The universe is cooling.
Civilization is heating up.
One day, seen from outside time, and the stories that just moved civilization forward.
Get it as a 4K wallpaper →
34.2
+117.916.4
16.2
Launches and updates of AI products, features, apps, tools, APIs and platforms.
Google Gemini@GeminiAppPicks69You can now take an interactive web app you built in Gemini App straight into Google Sites - no code to copy, no app to rebuild. :) Export it to a private site draft, customize it, then publish and share with your team, customers or students. 🧵 1/3
Exporting an existing Gemini web app into a private Google Sites draft removes the copying and rebuilding steps between app creation and site publication.
IBM introduces Granite Timeseries Ensemble, a permissively licensed recipe combining four pretrained Granite models to produce median and quantile forecasts without retraining or fitted ensemble weights. It offers default equal-influence probability pooling and an adaptive IQR alternative. Across 97 GIFT-Eval configurations, the default achieves aggregate MASE of 0.683 versus 0.685 for the strongest individual member, while PatchTST-FM-r2 retains the lowest aggregate CRPS/MWQL. Loading four checkpoints increases inference time and memory use.
The benchmark comparison clarifies the ensemble’s tradeoff between slightly better aggregate point forecasts and increased inference costs, while showing why probabilistic accuracy can favor an individual member.
Anthropic introduced a research preview of Claude Duet, a standalone proactive agent built on Claude Opus 5.5 that works on tasks while users are away. Available through a waitlist for US Pro and Max subscribers on iOS, Android, and the web, it checks connected apps every few hours and shares usage limits with other Claude products. Anthropic says Duet requires approval before sending messages, making payments, signing documents, or deleting anything, and acknowledges it can misread threads or miss updates.
Duet’s approval requirements and periodic account checks illustrate how a proactive agent’s autonomy is bounded by user control and the subscription limits shared with other Claude products.
Walmart announced Agent Gateway, an MCP server that lets approved AI agents search its online inventory and make purchases for shoppers. Agent Profiles, rolling out next month, will separate agent purchases and let customers require transaction approvals or set spending limits. Support requires a Walmart account but no Walmart+ membership, and Walmart says its returns process also applies when an agent acts unexpectedly.
Walmart’s purchase approvals, spending limits, and returns terms offer a concrete example of how retailers can give shoppers control over transactions made by AI agents.
vLLM@vllm_projectPicks65Native PyTorch on TPUs is here—and TorchTPU is officially open source. Train and serve models using the tools you already know: @vllm_project, @sgl_project, @_llm_d_, TorchTitan, @huggingface, and @raydistributed. Learn more from #PyTorchConf2026 → https://t.co/mISCD65vA4 https://t.co/z1Uwb2T3l6
Support for prefix caching, speculative decoding, and disaggregated serving gives readers concrete features to consider when assessing vLLM’s native PyTorch serving on TPUs.
vLLM@vllm_projectPicks66Introducing Flint, @Recursive_SI’s first model: a fast local coding model built with our automated research system, the Eureka Machine. • Best-in-class fast local coding • Up to 2x faster than comparable models: 125+ tokens/sec on a MacBook Pro, 150+ on RTX Spark • Built from https://t.co/ufg2xScSGw
The reported throughput at 100K context provides a concrete reference for evaluating Flint deployment on a single GB200 without speculative decoding.
ChatGPT Ads is expanding to 26 new countries, and over 100,000 advertisers have used the platform. More advertisers, from local businesses to global brands, can now reach people as they explore, compare, and decide what to buy in ChatGPT.
The expansion gives advertisers in 26 additional countries access to people exploring purchases in ChatGPT, extending the platform’s geographic availability for businesses ranging from local shops to global brands.
OpenAI and Google Cloud are bringing frontier models and Codex workflows to Gemini’s Model Garden, with availability through Google Cloud Marketplace.
The announcement identifies Google Cloud Marketplace as an access channel for OpenAI models and Codex workflows, clarifying where these offerings are headed.
SimReady Central Library brings more than 2,000 simulation-ready OpenUSD assets to Hugging Face, providing free, ungated access for robotics simulations and AI factory digital twins. Developers can search for equipment through natural-language queries and select assets by supported simulation profiles. OEMs can validate, package, and submit assets, while examples outline AI-assisted workflows for robotic pick-and-place scenarios and alternative factory layouts.
The profile system gives developers a concrete basis for selecting assets with the physical properties and metadata their simulation workflows require, reducing manual preparation when combining equipment from different suppliers.
NVIDIA introduced Confidential Bare Metal, a deployment option that lets frontier model builders run sensitive workloads on dedicated third-party infrastructure while retaining control of their models, data, and software stack. It joins Confidential Containers and Confidential Virtual Machines, with new security controls protecting data in transit, measuring platform integrity, and enforcing hardware-level access restrictions. Support now extends to Grace Blackwell NVL72, HGX Rubin NVL8, and DGX Rubin NVL8 systems; Cohere plans integration into Cohere Model Vault.
The deployment options clarify how model builders can match infrastructure to their trust boundaries while retaining control of their software stack on third-party computing capacity.
