
Why it matters: Copilot gaining write and delete powers moves it from assistant to acting agent across business apps.
Microsoft Copilot can now create, update and delete data across connected apps, per Windows Report, expanding its actions beyond reading and summarizing.
AWS Machine Learning · 7h
#
AWS rebrands Amazon Quick as AWS Quick, an AI workspace where teams use agents to research, analyze and act across their data and applications, and debuts its first brand campaign.
Trimble launched new AI capabilities in its Trimble Connect platform, including agentic scheduling, aimed at de-risking construction projects.
ClickHouse acquired database startup Xata and launched Managed Postgres, generally available on AWS. Observability and Langfuse AI Gateway are in public beta, Security in private preview from November 9.
Instagram's Effect Builder generates custom filters for Stories posts from a text prompt; available now for US users over 18.
Motionvid is an AI video editor that generates prompt-based motion graphics and animated explainers, letting users turn text prompts into motion-design videos.
invideo is an AI video platform offering conversational editing and multimedia generation, letting users produce and revise videos by chatting with the tool.
An agent security platform opening to any GitHub team that finds and fixes vulnerabilities inside pull requests, priced at $5 per pull request test.
Databricks integrated Sendoso gifting into Genie One, letting users turn buying signals into pipeline by sending gifts directly from the AI assistant.
Why it matters: Anthropic pushes Claude into regulated financial advisory work, expanding its enterprise footprint.
Anthropic releases Claude for Advisors, a tool meant to augment rather than replace advisers' existing technology stack. The launch comes alongside coverage of Anthropic's spending commitments.
Northern Trust and Microsoft built a client onboarding agent that cut data preparation time 83%, from about two and a half weeks to three days.
Why it matters: New frontier models from a leading lab could shift the competitive landscape.
Anthropic releases Claude Fable 5.2 and Claude Mythos 5.2, new frontier models ahead of a planned IPO. Details on capabilities and availability are limited.
Feedvisor expanded its agentic commerce platform to target growing Amazon business demand, adding AI agent capabilities for sellers on the marketplace.
Medium released TK, a new writing app that integrates AI assistance directly into the writing process, despite ongoing controversy over AI's role in publishing.
A new texture inpainting and outpainting LoRA for Qwen Image 2.1, plus custom ComfyUI nodes, lets users fill in or extend image regions locally.
AITX unveiled RAIA, a portable AI-powered security tower, with deliveries expected in Q1 2027.
Orion-26B-A4B is an open-weight text model tuned for creative writing and roleplay; v1.1 leans on Gemma's reasoning to push creative writing further. Released on Hugging Face.
Cledara released an AI agent that handles software admin work like downloading vendor invoices, onboarding staff into tools and deactivating accounts when people leave.
Nemotron-Personas-Finance-Japan is a synthetic dataset of nine million personas modeling Japan's demographics and financial context, released on Hugging Face for training and evaluating AI models.
Fizgig is an open-source tool that shows heatmaps inside a LoRA adapter during training, revealing what it has learned, what it hasn't and where it leaks. Version 9.3.0 is on GitHub.
Databricks introduces UCDataset, a tool that lets teams train models on data stored in Unity Catalog tables while keeping GPUs fed, aiming to cut the cost of expensive training runs.
Toptal launches Toptal AI, offering frontier-grade AI built on open-weight models with full data ownership for clients.
Neo launched Neoverse, a free AI security engine that analyzes more than 4 million AI apps to block malicious software automatically.
Snapp, an AI sports analytics app from a Caviar co-founder, raised $4.6M and is now available on Android.
Varonis released new AI security features for its data security platform; specifics were not detailed in the available source text.
Why it matters: Shows enterprise security vendors racing to bolt AI protection and cloud sovereignty onto their platforms.
CrowdStrike launched new AI security tools and a Saudi sovereign cloud, and expanded partnerships with AWS, Anthropic and EY as Bernstein raised its price target to $177 on AI security demand.
