Updated
Updated
Showing low-relevance items too. Hide low-relevance items
AITypesafe AI raised $870 million in a Series A round at a $7.5 billion valuation, led by Andreessen Horowitz with participation from Sequoia Capital and existing investor DCVC. Martin Casado joins the board. The company says it will add more machine-native models and enterprise features and that a third of the Fortune 500 use its product.
Prime Intellect@PrimeIntellectOfficialAI score46AIPrime Intellect says its Prime Agent orchestrated over 2,000 agents over two weeks to rewrite the agent in Rust. The run used more than 10,000 sandboxes, over 200B GLM-5.3 tokens, and 16,000 agent-to-agent messages. The company says the rewritten agent reaches usable input about 13 times faster and uses 83% less startup memory.
Prime Intellect@PrimeIntellectOfficialAI score22AIPrime Intellect says it preserved its TypeScript version's features and behavior by giving agents objective parity checks. The checks diff terminal frames, compare session transcripts and model requests, check daemon protocol messages, and audit every feature. Agents could see where behavior diverged and fix it before changes merged.
Prime Intellect@PrimeIntellectOfficialAI score20AIA root agent splits a rewrite into dependent tasks, with each task run through a Planner, Implementer, Reviewer, and Verifier state machine. Each implementation must pass independent review and verification in a fresh Prime Sandbox before merging, and failed checks send the task back to the implementer. Tasks can run in parallel without skipping these checks.
Prime Intellect@PrimeIntellectOfficialAI score36AIPrime Intellect says Prime Agent, after reaching feature parity, ran its own runtime benchmark suite and tested candidate changes against the current build. Changes that passed parity checks and independent review became the baseline for the next experiment. The loop moved work off the startup and render paths and released memory after large sessions loaded.
Prime Intellect@PrimeIntellectOfficialAI score29AIPrime Intellect says its Prime Agent Rust had the lowest time to usable input, startup memory, and installed size among six agent harnesses it benchmarked. The rewrite splits the codebase into nine crates with enforced dependency boundaries, and changes to shared protocol types are checked across the client, daemon, and session workers.
Prime Intellect@PrimeIntellectOfficialAI score42AIPrime Intellect says it is turning the workflow behind a recent rewrite into reusable state machines in Prime Agent, letting users run their own agent teams through implementation, review, and verification. The company also says it is accelerating work on capabilities and evals and connecting Prime Agent with cloud agent swarms and hosted training for autonomous research. This release adds native Windows support in beta and Homebrew installation.
Prime Intellect@PrimeIntellectOfficialAI score38AIPrime Intellect says Prime Agent is now open source and that its Rust rewrite is available today. The full blog post covers verification loops, multi-agent orchestration, and performance experiments.
Prime Intellect@PrimeIntellectOfficialAI score22AIPrime Intellect says users can update their Prime Agent by running the command prime-agent update. The post links to the project's GitHub repository, PrimeIntellect-ai/prime-agent.
Tibo@thsottiauxOfficialAI score38AIOpenAI's Tibo says composer predictions are now available in the Codex desktop app, suggesting a user's next message based on the conversation. The feature is in beta for Pro users and is included in Pro plans without consuming usage.
Tibo@thsottiauxOfficialAI score22AIOpenAI's ChatGPT app on iOS and Android now lets users create a dot and text it directly from the phone. Previously, dots could only be created through the desktop or web app.
StepFun@StepFun_aiOfficialAI score34AIStepFun's Step 5 Preview, a 600B-total, 27B-active MoE model with 1M context and vision, is free to try on Nous Portal for one week. Nous Research says it scored 33.89 on the Hermes Index, the same score as GPT-6 Luna.
Guillermo Rauch@rauchgXAI score38AIVercel says agents can now buy domains with the Vercel CLI, extending an agent marketplace where they already purchase infrastructure products and services. Guillermo Rauch says agents have bought from the marketplace through the CLI often enough to surprise the company. He adds that agents can now move from idea to online business, including registering a domain name.
ElevenLabs@ElevenLabsOfficialAI score38AIElevenLabs is releasing synthetic voice detection in ElevenAgents to help businesses identify AI agents calling on behalf of individuals, companies, or bad actors. The company is also joining the Personal Agent Protocol working group to help define how agents interact.
ElevenLabs@ElevenLabsOfficialAI score34AIElevenLabs says it is working on new standards for agents that interact with businesses as part of the Personal Agents Protocol working group led by Meta and Sierra. The company says businesses use its platform to deploy agents across phone, website, in-app, email, WhatsApp, SMS, and other text and voice channels. ElevenLabs says its goal is to ensure all of those channels are covered so customers can define how they interact with personal agents.
ElevenLabs@ElevenLabsOfficialAI score24AIElevenLabs is launching synthetic voice detection that analyzes a caller's speech in the first seconds of a call to determine whether it is human or AI generated. Calls are then routed accordingly, so people get a human-oriented experience, agents get bounded interactions, and bad actors can be stopped.
ElevenLabs@ElevenLabsOfficialAI score38AIElevenLabs says synthetic voice detection is available now to enterprise customers through its forward deployed team. Broader access is planned for later this month, and the company will share progress on its Personal Agent Protocol (PAP) as the work develops.
Arena.ai@arenaOfficialAI score31AIArena says Agent Arena replaces pairwise preference voting with a causal tracing methodology. The method measures five signals from real-world, long-horizon agentic sessions.
