Updated
Updated
Showing low-relevance items too. Hide low-relevance items
Google Research@GoogleResearchAI score34AIGoogle Research will demo ScientistTwo, an autonomous multi-agent framework, at the COLM booth (#107) today at 2:00 PM PT. The framework analyzes research papers, identifies their limitations, and produces verified codebases. The accompanying paper is available on arXiv (2609.19644).
Tibo@thsottiauxAI score34AIOpenAI says it removed some friction for developers building on its API, while the main post gives no specific changes. Background from @OpenAIDevs says the five paid usage tiers become three—Build, Launch, and Grow—and Grow, the new highest tier, requires $500 in total API payments, down from $1,000 for the previous top tier.
SemiAnalysis@SemiAnalysis_AI score23AISenko showed an expanded beam backplane connector at CIOE, and SemiAnalysis estimates it carries roughly 1,500 or more fiber strands per connector. Optical backplanes become relevant once hyperscalers and neoclouds adopt in-rack optics to link GPUs and switches for scale-up, which remains far off on the roadmap. Tracking this ecosystem's maturity could indicate the pace of optical scale-up adoption.
AINVIDIA's DOCA GPUNetIO lets GPU applications control networking and data movement directly, rather than routing each transaction through the CPU. The source says host-driven network handling adds latency on the critical path and limits how quickly distributed applications can respond in real time. The provided text is truncated, so details of the unified software stack are not available.
Databricks@databricksAI score10AIadidas is using Databricks Genie to reduce the time analysts spend on data crunching, freeing them to uncover insights that inform smarter decisions and better product design. The post is a promotional brief pointing readers to a fuller account of how adidas' product innovation runs on data.
Google Cloud Tech@GoogleCloudTechAI score20AIOn a livestream, Arthur Soroken interviews Amir and Ali Sadeghian, co-founders of Astrocade, about how generative AI and Gemini multimodal models are reshaping game development.
Google Gemini@GeminiAppAI score22AIGoogle Gemini's Guided Vision supports everyday tasks such as reading fine print, finding objects, describing items, and exploring one's immediate surroundings. The feature is also presented as practical visual assistance for older adults, people with low literacy, and anyone reading in low light.
Google Gemini@GeminiAppAI score29AIWith Guided Vision turned on, Gemini Live verbally prompts users to pan, tilt, or step back when their camera framing is too high, too close, or off to the side. These prompts help Gemini get the clear visual context it needs to answer questions.
Google Gemini@GeminiAppAI score34AIGuided Vision in Gemini Live is now available on Android devices running Android 9 and above, where Gemini Live is supported. Users need the latest Gemini app and Android system software to get started.
Sophia Yang@sophiamyangAI score28AIA 9B model trained with SFT and prompt repetition at about $25 of compute reaches roughly 79% of Jev's Decision Index score. It ranks second among all models at 10B parameters or below and 14th of 70 overall.
Anthropic@AnthropicAIAI score49AIAnthropic is expanding its Cyber Verification Program to give verified security professionals broader access to its most capable models. Through the program, they can use Claude Mythos 5.1, Opus 5.5, and Sonnet 5.5 with safeguards designed for defensive work. New tiers will also allow authorized offensive work such as penetration testing and red-teaming.
Google Gemini@GeminiAppAI score46AIGoogle's Gemini Live now includes Guided Vision, built with blind and low-vision community input, offering conversational real-time visual assistance. Users can share their camera to receive dynamic audio descriptions and verbal cues to help them explore their surroundings.
Google Cloud Tech@GoogleCloudTechAI score15AIGoogle Cloud Tech is hosting a live broadcast with the founders of Astrocade, a platform for building a video game from text alone. The post does not provide further details about the product or its features.
AIClaude Code v2.1.292 adds a --marketplace option to claude plugin install, which adds the marketplace if needed and then installs the plugin from it. The release also adds an effort parameter to the Agent tool and fixes several security issues, including permission prompts bypassed for network (UNC) file reads and a sandboxed read path that could return files outside approved access.
LangChain@LangChainAI score22AILangChain is promoting a LangChain Academy course that teaches developers how to build long-running agents for complex workflows using Deep Agents. The post links to the course page and invites readers to enroll today.
LangChain@LangChainAI score16AILangChain promotes a 90-second video explaining Deep Agents. The post provides no further details about the architecture or features.
Amjad Masad@amasadAI score22AIAmjad Masad shared a link to a full conversation, with no further details provided in the post itself. The linked video is YouTube content, and the post offers no further specifics about its subject or claims.
OpenRouter@OpenRouterAI score32AIOpenRouter now offers google/gemini-nano-banana-2.1, which accepts up to 14 reference images and supports Grounding with Google Search. The model produces cleaner 1:4, 4:1, 1:8 and 8:1 panoramas at 2K and 4K. Pricing is $0.0336 per image at 1K, $0.0504 at 2K and $0.1134 at 4K.
Lydia Hallie ✨@lydiahallieAI score23AIAnthropic's Lydia Hallie says users can still claim a one-time bonus credit for Claude Code cloud sessions until October 7 by running /claim-credit. Cloud sessions run Claude Code on a fresh VM per task, letting several run at once even after the laptop is closed.
Philipp Schmid@_philschmidAI score22AIPhilipp Schmid shares a Hugging Face Space that runs Gemma embedding models in the browser using WebGPU. The demo, a webml-community project, lets users generate embeddings locally without server-side inference.
ClaudeDevs@ClaudeDevsAI score20AIAnthropic's ClaudeDevs account says Pro or Max plan subscribers on September 23 can still claim a one-time bonus credit for cloud sessions by running /claim-credit in Claude Code by October 7 at 11:59pm PT. Cloud sessions draw on this credit first before counting toward plan limits, and the credit expires November 4.
