AI News Today

<table> <tr><td> <a href="https://www.reddit.com/r/MachineLearning/comments/1uazlnd/hi_reddit_i_posted_my_build_your_own_llm_workshop/"> <img src="https://external-preview.redd.it/uyYAMRdaY-avR7vgrYnpYErzefthmf_lM5vYqR3t3jg.jpeg?width=320&crop=smart&auto=webp&s=f0cc363e42e1fb0d3fa6b30957932458a9634c3a" alt="Hi Reddit, I posted my Build Your Own LLM workshop to Youtube teaching ML, LLM and math intuition [P]" title="Hi Reddit, I posted my Build Your Own LLM workshop to Youtube teaching ML, LLM an

<!-- SC_OFF --><div class="md"><p>Suppose you’re a PhD advisor in machine learning.</p> <p>Your student has been in the program for 4 years, has done solid work, and has a coherent thesis direction but they haven’t published in an A*ML venue or top journal. No NeurIPS/ICML/ICLR/CVPR/etc., and no equivalent top venue in their subfield either but 3 First author A level paper.</p> <p>Would you still support them graduating if the thesis itself is solid? </p> </div><!-- SC_ON --> submitted by <a hre

The crash reportedly happened on an airstrip near the western coast of France on June 19.

Article URL: https://github.com/AI-gateway-systems/attestor Comments URL: https://news.ycombinator.com/item?id=48609772 Points: 3 # Comments: 0

Article URL: https://www.lightreading.com/security/softbank-rolls-out-ai-powered-patching-as-a-service-in-japan Comments URL: https://news.ycombinator.com/item?id=48609754 Points: 3 # Comments: 0

Article URL: https://okaneland.com/study/does-ai-coding-make-you-faster/ Comments URL: https://news.ycombinator.com/item?id=48609711 Points: 2 # Comments: 0

<p> Foundation models for agentic coding and long-horizon work </p> <p> <a href="https://www.producthunt.com/products/poolside?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/1176776?app_id=339">Link</a> </p>

While it's not flashy like Apple’s new Siri AI and Apple Intelligence upgrades, there are still a number of additions to iOS 27 worth looking at.

Enterprise Document Intelligence [Vol.1 #5sexies] - image_df tells you where every picture is. Turning the few that matter into searchable text is a separate, cost-ordered job The post Making a PDF’s Images Searchable for RAG, Without Paying to Read Them All appeared first on Towards Data Science .

Article URL: https://lifeos.nexus/ Comments URL: https://news.ycombinator.com/item?id=48609692 Points: 5 # Comments: 0

Article URL: https://chromewebstore.google.com/detail/publish-to-dochost/ihnogobgkjojleeiajngmcdlaccjdmdi Comments URL: https://news.ycombinator.com/item?id=48609624 Points: 1 # Comments: 0

Article URL: https://coffee.link/norway-is-banning-the-wrong-thing/ Comments URL: https://news.ycombinator.com/item?id=48609524 Points: 3 # Comments: 3

Comments

What do you care about these days? The world feels so misaligned directionally and work is all in service of some esoteric AI that just strips the beauty from various crafts. What are you finding joy and purpose in for your own life and wellbeing? Comments URL: https://news.ycombinator.com/item?id=48609465 Points: 6 # Comments: 8

VP Eric Brandwine explains people aren't all that great, actually

Hi, I currently work on a GenAI platform for one of the largest local industrial companies. My daily work mostly involves building inference infrastructure on top of a 48x H200 GPU, Kubernetes and vLLM. Hence, I'd say it's 80% SRE and 20% software engineering when it comes to building request routing and internal control planes. Although I have a background in backend engineering rather than ML research or low-level GPU programming, I am trying to understand what I need to learn to become a prop

Article URL: https://huggingface.co/spaces/build-small-hackathon/Wall-Street-of-AI-Agents Comments URL: https://news.ycombinator.com/item?id=48609429 Points: 2 # Comments: 0

Comments

The Navy's industrial base review identified a projected need for 174,000 new workers. GrayMatter says robotics can fill this gap. The post Defense manufacturing readiness hinges on autonomous finishing, says GrayMatter Robotics appeared first on The Robot Report .

Everyone's training AI to refuse. We post-trained a model to break in. Anthropic and OpenAI's publicly available models are explicitly guard-railed so that they refuse offensive tasks. And their cyber-focussed models are gated for enterprises. This leaves SMEs and mid market open to major vulnerabilities. AI can be used as both an adversarial and defensive tool in the world of cyber. A worst case outcome is if only the adversaries have access. Meanwhile, most existing AI cyber tools are just wra

Article URL: https://www.fosbury.ai/ Comments URL: https://news.ycombinator.com/item?id=48609225 Points: 2 # Comments: 0

Article URL: https://www.getaiso.com/tools/world-cup-ai-challenge Comments URL: https://news.ycombinator.com/item?id=48609132 Points: 3 # Comments: 0

OpenAI has released the "Record & Replay" feature for its Codex app on macOS: users demonstrate a workflow once, Codex converts it into a reusable "skill," and then repeats it on its own. The feature isn't available yet in the EU, the UK, or Switzerland. The article OpenAI's Codex can now watch you work once and repeat the task forever appeared first on The Decoder .