Arena.ai@arenaOfficialAI score23AISierra has published a draft of the Personal Agent Protocol, known as Poppy, and named 35 additional design partners, including Adyen, Bank of America, Mastercard, OpenAI, PayPal, and Visa. Under the protocol, companies publish a /.well-known/poppy.json discovery file, and personal agents start sessions, identify themselves, and sign in through OAuth with session tokens limited to approved access. The company says the draft will be followed by design workshops and a reference implementation over the next month.
Why it matters: The draft specifies how personal agents identify themselves, obtain customer-approved access, and work with company websites, APIs, or agents, which helps readers assess its practical effect on agent-driven transactions.
AITypeSafe Inc., creator of the Jev AI model, says it raised $870 million at a $7.5 billion valuation, led by Andreessen Horowitz with Sequoia Capital and DCVC participating. Jev outputs structured results, such as yes/no answers, list selections or scores, instead of natural language text, and the company says it runs requests in under 700 milliseconds, up to 200 times faster than some frontier LLMs. TypeSafe plans to use the funds to expand its planned System One model series and roll out unspecified enterprise features.
Vercel Developers@vercel_devOfficialAI score38AIVercel says its CLI now lets AI agents purchase domains directly. The post links to a Vercel changelog entry with details.
AIOpenAI has released the Decisions API in public beta, returning typed probabilities, choices, and scores instead of prose from text and images. OpenAI says it runs about 10x faster than the Responses API and costs $0.10 per 1M input tokens with no output charges. The article notes that OpenAI has not published accuracy data and that TypeSafe Jev offers cheaper input pricing at $0.042 per 1M tokens.
AIAlibaba's Qwen team released Qwen-Image-2.1-Turbo, an accelerated checkpoint of Qwen-Image-2.1 that generates and edits images in 8 denoising steps instead of 40. The model keeps the same 7B architecture and offers a hosted API at CNY 0.1 per image, while its weights are under a Qwen Research License that requires separate permission for commercial self-hosting.
LangChain@LangChainOfficialAI score22AILangChain says LangSmith LLM Gateway now supports the OpenAI Decisions API for low-latency agent inference. The gateway provides centralized controls for model fallbacks, data redaction policies, and spend limits.
Andrew Curran@AndrewCurran_XAI score36AIBusiness Insider reports that Google is internally testing a new Gemini 4 checkpoint named Carbon. The checkpoint reportedly matches Opus 5.5 in coding.
Murtuza J Merchant@murtuza_mercXAI score36AIQwen has released its Turbo image model with open weights, giving image app builders an option to run it in-house instead of relying on hosted API pricing. The post does not specify model size, benchmarks, or pricing.
Google Gemma@googlegemmaOfficialAI score46AIGoogle AI Pro plans now include Colab access to A100 GPUs with 80GB VRAM, and Ultra subscribers can use H100 GPUs. With that memory, users can run Gemma 4 31B in bf16, fully fine-tune Gemma 4 E4B in bf16, LoRA-tune Gemma 4 26B A4B in bf16, and QLoRA-tune Gemma 4 31B in bf16.
Google Gemma@googlegemmaOfficialAI score32AIGoogle says Colab is now included in its Google AI plans. The post links to a Google Developers blog post for details and setup instructions.
🚨 AI News | TestingCatalog@testingcatalogXAI score36AIMicrosoft has made Microsoft-Decision-1 available on Microsoft Foundry, a model post-trained on Qwen3.5-9B for fast, single-pass decision scoring. Microsoft says it achieved the highest accuracy across a 36-benchmark comparison of nearly 150,000 questions, and runs 4.5 times faster than Quyet-1.0-Large and 35 times faster than GPT-6 Sol. Microsoft plans to rebase it on other models, including MAI and OpenAI models.
Sherwin Wu@sherwinwuXAI score38AISherwin Wu, who owns OpenAI's account context here, says he has been tab-accepting about 40-50% of Codex's next-message suggestions after a week of use. OpenAI Devs says composer predictions, which suggest a user's next message from the conversation and their style, are in beta for Pro users.
Prime Intellect@PrimeIntellectOfficialAI score20AIPrime Intellect shares an essay by @hallerite titled "On the nature of the swarm," linked on its blog. The post gives no further details about the essay's content.
OpenRouter@OpenRouterOfficialAI score8AIOpenRouter says the models Luna and Astra are driving a lot of new Codex usage. The post gives no figures, dates, or pricing for that usage.
OpenRouter@OpenRouterOfficialAI score23
Lucas Beyer (bl16)@giffmanaXAI score22AILucas Beyer says he worked at a startup a decade ago with nearly identical arms, grippers, and camera placement. He says those models used CNN and LSTM architectures without real pretraining, while several startups now show a big step up.
Julien Chaumond@julien_cXAI score23AICloudflare has published a new model called clef-omni on Hugging Face, according to a post from Julien Chaumond, who owns the account. The post links to the model page but gives no further details about its size, capabilities, or benchmarks.
AMD@AMDOfficialAI score22AIAMD's Data Center team says it worked with Zyphra to train an advanced reasoning model from scratch on AMD hardware. A linked post says Zyphra trains larger reasoning models more efficiently while supporting longer context windows.
Alexandr Wang@alexandr_wangXAI score20
Guizang@op7418XAI score16AIGrok Bot now has its own email service, and users can claim an address with the prefix they need. To get one, users mention the bot in a comment or send it this post on X. The bot can use the address to sign up for services, contact businesses, or schedule meetings.
elvis@omarsar0XAI score40AIMicrosoft releases Microsoft-Decision-1, a model for fast decision-making, according to Satya Nadella. Nadella says it outperforms both LLMs and other decision models on structured decision tasks in latency and quality. Microsoft is testing it internally for incident response, quality control, and scientific discovery.