ClaudeDevs@ClaudeDevsAI score38AIAnthropic's ClaudeDevs says Claude Code cloud sessions run each task on a fresh VM, so users can start several at once. The sessions keep running after the user closes their laptop. A field guide covers seven suitable workflows and how to connect GitHub.
Pika@pika_labsAI score22AIPika announces that users can try Nano Banana 2.1 through its create.pika.art app page. The post provides only a link and no details on features, pricing, or capabilities.
Pika@pika_labsAI score13AIPika is making Google's Gemini Nano Banana 2.1 image-to-image model available for developers to build with through its Pika API Club. The post links to a playground page for trying the model.
Pika@pika_labsAI score42AIPika has made Nano Banana 2.1 available on its platform, citing improved visual quality and faster generation speeds. The post says the model incorporates the real-world intelligence of Gemini, though it gives no benchmark figures, pricing, or technical specifications.
Philipp Schmid@_philschmidAI score22AIA Hugging Face Space lets users try EmbeddingGemma 2 directly in their browser using WebGPU. The post shares only a demo link, with no further details on model size, benchmarks, or pricing.
Google Research@GoogleResearchAI score36AIGoogle Research will demo Co-Director, a hierarchical multi-agent framework that optimizes video generation and consistency for long-form storytelling, at the #COLM2026 Google booth #107 today at 1:00 PM PT. The demo showcases interactive cinematic narratives, and the team's blog post details the approach.
GitHub@githubAI score39AIGitHub has released ReviewBench, a new open benchmark for evaluating AI code review agents, shaped by analysis of 103.9M GitHub pull requests. It contains 219 PRs across 19 languages, and developers can bring their own code review agent, evaluate it, and submit their results.
Philipp Schmid@_philschmidPickAI score70AIGoogle releases EmbeddingGemma 2, its first native multimodal embedding model, built on Gemma 4 under Apache 2.0. It embeds over 100 languages, code, images, audio, and video into one vector, with an 8,192-token context and four sizes from 270M to 740M parameters. Matryoshka output dimensions of 768, 512, 256, or 128 are supported, and the model is available in Sentence Transformers and LiteRT-LM, with a reported 14% gain on MTEB Code.
Why it matters: The release extends an embedding model to text, code, images, audio, and video in one vector, a useful option for retrieval systems that mix media types.
Boris Power@BorisMPowerAI score22AIResearch by pzeroresearch estimates that frontier models' experimental research taste has doubled roughly every three months since December 2025, with Opus 5.5 now exceeding their expert human baseline. The author of the main post, Boris Power, calls the plot very interesting for recursive self-improvement implications, while noting that the details matter for doing useful work at frontier labs.
NVIDIA AI@NVIDIAAIAI score16AINVIDIA AI announced a live broadcast titled "DGX Spark Live: Smart Routing for Hybrid AI" on X. The post provides no further details on the routing approach, specifications, or results.
Gergely Orosz@GergelyOroszAI score7AIGergely Orosz says he paid a lot for a book about how 1982 technology advances supposedly meant software no longer needed programmers. He notes the topic is newly relevant given today's rapid progress in technology and computing.
vLLM@vllm_projectAI score60AIvLLM announced day-0 support for EmbeddingGemma 2 from Google DeepMind, a bidirectional omni-modal embedding model that maps text, image, audio, video, and interleaved inputs into one vector space. Users can try it with the latest vLLM nightly build using the command vllm serve google/embeddinggemma-2 --runner pooling. The quoted Google post says the model is built on the Gemma 4 architecture and released under Apache 2.0.
elvis@omarsar0AI score41AIParsewave audited all 600 public tasks in Zapier's AutomationBench and human review confirmed 206 real verifier bugs, all of which were fixed in AutomationBench Verified. Replaying 1,235 Kimi K3 runs on the old and fixed verifiers changed 27.9% of grades, with pass rate rising from 18.8% to 43.8% where verifiers were too strict and falling from 60.2% to 49.7% where they were too lenient.
Matt Shumer@mattshumer_AI score8AIThe post observes that ambitious people are working longer hours than ever, while those on standard 9-to-5 schedules are happily coasting more than before. It frames this as a clear AI-related trend without offering further data or explanation.
Logan Kilpatrick@OfficialLoganKAI score40AIGoogle released Nano Banana 2.1, an update to its image generation and editing model, offering higher-quality images and bug fixes from previous versions. The update also comes with a new lower price point. It is available through the API, AI Studio, and the Gemini app.
WorkBuddy@WorkBuddy_AIAI score11AIWorkBuddy says students at MIT, Yale, and Harvard asked for fewer AI tools, less context switching, and smarter workflows. The company says it is building one place combining models, Experts, Skills, and Connectors to meet that need. It invites readers to suggest their school or city for a future visit.
Mistral AI@MistralAIAI score40AIMistral Large 4 reverse-engineered an unknown binary end-to-end, concluding it was Cobalt Strike and extracting its IoCs and malware configuration. The model also wrote a report with a YARA rule for detecting future incidents. The task, which could take a day of analyst work, was completed in 12 minutes.
Mistral AI@MistralAIAI score47AIMistral Large 4 solved 18 of 19 challenges in a CTF speedrun, with tool calls and solve times drawn from actual runs. The post frames the model as efficient at reasoning over diverse complex challenges compared with other models.
Mistral AI@MistralAIAI score31AIMistral AI says its Mistral Large 4 is one of the world's strongest AI models for cybersecurity. The post gives no benchmark scores, comparisons, or availability details.
