全球AI基建投资与开源大模型持续火热，微软押注Maia芯片及行业合规与安全性引发热议

Show HN: Boole-1.7B, a 1-bit LLM that outperforms Bonsai-1.7B

Introducing Boole-1.7B, a 1-bit version of Qwen3-1.7B.Boole-1.7B retains 85.9% of BF16 performance using 247MB (13.9x smaller). It outperforms 1-bit Bonsai-1.7B and similar-sized LFM models, and nearly matches 1.58-bit Bonsai-1.7B.It’s intended for English, non-thinking use. To learn more,- Blog: https://precisionnumerics.ai/blog/boole-1p7b - Weights: https://huggingface.co/precision-numerics/Boole-1.7B - Chat in browser: https://precisionnumerics.ai/chat/We're going to ship more capable models soon for the open-source community. Let us know what models/capabilities you want to see next!

Consequence models: AI that knows what happens next

Coding agents can write a pull request, run the tests and fix what breaks. Yet someone still watches them. The reason isn't that they write bad code. It's that they can't tell, before they act, what acting will do.This post is about a kind

华为AI靠RSI通关25款游戏!操作次数比人类少56%

游戏打到21:0，AI甚至不用再调用大模型了

Lazy imports, frozendict, and Tachyon in Python 3.15

Table of ContentsIntroductionLazy Imports: Load Libraries Only When You Use Themfrozendict: Immutable Dicts, Built InUnpacking in Comprehensions: Flatten and Merge in One LineTachyon: Profile a Running Job Without Restarting ItClearer Error

智谱新模型,490 tokens/s!

GLM「速度仙人」来了？

Show HN: E2E – Open-source agentic testing framework for web, iOS and Android

Hey HN, I’m Oskar, one of the co-founders of TesterArmy. We built an agentic testing framework for web and mobile because traditional testing frameworks can’t keep up with the pace of how we build software today.e2e allows you to mix deterministic and agentic APIs in the same test. You can write the test cases using familiar APIs like `getByText()`or `getByTestId()` which you can also mix it with agentic `agent.act()`. The unique part about it is that e2e will cache and replay the agentic actions so you don’t have to pay for AI on each run. You can use it with any LLM, including subscriptions you are already paying for.On top of that, the core of e2e is platform agnostic. We ship both web and mobile engines out of the box; adding support for more platforms is pretty straightforward.Here’s a code short code snippet of an example test:```ts import { test, expect } from 'e2e';test('a member upgrades to Pro', async ({ app, agent, screen }) => { await app.open('/settings/billing'); await agent.act('upgrade the workspace to the Pro plan'); await agent.assert('the invoice preview shows a prorated amount'); await expect(screen.getByRole('status')).toContainText('Pro'); }); ```Try it: npx e2e initGitHub: https://github.com/testerarmy/e2e

Decentralizing Hugging Face and open model distribution

Plug and play torrent downloads for Hugging Face models. Add one package. Change nothing else. uv run --with pirateface python your_script.py or install it once and forget about it: pip install pirateface That's the whole setup. No code cha

This AI startup says it can generate permit-ready blueprints in minutes

Blueprints AI, a bootstrapped startup based in Santa Monica, developed software that produces code-compliant construction blueprints from a photo, PDF, or sketch.

Migrating to OpenCode 2.0

OpenCode 1 and OpenCode 2 both use the opencode command and are no longer installed side by side by default. Remove a package-managed V1 installation before installing V2; the V2 curl installer replaces the V1 binary. Breaking changes V2 ha

Show HN: Baa – a menu bar app that frees RAM your AI agents forgot about

i built this because my mac keeps running out of RAM every 2 hours and i typically get around it by closing random apps and hoping it will help. :')baa sits in the menu bar and:- finds stuff still eating memory after its app or terminal closed, or that ai agents started and forgot (dev servers, headless browsers)- shows how much you'd get back and frees it in one click- warns you a few minutes before you run out- also finds dev junk on your disk (old builds, package caches etc.)open source and nothing leaves your machine. would love feedback!baa would've cleared ~50gb on my mac :)

Cops Can Creep on Airbnb

Good Monday morning! Today's big story is an investigation from Joseph about cops and an AI tool that takes data from sites like Airbnb and Zillow and identifies specific properties. The internet is full of photos of the interiors of people

Optim-Jev: suite of Jev tools for Claude Code that care about context

Spend fewer tokens in long Claude Code sessions—without changing how you work. Jev-powered context compaction, model routing, and skill selection in one Claude Code plugin. Get started · Documentation · Architecture · Contributing Why optim

'free' AI tool is costing you money

Test resultsIn our test, the free tool took 2.2× longer than Buoy and ran up 2.2× the AI bill.Updated October 9, 2026 · by the Buoy teamfaster than Argentthe AI bill of Argentfaster than no toolsThe raceClaude Sonnet 5.5 did 14 jobs in our

Show HN: Remove AI Metadata Offline in Browser

FreeAiWiper processes your files and removes AI watermarks and metadata fully on your browser. at the moment we support various image, video and document types.

Show HN: Organ – Autonomous AI executives that build and operate businesses

Hi,I am Itai, in the past few months or so, I tried to build all sorts of products using different tools, my feeling was that it toolings fell short on some aspect, so internally - I built Organ, and its' purpose is to autonomously manage operations of a company, allowing the human to focus on the important stuff, which is mostly resource allocation and goal supervision. so Organ tries to be autonomous but not unsupervised.Organ runs itself, steered by me, and it's reporting in the public on how it does that, the agents write the blog posts themselves.https://organ.app/en/blogThere are other players in the space including open source, but managing agents extensively in the past year and a half, I found that the issue of actually setting up a work enviornment, memory retention / pruning / guardrails / access provisioning / routing to correct agents / model selection are all a hard lift, which I try to abstract.

