AI News Today

A single click on a trusted Microsoft link could have let an attacker pull emails, calendar details, and indexed files out of Microsoft 365 Copilot Enterprise Search. Researchers at Varonis Threat Labs chained three bugs into a one-click exfiltration path they call SearchLeak. Because the link pointed to a real microsoft.com domain, traditional anti-phishing and URL filtering tools were

<pre style='white-space:pre-wrap;width:81ex'>feat(telegram): send rich messages as rich html (#93286) * feat(telegram): render rich messages through rich html * docs(telegram): teach agents rich formatting * fix(telegram): bound rich draft payloads (#93286) * fix(telegram): narrow rich draft payload type (#93286) * fix(telegram): preserve rich table cell formatting (#93286) * fix(telegram): honor rich table mode config (#93286) * fix(telegram): default rich markdown tables (#93286) * fix(telegra

PARTNER CONTENT: Onix's Wingspan platform promises to move enterprises from pilot purgatory to governed, enterprise-wide AI deployment in weeks, not years

A detailed look at MCP that turned my scattered tool definitions into a stable, discoverable server The post The Protocol That Cleaned Up Our Agent Architecture appeared first on Towards Data Science .

I have to admit that newest models and harness tools have gotten really good and produce working code really fast with even mediocre prompts "wish programming". The problem is more on my side than the LLM side. I feel I am 1) losing control of my project, even my hobby ones, because LLM outputs code so fast that I am 2) too lazy to manually review all of them, even if, indeed, I had the overall architecture design choices and sometimes straight up pseudo-code in natural language. But NL is, by d

<p><strong><a href="https://www.axios.com/2026/06/15/anthropic-white-house-fable-mythos">"They screwed us": Personality clashes sent Anthropic's models offline</a></strong></p> Lots of "source familiar with the administration's thinking" and "source close to Anthropic" in this Axios piece, which is the best collection of behind-the-scenes gossip I've seen about the US government <a href="https://simonwillison.net/2026/Jun/13/us-government-directive-to-suspend-access/">export control Mythos/Fable

Article URL: https://barvhaim.github.io/visual-ai-metrics/ Comments URL: https://news.ycombinator.com/item?id=48542226 Points: 2 # Comments: 0

Generate graphics and audio for game development. Graphics include pixel-art sprites, icons, HUD kits, and parallax backgrounds. Audio includes voice lines, sound effects, and music — all from a text prompt. Assets export engine-ready for Unity, Godot, and GameMaker. Available via MCP server (works with Claude Code, Cursor, etc.) or REST API. Pay-as-you-go per art asset. Comments URL: https://news.ycombinator.com/item?id=48542207 Points: 1 # Comments: 0

Has anyone here fully swapped Claude/GPT for a local model as their main coding tool, not just for side experiments? If so, please share your setup and performance (e.g tok/s) Comments URL: https://news.ycombinator.com/item?id=48542100 Points: 2 # Comments: 0

Salesforce says it wants to use Fin's team and technology to improve Agentforce, its existing enterprise platform that businesses can use to build custom AI agents that automate tasks.

Article URL: https://tectonforgeai2.vercel.app/ Comments URL: https://news.ycombinator.com/item?id=48541878 Points: 1 # Comments: 0

Compared against MCPs, CLIs, and SKILL.md files Comments URL: https://news.ycombinator.com/item?id=48541874 Points: 1 # Comments: 0

Support bot maker claims its AI agents can resolve three-quarters of customer queries without human help

The best deals this week. Mashable's shoping team finds deals on AirPods, Pokémon cards, robot vacuums, and so much more.

The government is using export control law to force Anthropic to cut access to its most powerful models. The legal authority is plausible but the facts remain murky.

AI does have its places, and one of them could be in helping you manage your Linux systems, be they desktops or servers.

The Sunseeker S4 robot lawn mower is on sale at Amazon for $984, down from the list price of $1,799.99. That's a 45% discount.

<p> AI customer support agent, live on your website in 5 minutes </p> <p> <a href="https://www.producthunt.com/products/botme-io?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/1172438?app_id=339">Link</a> </p>

Article URL: https://github.com/brandoncarl/llm-jailbreaking/blob/main/On%20the%20Impossibility%20of%20Perfect%20Universal%20Guardians%20Against%20LLM%20Jailbreaks.pdf Comments URL: https://news.ycombinator.com/item?id=48541538 Points: 3 # Comments: 2

An "exciting hypothesis". ScienceAlert stories are written, fact-checked, and edited by humans, never generated by AI. Don't miss a story, subscribe here.

