AI News Today

The startup has reached a $120 million annualized revenue run rate and more than 200,000 paying customers.

The guy behind TCP/IP is working on a standard for identifying AI agents in the wild.

Enterprise Document Intelligence [Vol.1 #7quinquies] - Hallucination is usually garbage-in. Fix retrieval, and the model has nothing left to make up The post Most RAG Hallucinations Are Retrieval Failures: How the Retrieval Brick Decides What the Model Can Invent appeared first on Towards Data Science .

A short time ago in a galaxy not so far away (because it's ours). ScienceAlert stories are written, fact-checked, and edited by humans, never generated by AI. Don't miss a story, subscribe here.

Gidi Littwin’s new AI startup, Hemispheric, makes diagnostic brain scans for conditions like depression, PTSD, and Parkinson’s. He wants the technology to be as cheap and easy as for a blood test.

OpenAI outlines a “reverse federalism” approach to AI governance, where state laws help build a national framework for safe, democratic AI.

For years, routing traffic through cloud proxies was good enough. Then work moved to the browser, AI entered the workflow, and the inspection model stopped keeping up. Enterprise workflows now live across SaaS applications, browsers, and an expanding ecosystem of generative AI tools, unsanctioned browser extensions, and autonomous agents. Employees routinely paste intellectual property into

Article URL: https://okaneland.com/study/unit-economics-of-wrapping-an-llm/ Comments URL: https://news.ycombinator.com/item?id=48919292 Points: 2 # Comments: 0

I spent weeks reading about how Cognee, Graphiti, and Neo4j's `agent-memory` build their agent memory architectures. They converged on the same heavy knowledge-graph design: an ontology, LLM extraction pipelines, deduplication, the works. I really wanted to use them for my personal use case, but that looks like such a heavy setup that adds a lot of friction and silos. Plus, it feels like I just get my data trapped in their service, for not a ton of value. That's why my "long-term memory" still l

A single approved marketing tag can quietly load fourth-party code your security team has never seen, granting full access to your forms, customer data, and checkout pages. This on-demand webinar reveals how this Approval Gap forms, and gives your team the blueprint to close it before an auditor, regulator, or attacker finds it first. The Reality of the Approval Gap It's a pattern every

Article URL: https://www.lesswrong.com/posts/hcq4ZDoijSjy3Wrba/how-much-of-ml-research-is-about-ai-safety-what-is-it-about Comments URL: https://news.ycombinator.com/item?id=48919012 Points: 2 # Comments: 0

Article URL: https://dl.acm.org/doi/10.1145/3800645.3813051 Comments URL: https://news.ycombinator.com/item?id=48919004 Points: 4 # Comments: 0

<p>Steve Feldstein argues that the conventional wisdom about AI warfare has it backwards.</p>

<div class="hs-featured-image-wrapper"> <a href="https://blog.hubspot.com/marketing/profound-vs-peec" title="" class="hs-featured-image-link"> <img src="https://53.fs1.hubspotusercontent-na1.net/hubfs/53/profound-vs-peec-ai-for-aeo-1-20260710-7963660.webp" alt="profound vs peec ai for aeo" class="hs-featured-image" style="width:auto !important; max-width:50%; float:left; margin:0 15px 15px 0;"> </a> </div> <p>According to McKinsey, <a href="https://www.mckinsey.com/capabilities/growth-marketing-

But with great power comes a great price.

Co-founded by serial entrepreneur Shai Morag, Israeli identity management startup Oak is emerging out of stealth with $60 million in seed funding.

Open a repository in Cursor on Windows and, if a file named git.exe is sitting in the project root, Cursor runs it. No click, no approval dialog, no warning that anything in the folder is about to execute. Whatever that binary does, it does as you, with your source, your SSH keys and your cloud tokens. Cursor keeps re-running it for as long as the project stays open. No prompt

Article URL: https://www.politico.com/news/2026/07/15/inside-anthropics-state-by-state-plan-to-ratchet-up-ai-rules-00998415 Comments URL: https://news.ycombinator.com/item?id=48918906 Points: 1 # Comments: 0

Home to leading manufacturers, robotics pioneers and infrastructure builders, Japan is one of the world’s centers of AI — building across the full stack with NVIDIA technologies. NVIDIA and its partners in Japan are this week showcasing the AI ecosystem’s latest advancements. Check back here for updates. NVIDIA and SEGA Celebrate 30 Years of […]

Article URL: https://promptman.app/ Comments URL: https://news.ycombinator.com/item?id=48918897 Points: 1 # Comments: 0

openai mfs were like yall liked subagents? introducing: subagents 2!!! and then its just the same thing with fully removed transparency and all outputs encrypted that is so cool Comments URL: https://news.ycombinator.com/item?id=48918863 Points: 2 # Comments: 0

Article URL: https://github.com/vkataev/ai-cli Comments URL: https://news.ycombinator.com/item?id=48918840 Points: 1 # Comments: 1

<p> The memory layer for AI agents </p> <p> <a href="https://www.producthunt.com/products/kit-for-ai?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/1197148?app_id=339">Link</a> </p>

seaticket.ai Comments URL: https://news.ycombinator.com/item?id=48918690 Points: 2 # Comments: 1

Article URL: https://lowlatencyclub.ai/blog/posts/ai-phone-story-hotline-python Comments URL: https://news.ycombinator.com/item?id=48918668 Points: 1 # Comments: 0

<p>For months, artists and activists have been calling on the prime minister, Anthony Albanese, to act on the AI boom as datacentres pop up around the country. Today he made a major speech at the University of Sydney addressing copyright, the regulation of datacentres and the future of AI in Australia</p><ul><li><p><a href="https://www.theguardian.com/technology/2026/jul/15/office-of-ai-artificial-intelligence-copyright-australia-government">'Not up for grabs': Albanese establishes AI office and