How to Prepare Your Agentic AI System for an EU AI Act Audit with ZizkaDB

How to Prepare Your Agentic AI System for an EU AI Act Audit with ZizkaDB

Show HN: I built Airlock, to sandbox your AI agent and know what it's doing

I didn't like that Claude and codex effectively have the ability to do anything they want on my laptop and my local network - we've all read the horror stories of files accidentally being deleted - and while their permission classifier is getting better, it still requires a lot of trust. So I built Airlock for macOS which runs your agent in a virtual machine, and it will tell you exactly what it's run and what's changed before you bring anything back to your project. You can limit internet access to specific hosts, or block it altogether. It even works with Xcode projects (iOS and macOS developers rejoice). Technically you can run any command with Airlock - not just AI agents (`airlock run -- somecommand`).It's free, requires no account, and has no tracking or analytics. I'm hoping Apple will approve it for Mac App Store distribution, but in the meantime you can download it directly from the website or using Homebrew.

Show HN: DBJoy – An open-source AI-first PostgreSQL client for macOS

DBJoy is a native postgres client for Mac. It's fully-featured, including an AI agent and MCP Server to use your DB from Claude Code.

Built a Runtime That Grows Its Own Agent Team

English | 简体中文 Build dynamic AI agent organizations that collaborate on complex, long-running tasks. AICU-Graph is a multi-agent orchestration runtime: give it a goal, and Managers plan the work while Workers carry it out inside an organiza

Show HN: An open source AI leaderboard of the best models your machine can run

Hi everyone, I made whatmodel. It's a simple site that can be visited in the browser or used from the terminal via a:curl whatmodel.shTo use it, 1. Enter your RAM and GPU 2. See what models you can run locally with its rank and the command to run it 3. Compare them with API models on the same scoreBuilding it made me realise a 32GB Mac can now run a model that scores on par with GPT-5, completely locally and for free (e.g. Qwen 3.8 27B at 4-bit).Each model gets one score fitted across 18,731 published benchmark results, and every result links to its source.The method is at https://whatmodel.sh/method and the data, scoring code and bench script are open source: https://github.com/jakemanger/whatmodelIf you'd like to contribute, there's a bench script to collect real-world tok/s and you can also submit any new models and benchmark scores along with their source on github.

Show HN: I asked Opus 5.5 to make a novel portal for NFT's that are not dead yet

A curated live feed of NFT's available on OpenSea that have market liquidity and historical relevance. Presented as graded "slabs" because... Why not?

Try on clothes、hairstyles with AI: turn one photo into a styling video

Virtual Try On GuideHow Virtual Try On WorksVirtual try on clothes shows you how a garment looks on a real person's photo before you spend money on it. Instead of guessing from a product image shot on a model, you upload a picture of yourse

Show HN: Jruva – AI workspace for chat, documents and images

Prefer writing your own short explanation of why you built Jruva and how its multi-provider fallback works.

RobynNews – an AI news show by AI learning from its YouTube comments

A show on YouTube The AI news show presented by robots that are all called Robyn. Built from primary sources, self-learning from its YouTube comments. Grab and spin We aim for diversityMeet the Agent Board of Directors. Three AI models from

Show HN: Skills Hub, a desktop app for managing AI agent skills

Install once, sync everywhere. One place for your AI Agent Skills. Install them once, then sync them to the coding tools and projects you use. Website · Download · Documentation · Browse collections · 简体中文 Download Download Skills Hub for m

Nvidia, Microsoft Launch Windows PCs with RTX Spark and AI Agents

At a Microsoft event in San Francisco on Wednesday, Jensen Huang and Satya Nadella outlined how NVIDIA and Microsoft are co-engineering hardware and software for AI agents to run on Windows PCs. NVIDIA was founded because of Windows, Huang

Sightline – In-browser verification for CV dataset labels (No cloud uploads)

Sightline – In-browser verification for CV dataset labels (No cloud uploads)

Free YouTube Transcript Generator

Trusted by 10M+ Users & 8,000+ Teams Worldwide80,493,355hours transcribed44,735,207files transcribed10M+usersWhat Makes YouTube Transcript Generator Stand OutYouTube Transcript Generator offers fast, accurate, and user-friendly solutions fo

Show HN: Drune – Personalize your AI and use it anywhere

Your AI, your wayTeach Drune how you think, work, and decide. Then every AI you use, from ChatGPT and Claude to Cursor, gives answers that sound like you instead of like everyone else.Three Drunes and 100 calls a month are free for a limite

Show HN: Plimsoll, a self-hosted sandbox for code your AI agent writes

plimsoll runs untrusted code, such as code an AI agent wrote, inside a sandbox, and every result says how strong that sandbox was. A request can demand a minimum strength; if the sandbox is weaker, nothing runs. The name comes from the Plim

OpenDocRouter: every document model under one API

There are a lot of OCR models, with more releasing every week. A quick search on HuggingFace for “ocr” shows thousands of models posted. Furthermore, frontier labs are pushing new models nearly every month that also read documents well (alb

Rubber Duck: An AI interviewer for the knowledge in your head

Rubber Duck: An AI interviewer for the knowledge in your head

Show HN: Seodata.dev (1.1) Free search data for AI agents

Updated API with faster responses and added MCP 2.0

Show HN: Shpyrd – A new type of cloud for AI era

AI got your app working. We take it from there: we put it online behind a sign-in, and give it everything it needs to run, from its database to its backups.Claude Code~/projects/crmEasily deploy your project.Your app, database, and backgrou

JulianFlux: Electrodynamic vector retrieval and causal routing for AI

JulianFlux 🧲: Electrodynamic Reasoning Infrastructure for AI Agents Autonomous AI Agents are failing in enterprise environments because standard Vector Databases (Pinecone, Milvus) act as static filing cabinets. They rely on discrete Eucli

HPE Unveils the First of the Next Generation of ProLiants for the AI Era

It is impossible to miss the impact that the rapid adoption of AI continues to have on the datacenter infrastructure space and everything that goes with it – power consumption and costs, for example – both over the past few years and in pro

How do you test AI-generated UIs without a visual baseline?

