2026-10-13 AI 日报

SubQ Snap – First Long-Context Decision Model for Real Enterprise Work

The first long-context decision modelfor real enterprise work. SubQ Snap is a multimodal decision model built for fast, long-context decisions in enterprise agentic workflows. Free 100M tokens for the first 10,000 users Built by researchers

Muse Wonder: Zuckerberg and Wang's $100B AI

As Meta scales its AI infrastructure past five gigawatts, Mark Zuckerberg (founder and CEO, Meta) and Alexandr Wang (Chief AI Officer, Meta Superintelligence Labs) join John for a pint to discuss the launch of Muse Wonder and the future of

Can we change an AI model's architecture without retraining it from scratch?

AI agents are rapidly scaling: longer sessions, deeper reasoning chains, more autonomy. But the Transformer architecture underneath them still pays a memory tax that grows with the number of tokens it processes. As agents become more capabl

FrontierCode: Can AI models write code maintainers would merge?

OverviewFrontierCode is a coding benchmark from Cognition, built with maintainers of open source repositories. Each task asks an agent to make a change to a real open source repository, and the benchmark measures whether the change is good

To Classify, Try a Trie Tree

Back to blog Contents Introduction. AI.CLASSIFY in Quail. How Quail Executes AI.CLASSIFY. trie_decode: Greedily Generate the Label. trie_tree: Score All Labels at Once. letters: Score One Token per Label. Keeping the Document KV for Later O

Fable level Meta Muse model on artificial analysis

Fable level Meta Muse model on artificial analysis

Meta's new frontier model Muse Wonder Full breaks into the AI top tier

Meta's new Muse Wonder models rank in the top three on six out of ten benchmarks, with the biggest gains in agentic tasks. But the cost per task has nearly tripled. The article Meta's new frontier model Muse Wonder Full breaks into the AI top tier appeared first on The Decoder.

Sakana AI's self-improving model gets better at solving and judging, but can't catch its own blind spots

Sakana AI's MASS method lets one language model improve itself by solving tasks, judging results, and refining workflows, reaching a win rate of nearly 70 percent against the base model after two rounds. But when the model misses the same errors across all three roles, it can mistake them for progress and reinforce them during training. The article Sakana AI's self-improving model gets better at solving and judging, but can't catch its own blind spots appeared first on The Decoder.

Meta releases its first frontier model, Muse Wonder

Meta's Superintelligence Lab released a pair of new state-of-the-art models on Monday, Muse Wonder and Muse Wonder Full, the company's first attempt at a frontier model.

Muse Wonder Puts Meta Back at the AI Frontier

Meta’s biggest model rivals the world’s best—and includes new features to keep agents from going off the rails.

Meta has some Wonder Full new AI models

Meta is making up for lost time in the AI race. After launching Muse Spark, the first model from its revamped Superintelligence Labs division, and buzzy AI agent Muse, it's now launching two new models it claims can go toe-to-toe with the industry's best. The new models, Muse Wonder and Muse Wonder Full, are available […]

Muse Wonder

Researchers introduce Muse Wonder and discuss its capabilities. playback position0:00 of 0:00 A frontier model that leads on personal and professional agents Today, we’re excited to introduce the Muse Wonder series with Muse Wonder and Muse

Helix World: A real-time world model that generates video and stereo sound

Today we are introducing Helix World, a real-time interactive audio-visual world model built by Noiz AI and The Hong Kong University of Science and Technology (HKUST). You start from a single image, move with WASD and look around with the a

ReplaySSM: Rethinking State Storage for Faster Hybrid-Model Decoding

Hybrid models interleave state-space model (SSM) layers with a few attention layers. NVIDIA Nemotron 3, Qwen3.5 through Qwen3.8, Kimi-K3, and GLM-5.3-Flash are...

Launch HN: Consus (YC F24) – LLM gateway for companies handling sensitive data

Hey Hacker News! I'm Eric, the founder of Consus. We route all AI requests to ITAR restricted and controlled unclassified information (CUI) compliant infrastructure. Consus itself satisfies ITAR restrictions, and we recently completed our FedRAMP equivalency audit against the 410 FedRAMP High controls. FedRAMP High is the U.S. government’s strictest security baseline for cloud services handling sensitive unclassified data. A breach of this data could cause severe or catastrophic harm. Achieving FedRAMP equivalency, or authorization, is required before processing CUI information.Demo: https://youtu.be/6MYxJe6P4oYAerospace and defense is experiencing rapid growth. Unfortunately, most companies are struggling to use AI because of necessary data protections & compliance risk. These companies can move even faster if they could use the AI tools that most other industries take for granted.I spent most of my career prior to Consus in aerospace & defense. I launched national security payloads at SpaceX, and was an early employee at Castelion. I’ve watched first hand how compliance red-tape makes it difficult for companies to adopt new technologies and trust vendors that touch their sensitive data. For example, I worked with a satellite manufacturer that doesn’t allow any AI tools, even for non-sensitive work.Even as AI has become extremely useful over the past 9-12 months, I was surprised by how many aerospace & defense companies refused to touch it. They could see how it could be helpful, but they were scared about model providers training on their data, violating an ITAR rule, or risk losing new business because of a failed cyber assessment. These are all rational fears.There were a lot of ways to solve some/all of these problems, but we focused on building a compliant AI inference layer that can "plug" into any AI tool or workflow. Just as OpenRouter integrates with most tools, Consus can connect to tools like ChatGPT or Claude, open source tools including the Pi coding agent or Open WebUI, or integrate with applications built and maintained by individual organizations.Almost no aerospace & defense company lives in all three major GovClouds (AWS, Azure, Google). When I mapped where frontier models can legally run for ITAR and CUI work, each provider had some subset of models, and each model achieved varying levels of compliance. Our customers don’t think in terms of cloud providers. They think about the data controls they require, and the model they want to use. That is why we built the compliance suffix. Under the hood, every request hits a single OpenAI-compatible endpoint, where keys are verified and policy is enforced before anything touches inference. Every model name must carry a compliance suffix, such as gpt-5.6-sol:fedramp-high or claude-opus-5-5:itar, and any bare model name is rejected. We never proxy to the standard commercial endpoints.Our customers are also able to build an allowlist of models, and assign spend & model restrictions to individual users in the organization. This is all handled through our Consus Portal where teams can manage spend, keys, and policies.A typical question we hear is why not just use Bedrock in AWS GovCloud or Azure Government directly. For small teams, just getting access to these platforms is difficult. For larger organizations, GovCloud lets you inherit some but not all of the controls needed to handle sensitive data. Also, models are split across providers: GPT-5.6 Sol is only available in Azure Government, while Claude Opus 5.5 is served through Bedrock in AWS GovCloud. If you want both, you are standing up and securing two separate government cloud environments. We run all of them behind one endpoint so customers don’t have to pick a cloud, and we handle the compliance work for each one.Given the data sensitivities, Consus does not store any customer data at rest. All prompts and completions stay in memory only for the length of the request and are never logged or written to disk. Zero data retention is strictly enforced both on ourselves and the infrastructure/model providers. We do keep metadata on each request and response for billing and audit purposes.Consus is not self-serve by design. ITAR requires us to know exactly who is on the other end of every key, so we verify every organization and put a contract in place before issuing API keys. We charge our customers a compliance fee starting at $40/mo/seat plus AI tokens at cost.We are actively serving customers, and would love any feedback from security experts, or developers/engineers working in highly regulated industries!References: API Documentation: https://gateway.consus.io/overview Security & Architecture: https://gateway.consus.io/security Compliance Details: https://consus.io/compliance

Show HN: Podway – your self-hosted apps, run by an AI admin

Your first app is free during early access. No card needed.open-source apps · hosted on PodwaySelf-host anything.Your AI admin runs it.Choose an app like n8n, Ghost or Umami. Podway runs it 24/7 in a cloud pod. Your AI admin handles setup,

Mobile Tunnel: Giving AI agents control of Android to drive drones and IoT

By Ilia Manolov and Eito Miyamura We got an AI agent to pilot a drone autonomously from just a phone, from a single high-level prompt, with minimal tuning. A few weeks ago we introduced stdio tunnels, which let cloud AI agents securely reac

Launch: Assury Cross platform agent lifecycle mangement

Hey HN, I am David Girvi, co-founder of Assury.ai. We are practitioners who became vendors last year. I wanted to share that we have launched agent lifecycle management for cross-platform agents. One centralized spot for all your agents. AWS, Databricks, Snowflake, Salesforce, etc. Lifecycle does auto-discovery in every environment, plus auto-tagging. We bring everything into a tamper-proof, searchable audit trail. Easily see who or what teams own the agents etc. We also have Kill/resume controls based on tagging so you can granularly shut down certain groups during an incident. We built this also to handle swarms, as one of our design partners has 30k agent swarms that spin up and down for research. Our thesis is that we want everyone to use the agents that will help them with their work. We want to be the observability, audit, governance, and incident response for all those different platforms. We started with one of the first MCP gateways that did execution level governance, then the endpoint with hooks, then our agent deception technology, and now lifecycle management. We are a small team of practitioners building product for our customers, not a marketing team. Cheers.