<p>I started my career as a “Dessinateur Projeteur” (Design Drafter) in mechanical engineering. In that world, if your plan is off by 1mm, the part is scrap. There is no “hotfix” for a piece of steel once it’s machined.</p> <p>When I started building my own software tools using LLMs, I brought that same paranoia with me. And honestly, it’s the only thing saving my codebases right now.</p> <p>Lately, I’ve been fighting a specific enemy: Thinkslop.</p> <p>It’s that code that looks “clean,” passes

<h1> Background Agents: The Open-Source System That Lets AI Code While You Sleep </h1> <blockquote> <p>⚡ <strong>Quick Answer:</strong> Background agents are autonomous AI systems that work in isolated environments, executing multi-step coding tasks without human supervision. Think of them as "digital employees" that can fix bugs, write tests, refactor code, and even deploy changes while you sleep. With over <strong>382,000 GitHub stars</strong>, the open-source <code>openclaw</code> project is

<pre style='white-space:pre-wrap;width:81ex'>fix(doctor): ignore rotating Claude CLI OAuth expiry (#108205)</pre>

Article URL: https://roboticsregistry.net Comments URL: https://news.ycombinator.com/item?id=48918580 Points: 1 # Comments: 1

<!-- SC_OFF --><div class="md"><p>Building a sports prediction model ,I found consistent edge when backtesting against closing lines.<br/> At inference time tho, I predict 12-24 hours before the event where closing lines don't exist yet. I use the current line instead.</p> <p>My strongest feature is line movement (opening to closing implied probability). At prediction time this feature is incomplete as the market hasn't fully moved yet.</p> <p>This creates a paradox:</p> <p>Closing lines are con

<p>Solana processes 400ms finality and handles thousands of transactions per second. Smart contracts on this chain are fast, deterministic, and predictable — which is exactly why they're limited.</p> <p>Deterministic logic works when every edge case is known. But on-chain markets, MEV dynamics, and cross-protocol arbitrage are anything but predictable. The gap between Solana's execution speed and the intelligence needed to navigate real-time markets is where AI agents enter the picture.</p> <h2>

<p>GitHub just shipped the most useful security feature in the Copilot app since Code Scanning — and it is a slash command. No experimental-mode opt-in. No terminal incantation. No Copilot Business upgrade gate. Type the command in the Copilot desktop app, and an LLM scans your uncommitted diff for injection, XSS, weak cryptography, path traversal, and insecure data handling before a single character of vulnerable code reaches your repo. This is the first time GitHub Copilot's pre-commit vulnera

Article URL: https://www.pbs.org/newshour/nation/googles-ai-search-features-pose-unacceptable-risk-to-children-new-report-finds Comments URL: https://news.ycombinator.com/item?id=48918541 Points: 1 # Comments: 0

<p>A successful payment is not the same thing as an active product entitlement.</p> <p>That distinction explains a common class of subscription incidents: the bank shows a charge, the store shows a subscription, yet ChatGPT, Grok, Claude, or Gemini still shows <code>Free</code>, <code>Upgrade</code>, or an inactive plan.</p> <p>The failure is easier to diagnose when we stop treating “payment succeeded” as a single boolean and model the purchase as a state machine.</p> <h2> The subscription state

<p>If you're building automations with n8n, LinkedIn content is one of the more satisfying workflows to tackle because the manual version is genuinely painful at scale, and the automated version is actually possible to get right.</p> <p>This post covers the full technical architecture of a LinkedIn post automation workflow using <strong>n8n + Google Gemini + DALL·E</strong>, including the parts most write-ups skip: prompt design, error handling, approval gates, and what the workflow <em>can't</e

Welcome to Remarkable People. We're on a mission to make you remarkable. Helping me in this episode is Dana Suskind. Dana is a pioneering pediatric cochlear implant surgeon, researcher, and bestselling author whose work has transformed how we think about childhood development. She has spent decades helping children hear and communicate while studying how [...] The post Why Parenting Matters More Than Ever in the Age of AI with Dana Suskind appeared first on Guy Kawasaki .

<h1> Stop describing pixels to your AI agent — annotate them instead </h1> <p>If you pair-program with an AI agent — Claude Code, Codex, Cursor, whatever — you've lived this moment. You take a screenshot of your UI, paste it into the chat, and then you start typing:</p> <blockquote> <p>"See the button in the top-right? No, not that one, the <em>other</em> one, next to the misaligned icon — make it full width on mobile and fix the icon while you're at it."</p> </blockquote> <p>You're describing p

<p>Last week the agent wrote a 200-line CSV-to-JSON normalizer with edge cases I'd have spent forty minutes on, in about ninety seconds. The same afternoon, on a multi-tenant billing edge case, it confidently produced a query that was subtly wrong in a way that would have leaked one customer's invoice totals into another tenant's dashboard. I caught it because I knew the invariant. It didn't, because the invariant doesn't exist anywhere in its training data — it exists only in my schema and my h

<blockquote> <p>Originally published at <a href="https://htpbe.tech/blog/crypto-otc-proof-of-funds-pdf-fraud" rel="noopener noreferrer">htpbe.tech</a>. The version on htpbe.tech stays in sync with the latest detection algorithm — refer to it for the canonical text.</p> </blockquote> <p>A self-introduced broker emails the desk on Tuesday morning. The counterparty — a family office out of Singapore — wants to take a $25M position. Attached is a one-page PDF on what appears to be DBS Private Bank l