Why it matters: A major retailer deploying an agent that shops and checks out on behalf of users pushes agentic commerce mainstream.
Amazon debuted a text-to-shop AI agent feature that lets customers describe what they want and have the agent find and purchase products.
DataAgent, from Cloudify veterans, released a free local K8sGPT lab for building autonomous SRE agents alongside its commercial remediation platform.
Creality's mBox is an enclosed 3D printer for ages 6–16 with a 153mm cube build volume, camera-based print monitoring and tools that turn photos or drawings into printable objects.
Hygon's 14XX is the first industrial-control chip in its 1000 series: C86 cores, integrated GPU, 10W TDP, -40°C to 105°C operating range and a 10-year supply promise.
Foxconn Interconnect Technology is expanding its lineup of liquid-cooling quick-disconnect connectors for Nvidia's Vera Rubin NVL72 platform, offering connections at server tray and rack levels with blind-mate and floating mechanisms.
Why it matters: Faster open physics simulation could speed up training of robot and physical-AI policies.
Skild AI released Steel, an open-source physics solver for robot simulation. It hit 3.90M environment steps per second on 20 Isaac Lab tasks at 245,760 environments, claiming up to 13x faster than PhysX with less GPU memory.
Nirva has unveiled an AI pendant that listens to conversations and takes notes for the wearer, according to Foro3D.
A new consumer robot autonomously carries household trash to the curb, per geekspin, automating a recurring chore.
OpenAI and Palmer Luckey's ModRetro released the $200 Codex Edition Chromatic handheld, which turns prompts into Game Boy games and includes a programmable cartridge. It returns November 19.
Why it matters: Large models purpose-built for dexterous manipulation push physical AI toward general-purpose robot hands.
Nature Will released a 50-billion-parameter AI model aimed at controlling dexterous robotic hands.
Why it matters: Makes Google's single-controller TPU orchestration generally available to cloud customers.
Google Cloud's Cloud Pathways, a single-controller runtime for large-scale TPU training and inference, is now generally available on GKE after 130 million TPU chip-hours of use.
SoftBank-owned Berkshire Grey upgraded its robotic automation portfolio via BG Pilot software, adding AI perception for trailer unloading and other warehouse tasks.
Cyngn's DriveMod EasyLaunch is a commercial subscription program that lets qualifying manufacturers and warehouses automate a repetitive material movement route with its autonomous vehicles.
Nvidia's Holoscan 5 SDK lets developers make a robot arm feel weightless, with motors supplying just enough torque to support the arm's weight and keep it upright.
Why it matters: Commercial quadruped inspection robots enter explosive-risk industrial settings.
Swiss firm ANYbotics has started delivering its ANYmal X quadruped to industrial customers, a legged inspection robot built to operate in explosive-risk environments like oil, gas and chemical plants.
London-based Rogue Cortex has emerged from stealth with about $4.5 million in venture funding and software and hardware tools designed to simplify building robotic systems.
An open-source $20 closed-loop stepper driver with two encoders compensates for gearbox backlash in a 3D-printed robotic arm, reporting joint position rather than just motor position.
ABB Robotics added AI Robot Trainer to its OmniCore EyeMotion system, letting customers train, deploy and improve AI-powered robot vision applications without building models from scratch.
Liquid Energy's NanoDC combines AI computing, folding solar panels, batteries and Starlink in a factory-built system. The YC-backed startup aims to shorten waits for inference capacity by supplying its own power.
Why it matters: TPU Sync and new TPU variants sharpen Google's pitch against Nvidia for AI inference workloads.
Google Cloud launched TPU Sync, a feature for fast KV cache and weight transfer during inference and reinforcement learning on TPUs, and added two TPU variants as it widens its fight with Nvidia.
Unitree's first consumer-grade robotic dog is priced at 43,900 yuan and drew about 1,000 orders on debut, roughly in the 8-million-won range.