In this article, we’ll build time-series machine learning models in Python using sktime and explore its core data structures for forecasting workflows.

Today, I’m talking with Adam Bry, who is CEO of Skydio, the leading US maker of autonomous drones. Before we recorded this episode, I actually got to remotely operate one of Skydio’s drones in the Bay Area from Adam’s laptop in our podcast studio in New York and fly an indoor drone around our office. […]

Vercel Functions using the Node.js and Python runtimes now support execution durations up to 30 minutes for Pro and Enterprise teams, more than 2x the previous 800 second limit. Support for additional runtimes is coming soon. Use longer-running Functions for work that needs more time to finish, including: Long LLM reasoning and tool calls AI responses that stream for several minutes Document and media processing OCR and extraction Web scraping and browser automation Complex Workflow steps or Que

In this post, you'll build a competitive research agent that demonstrates this pattern end to end. This walkthrough targets developers building multi-step AI workflows who need isolated execution environments for their agents. In Part 2 of the notebook, you can deploy this same agent to Bedrock AgentCore Runtime using the AgentCore CLI, so it runs as a managed, session-isolated service.

Stuff broke again. Not in a movie way. An old tool was left exposed. An abandoned package was abused. A deprecated feature was still running in prod. This week is the same lesson in a new form: phishing kits are easier to rent, AI names are useful bait, old login paths still fail, and forgotten software keeps becoming someone else's entry point. Scroll through the full Monday Cybersecurity

Indian IT services company HCLTech is investing $150 million in the Bengaluru startup.

The real reason Anthropic disabled Fable 5 for its users and what it means for product designers Continue reading on UX Planet »

Comments

A single model hands you a single answer and no sense of how much it hinges on the dozens of choices buried inside it. The post I Built 11 Models to Predict the 2026 World Cup. They Crown Four Different Champions. appeared first on Towards Data Science .

<div class="hs-featured-image-wrapper"> <a href="https://www.marketingaiinstitute.com/blog/brand-reputation-ai-b2b-marketers-summit" title="" class="hs-featured-image-link"> <img src="https://www.marketingaiinstitute.com/hubfs/Lovett.png" alt="Your Brand Reputation Precedes You With AI, Whether You Like It or Not" class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"> </a> </div> <p style="font-weight: bold;"><span>Marketers, take note: New res

Volunteer AR scans from Pokémon Go players fed into Niantic's spatial AI models. That technology is now being combined with a US defense contractor's software for GPS-free navigation. The article Pokémon Go data helped train AI now linked to military drones appeared first on The Decoder .

WSL 3 makes staying on Windows easier, especially for developers building or running Linux-based AI, container, or dev workloads.

The more I use Claude Code to generate large swaths of systems, the more I feel like we are missing a lot of practices and tools. The first bit that really annoyed me was the lack of tracking prompts. There are lot's of little inputs that inform the what the system should do and it seems to be very ephemeral. I've since started tracking the prompts in the codebase. Another observation is that if you want to have good agentic output, you need to provide lot's of signal into the context window. It

Waze is ideal for quick reroutes and real-time alerts, while Google Maps offers deeper Gemini integration and more features. Here's my choice.

Comments

Brussels says access curbs prove Europe needs greater technological independence

Claude Code is a genuinely good agent builder. You describe what you want, it reasons through the problem, picks tools, and ships working code. For greenfield projects against well-documented libraries, the experience is close to magic. Where it gets harder is the same place every coding agent struggles: building on a specialized platform with its... The post The DataRobot platform as skills in Claude Code appeared first on DataRobot .

NewCore argues the next challenge in enterprise security will be managing AI agents, not people.

Cloudflare is deepening our investment in AI with the addition of team members from Ensemble AI, focusing on machine learning infrastructure and efficiency.

<p> Give your AI coding agent the web as a command line </p> <p> <a href="https://www.producthunt.com/products/agentbrowse?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/1172386?app_id=339">Link</a> </p>