AI News Today

For starters, don't just hand over the keys to AI agents. Any endeavor needs to remain a human-instigated, human-led endeavor.

Hi HN, I built Language1 ( https://language1.app ), a word game where you play "reverse Taboo" against an LLM. How it works: You are given a target word (e.g., "Apple") and a list of forbidden "taboo" words (e.g., "fruit", "red", "tree"). Your goal is to write a prompt that guides the LLM to output the exact target word, without using any of the forbidden words. The Benchmark Goal: I am developing this project with the plan of using the gameplay data to build a benchmark dataset. The goal is to

Article URL: https://github.com/Ntooxx/Sentinel Comments URL: https://news.ycombinator.com/item?id=48586242 Points: 1 # Comments: 0

<h2> What does 80% AI authored code mean for solo devs? </h2> <p>In June 2026, Anthropic stated that about 80% of its new production code is authored by Claude. When a major AI vendor hits that volume, the shift is undeniable. For a solo developer or a one-person holding company, this changes the math entirely. The bottleneck is no longer typing characters. The bottleneck is review and ownership.</p> <p>When you run a solo shop, you do not have a team to absorb the review burden. If your agents

<p>Salesforce has shipped around 20,000 Agentforce deployments. ByteByteGo published a writeup of what they learned, sourced to John Kucera, the CPO of Agentforce. I run a one-person agent fleet, which is about as far from Salesforce scale as you can get. The lessons still translate. Better than I expected, actually.</p> <p>Short version: 90% of agent work happens after launch, not before. The failures cluster into three patterns. Putting deterministic logic inside an LLM loop, prompting harder

<p>I built a scanner that fires prompt-injection probes at a self-hosted AI agent and checks whether it leaks (a) real secret-shaped strings (API keys) or (b) the content of its own system prompt. Then I ran the same agent across 5 model backends. The leak rate ranged from 0% to 90% depending only on the model.<br> Here's what I found and how it works.<br> Why this matters now<br> Prompt injection is #1 on the OWASP 2025 LLM Top 10. It's not theoretical anymore:</p> <p>EchoLeak (CVE-2025-32711,

Two new studies published in Nature show that specialized AI systems diagnose diseases and make treatment decisions as well as physicians in simulated patient cases, sometimes even better. Both systems run on base models that are already outdated. The article AI systems rival doctors in new Nature studies, but one result suggests the tech won't age well appeared first on The Decoder .

<p>The npm account <code>ai</code> publishes seven packages. Combined, they install 964 million times per week:</p> <div class="table-wrapper-paragraph"><table> <thead> <tr> <th>Package</th> <th>Weekly downloads</th> <th>Publishers</th> <th>Risk</th> </tr> </thead> <tbody> <tr> <td>postcss</td> <td>245,612,332</td> <td>1</td> <td>CRITICAL</td> </tr> <tr> <td>nanoid</td> <td>206,588,788</td> <td>1</td> <td>CRITICAL</td> </tr> <tr> <td>caniuse-lite</td> <td>173,435,668</td> <td>1</td> <td>CRITICAL

The brain runs on 20W. It doesn't do matrix math. It routes sparse signals through a graph where each node makes a local weighted decision. Intelligence emerges from the path, not from a global forward pass. GPUs were adopted for AI because matrix multiplication is embarrassingly parallel. But is matrix multiplication actually required for intelligence — or is it just the only method we've tried at scale? Comments URL: https://news.ycombinator.com/item?id=48586030 Points: 1 # Comments: 0

<p>Marc Isaacs’ film Synthetic Sincerity may look like a documentary, but its fictional premise – a lab that scrapes movies to harvest human emotions – shines a hard light on just how far AI can go</p><p>In Marc Isaacs’ latest film, the subversive documentary maker reveals that an AI research laboratory recently licensed his entire body of work. That’s a quarter-century of droll, deadpan studies of ordinary life in Britain – from the poetic <a href="https://www.youtube.com/watch?v=FJNAvyLCTik">L

Adobe has announced a major expansion of its "creative agent" across its flagship Creative Cloud suite and upgraded Firefly AI studio. Available in public beta starting today across Premiere Pro, Photoshop, Illustrator, InDesign, and Frame.io, the agent is designed to serve everyone from individual creators to enterprise marketing teams. Unlike first-generation generative AI tools that simply output flat media from a chat interface, Adobe’s embedded assistant acts as an orchestration layer. It i

Article URL: https://hermes-agent.nousresearch.com/docs/guides/migrate-from-openclaw Comments URL: https://news.ycombinator.com/item?id=48586005 Points: 70 # Comments: 58

<pre style='white-space:pre-wrap;width:81ex'>refactor(copilot): remove unused usage total helper</pre>

<p>Most of the debt that actually takes a codebase down was never written down anywhere. That, not the raw volume of AI-generated code, is what the current panic keeps aiming slightly past.</p> <p>We borrowed "debt" from finance on purpose. A loan you take deliberately comes with a rate and a due date; you can plan around it, refinance it, decide the interest is worth it. Recorded debt is a tool. What sinks people is the liability they didn't know they'd signed for, and that was true of software

<p><strong>The problem</strong></p> <p>Every time I found an accessibility issue in a PR review, the same thing happened: tab out to the WCAG docs, figure out the right fix, come back to the code, apply it, re-verify. For every single issue.</p> <p>It wasn't hard work. It was repetitive work. And it was eating hours every sprint.</p> <p><strong>What we built</strong></p> <p>A11yResolver is a VS Code extension that brings that entire loop inside the editor.</p> <p>It runs as an AI agent. It flags

A Match Group survey found that companion AI use is a red flag for certain daters.