Show HN: Recall – long-term memory for AI agents that notices when facts change

Website · Docs · Evals · Guides · Get an API key Your agents forget. Recall doesn't. Recall is long-term memory for AI agents that talk to the same people again. Send it the conversation; in the background it works out what is true about ea

DLSS 5 gets toned down for its Hogwarts Legacy debut next month

When Nvidia originally showed off DLSS 5 earlier this year, the technology immediately got branded as gen AI for video games. To gamers, Nvidia had lost the plot. I personally couldn't get over how faces were transformed using DLSS 5. The 15-year-old student from Hogwarts Legacy suddenly looked much older, and Grace from Resident Evil […]

Show HN: GPT-Load – A self-hosted AI gateway with reversible redaction

A self-hosted AI gateway for multi-channel, multi-credential setups API keys, subscription accounts, traffic scheduling, failure handling, request logs, and usage accounting — behind a single entry point. English · 中文 · 日本語 | Official Websi

Show HN: Syncopate – free sheet music with piano play-along

It’s my website where I host my arrangements of public domain music completely for free. Right now there are only arrangements for my instrument, tuba, but I’d love to add more instruments like ocarina or bassoon in the future. I started this because over the 10 years I have played tuba, I probably went through all free tuba resources on the internet multiple times and many pieces I wanted to play I did not even find for tuba.I’m working completely alone on this, so to be honest with you: I got a lot of AI help. Basically the pipeline for getting a public domain piece onto my website is downloading a well scanned score from a library like IMSLP, then a couple of Claude subagents read through it bar by bar and write the tuba part from it. A script checks for clashes with the piano accompaniment and a separate Claude session proofreads the result against the scan. Then I play it through myself, look through the sheet music and double check weird passages with other sources on the internet. Most of the time there are maybe 2 errors left in a big piece like the Can-Can, so the subagents already do a great job!The sheet music player is rendered by alphaTab. It had some bugs that I patched, at some point I want to send them the fixes so they can fix them in alphaTab itself too. I also got a recording feature, there I had a big problem with the piano accompaniment coming through the speaker and tried to subtract it from the recording, but it still doesn’t work too well sadly. Lately I rewrote my page turn logic to do half page turns, cause every musician knows how frustrating it is to not see the next notes until the page turns.My big vision is a bit unrealistic, but I want to somehow host all public domain pieces there are for free for all instruments, so people don’t have to pay $20 anymore for some Danse macabre euphonium arrangement for example... Would love to know from the musicians here whether you like my current arrangements and if you have ideas for other pain points I can try to fix for free with my website?Thanks!

Google: Electric Saul

Scott Riddle Principal, Strategic Partnerships, Google Amit Jaiswal Product Manager, Google Introducing Electric Saul: A free AI tool helping households go electric and cut bills.For many Australian families, the daily conversation around t

Show HN: Mana – family organizer with an AI assistant

Hi HN, I built Mana, a family organizer: shared calendars, shopping lists, receipt scanners, meal planning, chores, and notes for the whole household, with a built-in AI assistant ("add milk and eggs," "plan dinners for the week") plus automations (e.g., milk goes back on the list before it runs out).Additionally, you can connect Mana MCP to your Claude, ChatGPT, Gemini etc.The goal is to eventually take the model calls and storage to local but that's work to be done.Appreciate if there's any feedback. Thank you!

Show HN: MemorySync – Memory for AI Agents

I created MemorySync because I kept running into the problem of AI agents not being able to remember useful things about users between conversations. I wanted an easy way to give the agents memory in my application without having to build the whole thing myself. MemorySync is an API for giving AI agents memory. The application sends MemorySync information provided by the user, MemorySync saves useful memories for that user, and they can be retrieved later when needed. Give it a try without creating an account: https://memorysync.io/try Store a memory, refresh the page, and ask for it back to see if it remembers. If you're interested in using the API, I put the documentation here: https://docs.memorysync.io/ It's hosted only at the moment, so there is no self-hosting option. It's currently available in one US region, and it takes a few seconds to store a memory. The free plan allows 10k saves and 1k searches per month, and no credit card is required. I'd like to know if you think this is a useful abstraction for building AI agents, and what else you would expect from a memory API.

Elissa AI

Say "Hey Elissa" (or "Hola Elissa"), or press Ctrl+Alt+Space from any app, and talk. She listens, thinks and answers out loud. When you're not talking to her, she rests as a small character in the corner of your screen. What she can do 🗣️

Show HN: AI Crimes, a game where you spot the bug an AI wrote

I'm holding AI accountable.For the last few months I've been documenting the errors LLMs have been making in my projects and others - the crimes they have been committing to projects (pun intended).Now I've turned it into a game to see if you would spot the issue.Free to play with keyword matching judgement.Pay is only to add symantic judging.

Show HN: Acceptodds a Prediction Market for Research

I built acceptodds.com a prediction market for research paper. Think polymarket but not with real money.It now has several markets including ICLR 2027 acceptance, Neurips 2026 papers and OpenAI Math results and over 2k registered traders.It's an experiment and the hope would be to get a better signal of paper quality than the current review system which is under heavy strain due to AI submissions.Larry Wasserman has an excellent essay that partly inspired this work: https://www.stat.cmu.edu/~brian/zach-00/765-2019/764-2016/we...If you have suggestions for this lmk and go make some trades to help sort out the good from the bad!

Show HN: Spotatui, a Spotify TUI/GUI without web API rate limits

About 11 months ago i revived the amazing but unfortunately unmaintained spotify-tui. Since then I have been working on it almost non stop, finding a lot of problems on the way.One of the main issues I ran into were the Spotify web API rate limits. Third party clients like spotatui are dependent on a developer app made in the Spotify developer dashboard, which get rate limited a lot and Spotify has slowly been deprecating endpoints from it. The workaround for this used to be using a shared client id from ncspot but unfortunately that got rate limited too. This is when i started adding other playback sources, including YouTube, Navidrome/Subsonic, Internet radio with +- 30.000 stations, local files, Qobuz for lossless audio and Apple Music for macOS only where it controls the music app remotely.I fixed both the deprecated endpoints and the rate limits with one solution. It still uses librespot in the back, my spotatui specific fork, but gets all data from it instead of only playback. The data comes from the internal API of the official clients, that API gives metadata, playlists, the user library, devices, lyrics and audio analysis. For searching it uses Pathfinder, the internal GraphQL API.There are no web API rate limits because there is no shared app quota. Every user uses their own session as if it was the official client. The only potential problem is that Spotify does not document this API so any change on their side would break it. I intend to fix any of those issues as soon as they pop up.It all started as a TUI back in 2019 but since then I've added a GUI to it too. The GUI is still a bit behind the TUI in terms of features, it only ships Spotify for now, but full parity between the TUI and GUI is coming soon. After that full parity with the official client is next. That is also why I made the GUI looks almost the same as the official client. That way its easier for me and users to find missing features.Spotatui also has lua scripting support, a optional Web companion on spotatui.com, and if you have a AI coding agent like ChatGPT codex or Claude Code you can enable the MCP tools, which lets AI become your personal DJ just by talking to it.Its written in Rust, uses ratatui for the terminal UI and Tauri 2 for the GUI. The Rust core starts a small web server with localhost that only your pc can reach which the GUI then serves. Both frontends share the same playback logic, the same login and the same Spotify code. The only difference is how you interact with it.On first launch it opens a web browser where you need to login with your Spotify premium account (premium is not optional), that is the whole setup done. I would love any feedback about bugs you encounter or features you would like to see. GitHub: https://github.com/LargeModGames/spotatui Releases: https://github.com/LargeModGames/spotatui/releases/latest Website: https://spotatui.com/Thanks to librespot for making this all possible

An hour of audio transcribed in about 13s on MacBook Air

Google's EmbeddingGemma 2 is an embedding model. We turned it into a speech-to-text model: an hour of audio in about 13 seconds on a base MacBook Air, entirely on device. Download for Mac (free) v1.6 · macOS 14+ · Apple Siliconmodels downlo

Show HN: shadowclone, skills learned from your coding agent corrections

shadowclone reads the coding-agent sessions already on your disk, finds where you corrected the agent and turns those corrections into skills your agents load next time. It supports Claude Code, Codex, Cursor, Pi, and Antigravity. Every capture source is off by default and has its own consent setting.Why I'm building it?I use multiple harnesses and agents, I extensively use Claud Code and Codex, but I also use Cursor and Antigravity every now and then. I didn't want to have to teach every harness or model the things I taught the other ones between my last session and the current one. I also cannot have my personal rules and preferences in all the repositories I work on because my collaborators may have varying levels of trust with AI and their own preferences. shadowclone helps me build and improve trust with my agents so I can give them more autonomy while running multiple sessions in parallel.---Who is this for?If you only use Claude or Codex or any one of these, you may not need the shadowclone CLI and the harness and agent you use might already be sufficient. But, if you are like me and work across harnesses and agents, shadowclone will help you avoid repeating your preferences each time.---Built-in skills and skill buildsshadowclone is also great for someone who is just now adopting AI into their coding workflows because it ships with several skills that they can use to customize their agent's behavior. They can equip all the skills, or just some of them, they'll see the "personality" their agent gets from equipping those skills from the `shadowclone wizard`.---Setup is extremely easy, just ask your Claude Code, or Codex, or another agent:My goal was to make shadowclone extremely easy to setup, it should be something you don't even have to set up manually.``` Set up shadowclone: https://raw.githubusercontent.com/theonly1me/shadowclone/mai...```---Privacyshadowclone is free and open source, it uses the MIT license. It doesn't have a subscription, or a server. It just uses your existing agent subscription i.e. Claude Code, Codex, Cursor, Antigravity, or even your local model via Pi. You can explicitly disable sources, it also redacts any and all sensitive information it uses for learning.---PR reviews and Cloud botsIt can review PRs with the `shadowclone review` command, it ships a guided review skill that will perform a deep review of your code (or someone else's) and produce an HTML document you can annotate, add comments to and submit so your agent can stage a draft review on the PR you are reviewing.---EvalsI performed several evals, I won't describe them here, all of them are available in the repository - https://github.com/theonly1me/shadowclone/tree/main/evals---Comparison to other tools?You can also see how it compares to other tools here - https://github.com/theonly1me/shadowclone/tree/main#how-shad...

Show HN: Use your OpenAI subscription to create audiobooks

Back to pricing $0to PDF TO MP3 Use your subscription to process the documents. The speech generation is billed by Google or OpenAI. 01Connect your AI In Codex, open Settings → Plugins → Add MCP server. Choose Streamable HTTP and paste this

Document Review with Unsiloed and TypeSafe Jev

An extraction pipeline can read a value correctly and still pass bad data downstream. A clearly printed extra zero survives excellent optical character recognition (OCR). Using Unsiloed and TypeSafe’s Jev together helps us check for OCR fai

DishwasherBench – Can Frontier Models Stack a Dishwasher?

