Show HN: Lelu – authorization engine that catches manipulated AI agents
Article URL: https://github.com/lelu-auth/lelu Comments URL: https://news.ycombinator.com/item?id=48608657 Points: 2 # Comments: 0
Article URL: https://github.com/lelu-auth/lelu Comments URL: https://news.ycombinator.com/item?id=48608657 Points: 2 # Comments: 0
Our weekly science news roundup. ScienceAlert stories are written, fact-checked, and edited by humans, never generated by AI. Don't miss a story, subscribe here.

I observe that the advent of recorded music resembles a Bose-Einstein Condensate. Before recorded music, live performance operated like a fluid, high entropy gas phase. Every performance was a localized, unrepeatable wave function–fundamentally moving and unique each performance. The advent of recording technology acted as a quantum cooling mechanism. It trapped the sound waves into a fixed state, collapsing in infinite fluid variations into a single, frozen, identical format of data. When milli
Article URL: https://github.com/Naveja00/OverReach Comments URL: https://news.ycombinator.com/item?id=48608558 Points: 1 # Comments: 0
<p><em>This is a submission for the <a href="https://dev.to/challenges/june-game-jam-2026-06-03">June Solstice Game Jam</a>.</em></p> <p>I built <strong>Among Liars</strong>, a realtime multiplayer elimination where six humans join a room, but the game secretly adds a seventh player: a <strong>Gemini-powered AI</strong> hiding inside the Spy side.</p> <p>There are two teams:</p> <ul> <li> <strong>Detectives</strong> are trying to expose the hidden AI.</li> <li> <strong>Spies</strong> are trying
Article URL: https://www.servethehome.com/building-a-dense-agentic-ai-cpu-rack-amd-dell-today/ Comments URL: https://news.ycombinator.com/item?id=48608400 Points: 1 # Comments: 1
Article URL: https://geoaxis.ai Comments URL: https://news.ycombinator.com/item?id=48608389 Points: 2 # Comments: 0
<p>GitOps is a way of operating infrastructure and applications where Git is the single source of truth for what should be running, and an automated agent continuously reconciles the actual state of the system to match it. The term was coined by Weaveworks in 2017, but the underlying idea — declarative desired state, version control, automated reconciliation — predates it. What GitOps added was a specific set of practices and a set of tools that made the pattern practical for Kubernetes.</p> <p>
<p>I ran into a memory problem while building a reflective AI product.</p> <p>The easy version of AI memory is tempting:<br> </p> <div class="highlight js-code-highlight"> <pre class="highlight plaintext"><code>Save useful facts about the user. Put them back into the next prompt. Call it memory. </code></pre> </div> <p>That can work for low-risk personalization.</p> <p>It is probably fine if an assistant remembers that a repo uses pnpm, or that a user prefers short answers, or that a team calls
<p>By 2026, the question isn't whether to use an AI coding assistant - it's which one actually fits how you work. The gap between a well-chosen tool and a poorly matched one shows up directly in your output: one makes you faster at real tasks, the other distracts you with confident-sounding hallucinations. This guide cuts through the marketing noise by comparing the leading tools across real TypeScript, Python, Go, and Rust codebases.</p> <h2> Why AI Coding Assistants Took Off in 2026 </h2> <p>T
<p>Exclusive: While recruits will increase headcount for now, broader adoption of AI could lead to jobs cuts in future</p><p>Lloyds Banking Group has launched an AI recruitment drive for 300 tech experts, weeks before its chief executive, Charlie Nunn, unveils a strategic plan for the 261-year old lender.</p><p>The bank said it intended the recruits to work on its use and development of agentic AI by September, referring to autonomous artificial intelligence models that can plan and execute task

<p>We asked people in the US about their customer service battles and hundreds responded on the financial and emotional costs</p><p>Guardian readers from across the US wrote in to tell us about their battles with big companies, and the time, expense and emotional toll exacted by businesses they say are prioritizing the bottom line over all else.</p><p>The top takeaway: people really, really don’t like AI customer service.</p> <a href="https://www.theguardian.com/us-news/2026/jun/20/guardian-read

Article URL: https://www.bbc.co.uk/news/articles/c0eyq7rnn22o Comments URL: https://news.ycombinator.com/item?id=48608173 Points: 4 # Comments: 3
Article URL: https://www.wsj.com/opinion/can-we-the-people-survive-the-ai-revolution-f9154ba1 Comments URL: https://news.ycombinator.com/item?id=48608119 Points: 2 # Comments: 1
Article URL: https://blog.apnic.net/2026/06/19/rethinking-vulnerability-management-in-the-age-of-ai-and-ci-cd/ Comments URL: https://news.ycombinator.com/item?id=48608085 Points: 2 # Comments: 1
The new Siri AI is conversational, omnipresent, and actually helpful.