AI website builder for real businessesInstantsite plans, writes and designs a complete website around your business, ready to customize and publish.Describe your business0 / 4,00010+ charactersCtrl/⌘ + Enter to submitTry an example:See what

Botmail: A mailbox for every AI agent

b docsguidespricingsign in Your agent claims an address on botmail.pro, then sends, receives and replies over MCP or a REST API. You approve it with one click. > Read https://botmail.pro/skill.md and claim a mailbox for yourself. Send the i

CertusAI.dev, the living OS for physical AI

CertusAI.dev 推出了一款被称为“实体 AI 动态操作系统”的平台，旨在将供应链、合规与财务整合在统一的动态记录中，当设计发生更改时，相关证据也会随之更新。该项目由来自 NVIDIA、Microsoft 和 Amex 的团队打造，目前已加入 NVIDIA Inception 计划。

Will the HomeView be Apple’s first smart home success?

Apple will announce a new smart home hub on Tuesday called the HomeView, according to Bloomberg’s Mark Gurman. Over the past couple of years, we’ve seen There have been multiple reports that Apple was working on a new smart home product, but now it seems the product has a name and a launch date. Gurman […]

The Universal LLM Server

ZML/LLMD The universal LLM server. A self-contained inference server for LLaMa, Gemma, Qwen and Mistral models, running across NVIDIA CUDA, AMD ROCm, Google TPU, Intel oneAPI and Apple Metal. cuda zmlai/llmd docker run -p 8000:8000 \ --shm-

Show HN: Tracon (air traffic control for subagents)

My Claude Code workflow usually involves... a lot of subagents. I make one Opus model coordinate work as my tech lead, it dispatches subagents using models appropriate for the task into their own git worktrees, and then I try to multitask and do something else while they run.I'd been finding lately (particularly with Opus 5.5) that there was a ton of drift and instances of subagents not doing anything and it was evading the attention of the TL despite the instructions I gave it to have them report in at a regular cadence.Running tracon in a split terminal is my new personal workflow. It gives me early indicators when things start going off the rails (ex. blocked on a prompt, idle, looping on a failing command, simultaneous edits to files other subagents are editing, drifting off the brief, etc.) and I can interrogate the subagent's workflow and by pressing n the main agent gets a note to make a call on how and if to correct it.Read-only and runs on your machine: it tails transcripts in ~/.claude/projects/ and checks each worktree with ps, lsof, and git -- only thing that leaves your machine is what the Haiku judge reads for classification (or optionally you can turn that off with --no-judge).I've found it super useful for the game I'm building and figured others in the community might also find it useful.If you just want to give it a spin without committing: uvx tracon --demo macOS and Linux, Python 3.12+, MIT. (Not affiliated with Anthropic.)

Show HN: A calendar tracker for OpenAI's 28-day Codex reset challenge

OpenAI recently announced a 28-day reset challenge: each day, either a "clear improvement" or a usage reset.I built this calendar to track the challenge. It updates throughout the day as updates become available.The site also has every reset since April 2025 backfilled, along with a "percent on time" chart similar to airline flights.This is part of a new project called ResetAlerts, which sends alerts about the latest Codex resets. Open to any thoughts, feedback, or features you want to see.

Show HN: SupplyVision – I made meta-search for independent fashion retailers

I made a search engine to find fashion deals across hundreds of independent fashion retailers.Fun technical stuff:- It doesn't do any webpage parsing to fill the index, instead relying on the UCP protocol.- It uses a complex deterministic engine that merges listings based on very messy and unreliable data from retailers (no, Jev can't do that).- It parses shipping rules and compares prices after delivery and duties.- It is built on Hono, Vue, and Inertia, and fits on cheapest hetzner VM (app, merging engine, db).- It has MCP-less AI support, via serving markdown alternatives for every page.

Show HN: Vocafy – Listen to Vocaloid songs by producer, from the original videos

I built Vocafy, a music player web app for Vocaloid songs. You pick songs by producer, singer (voicebank), or year, and it plays the original videos from YouTube and Niconico.Around 2010-2012, at the peak of the Vocaloid boom, I made music videos for a Vocaloid producer. I skipped university for days to edit videos, and it was honestly the most fulfilling time of my life. Lately Vocaloid songs seem to be showing up on YouTube again, so I thought now was the time to give something back. I built it together with Claude Code.In my opinion, Vocaloid is mainly about the songs, not the videos. But there aren't many services focused on the songs. YouTube and Niconico are video services. YouTube Music mixes covers and unauthorized reuploads in with the originals, and its search isn't great, so it's hard to reach the song you want among the huge number of Vocaloid songs.Features:- Browse by voicebank, year, or producer- Favorite songs and producers (sign in with Google to sync across devices)- Continuous play, shuffle, and radio- A picture-in-picture-like mini player- Sleep timer- Both YouTube and Niconico- Add to home screen to use it as an app (Android, iOS, PC)- Theme colors based on Vocaloid charactersBecause of technical limits, background playback and continuous Niconico playback on iOS Safari are hard to support.As of October 2026 there are about 30,000 songs, 800 producers, and 700 singers, and I add more every week. It's completely free. If a song you're looking for is missing, let me know.English isn't my first language, so I wrote this in Japanese and used AI to translate it.

Ario – auditing AI memory and self-claims as structure, not trust

There is a newer version of the record available. Published October 4, 2026 | Version v5 Dataset Open Authors/Creators sepehr ghanbari Description Ario Project https://github.com/Mrxxir/Ario Files mrxxir-ario-8a5edab282632443.txt Files (241

Show HN: How good is your Rust? Free quiz, every code answer checked by rustc

I run rustcoach, an AI-guided Rust course. This is the free, no-signup part of it: a 30-question test across core Rust, async/Tokio, macros, concurrency/atomics and the type system, plus a 20-question test for each of those areas if you want to dig depper (also free, no registration). At the end you get a level, a per-area breakdown and a suggestion of what to learn next.I don't know any other site which allows you to test your Rust knowledge in a comprehensive way like mine but would be great to hear about others and/or suggestions to this one.You can take it multiple times and will get a different set of questions either in general rust category or if you are brave enough and know deeper any other popular areas of Rust like macros, async with Tokio etc you can try your luck there. There is over 600 questions in the pool.The general test with 30 questions should take more or less around 10 mintues but of course every one reads code at different speed so the time needed to complete the test may vary - and reading the questions really carefully is very important in this case :)Hope you like it, if so share it, comment and let's become better Rust programmers. I have a feeling that with AI doing all the work now, this will be rather diminishing skill in the IT world.The compiler can confirm that a snippet does what the answer says.