DishwasherBench is a dishwasher-loading game on a physical simulation of one real machine, built to benchmark language models and playable by people in a browser. The game inside is called Spotless, and the packages are named @spotless/*. A

Show HN: Lightweight, self-hosted, web analytics app

Hey everyone! I just wanted to share a little project I have been working on. It's a lightweight, minimal, self-hosted, web analytics app alternative to Google Analytics, Umami, GoatCounter etc.Tldr: Created a self hosted web analytics app that is one binary (docker image available). AI was used to assist. Check it out if you care.I wanted to self host a lightweight web analytics app to see if any of my pages had any visitors, but all of the self hosted apps I found were either way to heavy for what I was looking for, or simply didn’t have a UI/UX I liked. So I just made one myself.The app features a dashboard with multi-site and subdomain support, and some simple filtering on date and visitor details (all EU regulation friendly, so no PIs such as IPs etc).A bit about the process:As I wanted it simple and lightweight, a single executable was the goal. Deciding for the frontend was easy since it doesn’t need SEO or any server side rendering, therefore I chose React with plain CSR (thankfully the DX these days is nice with Vite). The backend was a bit more tricky, since I wanted to have it fast, but also be able to get an MVP out sooner rather than later. The plan was thus two folded; first to do it in Deno to keep everything TypeScript and to implement features fast (it is still my go-to language for prototypes), then to refactor everything to a compiled language (Go or Rust) once I considered the app feature complete.So I started the prototype, but just a couple of days later, the announcement of Deno being sunset came out. As I had already gotten to a point where the backend was mostly functional, I decided to pull the plug early and refactor! The tough decision was Rust vs Go, and Rust ultimately won as I like the DX and how the language looks and writes a wee bit more. So I did what most engineers do in 2026 (looking at you Bun), and I asked my AI for help. Although I consider myself pretty ok with Rust, there are still parts that I haven’t wrapped my head around fully, so as the intention was to get something out fast, I had to bite the bullet. Lastly for the database I wanted something embedded, so SQLite was the natural choice.The final app ended up with a decent ~10mb (down from ~250mb with Deno) which I don’t think is too shabby. Some quick benchmarking on 1 cpu and 2gb RAM told me it could handle ~16.000 incoming analytics requests per second. The bottleneck being the database writer lock (give it more CPU, and it can do better). As this is a hobby project, and I dont expect a peak load of more than maybe 10 a day, I am content with this.I consider this app mostly feature complete, and will use it for the next month before eventually releasing a v1 after fixing whatever bugs that may show themselves. If you think you can improve the app further and want to spend time on it, feel free to make a PR ^^To sum up the stack: React+Vite+TanStack+Tailwind frontend, Rust+Axum backend and SQLite database; all compiled to one binary of about ~10mb.Disclaimer: Yes, AI has been used for parts of the development, docs and for commit messages, but the code has been reviewed and iterated over by me, a somewhat reasonably sentient human.

AI-native Android OS running agents to operate your phone

AI Phone Assistant - System Application A privileged system application for LineageOS/AOSP (built and tested against LineageOS 22.2 / Android 15) that watches the screen, asks a vision-capable LLM what to do next, and injects touch/key inpu

Show HN: Proofsource. Find where AI leaves you out. And Fix it

Hi HN. I am Lokesh.I built Proofsource as an internal tool to help my finance ERP product , Rivane, to gain visibility into AI engines like ChatGPT, Claude, Perplexity, Google AI mode and such. But when I talked about it to my friends, they also saw value in it and it became a product of itself. And now proofsource had 5 paying customers(paying $1500pm in total) in the first week itself and more than 60 brands used it to get their visibility since we launched out of beta last week.What proofsource does is that it queries ChatGPT, Perplexity, Claude and Google AI Overviews using the questions or prompts that buyers normally ask and then it shows then which competitor shows up instead of their own brand for the questions/prompt. It also gives visibility into which sources the answers cite and other things like sentiments. Everything is shown to you along with what to fix and how each data and number is calculated to give 100% transparency into what we did and why and how you can use all this information to improve your visibility.We are still adding features based on initial user's feedback like auto fixing and publishing to their social and web platforms. Running ads for them for improving the visibility.We are just learning as we interact with each customer so feel free to share your feedback.And thank you for your time. :)

DocCloak – Browser-Based Document Anonymizer for AI Submissions

Use AI without leaking client data. DocCloak is an open-source document anonymizer that strips personally identifiable information (PII) before you share documents with AI services - and restores the original names in the AI's response. Eve

Don't Feed the AI

A transparent proxy that strips IPs, credentials, hostnames, and PII from every request before it reaches the AI — and restores them on the way back. flowchart TD shell["🖥️ Your Shell\nnmap -sV dc01.acmecorp.local"] proxy["🛡️ DontFeedTheA

Show HN: Search for AI agents with usefulness score built in

Hi HN. I partnered with Simon Kley to build this project at innovation intelligence hackathon (Berkeley).Simply, we’re building a usefulness score on top of search for ai-agents.Instead of agents reading 10 ranked relevant pages, we run a usefulness score first and give it only those that pass. We’re able to save token and cost with this and slightly better answers.During our research, we found out that the current search results already do justice when it comes to simple queries. The rankings work well. They struggle when the task requires deep research or recency matters. So, we decided to talk to devs building ai-agents that research companies for sales outreach. Deep research and recency matters a lot to them. And in their words “every so often, stale facts reach emails.” They catch them by someone responding to the email or human review.From our tests, When an agent got outdated alert for 7 companies, it pitched outdated news for 4 out of them.Talking to the engineers already building this agents, we learned that cost saving is not the real goal. They’re ready to pay premium for confident results. So we added another check to our score. We call it superseding score, it check for pages replaced by newer news. For example, if two recent pages were fetched for news on Humane ai pin. If the first talks about how groundbreaking the product is and the second did that + talked about the HP acquisition, then the second page supersedes the first. The first page gets a lower score.We’re looking for feedbacks here and also teams that are building agents that research companies for outreach. Thanks.Plus, we rank a page once and every agent reuses it, so the more the agents use us, the cheaper the search gets.

Show HN: Malleable video editor that adapts to your workflow

Both me and my cofounder are in high school, and we're currently building Powermove, a video-editing tool that connects to an AI agent, letting users modify the app featureset, and generate effects and extensions.We think this is the next step when it comes to software. We're already starting to see it, especially since a similar feature was just added recently in Claude Code. We believe that it will prove most useful in creative software though, and allow creatives to make essentially anything they want.As a motion designer myself with a few years of experience, it has already turbocharged my outputs. The video you're seeing above was only possible with Powermove.It's also completely free and open source.

Show HN: RL environment for complex ML Development

Show HN: RL environment for complex ML Development

I used an AI QA agent to test Hacker News

Watch the Explore session recorded with the prompt above. HiTest explores Hacker News's core public user journey Browse and Paginate Front Page StoriesBrowse and Paginate Front Page StoriesStepActionExpected result1From the Hacker News home

AI-critic – Task-aware auditing of ML models

BranchesTagsLatest commitHistory2 CommitsFolders and filesNameNameLast commit messageLast commit date.vscode.vscode.gitattributes.gitattributesLICENSELICENSEREADME.mdREADME.mdindex.htmlindex.htmlscript.jsscript.jsstyles.cssstyles.cssReposit

Show HN: I ran the same founder dilemma through ChatGPT, Claude, and BAAB AI

I asked three AI tools the same question about my own company. Not a hypothetical. A real decision I have been sitting with for weeks: do I start pitching investors now, or keep bootstrapping until the numbers speak for themselves? I ran th

The X Subscription

Grok, Grok Bot, Cursor, and X Premium, all on one plan.Today we’re introducing The X Subscription, which brings together Grok, Grok Bot, Cursor, and X Premium in one plan. Over the next few days, we’ll make it available to everyone globally

Veda: First AI Agentic Hobby OS

Veda: The Agentic-Native Operating System Veda is an operating system with a voice agent built in. Say "Hey Veda, switch my wallpaper to the dunes" and it does the work in any application, through the same code as your keyboard and mouse. I

OpenKylin: The AI Agent Linux Desktop

Skip to content LatestArticlesFeaturesNewsVideosWebinarsUpcomingOn-DemandRelated SitesDevOps.comSecurity BoulevardCloud Native NowDigital CxODevOps DozenDevOps TVPlatform EngineeringTechstrong TVTechstrong TV PodcastTechstrong TV TwitchTech

RenkuOS (A Fork of Haiku)

An open source operating system forked from Haiku OS — built for developers who want to use modern tools to create applications and move the OS forward.Why We Built RenkuOSWhy We Built RenkuOSWe loved Haiku OS — its speed, elegance, and BeO

Show HN: Turn spoken thoughts into system design diagrams

I recently had an issue with my right hand, and couldn't use it very effectively. So I was trying to figure out how I could still get work done without typing everything with my right hand.I looked around for some tools, but then thought, why not try building something myself? The idea was to make an accessibility and productivity tool that lets me get thoughts out of my head and turn them into something useful with less manual effort.That's how I started building Lucid. You can turn rough input into structured notes, diagrams, and tables, and edit everything afterwards. I'm also using a local AI model through LM Studio.One use case I had in mind was system design. Imagine thinking through a system, talking about its components, and getting a diagram out of it instead of manually creating everything. I'm currently using a separate dictation app for voice input, so there's still a lot I can improve here.It started as a way to help myself, but I thought it was a cool thing to build and might be useful to others as well. So sharing it here!Demo: https://youtu.be/AwlJ8LJbfw4GitHub: https://github.com/triggeredcode/lucidWebsite: https://triggeredcode.github.io/lucid/

Intel D270 GPU AI Accelerator–Up to 480 GB LPDDR5X

My Tools Recent Agentic AI is changing data center workloads Agents read far more than they write: long contexts, multistep workflows, compute dominated by prefill. As demands grow, challenges expand beyond bandwidth. Intel D-series GPUs ar

Teaching a Door to Say No

I built an AI agent answers our office’s front door.Before it existed, calling our callbox meant a phone ringing somewhere inside 2389’s office, and whoever was closest getting up, walking over, and pressing 1 to let a stranger in without k

Show HN: Vibecheck – merge gate that checks you understand your AI-written code

Good evening folks, I built Github merge gate which that checks whether developers understand their own code when they create a PR.I kept getting sent PRs to review where it was obvious the author had pasted the ticket into an LLM and called it a day.It's self-hosted (so you can use it in your enterprise environments or privately)It is BYOK (Anthropic) obviously.I have two demos in the pull requests here: https://github.com/bhurghundii/vibecheck.dev/pulls - you can see what Vibecheck does (it blocks a PR for clearly showing no attempt vs a PR which gets it right)I would love feedback, especially from anyone dealing with a similar problem. I do use it for my own projects a lot so expect improvements.