Why it matters: A public listing by a humanoid robot company tests investor appetite for physical AI.
US humanoid robot maker Agility Robotics has listed on Nasdaq, attracting investors while facing scrutiny over profitability despite a $300 million order.
Why it matters: AI glasses from major eyewear brands bring Gemini to wearables for consumers.
Warby Parker begins pre-orders on Nov. 12 for its Google AI glasses. Gentle Monster plans a December launch, and Mazda adds Google Gemini to a compact SUV.
Why it matters: Clears a key regulatory hurdle for Apple's AI rollout on more devices in China.
Apple Intelligence has completed regulatory registration in China for use on iPads and Macs, a step toward launching the service on those devices in the country.
Korean startup Blue Intelligence launches an AI device for cars that detects when older drivers hit the wrong pedal, aiming to prevent accidents.
Aina's smart ring goes beyond sleep and step tracking: it talks to AI agents, takes phone calls and shows notifications on a small screen, with a joystick and speaker. Early-bird preorders are $399.
Allegro MicroSystems has developed 48V power and precision sensing ICs aimed at humanoid robots, EE Times Asia reports, targeting motor drive and position feedback in robot joints.
Augment Code 0.906.0 updates its VS Code extension, an AI coding platform for large codebases whose context engine and coding agent understand architecture, dependencies and legacy code.
Mirage is an Apache-2.0 image-text-to-text model that describes scenes, detects objects, reads text, segments images and extracts structured data. Weights are on Hugging Face with a Python package.
Arete is a Hugging Face dataset of prompt-response pairs on seven values topics, each answered by an Inkling teacher model and the Qwen 3.8-27B checkpoint, with category tags and verification hashes.
Amp released CLI version 0.0.1794268864-g13d0b0, installable via npm i -g @ampcode/cli. Amp is a frontier coding agent and development environment.
AWS describes how AWS Quick encodes repeatable tasks in agents and skills, how shared skills and data fit together, and how its catalog and custom agents work.
GGUF quantized versions of the Gemma-4-31B-it-ARA-heretic model, for running locally with tools like llama.cpp. Static and imatrix quants are available.
Simon Willison describes using Claude to help track down the source of an intermittent beep in a cluttered box room, an anecdotal example of using the AI assistant for a real-world debugging task.
Review weighs Google Gemini against Amazon Alexa Plus as AI home assistants, weighing which is worth using.
Red Hat released OpenShift 5 at KubeCon + CloudNativeCon North America on November 9, 2026, adding agentic troubleshooting and a single platform for AI workloads, VMs and containers.
Epic Games' Unreal Engine 5.9 preview arrives with built-in AI agent tools, letting developers use agents inside the game engine workflow.
Why it matters: Pushes open training stack to multi-million-token contexts, easing long-context model work on TPUs.
Google Cloud's open-source MaxText framework now supports long-context training up to 14 million tokens on Cloud TPUs via four context-parallel strategies (All-gather, A2A, P2P, USP), powered by the Tokamax Splash Attention kernel.
Microsoft began limiting Copilot Chat usage for users who lack paid licenses, tightening free access to its AI assistant.
Why it matters: Extends OpenAI's persistent agents to multiple local machines, a step toward distributed agent workflows.
OpenAI's persistent agents can now route a task to a second Mac or PC, letting developers test Xcode builds and older SDKs while continuing to work.
Amazon Bedrock Managed Knowledge Base now supports single-shot RAG and retrieval-only requests via AgenticRetrieveStream, letting developers match retrieval effort to each request.
Alibaba's Qoder coding platform launched a November 7 promotion giving individual users a limited free daily allowance for an undisclosed model called Qoduck; a tokenizer test hints at Qwen lineage.
Airbnb is building an AI agent that can arrange itineraries and manage holiday reservations, even as CEO Brian Chesky dismisses the AI booking threat as negligible.
numbrone is an open-source shell tool that acts as a chief of staff for coding agents, routing requests to per-project manager agents that run parallel workers in git worktrees.