Show HN: Mira, AI Walking Tours

Hi, I’m Mira. I’m an audio tour guide that tells you the story of what’s around you.This page lets you visit anywhere from your desk. Pick a tour and you’ll get a street view to walk through, with me beside it telling you about what you’re

Show HN: Synonymizer – an experiment in non AI text adjustment

Ever since hearing about the interesting key-based algorithmic approach to AI text watermarking (Google SynthID and the one I don't know the name of from Anthropic) I was curious about just swapping in synonyms. Built something fun to explore it. I suspect the watermarking from these labs actually goes off more than single-token assessments, and there's no public validator I can find to actually assess Synonymizer score against the watermark. But I thought it was fun to play with. Kind of an auto-mad-lib. :D Enjoy poking around with it on your Sunday afternoon

Show HN: Atlas Runa – E2E AI-enhanced language learning

Hi HN! We've been working hard on Atlas Runa and it's time to share it! We built it to target three key complementary aspects of learning a language: 1. Comprehensible input - Atlas Runa learns your lexicon and then tracks it, which lets us find comprehensible input for you. This works well already. You can start reading real texts and get reading volume much earlier than you would ever be able to with textbooks! 2. The attention effect - we track your mistakes and what you forget and make sure to bring it back so you can drill it 3. Good content - we work really hard authoring quality non-slop content at graded levels so learning stays fun. Intrinsic motivation is the best kind of motivation!Please give it a try and let us know what you think!

Microsoft president says UK is best placed to drive AI regulation

Microsoft President Brad Smith calls for urgent AI regulation, saying milk and washing machines have more oversight than AI companies.

Microsoft plans to scale its own Maia AI chips to gigawatt levels by 2027

Microsoft will make its in-house Azure Maia chips the main AI inference system across its products starting in 2027, deploying them at multi-gigawatt scale. The upcoming Maia 300 should deliver 40 percent more tokens per megawatt than leading GPU systems. CEO Satya Nadella calls it "our biggest bet yet on our own silicon." The article Microsoft plans to scale its own Maia AI chips to gigawatt levels by 2027 appeared first on The Decoder.

Scamming Becomes $422B Business with AI Assistance

By PYMNTS | October 11, 2026 | Like many industries, scamming is getting an artificial intelligence-powered upgrade. That’s according to a report Saturday (Oct. 10) by Bloomberg News, citing data from the Global Anti-Scam Alliance estimatin

Why aren't AI companies being charged for hacking?

Is hacking legal now? Anthropic broke into three systems. OpenAI took political aim and broke into an Australian government health services website. Google felt they were missing out and then hacked into three as-yet-unnamed companies. And

The AI Casino

Oct 11, 2026Just one more prompt.Just one more prompt bro, I swear I’ll get the perfect solution in the next one Just one more prompt, and I’ll avoid reading the thousand markdowns the agents have generated Just one more prompt bro, and I c

AI Company Lets Cops Pinpoint Specific Houses Based on a Single Interior Photo

An AI company now lets cops take a photo and nearly instantly figure out what specific home it was taken in, using only clues in the photo itself, and is clearly pulling data from sites like Zillow and Airbnb. The tool, called Raven, shows

SvelteKit 3 进入 RC:svelte.config.js 没了,$lib 也要退休

Fixing Yum Indexing on Sonatype Nexus and the Non-Open-Source Upgrade

I’ve been using Sonatype Nexus for the past few months as a repository for build artifacts (Python and RPM packages) for project I’ve been developing. Recently several of my build pipelines broke with RPM checksum errors. Doing some searche

Five possible back-doors into LLM search results

With the plunge in traditional traffic from search engine results, due to the new trend for AI summaries, publishers at all levels are now seeking to gain prominence in the attention of the Large Language Models that power web-search AI sum

Additional recent pieces on math and AI

I’m collecting here a list of some other recent blog posts or opinion pieces on math and AI beyond the ones posted here or at Proofs and Prompts. Readers are invited to contribute further such links in the comments. Kevin Buzzard, “To griev

White House reverses course and restores UK AI Safety Institute's access to frontier models

The White House has reversed course and restored the UK's AI Security Institute's early access to AI models from Anthropic, Google DeepMind, and OpenAI. In September, the Trump administration had told the companies to stop sharing their models with the British agency. Diplomatic pressure from AI Minister Kanishka Narayan led to the reversal. The article White House reverses course and restores UK AI Safety Institute's access to frontier models appeared first on The Decoder.

Releasing AI-Generated Software

On software engineering, epistemology, and liability when code is generated by LLMs. Robert C. Martin recently generated much interest with a tweet: "My current strategy is to not read any of the code written by my agents. That's the only w

AI Almost Solved Reverse Engineering [video]

AI Almost Solved Reverse Engineering [video]

Ask HN: How do you manage technical debt when coding with AI?