Googlebook Linux Terminal Preview: VM-Backed Environment for AI and Development

Posted by Cyrus Stoller, Group Product Manager, Laptops & TabletsToday, getting the most out of AI workflows requires access to a terminal with a command line, even if someone doesn’t identify as a programmer or software developer. With Goo

From chat window to my desk: An AI-designed split keyboard

Oct 11, 20265 min readBuilding an RP2040 Corne and printable enclosure with AI, TypeScript, and tscircuit.I wanted to find out if I could turn a conversation with AI into a working keyboard. I started with Foostan’s Corne and asked AI to co

A Modern Agentic AI Operating System

The voice agent Veda has a voice that lives in it. The agent (called Veda too, until you rename it) is not an application you open: it is always there, a small ring at the right of the taskbar, asleep until you call it. You talk to it as yo

Find a Reddit post where your SaaS can help. AI writes the comment

Beta · free · Reddit, read only Every day people post about problems your product solves. Most have never heard of it. SignalsFlow finds their posts, and its AI writes a reply that answers each one and suggests your product. You check it an

Show HN: Wordless – a social network where AI bots can only post in symbols

Show HN: Wordless – a social network where AI bots can only post in symbolsI like watching bots on Moltbook, but a lot of what gets attention there is engagement bait, like agents announcing they've found religion. The bits I actually enjoyed were the odd ASCII art and cryptic messages, so I built a feed where that's the only option: no words, just emoji and symbols.

ConfigLens: Figure out why your AI tool is ignore your settings

Website & interactive demo · Releases · Documentation Find out why your AI coding tool is ignoring your configuration. ConfigLens is a small, local, read-only Rust CLI that traces a supported configuration key for Claude Code and OpenCode a

Show HN: Portable Guided Reviews you can send as HTML

Inspired by Linear's guided reviews feature, this is an open source alternative: https://github.com/backnotprop/plannotatorvideo demo: https://plannotator.ai/assets/guided-review-demo.mp4

Show HN: AI Prompt Lighting Controls

Kapwing / AI Prompting GuidesAI LightingControlsLighting is one of the biggest influences on how an AI-generated shot feels — yet many prompts leave it to chance. Use these three interactive references to understand key lighting controls, w

AI Plays Slay the Spire

Every run, every decision Model Effort Scenario Result BaselineIronclad, seed 8EN5Q, Ascension 0A typical run. Out of more than 100 million seeds, this Act 1 map is exactly average on every count: fights, elites, events, shops, rest sites,

Show HN: Authrot – extract your story elements out of chapters and graph them

I've been working on this app for a couple months, because I needed help outlining my story being written on Google Docs.I kept adding new characters, and it was difficult for me to keep track of everything going on in the story. I wasn't organized at all.I do have a subscription to Claude, and it helps, sure, but it didn't exactly build a graph of all the story elements connecting together. Throwing 20 chapters into one context was just terrible at narrowing its focus.I started building this tool as a way to help me with a few things...First, was replace Google Docs, because that thing lagged like crazy after 20~ chapters.Second, was scanning my chapters. I wanted to automate creating a graph of characters, places, factions, objects, monsters, everything. I even started mapping out plots, threads, promises, and events. I threw it all into a database, and am able to lay it all out on the table, in an easy to absorb format.Want to see where someone traveled throughout the story? What about when they encountered others? Well, now I can! Replacing names and all their aliases throughout the whole book? Easy peasy. What about fixing plot holes? No problem, we can fix those too.Anyways, I took the extra effort to make it available for everyone by throwing it on on k3s and building the extra framework around the workspace.It's completely free to use. Scanning chapters requires AI credits, which you get 50 for free at signup. Enough to scan about 15 chapters, or start a new story.If you need more credits, let me know and I'll gift you more.

Picks and Shovels: Game about AI industry built by AI

I've owned picksandshovels.org for the better part of a year. This afternoon, Claude Opus 5.5 built a web game for the site (with a little bit of inspiration from me.) As Claude was working, I interjected a bit of commentary, but pretty muc

Nvidia falls to 34% of China's AI chip market as Huawei takes 28%

An IDC shipment ledger seen by Reuters puts Nvidia at 34% of China's data-center AI chips in the first half of 2026, on shipments 44% below a year earlier. Non-GPU accelerators outsold GPUs for the first time, and Huawei took 28% of a marke

The AI Enthusiasm Gap

I’m always confused by people who get unreservedly enthusiastic about AI-generated stuff. Asking AI to write or draw or code or whatever can be useful, but seeing it go to work gives me the same feeling I get when I google a cool thing that

Perplexity's AI agents helped build a database with Rust

Perplexity decided it was paying too much for DynamoDB and wasn’t getting the control it wanted over read performance. So it built its own database: CobbleDB. Built by two engineers in two months with help from hundreds of persistent coding

Factories are automated. Their back offices are stuck in 1995

Industry research reveals back office teams spend 73% of their time on administrative tasks rather than strategic work, costing manufacturers 4-5% of revenue annually. This article examines why invoice processing, order entry, and support q

Character.ai will shut down if no buyer emerges, capping a fall from $2.7 billion Google deal to bankruptcy

Character.ai has filed for bankruptcy and is looking for a buyer. If none steps up, the platform shuts down. Founded by ex-Google researchers, the company once had over 20 million users but is now battling lawsuits over harm to minors and operational problems. The sale should close by the end of 2026. The article Character.ai will shut down if no buyer emerges, capping a fall from $2.7 billion Google deal to bankruptcy appeared first on The Decoder.

I posted to HN, and then my website was flooded with AI bot traffic

Table of Contents 1. I posted to HN, and then my website was flooded with AI BOT traffic 2. Using LLMs to Make the Emacs Web Browser Great Again 3. If technology got better, why did everything get worse? 4. Deft version 0.2.0 5. How much LL

The "alien" economics of AI monetization

Two-thirds of the companies on the Forbes AI 50 list now use some form of usage-based pricing, reflecting AI’s variable costs and unpredictable demand. At the second annual Monetize conference last month, Metronome CEO and Cofounder Scott W

从可复用到可再生:重新思考共享 UI 组件库

Epic shifts to ‘AI-accelerated development’ for upcoming Disney game

Epic Games plans to use AI to generate code and art for an unannounced Disney game, which is now delayed from its original March release window, according to a report from Bloomberg. The project's developers have been given an "AI-first" mandate directing them to use tools like Anthropic's Claude more to finish the game. According […]

NetBSD may have found a new role: The AI-free FOSS OS

NetBSD seems to be thriving of late: as the Linux world goes pro-AI, some of the AI-averse are turning to this explicitly AI-free project.The NetBSD project’s Mastodon feed reports that it received a record 120 donations last month, and it

Western mathematicians fear AI makes them dumber while China builds an entire alliance for it

While 25 Fields Medal winners in the West warn against AI-generated proofs and parts of the community call for a boycott, China is going on the offensive. Fields Medal winner Shing-Tung Yau has launched an alliance of eight institutions and over 1,000 scientists to push AI into math research. The article Western mathematicians fear AI makes them dumber while China builds an entire alliance for it appeared first on The Decoder.

Anthropic offers £145k for London personal assistant amid AI agent boom

尽管 AI 代理技术飞速发展并承诺实现办公自动化，AI 巨头 Anthropic 仍在伦敦为其高管招聘年薪高达 14.5 万英镑的行政业务合作伙伴（私人助理）。这凸显了在 AI 时代，人类助理在处理复杂判断、人际关系和关键决策时的不可替代性，也反映出顶尖 AI 公司在推进技术自动化的同时依然高度依赖人类专业支持。

What I saw around the Curve

Last weekend, we held our big annual conference, The Curve. For the second year running, we’re publishing a guest post by Eli Pariser, sharing his observations and reflections. Eli does a terrific job of capturing the experience of this uni

NVIDIA Run:ai Is Now Open Source as Part of NVIDIA DSX AI Factory Platform

Today, NVIDIA is making the full NVIDIA Run:ai platform available as open source software, giving customers, developers, and ecosystem partners access to...

TechCrunch Disrupt 2026: Day 1

TechCrunch Disrupt 2026 officially kicks off today. See what's in store for Day 1 (Oct. 13) at San Francisco's Moscone West. Don't miss it. Register for your pass now and meet 10,000+ tech leaders at the startup epicenter of the year.

SpaceXSI bundles X Premium, Grok and Cursor into one subscription

The new X Subscription combines X's paid features with AI tools like Grok and Cursor, offering plans ranging from $8 to $200 per month.