<blockquote> <p>Claude Code stores 43MB of your conversations. Cursor saves your passwords in plaintext JSONL. Cline is the only one that puts your API key in the OS keychain — but it still archives everything you say.</p> <p>Not one of them tells you this is happening.</p> </blockquote> <h2> Why I Did This </h2> <p>Two weeks ago, I discovered Claude Code had silently archived 43MB of my complete conversation history on my machine. I wrote about it. The comments asked: "What about Cursor? What a

Both Google and Amazon have generative AI-powered smart speakers for under $100, so which one stands out?

i've been dogfooding some version of this idea for a few weeks now and i'd thought i'd share my re-implementation from this weekend. basically it's a webserver that controls and unlimited amount of codex (preferred) or claude agents. it allows me to context switch more easily, as well as kick off tasks and instruct agents while i'm away from my desk. the storage is sqlite and persistent, regardless of agent status. for codex you can upload images and queue messages just like the CLI. i'm on mac

Our genetic heritage is not a blueprint or an algorithm, as many biologists have imagined, but something else entirely. The post Why the Human Genome’s Tangled Physicality May Confound AI first appeared on Quanta Magazine

Prove You Can Work with the Technology Behind Modern AI Linux Foundation Education and PyTorch Foundation have launched the PyTorch Certified Associate (PTCA), a new certification designed for early-stage practitioners...

Hybrid systems could bring efficiency gains at the edge, but conventional infrastructure isn't going anywhere fast

If you’ve been wanting to buy a robot vacuum but have been put off by how much it can cost to get a good one, now is not a bad time to start looking. Prime Day is next week, but many retailers are offering early deals on several models we’ve tested from brands like Roborock, […]

On today’s episode of Decoder, my guest is Hayden Field, senior AI reporter for The Verge. Often when Hayden comes on the show, it’s because something has gone wrong in the world of AI. Last weekend, that something was a pretty intense mix of Anthropic, the Trump administration, and Anthropic’s new AI model, Fable 5. […]

Article URL: https://github.com/tigerless-labs/cost-xray Comments URL: https://news.ycombinator.com/item?id=48585376 Points: 1 # Comments: 0

I'm building a project for my company and they want to use a local model (100b max size). I think i tried all the models, the problem is that do not give a good answer even tho i gave them everything they need in terms of context to tackle any question. Comments URL: https://news.ycombinator.com/item?id=48585308 Points: 2 # Comments: 1

The Eufy E25 robot vacuum and mop combo is on sale for $629.99 at Amazon, a 37% discount off its $999.99 list price ahead of Prime Day.

A lot of things are always happening.

According to Yann LeCun, AI labs like OpenAI and Anthropic are heading for a "big bubble explosion": Their operations are effectively subsidized by investors, and operating costs aren't dropping fast enough, he says. LeCun's criticism isn't entirely selfless: his own startup, AMI Labs, raised $1 billion for an alternative approach to AI. The article Yann LeCun warns AI labs like OpenAI and Anthropic face a "big bubble explosion" appeared first on The Decoder .

<p> Build unreal games with AI agents </p> <p> <a href="https://www.producthunt.com/products/unreal-engine?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/1175263?app_id=339">Link</a> </p>

Since Fable / Mythos was so capable, I've been wondering if Anthropic could build a JIT layer for python. Would it significantly reduce the cost of training and running LLMs. Comments URL: https://news.ycombinator.com/item?id=48584991 Points: 2 # Comments: 1

Enterprise Document Intelligence [Vol.1 #6c] - The decisions the parser makes on top of the user string, using the document’s profile: dispatch, activations, full schema, three approaches to deciding what fires, the audit _meta block, and a broker-corpus walkthrough The post Dispatching the Parsed RAG Question: Chunk Strategy, Model Tier, Activations, Audit appeared first on Towards Data Science .

Article URL: https://display.dev/ Comments URL: https://news.ycombinator.com/item?id=48584961 Points: 3 # Comments: 1

Rumors about Midjourney hardware have circulated for years, but nobody saw this coming. The AI image startup is building a full-body ultrasound scanner and opening its own spa in San Francisco to house it. The article Midjourney, known for AI image generation, unveils a full-body ultrasound scanner and its own spa appeared first on The Decoder .

Comments

At Cajal (YC W26) we’re excited to share Talos ( https://github.com/cajal-technologies/talos ), an open source framework for formal verification of WebAssembly modules in Lean. AI is now writing tons of the code that gets pushed to production. As code generation gets cheaper, verification becomes the bottleneck. We believe in a future where every piece of software comes with a mathematical proof that it does what its author intended - in doing so, eliminating many classes of exploits. Talos is p

The D585 Pro is powered by a proprietary Gen 5 system-on-chip and delivers more than 2x better depth quality than the previous generation of RealSense cameras. The post RealSense unveils AI-native D585 Pro depth camera for robots appeared first on The Robot Report .

Article URL: https://github.com/flexorch/flexorch-audit Comments URL: https://news.ycombinator.com/item?id=48584670 Points: 2 # Comments: 0

The digital era gave the advertising and marketing industry speed; the AI era is giving it autonomous operations. For companies building next-generation technologies for advertising and marketing, the question is no longer whether to adopt AI but whether their infrastructure can support it at the speed and scale the industry demands. At Cannes Lions, running […]

By mimicking how the brain operates, neuromorphic computing can use dramatically less energy than conventional electronic AI chips. However, even the most sophisticated neuromorphic devices today are still quite simple, using only a small fraction of the number of connections found in human neurons. Now a new study suggests that using sound waves, neuromorphic devices can better mimic biological neurons and operate faster and with greater energy efficiecy than their electronic counterparts. “Thi