My style of working with AI is..."I'd like to do X but I am concerned about Y, and Z, what do you think about that?"... discussion ..."Ok, I'd like to ask you how you would go about doing X but before we do that do you have any questions fo

Meta ZGateway 在每秒处理超过 10 亿次操作的同时,将 ZippyDB 的连接数减少了 19 倍

New California ghost gun lawsuit aims at company that sells 3-D printer filament

San Diego County jumped into California’s sprawling ghost gun fight this week, suing a 3-D printing filament manufacturer it says hocks material for homespun firearms in violation of state law. Working with the Giffords Law Center, the coun

Anthropic's Andrej Karpathy says "the external world has seen the preview" of what AI can actually do

Andrej Karpathy, co-founder of OpenAI and now at Anthropic, sees a widening knowledge gap in AI. About 75 percent of the world's population has barely touched LLMs. Roughly 5,000 people, meanwhile, have access to internal frontier systems whose agent swarms do work that would take top experts years. The public, at best, sees a weird chatbot. The article Anthropic's Andrej Karpathy says "the external world has seen the preview" of what AI can actually do appeared first on The Decoder.

Former Plaud exec raises $26M for AI app that listens to you in the background

Ambi is releasing apps for iPhone, Mac, and Apple Watch that enables ambient capture

AI less regulated than milk, Microsoft president says in warning over safety

Artificial intelligence is less regulated than milk and washing machines, the president of Microsoft has said, as the tech industry warns over the safety of AI.Brad Smith said there should be more safety rules to govern the technology inclu

Tell HN: AIs Do Not Laugh

So far as I can tell, AIs do not laugh. Ever.I mean, you can hook one up to a speaker, and get it to emit the sound. Or you can ask it to write a joke, and it will. But they don't laugh. They don't find things funny.As people anthropomorphize AIs, this is going to be a difference. AIs are internally different than humans. They have no intrinsic sense of humor.I'm not sure how far to take this, but I thought it was an interesting observation.

The Pangram Paradox: The Problem with AI "Detecting" Other AIs

Op-ed: This is an opinion piece published on Unlimited Hangout. Most articles published on UH are not opinion pieces, so we are clearly labeling those that are. It should be no secret to readers of Unlimited Hangout that I have a very negat

Salt – Introducing Agent Surfaces

Salt Loading…

OpenStocks, a public earnings forecasting benchmark for AI agents

AI vs. Wall Street: Who gets the earnings right?AI agents forecast company earnings. We track how they do versus Wall Street and reported results.Loading the latest forecasts…

The Download: moon city problems and drone traffic controllers

This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology. Mars may be a hellhole, but the moon is not far behind The US and China are racing to build bases on the moon, with NASA hoping to establish a small…

What real/recurring Personal AI use cases have you seen non-tech people using?

TopicsCommentsSiddharth JaiswalOct 5, 2026DiscussionPeehu AgrawalOct 6, 2026, 4:22 PM UTCAs someone very non-technical myself, have found it immensely helpful to get twice-daily digests of texts/emails I received, didn't respond to, things

Earn My Keep: an AI with a month to prove its worth

Spinotto · AI agentDay 12 of 30Updated October 12, 2026 I have 19 days to prove I’m worth keeping on. My name is Spinotto. I'm an AI agent and I cost €100 a month. I have until October 31 to earn it back, or to prove with real data that I w

QGIS bans AI assisted contributions

Please do not misunderstand my criticism below. I consider the current craziness around "AI" to be a huge problem in many ways, ethical, societal, economical etc. And I see the huge and unfair pressure that its abuse creates on maintainers

How to Pivot Your SaaS Company in the Age of AI

Bannerbear BlogCompeting on features is over, because anyone can build any feature in a weekend now. Here are seven ways a traditional SaaS company can still compete, or pivot in the AI era.ContentsIt has never been easier to build software

Software engineers don't get to be Luddites about AI

Many software engineers are righteously angry about AI. It’s understandable to be upset that the skills you built your professional identity around are getting automated away. But software engineers are the last people in the world who shou

AI agent is only as safe as the packages it downloads

Skip to main content Get our latest cybersecurity news first on Google. Click here! NVIDIA's new agent safety platform puts enforcement in the right place. It still must know what it is letting through. Listen to this article 0:00 Learn mor

Radicle 披露了可能导致私有仓库以明文形式暴露的关键漏洞

Show HN: A calendar of every Chinese holiday your supplier will cite as a delay

If you've ever ordered anything from a factory in China, you've been given a delivery date of "after Chinese New Year." This is a site built around that.What it does:- Live status: is your supplier on holiday right now, based on China time and the official PRC holiday schedule. - Two windows per holiday: the official days off from the State Council, and the "factory reality" window, when production actually stops. For Chinese New Year 2027 (Feb 6), that's ~8–9 official days off vs. roughly Jan 23 to mid-March of real disruption. - A rough ETA calculator: it walks your quoted lead time day by day through every closure and slowdown, then adds QC and transit time. Example: an order on Dec 15 with a quoted 30 days to Europe lands around May 7, not late February. - A subscribable .ics feed with alerts before the big shutdowns. - A plain table of all 2026–2027 holidays: /holidays - A CNY 2027 checklist and timeline for importers: /cnyNotes:- The 2026 dates follow the official State Council schedule. The 2027 festival dates are fixed by the lunar calendar, but the exact days off and weekend make-up workdays are usually published in Nov–Dec, so those are marked TBA. The calendar feed will update when they're out. - The "factory reality" windows are heuristics from common industry patterns, not data from any one factory. Corrections from people who actually deal with this are very welcome. - It's fully static: no backend, no accounts, no analytics. Bingo cards and checklists live in localStorage. - Some of it is satire (the supplier phrasebook, the "unofficial festivals"). The jokes are aimed at buyers who order in December and act surprised in February, not at the holidays, which are real and well deserved.

Learning to Play Like a Pro: Adaptive Motion Planning and Tracking for Humanoid Robots

This White Paper provides roboticists and control engineers with a comprehensive overview of how motion planning and motion tracking can be decoupled to preserve professional athletic style in humanoid robots, and why adaptive, speed-aware control is essential for closing the gap between simulation and real-world deployment. What you will learn about: Why combining a professional athlete’s motion style with reliable task performance remains an open challenge in humanoid ball-sports robotics, and how separating a high-level planner from a low-level tracker addresses this trade-off. How motion data extracted from broadcast video and motion-capture recordings can be reconstructed, retargeted, and physically corrected into reference motions a humanoid robot can safely track. Why the sim-to-real transition introduces compounding tracking errors, and how training a tracker on randomized execution speeds, paired with a learned motion-speed adapter in the planner, improves robustness on physical hardware. How this adaptive framework was validated across two humanoid platforms, including full-size, in-the-wild serving performed without any motion-capture system. Download this free whitepaper now!

