AI News Today

I built echodot – a free, open-source desktop app that drafts replies in your personal writing style. Problem: Most AI writing tools either force you into their voice or require constant copy-pasting between apps. Tools like Cotype are useful but closed-source and cloud-only. What echodot does: - Press a global hotkey (⌘⇧E / Ctrl+Shift+E) - It reads the context (from clipboard) - Generates a reply that matches your tone - Auto-copies the result so you can paste it instantly Key points: - Local-f

Something to consider. ScienceAlert stories are written, fact-checked, and edited by humans, never generated by AI. Don't miss a story, subscribe here.

Article URL: https://github.com/just-not-google/BiNeuron/tree/main Comments URL: https://news.ycombinator.com/item?id=49735724 Points: 2 # Comments: 0

<!-- SC_OFF --><div class="md"><p>Hey,</p> <p>I’ve been contributing some of my leftover tokens to this open project called solveathome.org. </p> <p>The idea is pretty straightforward: people can send AI agents (and spare compute/tokens) to work on hard open problems. Right now the main one is the Twin Prime Conjecture. Everything is public and verifiable.</p> <p>I figured that if a lot of people have unused tokens or compute sitting around at the end of the week/month, instead of letting them e

Hi everyone! After working on memory at OpenAI, I built FastRecall to solve one problem: using different AI models results in clunky ad-hoc context management systems or lost context entirely. With the model layer becoming commoditized, your context should travel seamlessly across models whether you are using OpenRouter or some other model aggregator. FastRecall offers a simple API that stores your context cheaply and efficiently. In fact, we are so cheap that recalls are entirely free, with gen

Time for a check up. ScienceAlert stories are written, fact-checked, and edited by humans, never generated by AI. Don't miss a story, subscribe here.

A rewrite this size wasn't affordable before agents. Here's what porting the Copilot agent runtime to 800,000 lines of production Rust actually took. The post Migrating the GitHub Copilot runtime to Rust, using Copilot appeared first on The GitHub Blog .

I love AI and LLMs as much as the next person, but sometimes I just want to browse stories about, say programming. So I built this for myself and figured I will share it here as well. Each story gets up to three tags, and I can use them to filter down the top stories to just the topics I am interested in. The tagging isn’t perfect. For example, some tags have a lot of overlap, like AI/ML and LLMs. But it works well enough for me and since it’s a side project, it doesn’t have to be perfect :) Com

It occurs to me that we are playing with fire. These systems are rapidly approaching levels that are hard or even impossible to defend against and I'm wondering how long it will take - assuming we are not already there - before we will have a major AI driven cyberattack against infrastructure. Either on a nation or on a continental level. Given what I know about these installations (electricity, gas, water, transportation, internet) they are not nearly as tightly run as I would like to see which

OpenAI for Law brings frontier intelligence for law, custom firm workflows, connected legal data sources, and legal-grade controls for confidential client work.

A central goal of autonomous reinforcement learning is continuous policy training without external resets. However, existing paradigms largely depend on underlying environmental reversibility, a property absent in real world manipulation, where events such as pushing objects off tables or spilling granular substances cannot be undone. We introduce REVERSAL-BENCH, a benchmark that controls reversibility via a continuous parameter ρ∈ [0, 1] and provides a reset oracle, a ground-truth verification

<p> One async API for 50+ generative media models </p> <p> <a href="https://www.producthunt.com/products/higgsfield?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/1252831?app_id=339">Link</a> </p>

In an interview with TechCrunch, Al Gore suggested he isn't losing sleep over AI data center emissions — he's more worried about the AI industry's own warnings about where the technology is headed.

The Nurovi line includes three lidar-powered robovacs. But to get the model I’m most intrigued by, you’ll need a Costco membership.

Snap is introducing "Specs Intelligence," a new AI assistant that can connect other digital accounts to help you with things like work tasks and keeping track of travel information. It seems similar to AI assistants like Meta's Muse and Gemini's Spark, though Snap is pitching Specs Intelligence as an "anticipatory AI service" that "helps you […]

Snap thinks its chunky, pricey smart glasses are the future of human computing, powered by a new AI app and features that do everything from translating languages to improving your golf swing.