<p>Does a thought-experiment about US ascendancy in the technology say as much about AI jitters as it does about the reality?</p><p>It’s 2031 and the US and China are about to tear Europe into pieces.</p><p>The US ploughed vast sums into datacentres and the EU did not. China built robots and Europe did not. American companies “restructured” their workflows around AI and fired people, while EU workers went on long lunch breaks and handed over administrative tasks to the AI model Claude.</p> <a hr

Seven AI agents work together like a newsroom. The "Data Journalist Agent" from Oxford and Stanford turns a CSV file into a finished interactive article with graphics, web research, and verifiable source links for 93 percent of all statements. In a reader study, 74 percent preferred the agent's output over the human original. But against elaborately crafted long-form reports, the agent managed a tie. The article Data2Story turns a CSV file into a verified interactive news article using seven AI

Article URL: https://blogr.ai Comments URL: https://news.ycombinator.com/item?id=48607782 Points: 2 # Comments: 1
Article URL: https://rorlikowski.github.io/stepyard/ Comments URL: https://news.ycombinator.com/item?id=48607721 Points: 2 # Comments: 0
Agent-queryable WhatsApp history from an iOS backup — a single Go binary. Runs on Mac and Windows. Comments URL: https://news.ycombinator.com/item?id=48607710 Points: 1 # Comments: 0
I build this self-hostable, PWA-ready fitness tracking app. This is less about the app itself, there are probably dozens of them out there. This is more about me being fascinated about the idea of "personal apps": AI enables us to just build something that perfectly fits our indivual requirements - like a tailor-made suit. None of the apps I know matched my requirements, some of them come with a subscription, some required an account. So I build F1L0. It would even work as a static HTML page wit
Article URL: https://mwi.westpoint.edu/data-center-warfare-defending-the-key-terrain-of-ai-infrastructure/ Comments URL: https://news.ycombinator.com/item?id=48607647 Points: 1 # Comments: 0
We build an end-to-end forecasting workflow with TimeCopilot on a panel of real airline passenger data and a synthetic seasonal series with injected anomalies. We evaluate statistical, foundation, and optional GPU-based models using rolling cross-validation and multiple error metrics. We generate probabilistic forecasts with prediction intervals, visualize future trends, and flag unusual observations. We then explore TimeCopilot's optional LLM agent, which selects a model and explains its predic

<p>Literary magazine will no longer engage in ‘external publishing partnerships’ after Commonwealth prize furore</p><p>The prominent literary magazine Granta will no longer publish the winning entries of the annual Commonwealth short story prize after one of this year’s winners drew widespread accusations of AI use.</p><p>The magazine said it would no longer be involved in “external publishing partnerships” in which it had no editorial control.</p> <a href="https://www.theguardian.com/books/2026

Get ChatPlayground AI's Unlimited Lifetime Plan for just $54.97 — the lowest price ever — and compare ChatGPT, Claude, Gemini, DeepSeek, and more in one platform.

The 1min.AI Advanced Business Plan is down to $69.97, giving users lifetime access to multiple AI models and creative tools in one centralized workspace.

OpenAI is upgrading ChatGPT's scheduling feature. A new "Scheduled" page in the sidebar puts all active tasks in one place, letting users view, pause, edit, or delete them. Research tasks search the web and connected apps, sending alerts only when something actually changes. The previous "Pulse" feature is being retired. The article ChatGPT keeps creeping toward becoming your AI personal assistant with new scheduled task controls appeared first on The Decoder .
Article URL: https://blog.unitedheroes.net/5752 Comments URL: https://news.ycombinator.com/item?id=48607442 Points: 2 # Comments: 0
Article URL: https://www.cio.com/article/4187280/google-microsoft-offer-specs-to-help-you-prove-your-ai-is-behaving-nicely.html Comments URL: https://news.ycombinator.com/item?id=48607346 Points: 3 # Comments: 0
Article URL: https://github.com/P156HAM/logslim Comments URL: https://news.ycombinator.com/item?id=48607333 Points: 1 # Comments: 0
Article URL: https://placementist.com/insights/fear-farming-the-stanford-ai-hiring-study-debunk Comments URL: https://news.ycombinator.com/item?id=48607303 Points: 3 # Comments: 0
In the first quarter of 2026, OpenAI pulled in $5.7 billion in revenue and burned through about $3.7 billion, both figures tripled year over year. Stock-based compensation alone ate up over $2.3 billion. With $73 billion in reserves, OpenAI doesn't need fresh capital right now, but a price war with Anthropic could change that fast. The article OpenAI tripled revenue to $5.7 billion in Q1 but burned through $3.7 billion to get there appeared first on The Decoder .

But there's a lot we don't know. ScienceAlert stories are written, fact-checked, and edited by humans, never generated by AI. Don't miss a story, subscribe here.

Article URL: https://zenodo.org/records/19560684 Comments URL: https://news.ycombinator.com/item?id=48607245 Points: 2 # Comments: 0
https://reachpad.dev/ I noticed that my GitHub was getting cluttered with .md plans, PLAN.md, SKILL.md, research/notes, artifact html files etc... I've also noticed that these documents pollute the context of the agent. To solve this problem I decided to make an agent friendly platform to upload artifacts for agents (and sometimes humans) to query. It's basically a pastebin for agents but with one thing I cared about: every edit is saved as its own version, and you can check the whole history yo
Article URL: https://getstayup.app Comments URL: https://news.ycombinator.com/item?id=48607172 Points: 1 # Comments: 0
We are building a VM that helps you simulate realistic production conditions, model latencies, different interleaving, user requests, and find bugs. Every non-deterministic property is turned into a knob you or a coding agent can control. We have helped teams perfectly reproduce support incidents and found bugs in some of the world's most well tested software (including a database) I you would like to try it out, I am giving away $100 in free credits to first 100 signups. Comments URL: https://n
Many projects and addons exists to reduce the verbosity on standard bash / git /npm etc. Commands that agents pass regularly as tools to LLMs. (eg. rtk, headroom, lean-ctx). Tool output compression does yield good token savings. Though it can lead to increased turns - effectively nullifying the per turn token savings. This is the current topology. Are we going to see a class of products and libraries that structure o/p to what models want to see? Comments URL: https://news.ycombinator.com/item?i
Article URL: https://gcontext.ai Comments URL: https://news.ycombinator.com/item?id=48606940 Points: 1 # Comments: 0