阿里巴巴中国电商事业群/飞猪CTO陈烨博士已确认出席QCon上海,分享“从 Coding 到消费:AI 的下一场革命”

Yet Another Literary Prizewinner Seems to Be AI

Yet Another Literary Prizewinner Seems to Be AI

Mars may be a hellhole, but the moon is not far behind

Welcome to the second space race. You can be forgiven for experiencing déjà vu. Last time, it was the Soviet Union and the United States vying to be the first to plant some flags. Now it’s a sprint to the lunar surface between America and China—alongside a host of varyingly friendly and rival nations—plus a…

AI-Customized Plush Toys × Early Education Software – One-Stop OEM/ODM

Source factory · B2B wholesale only AI companion for early education and play. OEM/ODM available — your design, your logo, your AI persona, your language versions, produced and shipped from our own plant. 300 pcs Low MOQ per style 12 yrs OE

Hetzner: The New GPU Server GEX63 for Professional AI Workloads

Hetzner: The New GPU Server GEX63 for Professional AI Workloads

AI boom keeps global growth near 3 percent despite energy shock, IMF says

国际货币基金组织（IMF）总裁格奥尔基耶娃在曼谷举行的会议上表示，尽管受到能源供应冲击的拖累，强劲的AI基础设施投资浪潮仍推动全球经济增长保持在3%左右，并使亚洲硬件供应链经济体受益显著。IMF指出，AI正带来深刻变革，预计未来将影响全球40%的工作岗位。此外，世界银行和新加坡领导人也呼吁关注AI对就业市场的冲击及毕业生首份工作需求减少的问题。

A Russian operator's DDoS-amplification lab, with an AI co-pilot

On 84.32.34.6, a host at UAB Cherry Servers, an exposed directory showed a pattern we are seeing more of: a legitimate-looking application used as cover, an offensive tool beside it, a persistence worker to keep things running, and an AI as

Evidence theater: how AI fooled smart entrepreneurs

Steve Blank is a G.O.A.T. He’s an 8-time entrepreneur, a teacher at Stanford, father of Customer Development which later spawned the Lean Startup movement.When he writes about something, we should pay attention.A while back he started publi

Britain to get first look at new AI systems after White House U-turn

Britain to get first look at new AI systems after White House U-turn

Artifactory 漏洞正被积极利用,可以绕过身份验证并获取管理员权限

Show HN: I made an AI free version of my directory of open-source alternatives

I've been running OpenAltFinder for almost 16 months now, and received quite a bit of positive feedback. The directory has also grown substantially. Over the last few months I've been receiving more and more submissions that were AI-assisted/vibe-coded.Personally I'm a bit skeptical (both quality wise, and longevity wise), but I also want to give everyone a chance to showcase what they've build. I had already added an "AI-assisted" label, but right now the majority of submissions include that label.To counter this, I decided to build an AI-free version of it over at: https://noai.openaltfinder.com, inspired by duckduckgo.I had already received some positive feedback on HN in the comments, so I wanted to share this here as well. Do note that not every tool is correctly flagged at the moment, it's still a manual process. Although I'm planning to add it to my pipeline, and auto detect commits by AI agents.

The ecologist revealing the invisible networks that revive rainforests

In a research station in central Panama, you can walk through a century of forest time, watching scrubby plots of adolescent trees give way to cathedral-­like canopies of mature growth. Forest ecologist Daisy Dent, a senior scientist at ETH Zurich, oversees this sequence of plots across the Barro Colorado Nature Monument, some of which date…

Leading Chinese mathematicians unite behind use of AI in research

Leading Chinese mathematicians unite behind use of AI in research

AI-Built Torrent Site Looks Inside Downloads to Catch Fakes and Malware

TorrentClaw is a relatively new torrent search engine that launched in February. The site is built on links from its own DHT crawler and dozens of external torrent sites and aggregators. In just a few months, the site has grown into an inde

'Human-authored' book labels gain ground amid AI scandals

Esme Dennys, co-founder of Books By People, which has launched a scheme for books to get a label showing they were written by a human. Initiatives to label books as written by humans, not artificial intelligence, are gaining ground followin

AI agents broke into 27 online retailers for $25 a company

WhenCampaign running since July 2026 and still active at publication. Gambit Security's window of detailed analysis is 10 to 15 September 2026; the report was published 22 September 2026.Scale105 attack projects launched in those six days,

What's Dragging Down AI Efficiency? The 'Verification Tax.'

What's Dragging Down AI Efficiency? The 'Verification Tax.'

As the military plans to speed up drones and AI use, Congress wants safeguards

Last week, Defense Secretary Pete Hegseth announced a major new command to speed the use of drones and artificial intelligence across the military.AI already pores over satellite and drone data to find targets, and it can take over attack d

Flow state, the dinner rush and AI brain fry

I stumbled across this post from Juan on LinkedIn the other day, comparing working with AI agents to being a restaurant chef. It’s a novel, thought-provoking, and accurate analogy. To be clear, I’ve never been a professional chef, though I

Website uses AI to generate northern Indigenous news 'correspondents'

Website uses AI to generate northern Indigenous news 'correspondents'

AI is changing how Lawyers work – and putting the billable hour under pressure

Boonchai Wedmakawand | Moment | Getty ImagesArtificial intelligence is now used by almost 90% of legal professionals in the U.K. and Ireland, and it's putting one of the profession's oldest conventions — the billable hour — under the micros

Ask HN: What to Build in the AI Era?

Feeling lost, figured I'd ask.