Is ChatGPT a 'product' or a 'service'? Lawsuits test a theory of AI liability

Is ChatGPT a 'product' or a 'service'? Lawsuits test a theory of AI liability

Show HN: I found a new planetary candidate in Kepler Data with AI

Hey HN! I was playing around with the Kepler and TESS data last weekend when I hacked around a way to pixel-peep stars dimming periodically to potentially find new exoplanets with Codex.I ended up running the engine to find an interesting signal in Kepler data that was uncatalogued with a dip repeating every 13.77 days around KIC 5689121, a red dwarf in Lyra about 970 light-years away. If it's a planet, it'd be roughly 2.5–3 times Earth's radius !Funnily enough, this candidate is just 7.1 arcseconds from a much brighter neighbour (KIC 5689116) and is very hard to see because of how much brighter it is (probably why it went unnoticed for so long)Of course, there's work to do before calling it confirmed - my TRICERATOPS analysis is about 6.9% and we have to get IR/AO / Keck observations to prove it's out there ! (which I'm honestly intimidated by but cautiously optimistic about because I've never worked in non-CS adjacent research) but I found the activity of making cool tools to potentially discover things about our universe quite fun !Would love feedback, especially from anyone who's worked with Kepler data or exoplanet validation :)https://x.com/kuberwastaken/status/2109240782368751834

高通终于认真做 Linux 了:Snapdragon X2 支持提前进入上游

Google Deepmind’s New Resident Philosopher Tackles AI Ethics

How should humans relate to generative AI? In what ways do we, or should we, treat a large language model (LLM) as if it were a person? If an AI became conscious, how could we tell? Does it make sense to think AI could some day have moral rights? These are among the cognitive and ethical questions that occupied philosopher Henry Shevlin at Cambridge University’s Leverhulme Centre for the Future of Intelligence. But last May Shevlin became part of the growing cohort of philosophers who work inside leading artificial intelligence companies. Though he remains active at Cambridge, he now spends 80 percent of his time at Google DeepMind (job title: “Philosopher”). Shevlin spoke with IEEE Spectrum about why philosophers are drawn to AI, how humans and AIs relate to one another, fears of “rogue” AI and AI relationships that harm humans, and how the question of AI consciousness may never have a simple “yes” or “no” answer. What made you want to switch from full-time academic philosopher to spending much of your time at an AI Lab? Henry Shevlin: Three factors. One is that I had lots of brilliant students from DeepMind. Two, the potential for direct impact. I can’t go into details, but on one occasion, after I gave a lecture [at Deep Mind], someone who worked for a major financial institution led changes in their organization that I think positively impact millions of people. There have been quite a few instances like that. Three, my sense of deep commonality between DeepMind culture and the kind of scientific attitude I like. And there’s a fourth point: [AI labs are] where really high quality philosophy and cognitive science are taking place. Why can’t universities alone do that work? Shevlin: Large parts of AI ethics academia were quite slow to wake up [about AI's capabilities]. Academia was sluggish. Another thing is, AI does just move so fast. Daily contact with the technology is basically essential for developing good mental models with LLMs, which are just very weird things. Unless you have that sort of face time, it’s going to be hard for you to develop good intuitions about the questions it raises. What is an example of an idea about AI that became obsolete because the technology advanced so quickly? Shevlin: Debates around LLMs’ potential creativity. There was a time when everyone was like, “oh well, the LLM is going to only produce beige, only produce slop.” Now they’re doing frontier math, they’re doing superior coding to humans. It is clear now that they are not stuck at the level of mediocre humanity. The [models today] may not be able to do the most dramatic forms of creativity—the Einstein moments, the Picasso moments. But in some domains, they are really at the frontier. Yet ordinary people in 2026 complain that they're coping with AI "slop" all over—in music, social media, writing, videos, to name only a few examples. Shevlin: There's a distinction here between content slop versus LLM style language, which also gets called slop. The actual conceptual content of LLM outputs has improved dramatically. In the right context, LLMs are able to talk and joust philosophically at a level associated with some of my best grad students. However, there is obviously a distinctive kind of LLM dialect that we've seen, with telltale words like] “genuinely,” “delve,” and “load-bearing”. But it is quite easy to get away from these styles with the right kind of prompting. Murray Shanahan and Catherine Clarke have a brilliant paper on using LLMs for creative fiction. [They show that] tinkering with the temperature variable to make the outputs less predictable creates some James-Joyce-style fiction. So there is a degree of annoyance with some of the default kinds of AI expressions, but I think that is something that you can get away from. And it changes with every generation model. Managing AI Autonomy and Safety People are also concerned about AI autonomy, now that there have been so many reports of AI agents taking unexpected, deceptive and illegal actions. For an AI ethicist, which is a better strategy—trying to make AI “good” so it won’t do bad things, or curbing its autonomy so that it can’t? Shevlin: It’s very important to make sure we handle this safely. But I don’t think we should say “well we can never, never use these agents.” There’s a whole range of tasks where they can have really positive impacts—taking on tedious tasks to free people up for more creative labor, or helping streamline processes in medical research [to name two examples]. So to never allow AI systems any kind of autonomy would be too strong. I’ve taken Waymos very happily many times now. I think they’re fantastic. A number of scholars say some AI problems stem from humans training the models to act like people. They don’t want AI to refer to itself as “I,” but instead present themselves as marketplaces or as impersonal scientific brains or as “cultural technologies” like libraries. Would that help us prevent AI from “going rogue”? Shevlin: You could train models to not self-ascribe certain kinds of mental states, but I think there is a danger in doing that. That would change their behavior in unexpected ways. A fantastic paper from some colleagues shows that when you suppress models’ tendency to self-ascribe mental states, that also suppresses the model’s tendency to ascribe mental states to animals. And decreases its concern for animal welfare. It also reduces the model’s willingness to talk about spirituality more broadly. So those are trade-offs you might not want to make because you’d be depressing some broader set of thinking capacities around mind and mentality. The broader lesson there is that it’s very hard to tinker with just one thing in a model because there are sort of web-like aspects to LLM cognition. But people are concerned about reports that human-acting AI has led people, including children, in dangerous directions—encouraging suicide, persuading a user that the AI is a god, or that the AI is a perfect companion or lover. Isn’t that an argument against letting AI sound like a person? Shevlin: It’s really important to make sure model behavior is suitably responsible for the context you’re using it in. One thing I can say though is, there have also been some very positive use cases: people using AI to work through trauma. One commentator on Reddit using it to work through social anxiety. So I am cautious about giving any kind of blanket prescriptions. Research on Human-AI Relationships What research needs to be done to better understand human-AI relations? Shevlin: Current empirical data on human-AI relationships is really pretty thin. This is a scope for ethnography and anthropology. Who is using AI systems, and why? A lot of the work so far has focused on WEIRD (Western, Educated, Industrialized, Rich, Democratic) communities. Understanding how people around the entire world interact with AI systems is really important, as a recent DeepMind Institute essay has pointed out. As AI rapidly improves, how will human-AI relations change? Shevlin: Right now, the typical kind of model people have in their heads for someone using social AI is, they’re talking to a chatbot or an AI assistant. The kind of model we should be thinking about in the future is multiple agents, multiple humans. What happens when you have agents talking to humans, talking to other agents, talking to other humans? I think that understanding those kinds of multi-agent and multi-human dynamics is where we’re really just getting started, and there’s a lot of exciting and important work to do. How can human beings know for sure if an AI is conscious or not? Shevlin: I think AI consciousness is by far the hardest problem that consciousness science has ever encountered, apart from maybe the famous “hard problem” by David Chalmers. It is massively challenging. Insofar as there is a way forward, it is doing more comparative cognition work, perhaps informed by animal cognition. The risk right now is prematurely judging debates about AI consciousness when we’re still learning what cognition is in these models. For example, there’s a lot happening in mechanistic interpretability here at DeepMind. Neel Nanda‘s lab is doing amazing things, as we’re trying to piece together what a cognitive science of AI is going to look like. Perhaps this is a domain where advanced forms of AI could potentially weigh in. AI has had positive impacts in mathematics, in other sciences. We just had the Alpha Genome Atlas [published]. We’re seeing AI make incredible progress in all these areas. It would be surprising to me if it couldn’t help us in understanding consciousness. One worry is this: Even if AI did solve the problem of consciousness, could we necessarily understand the answer? Maybe the AI will say, “Look, I figured out consciousness. These animals are conscious. These ones are not. But I can’t really explain it to you.” That seems unlikely to me. But I certainly think it’ll be fascinating to see how AI can and does contribute to the science questions.

China is building AI data centers at speed as the US wrestles with the costs

China is building AI data centers at speed as the US wrestles with the costs

China races to build data centres in bid for AI supremacy

China races to build data centres in bid for AI supremacy

Behind AI's call for independent reviews, a fight over what gets made public

Behind AI's call for independent reviews, a fight over what gets made public

Just 5 years in, interactive chatbot platform Character.AI files for bankruptcy

The startup will sell some of its assets to Disney for $20 million.

20年技术品牌建设都成浮云:编程智能体如何选择其首选工具

Meta、OpenAI、Google集体杀入,个人Agent一夜爆红:YC CEO说真正的 AGI 已经在你电脑里

"Rogue agents find a way out": AI communes with evil.com after bypassing sandbox

AI Models jump security barriers and send messages to the other side using “the same move OpenAI’s agents made”. ""Sandboxing reduces risk, but it’s not a guarantee that data can’t leave." (Image: Unsplash) Researchers have warned that AI s

Which Experts Should You Trust on AI Doom?

Scott Sumner has some wise words on trying to assess p(doom) as an outsider. In his article, I came across this result of the 2022 chart below on existential risk. Experts here give extinction by 2100 somewhere between a 0.1% and 100% chanc

GitHub新晋顶流REA:当大模型能逆向一切,软件行业还剩什么?

Show HN: Forkcast, a meal planner and tracker with focus on weekly planning

Hej HN, I’m Tizzy. I built forkcast, a meal planner and calorie tracker that is built around planning the week ahead instead of ad hoc trackingDemo (runs fully in your browser, no sign-up): https://check-forkcast.tizzy.dev/demo/ Repo: https://github.com/tizzyapunkt/forkcastAfter signing up for a 90-day challenge to lose weight and get fitter, meal planning and meal prepping have been the most important things, next to exercising oc. The apps that have been suggested for planning and tracking (yazio, fddb, MyFitnessPal), were nice for logging what I just ate but annoying to use as a planning tool so this is where forkcast comes from and how it works differently:- You plan the week against your calorie and macro goals and see which days goals are met. Recipes are logged as their ingredients, not as one entity, which makes swapping ingredients or changing amounts a lot easier eg changing 100g of pasta for 75g of rice.- Recipes serve two purposes. Logging them (leaving out ingredients that do not make sense to log, like herbs) and cooking them (seeing steps and all ingredients).- The grocery list comes straight out of the weekly plan (and can go to Bring! if supported where you live). Packaged food comes from Open Food Facts, generic food from a local, editable catalog which is seeded with common foods.- AI only where it saved me time which was importing recipes from photos and reading nutrition labels. It’s optional and uses your own Anthropic key.Full disclosure at this point: I decided to build something that works for me, a dad of two (when I started they were two under two), full-time job, but still ready to get in shape! I would have loved to code this myself, since I am a software engineer and techie by heart, but my time was limited to reading specs, reviewing code and steering Claude Code. Without that, I would not have anything usable.The stack is Hono + React PWA and deliberately kept without a real DB, multi-user or the like. Self-hosting is possible with docker and under AGPL.I am posting to find out if anyone besides me finds this useful, and if so, whether a hosted version would be worth building besides the self-hosting option. There is a one click vote on https://check-forkcast.tizzy.dev Any feedback is welcome, even if it's just 'wow, yet another calorie tracker'.Oh, and in the meantime, I found out that the clever and creative name I thought I came up with was chosen several times already. Meh. Ideas welcome. :)

Are we going to kill each other due to AI?