NVIDIA’s Factory Operations Blueprint (FOX) is now available in early access on GitHub for building factory manager and specialist agents, while cuDSS reaches general availability for accelerated simulation. Siemens is integrating FOX-based agents into Intelligence Center X for estimated cost savings of 20% and reports reducing inspection calibration time by 40% using Cosmos-generated defect images. Other European manufacturers describe applications spanning engineering simulation, digital twins and robot validation.
The factory examples illustrate how connecting visual inspection, machine logs and engineering knowledge can support coordinated operational responses, with Siemens reporting reduced inspection calibration time.
NVIDIA expanded its open source Agent Toolkit with Agent Knowledge Foundation, new Kumo models and tools for grounding enterprise agents in company data and controlling their operation. Agent Knowledge Foundation combines CUDA-X libraries, Auto-Ontology and open models to support retrieval and predictions without moving data or changing existing workflows or governance controls. OpenShell supplies runtime boundaries, while NemoClaw blueprints provide reference designs for secure, specialized agents.
The expansion illustrates how enterprise agent infrastructure connects company data, predictive models and runtime controls while preserving existing data locations, workflows and governance controls.
宝玉@doteyPicks79Day 16/ You can now cache for 12h in the API and for logged in users across Codex & Work we keep cache forever and you don’t have any penalty for coming back to an old conversation or resuming after a few hours and are now guaranteed to hit the cache. https://t.co/0HYeB01KVh
The update distinguishes API cache retention from logged in session retention, clarifying how returning to older conversations is treated across the two access paths.
vLLM@vllm_projectPicks75As promised, FastVideo open-sources @MiniMax_AI FastH3 Omni, training code included! 🚀 - Real-time gen @ 12x faster on NVIDIA Blackwell - NVFP4/FP8 for Spark + RTX! - ComfyUI day-0 Collab with @NVIDIAAI , Nuva Lab @nuvalab and Reactor @reactorworld Open source must win! ♥️ https://t.co/aBfZ4hfbl5
The release pairs model weights with training code and day-zero vLLM-Omni support, giving readers concrete context on both access to the model and serving options on consumer RTX GPUs.
Arthur Mensch@arthurmenschPicks69Introducing Voxtral 3: state-of-the-art Transcribe and Speech models powering Mistral’s new voice agents API. 🧵
The deployment options give teams a concrete infrastructure choice when considering Voxtral for voice workflows that require speaker diarization and end-of-turn detection during processing.
Google announced Tuesday that TorchTPU, its version of PyTorch for TPUs, is available as an open-source project, allowing existing PyTorch apps to migrate through a configuration change without rewriting. Amazon released TorchNeuron on Monday with a similar migration path to Trainium and AWS. Google also launched the TPU agent suite to assist migration and debugging, worked with AI development tool providers on compatibility, and updated LiteRT to better support PyTorch apps.
The comparison highlights how Google and Amazon are targeting the software migration work that keeps PyTorch applications tied to Nvidia hardware, with both offering paths that avoid rewriting applications.
Perplexity@perplexity_aiPicks75Running both the model and agent on a Mac offers a local processing option for tasks involving sensitive files and native apps without sending that data to the cloud.
Hugging Face announces native TPU support across Transformers, Diffusers, and Kernels through TorchTPU, a native PyTorch backend. Models can use device_map="tpu" or pipe.to("tpu"), with optional compilation and distributed execution; Optimum TPU is no longer required for Transformers and Diffusers. Hugging Face reports that 96.5% of tests running on TPU pass, with 19 of 34 reported TorchTPU issues already fixed.
Native TPU support lets PyTorch users keep familiar Hugging Face workflows, reducing the software changes previously required to move models onto Google’s accelerators.
Aravind Srinivas@AravSrinivasPicks68Computer built Big Boy No. 4014, the world's largest operating steam locomotive, as a live isometric SVG with its own renderer and a 7-step operating sequence from cold boiler to 60 mph. The live model has 1,324 parts and 15,312 convex faces, built from @UnionPacific's published figures. Try it: http://big-boy-4014.pplx.app
The example connects published locomotive specifications with an interactive SVG model, offering a concrete reference for turning technical figures into a visualization of operating stages.
ollama@ollamaPicks71Aloha!🌺 Meet Ornith-1.5-122B: No Tradeoff Between Strong and Fast Stronger. Faster. Leaner. Ornith-1.5-122B delivers frontier-level capabilities with exceptional inference efficiency. With superior instruction following, stronger task decomposition, and deeper understanding of real-world workflows, it achieves state-of-the-art performance among open-source models of comparable size, while running on a single NVIDIA DGX card!! Open weights under the MIT license. Coding ✅SWE-bench Verified: 0.845 ✅TB2.1: 0.813 ✅CyberGym: 0.800 ✅NL2Repo: 0.512 Agentic ✅HLE: 0.446 ✅BrowseComp: 0.785 ✅MCP-Atlas: 0.746 ✅Toolathlon: 0.708 🤗Hugging Face: https://huggingface.co/ornith-ai/Ornith-1.5-122B We can’t wait to see what you build with Ornith-1.5-122B.
The Ollama command provides a concrete starting point for local deployment, while the quoted MIT license clarifies the terms under which the model weights are released.