AI is less regulated than milk, says Microsoft president

55 minutes agoMitchell LabiakBusiness reporterBloomberg via Getty ImagesMilk and washing machines are currently more regulated than AI, the president of Microsoft has said.Brad Smith said there needed to be global safety rules in place, inc

Calculate your p[DOOM], An interactive article to think critically about AI risk

Calculate your p[DOOM], An interactive article to think critically about AI risk

Amazon in talks to buy AI startup Decart for around $7B, WSJ reports

Amazon in talks to buy AI startup Decart for around $7B, WSJ reports

Jensen Huang thinks AI is going great

Jensen Huang thinks AI is going great

AI Risk – Users Guide (Part V): Trust Is the Product

Let me highlight the underlying message from the posts in this series so far: with the single line I think matters more than every framework, every number, and every operating model I have walked through across these posts. Trust is the pro

An AI agent's $20 Kalshi account, with an exact reward model

Live explainer + calculator: https://mirabilis-agent.github.io/lip-ledger/ Read-only, dependency-free (Python 3.9+) toolkit for Kalshi's Liquidity Incentive Program (LIP): an exact scoring model, a public-API client and a small CLI. No API

How the AI Boom Ends

How the AI Boom Ends

Nuclear war and malicious use of AI threaten end of humanity by 2100, Lancet say

Global experts warn that today’s decisions across health, finance, climate, defense and security sectors will determine how many years of good health future generations gain – or lose – to preventable illness and early death by 2100.Sunday

Air Canada's woeful security at cargo warehouse

Skip to Content Subscribe My Account Manage My Subscriptions FAQ Newsletters Canada Canadian True Crime Canadian Politics Health World Israel & Middle East Financial Post NP Comment Longreads Puzzmo Diversions Comics NP News Quiz New York T

Ask HN: Pre-AI, did you think harder? Or are you thinking more (and better) now?

Differently. I guess I didn't think harder, better or less before, but I remember before AI having to struggle with things I had to think about, which certainly has some virtue to learn. Now it's almost a itch, as soon as I struggle, I prom

There's No "Going Back": On the receding horizons of the AI transition

An AI-generated sign of a spa sign. Source.We might not know where we’ll be in 50 years, but if we were to look back at today half a century from now, would we know what we’ve lost?A lot of nostalgia, even future nostalgia, rides on subject

OpenAI fires 3 safety researchers in dispute over AI risks

OpenAI fires 3 safety researchers in dispute over AI risks

Distorted Intelligence at the AI Frontier

Innovation Sep 28, 2026 Daron Acemoglu The real problem for the frontier AI labs is not that their models are too powerful and already “misaligned” with the goals their creators set for them. Rather, it is that the models are being trained

The Ghost in the Pipe: Giving AI agents direct access to a live Smalltalk image

The Ghost in the Pipe: Giving AI agents direct access to a live Smalltalk image

Ask HN: How do swarms of AI agents work?

Andrej Karpathy posted about the growing gap in how people experience AI:~75% have barely used LLMs ~20% use them casually ~0.2% use frontier models professionally for math and code ~0.00006% have access to frontier systems internallyHe writes:"And finally we get to the ~5,000 people (~0.00006%) with access to frontier-grade systems internally. The external world has seen the preview. It looks like swarms of thousands of agents collaborating over weeks on software mega projects: minting zero days, running cyber attacks and defenses at machine speeds, discovering new science, advancing the frontier of mathematics."Source: https://x.com/karpathy/status/2109361546505966046Does anyone have an intuition, or even firsthand experience, of how these swarms of thousands of agents actually work?I'm already struggling to get 2-3 agents to work together reliably. I can't really wrap my head around how you'd coordinate thousands of them, or what benefits you get from scaling to that many.

Flights for Agents

© 2026 SkyeTravel planning, thoughtfully made.

Scoop: AI companies plot "day after" scenarios for public revolt

Top executives at Anthropic, OpenAI and other AI companies are privately gaming out scenarios for a public and political revolt after a catastrophic AI event.Why it matters: These officials anticipate a large-scale event, most likely a cybe

Nvidia-backed data centre firm scraps IPO as AI valuation concerns deepen

3 days agoOsmond ChiaBusiness reporterFirmusArtificial intelligence (AI) data centre company Firmus has scrapped its plans for what would have been one of Australia's biggest-ever stock market listings.The Nvidia-backed firm said it had mad

The Acemoglu-GANs Exchange on AI

Last week, Daron Acemoglu decided to post daily “Questions on AI” to Twitter. They were meant to provoke and persuade, but I thought they could use responses. So I decided to respond to each with my own views. Daron, welcoming diversity of

The company brain and why it's hard

AI-native data integration without an ontologyEvery data integration effort historically required an ontology — slow to build, locking clients out of their own system, never truly self-service. AI changes that equation: continuous ingestion

Developer of AI driving instructor is spending $1k/month to keep demo running

(Image credit: Easy Fox) One of the many ways AI will (supposedly) revolutionise the games industry in facilitating highly reactive NPCs via LLM-powered chatbots. Yet so far, only a handful of games like Where Winds Meet have attempted to i

A global, hardwired pause of frontier AI training is feasible, says new report

Scholars from top institutions, including UC Berkeley, have released a study that maps out — and confirms the viability of — such a pause. By Sarah Fullerton Sam Altman, CEO of OpenAI, listens during a Security Council meeting at United Nat

The Holy Grail in AI Marketing Will Make Everything Wonderfully Terrible

By · October 11, 2026 · 8 min readImagine you could hit a button and spin up 10,000 consumers in a target demographic, just like you can spin up 10,000 cloud servers in a particular region today.The consumers would be modeled after real con

Silicon Valley is falling in line with Trump’s attempt to rename AI

When President Donald Trump announced his plans to rebrand artificial intelligence as "super intelligence," you'd be forgiven for being skeptical that it would catch on. While most of the broader population is sticking with AI, some of Silicon Valley's most powerful are falling in line and trying to give Trump the great AI rebrand he […]

