AI News Today

In this tutorial, we build a multilingual ASR and speech translation pipeline with NVIDIA Canary-1B-v2. We load the model on a GPU-enabled runtime, prepare audio into 16 kHz mono, and run English ASR. We then translate speech into French, German, Spanish, and Italian, and extract word and segment timestamps. We export translated subtitles as an SRT file, test long-form transcription, run batch processing, and benchmark inference speed. The post How to Use NVIDIA Canary-1B-v2 for ASR, Translation

<p><em>Originally posted on <a href="https://devblogs.microsoft.com/cosmosdb/deep-agents-to-plan-act-verify-against-operational-data/" rel="noopener noreferrer">https://devblogs.microsoft.com/cosmosdb/deep-agents-to-plan-act-verify-against-operational-data/</a></em></p> <p><a href="https://docs.langchain.com/oss/python/deepagents/overview" rel="noopener noreferrer">Deep Agents</a> is an agent harness built on <a href="http://docs.langchain.com/oss/python/langgraph/overview" rel="noopener norefer

The new post-quantum executive order sets a 2030 migration deadline and establishes a powerful foundation for post-quantum resilience. We look at what it gets right, where it can go further, and our migration playbook for government and industry.

Article URL: https://arxiv.org/abs/2606.13175 Comments URL: https://news.ycombinator.com/item?id=48649183 Points: 13 # Comments: 10

Article URL: https://www.youtube.com/watch?v=OQV0CcjBF2Y Comments URL: https://news.ycombinator.com/item?id=48649151 Points: 2 # Comments: 0

Article URL: https://gizmodo.com/yet-another-piece-of-ai-pilled-speculative-fiction-has-gone-dangerously-viral-2000774895 Comments URL: https://news.ycombinator.com/item?id=48649141 Points: 2 # Comments: 0

Comments

Article URL: https://www.youtube.com/watch?v=d5x38FCSCyM Comments URL: https://news.ycombinator.com/item?id=48649129 Points: 3 # Comments: 1

Hey folks. We built SAA (Selective Auditory Attention) after trying to find ways to make a good experience with multiple robots/multiple agents. What typically ended up happening is they'd never stop talking. This is an SDK you can put before your STT. It lets you know when your device is being spoken to or not without a wakeword. You can use it for: -Single AI, Multi human -Multi AI, Single human -Multi AI, Multi human (we recommend also adding a wakeword on top for a better system) There are t

<p>The Security Bottleneck Nobody Talks About</p> <p>Here's a conversation I had with a security engineer last month:</p> <blockquote> <p>"We find maybe 5-10 vulnerabilities per week. Sounds good? Wait for it..."</p> <p>"It takes 4+ hours per vulnerability just to analyze the impact. Which services? Which teams? What's the real risk? Then we write the fix, create the MR, get reviews..."</p> <p>"By the time we're done, it's been 2 weeks. Vulnerabilities pile up."</p> </blockquote> <p>This is the

Article URL: https://lethe.gg/ Comments URL: https://news.ycombinator.com/item?id=48649054 Points: 2 # Comments: 0

<p> Curated, human-written articles about agentic coding </p> <p> <a href="https://www.producthunt.com/products/off-autopilot-2?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed">Discussion</a> | <a href="https://www.producthunt.com/r/p/1179366?app_id=339">Link</a> </p>

Article URL: https://www.nrc.nl/nieuws/2026/06/21/het-belangrijkste-element-van-onbemande-oorlogvoering-is-dat-de-risicos-voor-onze-militairen-omlaag-gaan-a4930367 Comments URL: https://news.ycombinator.com/item?id=48649019 Points: 2 # Comments: 0

<p><em>Originally posted on <a href="https://devblogs.microsoft.com/cosmosdb/introducing-omnivec-an-open-source-embedding-platform-for-ai-apps-on-azure/" rel="noopener noreferrer">https://devblogs.microsoft.com/cosmosdb/introducing-omnivec-an-open-source-embedding-platform-for-ai-apps-on-azure/</a></em></p> <p>Today we are open-sourcing <a href="https://github.com/AzureCosmosDB/OmniVec" rel="noopener noreferrer">OmniVec</a>, a platform for building and operating the embedding pipelines that keep

<p><strong>TL;DR:</strong> I priced 8 local Ollama models by <strong>€ per 1,000 correct answers</strong> — metered GPU energy ÷ correct answers, on one RTX 3090. <code>gemma4:26b</code> won at <strong>96.9% accuracy for €0.013/1k-correct</strong>. The most expensive model (<code>qwen3:8b-fp16</code>) cost <strong>€0.239/1k</strong> and scored <em>worse</em> (66.7%). Reasoning tokens and full precision both cost a lot and bought nothing here. Every cost comes from real metered kWh via the open-s

Article URL: https://semgrep.dev/blog/2026/introducing-semgrep-guardian-real-time-security-for-ai-written-code/ Comments URL: https://news.ycombinator.com/item?id=48648871 Points: 1 # Comments: 1