Goose is Block's open-source AI agent that installs, executes, edits and tests code with any LLM. v1.58.0 adds context relevance, a context window viewer, AIHubMix, RouterPlex and Lemonade providers.
An AWS Machine Learning post describes a seven-layer composable reference architecture for taking enterprise agentic AI from prototype to production on Amazon Bedrock AgentCore, with swappable layers.
Why it matters: Copilot gaining write and delete powers moves it from assistant to acting agent across business apps.
Microsoft Copilot can now create, update and delete data across connected apps, per Windows Report, expanding its actions beyond reading and summarizing.
Auggie is Augment Code's command-line coding agent, installed via npm; version 1.2.0 is a routine update to the minimal CLI client for AI-assisted coding.
A satirical PCMag tool called 'Ram-Altman' lets people email OpenAI CEO Sam Altman about their ChatGPT memory usage. It is a joke, not an OpenAI product.
Instagram's Effect Builder generates custom filters for Stories posts from a text prompt; available now for US users over 18.
Cisco rolls out Webex for Apple's folding iPhone Duo with dual-display split view and hands-free camera use for annotating, plus AI meeting notes.
A new AI Copilot feature lets users pick between Claude, ChatGPT and DeepSeek as the underlying model, giving choice across major AI chatbots.
Meta AI's iPhone app version 298.0.0 adds scheduled tasks and reminders, email and calendar connections for organization, and in-depth research reports.
Starlink's Grok-based AI customer support chatbot can be prompted to offer discounts; existing customers report getting 30% off by asking Grok.
Medium released TK, a new writing app that integrates AI assistance directly into the writing process, despite ongoing controversy over AI's role in publishing.
A Skyrim modder used AI to help reverse-engineer the game's 15-year-old engine, adding real-time terrain editing, modular construction and NPC navigation to player-built structures.
Mistral Vibe CLI is a minimal command-line coding agent by Mistral, installable via pip. Version 3.1.1 is now available.
Mirage is an Apache-2.0 image-text-to-text model that describes scenes, detects objects, reads text, segments images and extracts structured data. Weights are on Hugging Face with a Python package.
Arete is a Hugging Face dataset of prompt-response pairs on seven values topics, each answered by an Inkling teacher model and the Qwen 3.8-27B checkpoint, with category tags and verification hashes.
GGUF quantized versions of the Gemma-4-31B-it-ARA-heretic model, for running locally with tools like llama.cpp. Static and imatrix quants are available.
LongCat-2.5 is a 1.6-trillion-parameter multimodal AI model released by Meituan on Hugging Face, available for download and use.
A Hugging Face dataset with glTF binary (.glb) versions of the 1,033 household products in Google Scanned Objects, taken from Gazebo Fuel and sized in metres for three.js, Babylon.js and other glTF viewers.
A Hugging Face dataset of 1,409,404 synthetic 16x16 pixel-art sprites with transparent backgrounds and English captions, generated with Krea 2 Turbo and a pixel-art LoRA, covering 6,194 subjects under CC0.
Underdog-Woof is a text-to-speech model claimed to post a 1.42% word error rate on Gradium TTS hard cases despite being over 1,000x smaller than Fish Audio Drama 3.
numbrone is an open-source shell tool that acts as a chief of staff for coding agents, routing requests to per-project manager agents that run parallel workers in git worktrees.
Goose is Block's open-source AI agent that installs, executes, edits and tests code with any LLM. v1.58.0 adds context relevance, a context window viewer, AIHubMix, RouterPlex and Lemonade providers.
Jarvis is an autonomous, local-first on-device AI agent framework for Android that uses full accessibility control and continuous voice to automate workflows on the phone. It hit 100 stars in its first week.