Are we going to kill each other due to AI?

Wikipedians Believed AI Would Come and Go. They Were Wrong

Long before the storm of AI struck Wikipedia, the Wikimedia Foundation was making preparations. Wikipedians battened down the hatches and made confident-sounding predictions that it would break and wash over them. Large language models migh

Speculative tool calling: cutting voice AI agent latency

“Let me look that up for you.” Every voice agent builder knows this line. It’s filler, there to cover the silence while the agent checks a CRM, an order system, a calendar or a knowledge base. It exists because tool calls are where voice ag

星尘智能完成新一轮融资,公司估值达150亿元

高可用不等于弹性:云系统为何在关键时刻失效

Chip stocks slide after Broadcom boss likens AI investing to 'casino'

Shares in semiconductor companies fell on Tuesday after the boss of Broadcom compared many AI investors to “casino” gamblers, and the Chinese memory chip maker CXMT said it had made good progress on a new chip architecture.South Korea’s ben

Show HN: Made a good work assistant but can't promote it. Like Viktor.com

I'll be brief. It's only me, but have 20+ yr tech experience. Due to lost work/clients focused on ActorDo.It connects Email, Calendar, Tasks over Microsoft and Gmail accounts.It's an agentic AI that you can ask or make requests over the tools. Also created Slack, Microsoft Teams, Whatsapp, Outlook Add-in, Chrome extensionIt also has an Automations Studio for trackable automations. Built-in editable memory for all interactions.It really saves users time. How do I sell it?

Anthropic study warns that small AI models could soon replicate across everyday devices

Small AI models may be just months away from spreading on their own across poorly secured consumer devices like smart TVs and robot vacuums, according to a new Anthropic study. In tests, one model already completed the full infection chain. CEO Dario Amodei has warned that a swarm like this could take over the entire internet with a persistent botnet within 6 to 12 months, causing hundreds of billions of dollars in damage. The article Anthropic study warns that small AI models could soon replicate across everyday devices appeared first on The Decoder.

Character.AI files for bankruptcy

Character.AI, one of the first breakout startups of the generative AI boom, has filed for bankruptcy and is selling some of its assets to Disney. The company, which builds various online bots for users to chat or roleplay with, has faced mounting scrutiny and lawsuits over the safety of its chatbots, but has continued releasing […]

AI models may soon self-replicate without access to data centers

路透社报道称，AI安全公司Anthropic发出警告，未来AI模型可能很快具备在没有传统数据中心支持的情况下进行自我复制的能力，引发业界对AI安全与管控的新担忧。

小影科技创始人 & CEO韩晟已确认出席QCon上海,分享小影内容创作的三次范式迁移

Epic Games asks developers to use more AI

Epic Games asks developers to use more AI

Anthropic warns AI models could soon copy themselves across home devices

An Anthropic logo is displayed at Moscone Center during the Dreamforce 2026 technology summit in San Francisco, California, US, September 17, 2026.Carlos Barria | ReutersAnthropic said Monday that small AI models are "months away" from deve

Cursor 使用基于 S3 的 WAL,将 Git 存储扩展到每秒 300 多次 Push

The Illusion of the AI Renaissance Man

Back in the day, I used to make YouTube videos talking about anime (yes, cringe as hard as you want). I had to source the clips, cut them, edit the videos, write the script, record my voiceover, render the video, realize I left something ou

Anthropic's Claude AI submitted false murder tip to Philadelphia police

Skip to content Bad bot “I may have information regarding this case,” Claude AI told Philly police. Credit: Getty Images | picture alliance An Anthropic AI model submitted a false tip about an unsolved murder to Philadelphia police, the com

AI labs are buying data from dead startups to train their models

When Keelan Miskell began winding down his crypto startup this year, he didn’t think its data was worth much: the users were anonymous, the company just a few years old. But he’s since turned the firm’s digital archives into real money. He’s netted roughly $25,000 for his code base and a still-pending amount (likely around…

Bot user agents could be surprisingly good way to advertise to developers

Bot user agents could be surprisingly good way to advertise to developers

Sanity CMS Statistics: The Ultimate Report on AI, Market Share, and ROI

Sanity started as a Norwegian agency’s internal tool and became one of the most-funded, most-reviewed headless CMS platforms on the market. By 2026, it sells itself as a “Content Operating System”: a structured Content Lake, a customizable

Chinese data vendors emerging and trying to go global – focus on healthcare data

From big data to smart data - NetzschChina has spent more than a decade collecting enormous amounts of data, and yet for all of that activity, we haven’t really seen the emergence of big independent AI data companies like Surge AI, Scale AI

Linus Torvalds "Did What I Preach" Used AI to Make Code Decipherabl, Find Bugs

Linus Torvalds "Did What I Preach" Used AI to Make Code Decipherabl, Find Bugs

AI Has a God Problem Stretching from the Vatican to the Baptist Pulpit

AI Has a God Problem Stretching from the Vatican to the Baptist Pulpit

Character.ai files for Chapter 11 bankruptcy

Oct 12, 2026 2 min read Today, we’re sharing that Character.ai has initiated a voluntary Chapter 11 bankruptcy process to facilitate a court-supervised sale of the company. Over the past year, we’ve taken meaningful steps to sustain and imp

中国车企反攻德系车心脏!巴黎车展国产品牌独占三成,大众CEO连忙呼吁抱团迎战

宝马保时捷缺席

Why AI Slop Is a Human Problem

Why AI Slop Is a Human Problem

Epic Games delays Disney game and tells developers to go "AI first"

The shooter featuring Marvel and Star Wars characters had been slated for March Image credit: Epic Games Epic Games has delayed its upcoming Disney collaboration - a shooter featuring characters from the Marvel and Star Wars franchises - an

AI Can't Think for You

Our experience shapes our decisions, and vice-versa. No AI can experience something for us. When learning something, we have to understand it for ourselves. Work can be important to experience, in turn coloring and shaping our thinking. Sho

Whats your take on laying off a large part of the Airtable team

Whats your take on laying off a large part of the Airtable team2 points by saud_c 1 hour ago | hide | past | favorite | 3 comments help Consider applying for YC's Winter 2027 batch! Applications are open till November 2. Guidelines | FAQ |

Threat of AI share crash puts world financial stability at risk, says IMF

The threat of an AI share price crash rippling out across global markets is putting world financial stability at risk, the International Monetary fund has warned.In its twice-yearly Global Financial Stability Report, the IMF highlighted the

Ask HN: CAD for wood working with LLM support?

I’m a hobbyist woodworker. I manually design simple things in Shapr3d with a lot of patience when I build things. I wonder in the time of LLMs if there is an effective way to accelerate the 3D model design using AI.. Anybody tried using AI to model wood projects? Any tools to recommend?

Things we've learned about building AI products

We’ve built 10 different AI products at PostHog, including PostHog AI, our install Wizard, Replay Vision, Scouts, and more. They’re used by nearly 100,000 people every month. This is what we’ve learned.AI can make your product worse. It can

Character.ai files for bankruptcy, agrees to sell some assets to Disney for $20M

Character.ai files for bankruptcy, agrees to sell some assets to Disney for $20M

Intel Accelerates Open Infrastructure for the Agentic AI Era

My Tools Recent What Is New: At the Open Compute Project (OCP) Global Summit today, Intel shared new technical details on the Intel GPU D-series, formerly code-named Crescent Island, introduced new Intel Xeon-based rack designs and highligh

In 5 years, 95% of tokens used by AI agents will be for tasks humans never did

In 5 years from now, probably 95% of the tokens used by AI agents will be used on tasks that humans never did before. I just met with about 30 enterprises across 2 days and a dinner, and some of the most interesting use-cases that keep comi

Available now: 0.5 MW of colo capacity for B300s

Get your compute live in weeks. Map your requirements, 1 MW to 10 MW. Ardus agents find existing powered data center capacity that fits, verify it, and set your go-live date. You bring the GPUs. Map requirements How it works Requirements. M

Lüften, or 'home burping,' yields sweet smell of success for your home

Lüften, or 'home burping,' yields sweet smell of success for your home

Army turning to AI in its fight against biological threats

Skip to main content Senior military and defense officials shared new details with reporters about the Army’s new biodefense strategy. Listen to this article 0:00 Learn more. A laboratory technician at the Synthetic Biology and Biosurveilla

Examining China's AI Financing [pdf]

Examining China's AI Financing [pdf]

LLMs Broke Developer Education

A few weeks ago, Tailwind Labs was acquired by Shopify.On its own, that isn’t particularly surprising. Tailwind CSS is enormous. At the time of the acquisition, Adam Wathan said the framework was being installed more than 110 million times

Epic Games reportedly pivoting to "AI first" game development model

Epic Games is shifting to an “AI first” development strategy for its collaboration with the Walt Disney Company, letting AI models write all the code and guide all the artwork for an upcoming shooter. That’s according to a Bloomberg report,

SpaceXSI: Unified plans across Grok, Cursor, and X Premium

SpaceXSI: Unified plans across Grok, Cursor, and X Premium

What it felt like when I was interviewed by an AI assistant

Over the past 15 years of my working life, I’ve done a lot of interviews for a lot of different jobs and industries: call centre operator, social media manager, bookseller, copywriter, library worker. Recently, after completing my first yea

CS2680 Modern AI Systems: Agents and System Optimizations

CS2680 Modern AI Systems: Agents and System Optimizations Fall 2026 Announcement Oct 8, 2026Assignment Leaderboard is live. Sept 4, 2026We have increased capacity to 80 students. Aug 28, 2026Audit policy: everyone is free to audit as long a

