David Arnal

Claude Meetings brings Claude into team calls as an active participant, creating documents, interactive whiteboards, and other materials alongside the team in real time. 🎧 Team members can interact with Claude by voice or chat, request documents, and edit its work as it's being created. Claude also takes notes. After the meeting, its Claude page includes: - Discussion recordings - Action items - Full transcript - All generated materials Google Meet is supported at launch, with other video platforms coming soon. The beta is gradually rolling out to Team and Enterprise users.

Jensen Huang is stepping down as CEO of Nvidia. The company’s CFO will serve as interim CEO while the board of directors searches for a permanent successor.

Google is back in the game. The team has just released Gemini 4 Hydrogen, a new model with promising results. - Built on Carbon and optimized for high speed. - Outperforms Sonnet and Terra on benchmarks. - Scores 57% on Terminal-Bench 4.1. - Pricing: $1.50 per 1M input tokens and $7.50 per 1M output tokens. Currently available only to Pro and Ultra subscribers.

Just found a useful resource for vibe designers: DesignBookmark. It features 2,200+ design tools, websites, and resources across 59 categories. A new Motion section has just been added, featuring 76 animated components you can preview before adding to your project. Each component comes with a ready-made prompt. Click Claude or Codex on the card, copy the prompt, and let AI recreate the component in your project. designbookmark.com/motion

Chrome has updated its Modern Web Guidance skills to version 1.0. They give AI up-to-date knowledge of modern Web APIs, web development best practices, browser compatibility, and more—helping it avoid outdated approaches and use modern web platform features. The update also adds legacy code analysis. The project was developed in collaboration with the Firefox and Edge teams to ensure cross-browser compatibility. developer.chrome.com/docs/mo… github.com/GoogleChrome/mode…

Google now has its own Jev: Gemini Decisions. It's multimodal, accepting video, audio, images, and text as input. Pricing starts at $0.10 per million input tokens, and you can try it for free in Google AI Studio.

MiniMax has launched H3-Image, an image generator with native 4K output (4096×4096) in just 6 steps. It was built from MiniMax's video model H3 and supports prompts up to 32,000 characters and up to 16 reference images. Generation speed for 4K text-to-image: - Highspeed: ~10.3 seconds - Standard: ~16 seconds - Nano Banana 2.1 (Thinking Medium): 50 seconds in the same test Standard API generation costs ~$0.07, with a 50% launch discount. Until November 23, fal also offers 3–5 free generations. Under the hood, H3-Image uses a 33.1B diffusion transformer. The full model, including encoder and VAE, requires ~135 GB of downloads. SGLang supports offloading layers to RAM for 24 GB GPUs, and ComfyUI templates are already available. A two-step LoRA has also been developed, though its quality still needs testing.

OpenAI has launched a $200 retro console for vibe coders: the ModRetro Chromatic Codex Edition. It plays Game Boy and Game Boy Color cartridges and comes with a rewritable cartridge for your own games. With the ModRetro plugin, you can describe a game in Codex and have it build levels and animate sprites. In a demo, Codex created a snowboarding level that ran on the actual console. You can test games in the built-in emulator, then upload the finished ROM to a cartridge. This requires a firmware update and developer mode. OpenAI also released a browser-playable game gallery featuring titles like FlapGPT, with ROM downloads available.

OpenAI is turning ChatGPT into a new customer acquisition channel for businesses. Ads inside ChatGPT can now lead users directly to a company's AI agent instead of its website. Businesses will be able to: - Verify their business profiles and keep product information up to date. - Create custom AI agents to answer customer questions (for eligible US businesses). - Promote those agents through ChatGPT ads. Users click an ad and start chatting with the company's AI agent directly inside ChatGPT. OpenAI is gradually turning ChatGPT from a search tool into a platform where businesses can reach potential customers directly.

Gemini 4 Carbon is a beast. Using the new /swarm feature in Google Antigravity, a team of Gemini 4 Carbon agents turned historical photos and engineering blueprints into an interactive Three.js 3D exhibition — all from a single prompt. The project recreates the construction of the Eiffel Tower between 1887 and 1889. 😳

Alibaba has opened access to Qwen4-Max-Preview, a preview of its new Qwen 4 model. You can try it now through: - Qwen Chat - Agent Mode - Alibaba Token Plan The model targets complex coding tasks, office work, interface creation, and professional software development. No benchmark results are available yet. 👀 chat.qwen.ai/

Anthropic is turning Claude into a full-fledged office suite with three new tools: > Whiteboards — interactive boards for collaboration. > Sheets — spreadsheets with formula support. > Data — dashboards powered by Snowflake and Salesforce, with the ability to turn metrics into short narrated videos. Claude is also getting interactive connectors, allowing developers to embed app interfaces directly into the chat. You can now work with tools like Adobe and Atlassian without switching tabs. Coming soon: Claude will be able to join meetings, take notes, and update documents automatically. Whiteboards and Sheets are in beta for paid plans. Data is available to all paid users.

Gemini 4 Carbon is officially here! The model is now available in: - Google AI Studio - Gemini API - Antigravity - Gemini app - AI Mode in Google Search Unfortunately, access is limited to Google AI Ultra subscribers and developers with paid access. Disappointing that Google Pro users have been left out. 😭

An Anthropic engineer has built an operating system using Fable 5.2. The project took 8 days of continuous work, more than 1,000 subagents, and an enormous number of tokens. Soon, software will be created at a level beyond human capabilities.