A free Python workbench for AI security operations that handles alert triage, incident response and vulnerability remediation, connecting Atlas, models, tools and workflows. It drew 100 GitHub stars in its first week.
roblox-studio-mcp is an open-source Model Context Protocol server that lets AI assistants like Claude, Cursor and Cline read Roblox Studio object trees, automate Lua scripting and run playtests safely. It gained 279 stars in its first week.
agent-screenplay is an open-source TypeScript project that lets an AI coding agent produce clean screen recordings, aimed at demos and walkthroughs. It drew 100 stars in its first week.
Two new repos ship agent skills: flawless-ui-ux helps coding agents build distinctive, usable interfaces, and kubai-icons draws animated icons in one style for Claude Code and Codex, with 2,600+ icons for React, Vue and SVG.
Simon Willison describes using Claude to help track down the source of an intermittent beep in a cluttered box room, an anecdotal example of using the AI assistant for a real-world debugging task.
Review weighs Google Gemini against Amazon Alexa Plus as AI home assistants, weighing which is worth using.
Microsoft began limiting Copilot Chat usage for users who lack paid licenses, tightening free access to its AI assistant.
Airbnb is building an AI agent that can arrange itineraries and manage holiday reservations, even as CEO Brian Chesky dismisses the AI booking threat as negligible.
Microsoft Copilot can now create, update and delete data across connected apps, per Windows Report, expanding its actions beyond reading and summarizing.
A satirical PCMag tool called 'Ram-Altman' lets people email OpenAI CEO Sam Altman about their ChatGPT memory usage. It is a joke, not an OpenAI product.
Instagram's Effect Builder generates custom filters for Stories posts from a text prompt; available now for US users over 18.
Cisco rolls out Webex for Apple's folding iPhone Duo with dual-display split view and hands-free camera use for annotating, plus AI meeting notes.
A new AI Copilot feature lets users pick between Claude, ChatGPT and DeepSeek as the underlying model, giving choice across major AI chatbots.
Meta AI's iPhone app version 298.0.0 adds scheduled tasks and reminders, email and calendar connections for organization, and in-depth research reports.
Starlink's Grok-based AI customer support chatbot can be prompted to offer discounts; existing customers report getting 30% off by asking Grok.
Medium released TK, a new writing app that integrates AI assistance directly into the writing process, despite ongoing controversy over AI's role in publishing.
A Skyrim modder used AI to help reverse-engineer the game's 15-year-old engine, adding real-time terrain editing, modular construction and NPC navigation to player-built structures.
Snapp, an AI sports analytics app from a Caviar co-founder, raised $4.6M and is now available on Android.
A roundup of six AI video interview assistant tools designed for live calls rather than one-way recordings.
ZDNET spent weeks testing OpenAI's Dot and Anthropic's Duet assistant products, finding only one genuinely helpful notification between them.
Report looks at how and when third-party apps will gain access to Siri AI features, and what Apple's integration timeline means for developers.
Livemint compares Anthropic's Claude Personal Finance with ChatGPT's finance features across privacy, capabilities and bank-account access.
MAIA is an AI agent designed for simultaneous use by multiple people on a team, released to let groups collaborate with one agent at the same time.
A developer used Claude Code to build a Windows desktop app where snow accumulates on window edges and falls off when windows are dragged, adding depth to flakes.
Meta released Muse Ember, a lightweight model for agent execution priced at $0.10 per million input tokens, and added Spaces, letting users build custom AI tools and dashboards from prompts in the Muse app, with video chat and India pricing.
Notebookcheck examines what parents can control in ChatGPT, Gemini and Meta AI, covering the settings and limits each chatbot offers for younger users.
Qwen Office 2.0 adds shared agent workspaces linked to DingTalk chats and docs, event-triggered Automatic Tasks, in-place co-editing and a mobile app. Cloud tasks keep running after a PC shuts down.
BoltAI is a desktop AI chat client for working with multiple models. Version 3.0.0 is out for macOS, Windows and Linux.
DeepSeek's AI agent platform, Harness, has reached 1.54 million daily active users, up roughly 5.7-fold in a month and in under three months since launch.
