AI News Today

Welcome back to TechCrunch Mobility, your hub for the future of transportation and now, more than ever, how AI is playing a part.

Greetings, This is a link to our learning guide to make your own MCP server in C/C++ (with no other dependencies). If you wanted to make one from scratch but needed some insight into dealing with MCP messages: the I/O flow, JSON parsing, and request dispatch, grab the teaching version tagged mcp/teaching/v1.0.0 in the repository. The JSON parsing in C would have been a hurdle, but then there's this amusing story: Eric: "Let's create a helloWorld MCP server in C." Gemini: "That will be somewhat d

For my local project, I have been finding a good LLM or a SLM that consumes less than 2gb ram, can you get me any models that would aid me? Comments URL: https://news.ycombinator.com/item?id=48708262 Points: 1 # Comments: 0

<p>Tech and AI stocks now make up as much as 12% of most balanced superannuation funds, experts say</p><ul><li><p>Get our <a href="https://www.theguardian.com/email-newsletters?CMP=cvau_sfl">breaking news email</a>, <a href="https://app.adjust.com/w4u7jx3">free app</a> or <a href="https://www.theguardian.com/australia-news/series/full-story?CMP=cvau_sfl">daily news podcast</a></p></li></ul><p>Artificial intelligence and technology stocks have become a driving force on Wall Street and, unbeknowns

<p>ChatGPT relieves me of my discomfort, but in doing so it robs me of contemplation, of the holy ground between question and answer</p><ul><li><p><a href="https://www.theguardian.com/commentisfree/series/making-sense-of-it">Making sense of it</a> is a column about spirituality and how it can be used to navigate everyday life</p></li></ul><p>As a person of faith raised in a religious household, I have a fairly clear picture of what prayer means to me. Prayer is the practice by which I draw close

Eager to find more public AI-related companies that may do as well as Nvidia, Wall Street investors think they've found a winner with Micron.

Behind a customer's API, a high-quality answer isn't enough. It has to be usable, which means on time. Delivering that consistently is a problem about variance, not speed, and the fixes are counterintuitive. The post Tail Control: The Counterintuitive Engineering of Reliable Agentic Workflows appeared first on Towards Data Science .

Article URL: https://fuckui.com Comments URL: https://news.ycombinator.com/item?id=48707839 Points: 2 # Comments: 1

While we wait for a general release, the system card is the best hint as to what is going on with the new candidate for America’s Next Top Model, GPT-5.6.

Many chat AIs on characters.ai and other bigshot platforms have gotten very well traffic and fans all over the world utilizes them, ignoring small websites. We are creating an Anime AI assistant for Anime fans and we wanna add every possible feature a modernized AI can have, we have added very much of those features to our otaku AI but still it feels off. We know this is not the platform to ask this kind of questions but having an advice from experienced devs and founders is something very impor

Article URL: https://www.businessinsider.com/ai-tech-jobs-builder-seller-not-measurer-cloudflare-2026-6 Comments URL: https://news.ycombinator.com/item?id=48707663 Points: 1 # Comments: 0

Article URL: https://www.smh.com.au/politics/federal/global-recession-and-the-end-of-the-middle-class-what-ai-exuberance-could-do-to-the-world-20260626-p60ag6.html Comments URL: https://news.ycombinator.com/item?id=48707606 Points: 2 # Comments: 0

Article URL: https://www.cio.com/article/4189149/ai-coding-token-costs-are-on-track-to-rival-human-payroll.html Comments URL: https://news.ycombinator.com/item?id=48707589 Points: 2 # Comments: 0

Article URL: https://blog.1001ud.me/ideas/thoughts/the-panopticon-for-all Comments URL: https://news.ycombinator.com/item?id=48707566 Points: 1 # Comments: 1

Article URL: https://stefan.schueller.net/posts/making-a-label-printer-work-under-linux-using-agentic-ai/ Comments URL: https://news.ycombinator.com/item?id=48707513 Points: 1 # Comments: 0

Jonathan Rinderknecht was facing arson charges for setting a fire on New Year's Day in 2025, which became one of the deadliest wildfires in LA history. To make their case, prosecutors turned to location data from his iPhone, security camera footage, and witness testimony. But they also turned to his ChatGPT logs. Prosecutors said that […]

<p> Money movement for AI agents </p> <p> <a href="https://www.producthunt.com/products/sequence?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed">Discussion</a> | <a href="https://www.producthunt.com/r/p/1183031?app_id=339">Link</a> </p>

Article URL: https://www.businessinsider.com/satya-nadella-said-every-company-should-build-own-ai-model-2026-6 Comments URL: https://news.ycombinator.com/item?id=48707317 Points: 2 # Comments: 0

Article URL: https://github.com/thefullnacho/hestia Comments URL: https://news.ycombinator.com/item?id=48707227 Points: 2 # Comments: 0

Article URL: https://www.bloomberg.com/news/articles/2026-06-28/austria-lobbies-eu-to-host-anthropic-after-us-access-curbs Comments URL: https://news.ycombinator.com/item?id=48707146 Points: 42 # Comments: 30