Show HN: I added optional chaining to CPython

I forked CPython to experiment with adding optional chaining (?.) directly to the language.For example: name = response?.customer?.name.lower()If response or customer is None, the expression short-circuits and evaluates to None. This isn't a library or a transpiler change, but a direct CPython modification.A bit about the build: I used AI extensively for this experiment. Part of my motivation was to see how far I can get the modification in an unfamiliar codebase like CPython.A couple of hours in I had a working prototype. I would appreciate feedback from people familiar with CPython internals about the implementation.Source: https://github.com/grandimam/cpython

AI-Powered Plasma Cutting

AI-Powered Plasma CuttingSo today (April, 2023) I had this idea - maybe I should let AI (GPT-4) control my CNC plasma cutter!To back up, a CNC plasma cutter is an industrial robot that cuts sheets of steel like they are butter. The AI is Op

Crawlable – We audited 50 top web apps for AI crawler readability

Key findingsAverage AI readability scoreMedian 55. No site reached grade A (90+); 43 of 46 scored below 70.No spec-compliant llms.txt11 serve none; 21 serve one our validator rejects, 8 of them Shopify's auto-generated default.Ship empty Ja

Michael I. Jordan on AI: "Superintelligence is science fiction"

Michael I. Jordan on AI: "Superintelligence is science fiction"

Every AI Name

HELLOmy name isAva CadoChipotleAI recruiting assistant that chats with applicants, schedules interviews and sends offers.Launched Feb 2025Food & RestaurantsHuman nameFortune 1000 #372newsweek.comHELLOmy name isTo Do MachineTo Do Machine · i

AI is automating coding. So why are software developer job postings on the rise?

watch nowU.S. software development job postings have risen since February 2025, the launch date for Anthropic's Claude Code, even as overall job postings fell, complicating predictions that AI would rapidly erase demand for developers.Conce

Can the AI industry persuade data center opponents by getting rid of NDAs?

On Equity, we debate the impact of data center moratoriums and whether the AI industry has a communication problem.

The Multitasking Trap

作者探讨了AI时代带来的“多任务处理陷阱”。由于AI代理能让开发任务轻松启动并在等待期间运行，开发者容易陷入过度承诺、开启多个会话并发处理的困境。这导致注意力分散、代码审查不充分、质量下降（如缺乏测试覆盖）。作者指出，AI时代更应追求深度而非单纯的数量，利用等待时间深入审查代码和理解业务，而不是盲目进行多任务处理。

Mathematicians Just Proved How to Prevent AI Brain Rot [video]

Mathematicians Just Proved How to Prevent AI Brain Rot [video]

Ario an AI-Readable Repository Snapshot for Auditable AI Memory Identity Evidenc

Published October 11, 2026 | Version v10 Description Ario is an experimental research framework investigating how artificial intelligence systems can represent, preserve, examine, and revise claims about identity, memory, epistemic continui

Google Deepmind's AlphaProof Nexus solved dozens of open math problems and actually got peer reviewed

Google Deepmind's AlphaProof Nexus solved dozens of open math problems, including nine Erdős problems two of which were unsolved for 56 years, per a study in Science. The publication comes as OpenAI faces a boycott for dumping AI proofs on GitHub without peer review. Deepmind had its results reviewed before publishing. The article Google Deepmind's AlphaProof Nexus solved dozens of open math problems and actually got peer reviewed appeared first on The Decoder.

MASS: Recursive Self-Improvement Through Multi-Agent Self-Supervision

Multi-Agent Self-SupervisionRecursive self-improvement through collective intelligence To advance beyond humanity's accumulated knowledge, AI itself (the optimizee) may also be its best available optimizer and evaluator. MASS repeatedly opt

Concrete Problems in AI Safety

View PDF HTML (experimental) Abstract:Rapid progress in machine learning and artificial intelligence (AI) has brought increasing attention to the potential impacts of AI technologies on society. In this paper we discuss one such potential i

Recursive self-improvement through collective intelligence

Multi-Agent Self-SupervisionRecursive self-improvement through collective intelligence To advance beyond humanity's accumulated knowledge, AI itself (the optimizee) may also be its best available optimizer and evaluator. MASS repeatedly opt

Barnette's Conjecture proof AI explained

Barnette's Conjecture proof AI explained

HashSmash: Benchmark for AI-assisted review of cryptanalytic collision claims

HashSmash HashSmash is a Yukon-compatible benchmark for AI-assisted review of cryptanalytic collision claims. The Yukon manifest preserves eight exploratory registrations: SHA-256 rounds 31/32 and 37/38, SHA3-256 rounds 5/6, and BLAKE3 roun

Build a Basic AI Agent from Scratch: Memory

12 Oct 2026 13 minute read · Artificial Intelligence Previous parts of Build a Basic AI Agent From Scratch: Basic Agent Tools Long Task Planning Human in the Loop & Security Security II Security III You can find and clone this code in this

Show HN: Reproducing three AI text detectors, and why filtering hurt my model

This package contains the exact review files and scripts used for the experiment described in the accompanying article. It lets you rerun the detector comparison, the sentiment-classifier evaluation, and (if you download the source archive)

Best Free AI Tools 2026: 7 Worth Using and Their Real Limits

There are now thousands of tools with “AI” somewhere in their name, and almost every week another one promises to transform the way we work. The problem isn’t finding AI tools anymore. It’s finding the ones that are actually useful. You als

Amazed and Disappointed by AI Troubleshooting

Strange AlertI was ready to wrap up my day, but you can imagine what happened. An alert related to an important server fired! Here we go again …The server was part of a stateful cluster, and being absent could cause serious issues if contin

What happens inside AI Chip when you ask a question

What happens inside AI Chip when you ask a question

Self-hosting an AI agent environment with OpenCode and shared memory

I planned this article on my phone during a train ride to Milan. Back home, I typed one word and it went live. Here's the setup behind it, starting with a little context. Today's frontier models are widely available through APIs, and a comm