Patch is a new platform for creating games, modding existing ones, and launching finished projects. It brings together everything needed for game development, with tools accessible through CLI and MCP — including control by AI agents. What it offers: > 50,000+ open-source assets. > Modify any game published on Patch or create mods for games installed on your PC. > Publish your own games as open source so others can play and modify them. Patch is now available to everyone. patch.gg/ 🔥

AI has fundamentally changed software development in just one year. A year ago, we laughed at neural networks making junior-level mistakes. Entire companies pretended nothing was changing. Today, AI writes code, sends it through AI reviewers, tests it with AI testers, and ships it to production through AI DevOps. One person orchestrates the entire process, mostly because someone still needs to understand what's happening. Autonomous agents and overnight workflows are now routine. A bug reported to support can turn into a ready-to-merge PR within an hour, without human involvement. Companies are questioning whether they even need traditional UIs when users can interact with products through Claude. The conversation has moved beyond "AI in development." We're figuring out an entirely new way to build products. At our online conference on November 11–12, we'll explore: — How autonomous agents, multi-agent systems and humans fit into product development. — How engineering teams work when AI can handle so much of the job. — How to adapt company structures and processes for AI. — What skills matter, and whom companies should hire in this new environment. Speakers include: — Ilya Izmailov on building agent knowledge bases without drowning in hallucinations. — Kirill Mokevnin on introducing AI into legacy projects. — Danila Simonov on making companies AI-native by 2027. — Konstantin Doronin on quality assurance for agent-generated products. — Maxim Klyuchnikov on the human side of agentic development. — Alexander Polyakov on giving agents access to products through MCP, Skills and CLI. — Maxim Romanovsky on product thinking for developers. And more. November 11–12. Online and free with a subscription to the speakers. ai-dev.live/?utm_source=vibe…

Anthropic has published a guide to choosing between Fable 5.2, Opus 5.5, Sonnet 5.5 and Haiku. You can ask Fable 5.2 to audit your Claude Code configuration against its recommendations: claude.dev/blog/which-claude… One important change: Claude Code's built-in Explore subagent now uses your primary model instead of always running on Haiku. That means Opus 5.5 or Fable 5.2 can burn through your premium usage just reading files and searching code. The fix: create your own Explore subagent with model: haiku. You can also ask Fable 5.2 to audit three areas: 1. Subagent model selection. Lightweight agents handling file searches, logs and tests should use Haiku. Agents editing code or making decisions should retain your primary model. Check that CLAUDE_CODE_SUBAGENT_MODEL isn't forcing a weaker model. 2. Model and effort settings. Anthropic recommends: - Opus 5.5 (medium effort) for everyday development and debugging. - Sonnet 5.5 for well-defined tasks like bug fixes and feature implementation. - Fable 5.2 for large refactors, migrations and coordinating many subagents. Flag default effort above high and opusplan, which uses Sonnet for code editing. 3. Project verification. Check whether CLAUDE.md documents how to run tests and builds, and whether any validation is available. For every issue, ask for the relevant configuration and guide recommendation, the likely cost or downside, and a minimal copy-paste fix. Finally, inspect the project code and recent commits to build a table matching task types with models and effort levels. The recommended optimization order: make results verifiable first, increase effort one level second, and only then move to a more powerful model. Use /model and /usage to run the same task on two models and compare costs. Show recommendations first. Don't change anything yet.

Created by an Anthropic engineer using Fable 5.2. This is already on the level of big-budget AAA games! 🔥

Ethereum is seeing record activity, but Vitalik Buterin wants the market to focus less on ETH burn and more on its role as a long-term store of value. According to an unfolded chart using Coin Metrics data, Ethereum mainnet averaged 2.16 million transactions and 899,000 active addresses per day in 2026 through November 4. Yet higher network activity does not automatically mean more ETH deflation. Since EIP-1559, Ethereum burns part of each transaction's base fee. But as the network scales, processing capacity improves and transaction costs fall. More transactions can therefore coexist with relatively low ETH burn. In fact, EIP-1559 was primarily designed to improve transaction fee pricing. Supply reduction was a consequence, not its only objective. Vitalik's broader vision is for ETH to derive value from Ethereum's scalability, security, reliability and economic role. Proposals such as EIP-7999, FOCIL and Frame Transactions reflect research in these areas, though they have not all been implemented. The bigger challenge is competition. Tokenized assets, stablecoins and RWA are expanding across multiple chains, including Solana, Avalanche and Canton. Growth in these markets does not automatically translate into demand for ETH. For ETH to become a durable store of value, collateral asset and source of ecosystem liquidity, its role in payments, collateral, liquidity and network security matters more than burn alone. Burn remains important, but high transaction volume, fee revenue or token destruction does not necessarily create sustainable demand for a blockchain's native asset.

Full CGI movies could soon be created with Python for just a few thousand dollars. For example, Claude Fable 5.2 generated this entire video from scratch by writing Python code to control Blender. 🤯 xitter.cf/jeremyberman/status/21…