Article URL: https://www.cnbc.com/2026/06/28/google-limits-metas-use-of-its-gemini-ai-models-ft-reports.html Comments URL: https://news.ycombinator.com/item?id=48707103 Points: 25 # Comments: 8

Article URL: https://www.reuters.com/legal/litigation/chinas-360-says-it-has-developed-tools-match-anthropics-mythos-2026-06-24/ Comments URL: https://news.ycombinator.com/item?id=48707090 Points: 7 # Comments: 0

Sorry, I just realized, and I ask for your opinion: if now the policy of the administration at the USA is to mandate a staggered, conditional, restricted ("to trusted parties") release of the forthcoming SotA LLMs, could that mean that abating their market (users from a pool "of 8 thousand or 8 million instead of 8 billion") may cause an RnD slowdown, in fear that the fees to the customers will not be able to cover the expenses? This completely clashes with the "race to AGI", to the "race to geo

A concrete bias–variance lesson: why the smallest model had the best cross-validated fit, and how to know when to reach for the big hammer. The post I Pitted XGBoost Against Logistic Regression on 358 Matches. The Boring Model Won. appeared first on Towards Data Science .

Article URL: https://zenodo.org/records/20999945 Comments URL: https://news.ycombinator.com/item?id=48706896 Points: 2 # Comments: 0

A survey paper by Tencent and several Chinese universities traces the path from chatbot to "digital colleague." AI systems won't become reliable coworkers, the researchers argue, until they finish entire tasks in persistent work environments instead of just generating answers. The key lies in combining persistent workspaces with reusable skills. The article AI won't become a real coworker until it stops answering and starts finishing tasks appeared first on The Decoder .

Article URL: https://www.youtube.com/watch?v=uWnUnMphmPM Comments URL: https://news.ycombinator.com/item?id=48706777 Points: 2 # Comments: 0

<!-- SC_OFF --><div class="md"><p>I've been teaching myself how LLMs actually work, not at the API level, but down to the matrix multiplications. To force myself to really understand the forward pass, I first built a complete transformer by hand in a spreadsheet from embeddings through to the loss. Then I turned the forward pass into a web page so it's easier to share.</p> <p>It's a full transformer (single attention head, single block) shrunk to the smallest size where every single number still

Article URL: https://github.com/openai/codex/issues/2847 Comments URL: https://news.ycombinator.com/item?id=48706714 Points: 75 # Comments: 55

Article URL: https://twitter.com/DouglasYaoDY/status/2070904914050797582 Comments URL: https://news.ycombinator.com/item?id=48706690 Points: 21 # Comments: 22

Coinbase CEO Brian Armstrong is switching his company to Chinese AI models like GLM 5.2 and Kimi 2.7. An automated routing system picks the best model for each request based on task and price, and better caching pushed the hit rate from 5 to 60 percent. Coinbase has cut its AI spending in half even as token usage keeps climbing. The article Coinbase joins the rush to Chinese AI models as Western labs face a pricing stress test appeared first on The Decoder .

Article URL: https://pointer.kro-ncrv.nl/meerderheid-nederlandse-inspraak-op-strengere-eu-tabakswet-afkomstig-van-ai-tool-philip-morris Comments URL: https://news.ycombinator.com/item?id=48706616 Points: 5 # Comments: 0

There's a better option! ScienceAlert stories are written, fact-checked, and edited by humans, never generated by AI. Don't miss a story, subscribe here.

Article URL: https://geniusaidetector.com/ Comments URL: https://news.ycombinator.com/item?id=48706490 Points: 3 # Comments: 0

Article URL: https://furstenheim.substack.com/p/creating-svgs-with-ai Comments URL: https://news.ycombinator.com/item?id=48706394 Points: 2 # Comments: 0

Article URL: https://www.euronews.com/culture/2026/06/24/cate-blanchett-launches-free-tool-to-help-people-protect-identity-from-ai Comments URL: https://news.ycombinator.com/item?id=48706321 Points: 1 # Comments: 0

Hi HN, I’m a solo developer and built AgentWatch to solve a problem I kept running into while building AI agents: preventing runaway loops and unexpected LLM spend before requests reach the model. AgentWatch sits in front of OpenAI, Anthropic, Gemini, Bedrock, Azure OpenAI, Groq, and others to enforce budgets and runtime policies. I’d really appreciate your feedback. If you’re building AI agents, does this solve a problem you’ve experienced? I’d also love to hear what you’d improve or challenge.

Article URL: https://medium.com/@aishwaryaadyanthaya1994/yc-ai-investment-trends-2026-where-the-us-ai-market-is-heading-07ea828cecf8 Comments URL: https://news.ycombinator.com/item?id=48706228 Points: 2 # Comments: 0

Hey everyone, I'm launching role-model today: a routing protocol, a reference router runtime, and an extension for Pi that allows for better informed routing decisions. role-model is mostly deterministic, with fallback to a controller model, that routes requests based on a chosen routing strategy. the protocol is structured around assigning domains and roles to models, where requests sent by consumer applications like Pi have task types to enrich routing metadata and thereby accuracy. you can to