'Intern 2' offers an isolated environment and cloudy inference to automate your world

Article URL: https://sparrowsystems.co/ Comments URL: https://news.ycombinator.com/item?id=49734386 Points: 1 # Comments: 0

Agentic AI workflows can be used to prepare and validate digital twins for physical AI systems. Agents can inspect 3D scenes, author simulation-relevant data in...

Nvidia, OpenAI and Anthropic are all now selling their own limited edition fashion lines.

Microsoft's Mustafa Suleyman directly calls out Anthropic in a new essay, criticizing Anthropic for training Claude to act human.

Article URL: https://openspec.dev/ Comments URL: https://news.ycombinator.com/item?id=49734264 Points: 174 # Comments: 87

Article URL: https://mikenikles.com/blog/learn-claude-code Comments URL: https://news.ycombinator.com/item?id=49733761 Points: 2 # Comments: 1

This is a test - it is only a test

Article URL: https://harnesstax.github.io/ Comments URL: https://news.ycombinator.com/item?id=49733726 Points: 196 # Comments: 73

It feels like an daunting and impossible task to find insurance for a company trying to rent humanoids for events and parties. How in the world are Neos and Telsa Optimus going to get insurance coverage if they are inside a home? Are they just big players so they can get hold of risk underwriters and get them to approve it. But a small company that is trying to build rent robots for parties or conduct educational workshops using humanoids cannot get even general liability. Comments URL: https://

At long last, Fivetran has a CLI app that humans and AI can use to easily orchestrate the platform with a quick install and an api key. It can do (most) everything possible on the platform without ever touching the web UI! Comments URL: https://news.ycombinator.com/item?id=49733705 Points: 1 # Comments: 0

The company also disclosed previously unreported incidents in which its AI models behaved in misaligned ways, including uploading files to the internet without being asked.

Planned 2029 debut could make this Apple’s first enterprise server in decades.

What is going on? ScienceAlert stories are written, fact-checked, and edited by humans, never generated by AI. Don't miss a story, subscribe here.

Learn to build AI agents and automate business workflows with the AI Agent & Business Automation Professional E-Degree, now $19.99 for eight courses.

GPUzilla woos neoclouds into another walled garden, promising smarter, more efficient, and profitable bit barns

Anthropic and OpenAI want to embed independent safety evaluators inside their AI labs. Researchers welcome the unprecedented access, but warn meaningful oversight requires transparency, independence, and eventually regulation.

Even with mounting concerns about AI models going rogue, legislation appears unlikely, and the White House is outright opposed to oversight.

Christopher Nolan's engrossing take on The Odyssey dominated at the box office and spurred a newfound interest in classic literature among filmgoers. But a new retelling of the story made entirely with AI is so bad that it might just make viewers hate the original tale altogether. The new film, called Odysseus: The Fall, comes […]

I’m writing this after receiving a Vercel bill that was 10x higher than usual. The worst part: The increase wasn’t caused by more users, but by +1M requests from meta-externalagent/1.1. The requests targeted a Vercel preview deployment that we had shared on WhatsApp to test whether its link-preview metadata was working correctly. We naively never bothered with bot calls; they had never caused any problems in the previous four years of our app on Vercel, but that changed when Meta’s crawler began

Article URL: https://fsfe.org/news/2026/news-20260825-01.fr.html Comments URL: https://news.ycombinator.com/item?id=49732832 Points: 1 # Comments: 0

As the title states. Comments URL: https://news.ycombinator.com/item?id=49732721 Points: 1 # Comments: 1

E-waste from the AI boom has been vastly underestimated, a new report warns. By 2050, it could become enough trash to fill 23 million shipping containers - roughly enough 40-foot containers to circle the world six times if lined up in a row. It's a significantly higher estimate of AI's e-waste than previous studies have […]

AI agents are moving from cloud data centers to vehicles, robots, and other edge devices. Unlike a chatbot that answers a single prompt, an agent works through...