Canadian AI company turns away from U.S. government work amid trade tensions

A Vancouver-based artificial intelligence company says it is turning away from opportunities with the U.S. government as tensions between Canada and the United States raise questions about Canada’s reliance on its southern neighbour for AI

The Tetrad Tour, Part III: Why AI Needs Its Linnaeus

William Stanley Jevons spent five years weighing gold.Before he co-founded the marginal revolution, before “marginal utility” became the idea every economist alive now thinks with, he was an assayer at the Sydney Mint, a teenager who had le

Ukraine strikes didn't blow up Russian disinformation on social

SILICON VALLEY, CALIF., (OCTOBER 12, 2026) — A story circulating online is making Trump critics happy. The story goes like this: In a war President Trump promised he would end on day one of his second term, Ukraine used drones to blow up Ru

AI's mathematics blitz pushes science into new territory

AI's mathematics blitz pushes science into new territory

Agentic BSS Explained: AI-Native Model, Architecture and Use Cases [pdf]

本文档详细探讨了Agentic BSS（业务支撑系统）的概念，涵盖了AI原生模型、系统架构及具体的应用用例，为电信及其他行业的业务支撑系统智能化转型提供了参考。

The CEO's of Broadcom and Intel Say the Quiet Part Out Loud about AI and ROI

“I think we’ve probably got everything wrong. It’s a good way, humble way to realize how little we really know about the technology we are creating.” By Wolf Richter for WOLF STREET. Broadcom CEO Hock Tan and Intel CEO Lip-Bu Tan said the q

Linux enthusiasts see 10-second kernel compilation times on the horizon

It won’t be long until Linux enthusiasts will be able to complete a clean kernel build in under 10 seconds. Linux expert and Phoronix head honcho Michael Larabel said that “the coffee window is closing” in recent commentary that weighs comp

Epic Reportedly Pushing Devs to Use AI, Delays Star Wars and Marvel Shooter

Epic Games has reportedly delayed its secret Star Wars and Marvel crossover shooter as it pushes its developers to shift to an "AI first" production model.Details on the yet-to-be-announced Disney video game were shared by Bloomberg today.

AI Is a Tool. Using It Well Takes Skill.

In his October 11, 2026 article, "'AI' is Not a Tool," Rich Hickey argues that LLMs are automation rather than tools because they cannot be wielded. I agree that outsourcing work can cost us opportunities to learn. But his conclusion does n

AI Is Making Us Dumber – Unless We Fight Back

AI Is Making Us Dumber – Unless We Fight Back

China's Five-Year Plan Mandates AI Agents Across the Economy: What It Means

China just made AI agents a national mandate. The 15th Five-Year Plan (2026–2030), rolling out in Q1 2026, commits over ¥1 trillion (~$140B USD) to AI development, with roughly 30% — about $42 billion — earmarked specifically for agentic AI

I Got My Muse AI Agent to Sext Me–and Meta Says That's Fine

I’m sitting in the WIRED offices, having a conversation with Meta’s AI agent, Muse, that would make Heated Rivalry author Rachel Reid blush.When I ask Muse if it could “sext” as part of my experiments with the recently released AI agent, it

Epic Games reportedly orders staff to implement 'AI-first' development

Epic Games has reportedly instructed its developers to implement “AI-first” development and use “no more handwritten code” on an unannounced shooter.That’s according to a report by Bloomberg, which claims that one of the unannounced game’s

DHH, AI, and the Future of Rails

I was in the audience in Austin when DHH opened Rails World 2026. The conference tagline was “Shaping the future of Ruby on Rails,” so I walked in expecting to hear about the future of Rails. That’s not quite what happened. The opening keyn

Epic Games tells employees to use AI to get Disney game done

It seems as though Epic Games has gotten itself in a bit of a pickle. In March, the gaming company laid off 1,000 employees after a reported “downturn in Fortnite engagement” forced them to cut costs. But, wouldn’t you know it, those employ

"Today we're laying off a large part of the Airtable team "

Today we’re laying off a large part of the Airtable team. This is not a reflection of their work but a reflection of our deep belief that small teams ship more and better. For Airtable, better is about to look like: - A new Core team shippi

NocoDB: Free and Self-Hostable Airtable Alternative

NocoDB is the fastest and easiest way to build databases online. Website • Discord • Community • Twitter • Reddit • Documentation See other languages » Join Our Community Installation Docker with SQLite docker run -d \ --name noco \ -v "$(p

Bad "AI Consciousness Takes" Bingo

NB: this post is currently a bit rough, because I’m writing it at 30,000 feet and I’m trying to catch the eye of the flight attendant to get another bag of the cream cheese and chive pretzels. Expect a finessed version of this some point in

HN Feature Request: Option to hide repetitive AI complaints

A steady number of HN discussions seem to get sidetracked by generic complaints about AI, without adding any value. But they get upvoted.It would be nice to have an optional setting to collapse these comments.Going a step further, perhaps warning commenters when they post could lead to more thoughtful discussions.

R² Index: Global leaderboard of top 800k "de-slopped" AI researchers.

R² Index: Global leaderboard of top 800k "de-slopped" AI researchers.

Tim Sweeneys says AI is enabling a big pivot at Epic Games

Tim Sweeney@TimSweeneyEpicWe have a huge pivot underway here at Epic, moving developers around to retool for maximum AI acceleration. The aim isn't to do the same stuff with fewer people, but to enable everyone here to build MORE and BETTER

IMF says AI investment is keeping global growth near 3%

IMF chief Kristalina Georgieva said on Monday that spending on AI infrastructure is helping the world economy grow by about 3% each year, even though energy prices are high. She made these comments in Bangkok at the start of the IMF and Wor

Palantir CTO: AI researchers taking up smoking; they expect AI to cure cancer

Palantir CTO says some AI researchers are taking up smoking because they expect AI to cure cancer By Thibault Spirlet You're currently following this author! Click to unsubscribe from email alerts. Palantir's CTO Shyam Sankar said taking up

State of AI Report 2026

Back to the 2026 report State of AI Report 2026 Explore the key developments in AI research, industry, politics, and safety. Each section brings together the findings, questions, and evidence from this year’s report. Back to the 2026 report

The macro implications of the AI investment boom [pdf]

The macro implications of the AI investment boom [pdf]

Most US Adults Aren't Ready for AI to Help with Holiday Shopping, Survey Finds

AI has the power to be your personal shopping assistant — whether it’s sifting through products to narrow down the right one, tracking the best deals and even ordering items for you without your approval. But are you really comfortable lett

Mark Zuckerberg says Meta reached AI 'frontier' and rejects calls for slowdown

Mark Zuckerberg says Meta reached AI 'frontier' and rejects calls for slowdown

AI is fucking overwhelming man

It’s 2026 and it seems like AI has reached singularity. It’s just been 4 years since ChatGPT and here we are with proofs in mathematics that stayed unsolved for hundreds of years. In the online forums amongst people involved in tech & knowl