<p><em>Originally posted on <a href="https://devblogs.microsoft.com/cosmosdb/announcing-the-public-preview-of-integrated-embeddings-in-azure-cosmos-db-build-ai-apps-with-embeddings-that-stay-in-sync/" rel="noopener noreferrer">https://devblogs.microsoft.com/cosmosdb/announcing-the-public-preview-of-integrated-embeddings-in-azure-cosmos-db-build-ai-apps-with-embeddings-that-stay-in-sync/</a></em></p> <p>AI applications built on Azure Cosmos DB depend on embeddings for grounded results. Keeping th

Article URL: https://github.com/alehlopeh/hallu Comments URL: https://news.ycombinator.com/item?id=48648823 Points: 1 # Comments: 0

Article URL: https://github.com/anasfik/flutter_copilot Comments URL: https://news.ycombinator.com/item?id=48648808 Points: 1 # Comments: 1

Learn about the concept of loops to power your coding agents. The post How to Create Powerful Loops in Claude Code appeared first on Towards Data Science .

<p> Give all your AI agents one shared context </p> <p> <a href="https://www.producthunt.com/products/n71?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed">Discussion</a> | <a href="https://www.producthunt.com/r/p/1179347?app_id=339">Link</a> </p>

The company had started just two months ago tracking workers’ computer usage for AI training data.

Article URL: https://master.dev/blog/ai-generated-ui-is-inaccessible-by-default/ Comments URL: https://news.ycombinator.com/item?id=48648479 Points: 1 # Comments: 0

Article URL: https://github.com/DeepBlueDynamics/charon Comments URL: https://news.ycombinator.com/item?id=48648435 Points: 1 # Comments: 1

Article URL: https://github.com/hoppscotch/hoppscotch/security/advisories/GHSA-j542-4rch-8hwf Comments URL: https://news.ycombinator.com/item?id=48648267 Points: 1 # Comments: 0

Article URL: https://www.youtube.com/watch?v=N9lQaXt1eqQ Comments URL: https://news.ycombinator.com/item?id=48648156 Points: 1 # Comments: 0

Article URL: https://github.com/itsthelore/wayfinder-router Comments URL: https://news.ycombinator.com/item?id=48648150 Points: 1 # Comments: 0

Hi, over the past 6+ months I have built an ecosystem (libraries, CLI, Desktop) of tools to build or use AI agents. Depending on how close or not close you are with coding, you may lean towards the libraries (compilr-dev/agents and compilr-dev/agents-coding) to build your own AI Agents; you may use the CLI to manage your project and run your agents in the terminal when working, for example in VS Code or you may use the Desktop app (it shares the same core of the CLI but has for sure a richer UI)

CreateMe said the partnership aims to demonstrate how apparel can be produced faster, more locally, and with greater supply chain resilience. The post CreateMe partners with Avalo and Laguna Fabrics to bring resilience to apparel supply chains appeared first on The Robot Report .

Comments

Recent hype about loop engineering made me realize that im missing one thing, and that is an agent that should be doing my job at the end of my loop: real user test i developing a huge b2b ecommerce project codebase. and for every new feature or fixes my job is always to test it on the browser. yuck. so i have made some research and there is TestSprite, but its not what i really want, i cant even self host it. so i built OpenUser. now i just engineer the loop with simple prompt: complete {featur

SpaceX and other tech stocks continued to drop on Tuesday. Are investors starting to rethink their AI investments?

This article breaks down each essential math discipline, explains its role in data science, and maps out an efficient learning path you can start today.

Hello and welcome to Regulator, the newsletter for Verge subscribers chronicling the misadventures of their favorite tech overlords and Washington swamp creatures. ("Favorite" is, of course, subjective.) Not a subscriber yet? Sign up here, especially if you want the hot scoop on quality Amazon Prime Day deals recommended by the wonderful humans of The Verge's […]

GPT-5 Pro helped solve a 3-year-old immunology mystery, offering insights into T cell behavior. The breakthrough could support cancer and autoimmune research.

Anthropic has introduced @Claude for Slack.

Anthropic’s new Claude Tag brings an always-on AI teammate to Slack. But beyond productivity, the feature is a strategic play to capture organizational context, institutional knowledge, and enterprise workflows.

Anthropic on Tuesday launched Claude Tag , a new product that embeds its most advanced AI model directly inside Slack as a persistent, shared teammate that anyone on a team can delegate work to by simply typing @Claude. The product, available today in beta for Claude Enterprise and Team customers, replaces Anthropic's existing Claude in Slack app and represents the company's most aggressive move yet to colonize the enterprise collaboration layer — the place where decisions get made, work gets as

Article URL: https://github.com/almakit/alma Comments URL: https://news.ycombinator.com/item?id=48647761 Points: 4 # Comments: 0

This post shows you how to build a conversational protein research assistant that combines three capabilities: Natural language query parsing to extract structured search parameters, vector similarity search over protein embeddings using a specialized language model and ai-generated scientific summaries of search results.