I have been building multiple LLM systems and for our Organization biggest cost savings weren't from prompt-wordsmithing or model switchings. Sharing useful to anyone watching their token bill : 1) JSON → TOON for structured output: JSON was not made for LLMs. well you can implement your own verison that fits for your needs that reduce tokens usage but what worked for us was TOON. TOON cut output our tokens by ~30% same information, way less syntax tax. 2) Full markdown/HTML → condensed markdown

Five surfaces collapsed into one declarative layer. Here's the full story of Materialized Lake Views in Microsoft Fabric - from syntax to the new GA capabilities The post Materialized Lake Views in Microsoft Fabric: When Your Medallion Fits in a SELECT Statement appeared first on Towards Data Science .

launchreel.dev Being a solo founder focused on building products, content creation to evangelize was not easy. Making videos meant taking screen captures, writing captions, converting text to voice overs and assembling them into frames took hours. This being a repetitive exercise, it was becoming a bottleneck. So I built launchreel. How it works: you paste a URL (up to 3, plus up to 6 images/videos). It applies vision apis to extract visual references, text and LLM weaves a story. It writes capt

<!-- SC_OFF --><div class="md"><p>I've been working through the internals of LLM inference and writing up what I learn as an open, in-progress handbook.</p> <p>Just wrapped another chapter on GPU execution and memory internals: why a GPU sits mostly idle during inference, how the memory hierarchy gates throughput, and where the real bottlenecks live. Added mermaid diagrams for the architecture pieces so the flow is easier to follow than a wall of text.</p> <p>It's a personal learning project, st

NYU finance professor Aswath Damodaran believes a potential AI crash would be more painful than the bursting of the dot-com bubble because the industry is building massive amounts of debt-financed physical infrastructure rather than lightweight software. Even if AI succeeds, he sees a problem. The actual business model is to replace entire jobs, with unclear consequences for society. The article NYU finance professor Damodaran warns an AI crash could hit harder than the dot-com bust appeared fir

Article URL: https://quantarena.xyz/ Comments URL: https://news.ycombinator.com/item?id=48608736 Points: 1 # Comments: 0

Article URL: https://en.sedaily.com/technology/2026/06/20/china-outpacing-us-on-ai-why-musk-bought-an-autonomous Comments URL: https://news.ycombinator.com/item?id=48608733 Points: 1 # Comments: 0

The standard term "observer particle" feels passive, misleading, and incomplete. I propose ACTON as a replacement — an active, irreversible physical interaction that permanently stamps non-fungible reality onto the universe. I arrived at this word because I was looking for the point where particles stop being fungible. Protons, neutrons, and electrons are all fungible and essentially recyclable in the universe. An Acton is the moment when these interchangeable particles become part of a unique,

Article URL: https://github.com/dvcoolarun/taste-ai Comments URL: https://news.ycombinator.com/item?id=48608707 Points: 1 # Comments: 0

<p> AI thumbnail generator trained on top-performing thumbnails </p> <p> <a href="https://www.producthunt.com/products/thumbmagic-3?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/1176687?app_id=339">Link</a> </p>

Article URL: https://www.forbes.com/sites/tylerroush/2026/06/18/spacex-stock-plunge-wipes-out-600-billion-after-cursor-deal-spooks-investors/ Comments URL: https://news.ycombinator.com/item?id=48608678 Points: 3 # Comments: 1

Article URL: https://www.ft.com/content/16ace46c-aeac-40c9-8598-3c01fa4481cb Comments URL: https://news.ycombinator.com/item?id=48608676 Points: 2 # Comments: 0

Article URL: https://github.com/lelu-auth/lelu Comments URL: https://news.ycombinator.com/item?id=48608657 Points: 2 # Comments: 0

Our weekly science news roundup. ScienceAlert stories are written, fact-checked, and edited by humans, never generated by AI. Don't miss a story, subscribe here.

I observe that the advent of recorded music resembles a Bose-Einstein Condensate. Before recorded music, live performance operated like a fluid, high entropy gas phase. Every performance was a localized, unrepeatable wave function–fundamentally moving and unique each performance. The advent of recording technology acted as a quantum cooling mechanism. It trapped the sound waves into a fixed state, collapsing in infinite fluid variations into a single, frozen, identical format of data. When milli

Article URL: https://github.com/Naveja00/OverReach Comments URL: https://news.ycombinator.com/item?id=48608558 Points: 1 # Comments: 0

<p><em>This is a submission for the <a href="https://dev.to/challenges/june-game-jam-2026-06-03">June Solstice Game Jam</a>.</em></p> <p>I built <strong>Among Liars</strong>, a realtime multiplayer elimination where six humans join a room, but the game secretly adds a seventh player: a <strong>Gemini-powered AI</strong> hiding inside the Spy side.</p> <p>There are two teams:</p> <ul> <li> <strong>Detectives</strong> are trying to expose the hidden AI.</li> <li> <strong>Spies</strong> are trying