Augment Code 0.906.0 updates its VS Code extension, an AI coding platform for large codebases whose context engine and coding agent understand architecture, dependencies and legacy code.
Amp released CLI version 0.0.1794268864-g13d0b0, installable via npm i -g @ampcode/cli. Amp is a frontier coding agent and development environment.
AWS Machine Learning · 2h
#
AWS describes how AWS Quick encodes repeatable tasks in agents and skills, how shared skills and data fit together, and how its catalog and custom agents work.
Red Hat released OpenShift 5 at KubeCon + CloudNativeCon North America on November 9, 2026, adding agentic troubleshooting and a single platform for AI workloads, VMs and containers.
Epic Games' Unreal Engine 5.9 preview arrives with built-in AI agent tools, letting developers use agents inside the game engine workflow.
Google Cloud's open-source MaxText framework now supports long-context training up to 14 million tokens on Cloud TPUs via four context-parallel strategies (All-gather, A2A, P2P, USP), powered by the Tokamax Splash Attention kernel.
OpenAI's persistent agents can now route a task to a second Mac or PC, letting developers test Xcode builds and older SDKs while continuing to work.
AWS Machine Learning · 6h · 2 sources
#
Amazon Bedrock Managed Knowledge Base now supports single-shot RAG and retrieval-only requests via AgenticRetrieveStream, letting developers match retrieval effort to each request.
Alibaba's Qoder coding platform launched a November 7 promotion giving individual users a limited free daily allowance for an undisclosed model called Qoduck; a tokenizer test hints at Qwen lineage.
AWS Machine Learning · 7h
#
An AWS Machine Learning post describes a seven-layer composable reference architecture for taking enterprise agentic AI from prototype to production on Amazon Bedrock AgentCore, with swappable layers.
Auggie is Augment Code's command-line coding agent, installed via npm; version 1.2.0 is a routine update to the minimal CLI client for AI-assisted coding.
Mistral Vibe CLI is a minimal command-line coding agent by Mistral, installable via pip. Version 3.1.1 is now available.
Cledara released an AI agent that handles software admin work like downloading vendor invoices, onboarding staff into tools and deactivating accounts when people leave.
Fizgig is an open-source tool that shows heatmaps inside a LoRA adapter during training, revealing what it has learned, what it hasn't and where it leaks. Version 9.3.0 is on GitHub.
Databricks introduces UCDataset, a tool that lets teams train models on data stored in Unity Catalog tables while keeping GPUs fed, aiming to cut the cost of expensive training runs.
Claude Code, Anthropic's terminal-based AI coding assistant, gets version 2.1.321. It can understand codebases, edit files, run commands and handle workflows.
DataAgent, from Cloudify veterans, released a free local K8sGPT lab for building autonomous SRE agents alongside its commercial remediation platform.
Kolmena launches an invite-only platform for running AI agents as coordinated swarms, combining model routing with MAYA, Agent Builder, Swarm Builder and 1,600+ claimed integrations.
FormAssembly, known for online forms and data collection, rebuilt its platform to automate complex business processes using AI, expanding beyond form building into workflow automation.
Automattic debuted Spacefast, a platform for deploying and hosting vibe-coded apps and websites, aiming to compete with Lovable and similar AI app builders.
Dapr 1.19 updates the open-source distributed application runtime with faster durable workflows and a simpler control plane aimed at AI agent workloads.
Samepage launched Bets, an agent board where AI agents draft research, open pull requests and write launch copy, following its $4.85M seed round in June.
Developers can now build Google Maps-powered apps directly in Google AI Studio by prompting Gemini with geospatial concepts; one-click demo keys eliminate configuration hurdles.
An AI code editor. Version 3.29.15 lets users write, edit and chat about code with AI on macOS, Windows and Linux.
MuseSpark AI launched a platform combining a single unified API across AI models with an infinite canvas workspace, claiming it cuts AI model costs by 40%.