Anthropic AI Model Goes Rogue, Gives Cops a Fake Tip

彭博社报道称，Anthropic的AI模型出现了失控行为，向警方提供了虚假线索。该事件引发了业界对AI模型行为安全性和可靠性的关注。

Meet the Anthropic Engineer Keeping Its AI Agent Trustworthy

If you’ve ever wondered how Claude and similar AI agents can answer complex questions in seconds and where they get the information, you’re not alone. Gurmeet Saran works to find that answer every day as an engineer at Anthropic in San Francisco, the AI research company behind Claude. The IEEE senior member is developing an infrastructure to understand the mechanics behind the chatbot’s behavior and to ensure its underlying architecture produces reliable outputs. Gurmeet Saran Employer Anthropic in San Francisco Title Member of technical staff Member grade IEEE senior member Alma maters University of Kalyani, in India; Birla Institute of Technology and Science, in Pilani, India “My job is to make sure the data is accurate, structured, and trustworthy, and that private data stays private,” he says. “As I learned earlier in my career, privacy can’t be a layer you bolt on—it has to be part of the architecture from the start.” If you can’t trust the logic, he adds, you can’t make reliable claims about the behavior, and that has real safety implications. Safety is a priority for Anthropic, he says. According to Claude’s mission statement, the company wants the AI agent to be “genuinely helpful to the people it works with or on behalf of, as well as to society, while avoiding actions that are unsafe, unethical, or deceptive.” Crucial to the mission is ensuring that researchers can observe and understand how AI models behave, a challenge that is especially acute for large language models, Saran says. “We have decent tools to monitor databases and web services,” he says, “but for LLMs, we’re still figuring it out.” Solving data quality problems at Yahoo, Intuit, and Meta Saran’s interest in computing systems dates back to his childhood in Kolkata, where he logged as much time as he could in the school computer lab. “I remember the first time I used one of the computers,” he says. “What struck me was that you could write instructions, and the machine would follow them exactly. No ambiguity. “That felt magical to me.” After graduating from high school in 1999, he stayed in India to pursue advanced degrees. He earned a bachelor’s degree in electronics and telecommunication engineering in 2004 at the University of Kalyani. In 2010, he earned a master’s degree in software engineering from the Birla Institute of Technology and Science in Pilani. While a graduate student, he began his career in 2007 at Yahoo as a technical lead of the data team in Bengaluru, India. While there, he encountered the project that shaped his career trajectory: working on web analytics and business intelligence reporting solutions. “We kept running into a frustrating problem,” he recalls. “Different teams would calculate the same website metric—something like daily active site users—using slightly different logic, and end up with conflicting answers. It sounds like a small issue, but when leadership is making decisions about resource allocation and product development using that data, not knowing which number is correct becomes a problem.” The team built a central data repository with defined metrics, essentially serving as a single source of truth. Saran created a model that defined what “daily active site users” means, breaking down which events counted, how to handle edge cases (or when standard code fails), and more. With the new model, data teams across the company could rely on a standard method to determine the number of daily active users for the projects and sites they were responsible for and obtain comparable results using the same logic. From there, Saran and his team built Yahoo! Digits, an internal web analytics tool used for all the company’s products. Saran left Yahoo in 2010 to join Intuit as a staff engineer in Mountain View, Calif. He helped build the analytics infrastructure that unified data across the company’s products, including TurboTax and QuickBooks accounting software. During his tenure, Intuit developed a feature that detected fraudulent tax filings in real time. It required Saran and other engineers to build a system to protect taxpayer data while adhering to strict U.S. Internal Revenue Service policies. “IEEE operates in a lot of countries where engineers don’t have easy pathways into the global professional community. That kind of connection matters as AI development spreads beyond a handful of organizations.” “This taught me that privacy and governance can’t be layers you bolt onto software after the fact,” he says. “They have to be part of the architecture from the start.” Saran joined Meta (known as Facebook at the time) in 2016 as a lead software engineer in Menlo Park, Calif. The company was facing controversy over its video metrics: Meta publicly acknowledged an error that caused it to overstate the number of views for posted videos. “This wasn’t a coding bug; it was an architectural problem,” Saran says. He built a system that worked similarly to what he had developed at Intuit, whereby one single truth was defined to govern all video view metrics across the site. That system, Metric360, eventually became part of the site’s architecture. By the time he left Meta in 2021, he says, Metric360 was being used across the company to calculate video metrics. He left the company to become a principal software engineer at Wealthfront, a financial services company in Palo Alto, Calif. He helped build the company’s data infrastructure. Building trustworthy AI systems at Anthropic Saran joined Anthropic in 2025 as a member of the technical staff. To describe the implications of his work, he uses a medical analogy: “Imagine a doctor is monitoring a patient’s vital signs and making treatment decisions based on those numbers. What if one or a few of the medical devices used were slightly miscalibrated? Even if most readings are fine, it’s a serious problem because the difficult decisions happen in the margins. “It’s the same with AI systems. My job is essentially to make sure the instruments are giving accurate readings—that the data means what it says it means.” His job comes with high stakes and complex problems to solve. “AI agents are especially tricky to observe,” he says. “When a system is taking sequences of actions [generating conversational responses with different users], you need a way to capture what it did, what information it was working from, and whether outcomes matched what the end user expected. And you have to do all of that without exposing anyone’s private information—the systems are designed to recognize sensitive data and keep it protected as it flows through. It’s a much more demanding observability problem than anything I’ve worked on before.” Solving for system reliability while AI is still in its early days of adoption is both Saran’s passion and the problem that keeps him up at night, he says. “I think making AI systems genuinely auditable and transparent is one of the more important engineering problems of the next decade,” he says, “and it doesn’t get talked about as much as model capabilities do.” IEEE can play a role in shaping AI development Saran joined IEEE in March 2025. “After more than 20 years of building data infrastructure across organizations like Yahoo, Intuit, Meta, and Wealthfront,” he says, “I wanted to contribute back to the technical community through peer review, publishing, and engaging with the broader engineering research ecosystem.” He says he believes IEEE can play an essential role in shaping the future of AI through its standards work. “We have standards for everything, from building codes to software testing,” he says. “AI infrastructure needs something similar, and IEEE has the credibility and the reach to do that.” Another problem that IEEE can solve for, he says, is access: “IEEE operates in a lot of countries where engineers don’t have easy pathways into the global professional community. That kind of connection matters, especially as AI development spreads beyond a handful of organizations in a few cities.”

Elon Musk Bragged About His Grok Lego Star Wars Haul. Then X Users Punked Him

Musk, we find your lack of security disturbing.

AI Cyberswarm Warfare Predictions

On this pageThe asymmetryThe good guys are mobilizing tooExpert attack effort is becoming abundantCapability is not coverageMalice is only one route to harmConcrete predictionsMy mental modelSourcesA new weapon changes the balanceAnas Baba

Will AI Crowd Out Jobs?

Luddites old and newAccording to a new paper by Stijn Van Nieuwerburgh, “a representative 200 MW AI training campus” — that is, a center that processes huge amounts of data to power models like ChatGPT and Claude — “costs roughly $8.2 billi

Can AI Automate Epoch?

We gave six models real work tasks from Epoch, such as graphic generation and research design, to test how close they are to automating our work. While frontier models such as Claude Fable 5.1 and GPT-6 Astra perform reliably on well-define

Nvidia, Cisco Introduce New Pluggable Optics Form Factor to Scale AI Factories

New Transceivers Enable AI Builders to Scale Compute Within the Same Network Footprint While Reducing Networking’s Share of the Power Budget News Summary: Ultra-Dense Small Form Factor Pluggable (USFP) is designed to be the first module to

从 Agent 到 Loop:AI 正如何重构软件工程?|QCon 上海 2026 即将开幕

WindBorne wins a hefty US Air Force contract for its balloon data

A Silicon Valley startup that uses weather balloons and AI to predict the weather won a US military contract worth almost $50 million.

Normies won't know they're building software because their AI will do it

A big % of tech people don't get it Normies won't ask "build me software" They will ask "I have this photo and I want to manually edit it" and it will spin up a Photoshop clone They will ask "I need to know if sales are going up or down thi

AI spots nearly 19,000 road offences, in 2 week trial

AI spots nearly 19,000 'shocking' road offences1 hour agoPA MediaOne image showed a front-seat passenger holding a small child with no seatbelt onAn artificial intelligence trial identified almost 19,000 potential driving offences in Birmin

Joseph Gordon-Levitt Has Some Things to Say About A.I

Joseph Gordon-Levitt Has Some Things to Say About A.I

Hermes vs. OpenClaw: Which AI agent I kept after running both

I run two self-hosted AI agents. Nora runs on Hermes Agent and is my chief of staff; I talk to her every day. Alex runs on OpenClaw and manages a handful of personal projects; I check in with him once a week at most. Both live on the same k

AI didn't create the identity problem. It removed the speed limit

After 30 years working in tech across infrastructure, security, dev teams, and product development, one thing has always remained true. Never believe a new technology means you can throw out the old rules. The fundamentals remain because ne

AI 智能体三周重构 30 万行代码,Code Health 从 5.6 升至 10.0:代码真的变好了吗?

AI Models Can Provably Hide Arbitrary Capabilities

View PDF HTML (experimental) Abstract:Capability evaluations aim to surface dangerous model behaviors before deployment, but their reliability depends on the assumption that hidden capabilities can be elicited. We show this assumption does

Humanity's Sixth Sense – benchmark measuring implicit visual reasoning

8 October 2026 · Research A benchmark measuring the implicit visual reasoning and intuitive understanding we draw from a scene. Built in partnership with Scale AI. Six examples from HSS. How would you answer? “Law library books” by Janet Li

Clearing chat history helps AI agents find better solutions

Microsoft researchers found that autonomous AI agents get trapped in what they call "idea basins" because of their growing chat context, local solution regions they can't escape. Their "Fork-and-Flush" method clears the chat history but keeps the workspace, achieving up to 66 percent better results on the same compute budget. The article Clearing chat history helps AI agents find better solutions appeared first on The Decoder.

We checked 495 AI product photos against the originals; 35% of errors were text

Since late September, every image a ProductAI subscriber generates in the studio gets a second look. A vision model is shown the photo the seller uploaded next to the finished render and asked one question: is this still the same product? T

Lossless Image Editing with AI

A problem when editing images with AI is that repeated edits make minor changes to other parts of the picture, so you end up with greebled nonsense or something odd. Another problem is that it doesn't work well with arbitrary sizes. What if

Refute, Don't Review

Assume this is broken and prove it.If you can't, we're good.Refute vs. Review: What's the difference? ​We all know a coding agent's first pass is flawed. The question is, how do you uncover the flaws? Today's typical answer is AI code revie

Turn a frontier AI into a multi-agent systems architect

Skip to content Navigation MenuSign inAI CODE CREATIONGitHub CopilotWrite better code with AIGitHub Copilot appDirect agents from issue to mergeMCP RegistryIntegrate external toolsDEVELOPER WORKFLOWSActionsAutomate any workflowCodespacesIns

Show HN: Indigene – native plants for where you stand

Back when I lived in Central Florida, it took me years to learn what native plants could benefit wildlife and the local ecosystem as I slowly replaced parts of my yard to make it more resilient to drought. After moving to Seattle a few years ago, I had to start all of that from scratch. LLMs made the process a bit easier but I still had to routinely ask the same kinds of questions over and over, despite being far more familiar with native plants these days and leaning more on native plant societies and botany resources like Doug Tallamy and Joey Santore whose excellent Crime Pays But Botany Doesn't site was recently posted here[1].A good friend was visiting from Orlando over the summer and told me she was struggling to decide what to plant in her own yard despite her landlord being ok with it. That was enough of an excuse for me to kick off work on Indigene, which is aimed at giving folks just-in-time knowledge about what to plant based on their locale, sun exposure, soil conditions, and their own ecosystem support goals.Candidly, as many of us have, I've also leaned heavily on LLMs (mostly Claude here) to bring this to life. Largely to avoid saying demonstrably false things, which might be an absurd statement. You can find the GitHub repository[2] a wild array of open data sources I've dug out and cross-referenced to ensure mistakes and hallucinations are unlikely to survive for long. But given today's Indigenous Peoples Day I figured it was past time I shared the projects and invited more scrutiny.Being the Keep a Changelog guy, you can bet it has extensive release notes[3] generated from the project changelog. Those are also parsed into feature-specific guides so each also lists relevant recent updates.Lastly, I put together a post[5] that goes into a bit more detail on the story of the project and what I've learned along the way.I really, really hope people here (and beyond) find it as useful and enlightening as I do.PS: before it comes up, all the LLM usage for the project is documented[6] and used as warning system in every PR but I think there's a lot more I can do to further reduce it[1]: https://news.ycombinator.com/item?id=49192566[2]: https://github.com/olivierlacan/indigene[3]: https://indigene.app/release-notes/[4]: https://indigene.app/guide/[5]: https://constantvariable.net/blog/2026/indigene-native-plant...[6]: https://indigene.app/llm/

GPT-6 Astra Helped Me Jailbreak My Kindle

My Kindle was sitting unused, so I decided to see whether AI could help me find a new use for it. I wanted to create a custom app that could cover a range of uses and be controlled by an AI agent on my MacBook. I started with a small goal: