AI News Today

"The reality is, when you're optimizing for production, you start looking at a price/performance," Guillermo Rauch tells TechCrunch.

The next version will have 32,000 MCUs

Article URL: https://www.researchgate.net/figure/Gartner-Hype-Cycle-2017-11_fig1_322215706 Comments URL: https://news.ycombinator.com/item?id=48809439 Points: 1 # Comments: 0

Article URL: https://github.com/akahkhanna/groundtruth Comments URL: https://news.ycombinator.com/item?id=48809364 Points: 1 # Comments: 0

Article URL: https://www.kapa.ai/blog/how-we-prune-rag-context Comments URL: https://news.ycombinator.com/item?id=48809354 Points: 4 # Comments: 0

Article URL: https://www.thirstymachines.com/ Comments URL: https://news.ycombinator.com/item?id=48809253 Points: 3 # Comments: 2

A few simple adjustments to your settings will opt you out of Google's latest AI-training method.

Article URL: https://www.worksinprogress.news/p/ai-is-bottlenecked-by-the-grid Comments URL: https://news.ycombinator.com/item?id=48809084 Points: 2 # Comments: 0

The update is part of Apple's broader effort to make Siri feel more natural and personal, as it rebuilds the assistant around generative AI.

Article URL: https://github.com/Thomaszhou22/danger-guard-skill Comments URL: https://news.ycombinator.com/item?id=48809001 Points: 1 # Comments: 0

Article URL: https://www.revieve.com/job-board Comments URL: https://news.ycombinator.com/item?id=48808974 Points: 2 # Comments: 1

<pre style='white-space:pre-wrap;width:81ex'>fix(codex): preserve yielded native subagent delivery</pre>

Cloudflare is giving all customers granular AI bot controls. Site owners can now manage Search, Training, and Agent bots separately instead of blocking them all at once. Starting September 15, 2026, Training and Agent bots will be blocked by default on ad-supported pages. The article Cloudflare replaces its blanket AI bot block with granular controls for search, training, and agent crawlers appeared first on The Decoder .

Distributed AI training is notoriously fragile because losing a single machine typically crashes the entire multi-node job, forcing a time-consuming, full-workload infrastructure restart. To address this, Google’s JAX ecosystem utilizes elastic training via Pathways, which converts a hardware failure into a catchable Python exception so the running process can survive. When an unplanned failure occurs, the system automatically replaces only the broken worker, restores the last viable checkpoint

<pre style='white-space:pre-wrap;width:81ex'>fix(agents): bound tool_search_code stderr accumulation (#101007) * fix(agents): bound tool_search_code stderr accumulation * fix(agents): report latest bounded stderr --------- Co-authored-by: NIO <nocodet@mail.com> Co-authored-by: Vincent Koc <vincentkoc@ieee.org></pre>

Article URL: https://github.com/patrickjh/ssa Comments URL: https://news.ycombinator.com/item?id=48808667 Points: 1 # Comments: 1

A running look — in reverse chronological order — at the bigger tech companies that have announced significant layoffs this year with AI as a stated factor.

<pre style='white-space:pre-wrap;width:81ex'>feat(ios): add Listen action speaking assistant messages via gateway TTS (#100771) * test(gateway): shift advertised-method window for tts.speak * feat(ios): add Listen action speaking assistant messages via gateway TTS * docs(ios): document the chat Listen action * style(ios): keep outbox label doc comment attached and give speech chip its own * fix(ios): harden chat Listen playback * docs(changelog): note iOS chat Listen * fix(ios): unwrap speech MI

Article URL: https://github.com/josephsenior/Grinta-Coding-Agent Comments URL: https://news.ycombinator.com/item?id=48808609 Points: 1 # Comments: 0

Quantum plans to use the proceeds to expand production capacity, strengthen supply chain resilience, and more. The post Quantum Systems raises $1.2B to expand drone production and deployments appeared first on The Robot Report .

Zhipu AI is bringing GLM-5.2 to its ZCode development environment, pitching the model's long-context capabilities for complex coding tasks. New customers get a free five-day trial with up to 5 million tokens per day, and subscribers receive about 1.5 times more token quota through July 2026. The article Zhipu AI launches ZCode to challenge Claude Code and OpenAI Codex at a fraction of the cost appeared first on The Decoder .

Article URL: https://ramp.com/data/heavy-ai-adopters-hire-more Comments URL: https://news.ycombinator.com/item?id=48808498 Points: 2 # Comments: 0

<p> The feedback layer for anything you build with AI. </p> <p> <a href="https://www.producthunt.com/products/dot-7?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/1189620?app_id=339">Link</a> </p>

We use sklearn in production for time series forecasting in industrial settings. Pickle was causing us constant pain: refactoring our wrapper classes invalidated saved models, and the binary blobs were completely opaque. We built skeights to decompose a fitted estimator into two files. A JSON file for hyperparameters and structure, and a safetensors file for the numeric arrays. The JSON is human-readable and diffable, and safetensors is a well-established safe format from Hugging Face. This allo

When you build something people REALLY want, you can pull of ridiculous stunts like Claude's pricing: - Vague limits - Weekly/monthly usage stats ?? - 50% of your usage on Fable 5 - $200/mo subscription that doesn't include Fable 5 - But you can use it with "usage credits" I'm frustrated but am also hooked. Kinda of a lesson in startups in a way. Comments URL: https://news.ycombinator.com/item?id=48808413 Points: 3 # Comments: 1

hello HN, let me tell you why i built boilerroom ( https://boilerroom.ai ). throughout my past endeavours in sales, the one thing that thoroughly pissed me off was living in a million places to execute one thing; the outbound motion this includes everything top of funnel like finding leads (be it inbound or outbound), researching, filtering for qualification criteria, throwing them in a sequence, engaging them via dialer, email, and linkedin, gathering context, all the way through bottom of funn

Article URL: https://yl3469.github.io/uniboost-icml26/ Comments URL: https://news.ycombinator.com/item?id=48808378 Points: 2 # Comments: 0

<p>Documentation is often the first product a developer experiences.</p> <p>Before they see your architecture, your engineering culture, or your code quality, they interact with your documentation.</p> <p>If that experience is confusing, incomplete, or frustrating, many developers won't make it to their first successful API request.</p> <p>Over the past few weeks, I've been reviewing documentation from AI startups to understand what makes onboarding smooth—and where teams unintentionally create

<p>I wanted to give <a href="https://github.com/argenkiwi/ambler-ts" rel="noopener noreferrer">AmblerTS</a>, my Deno/TypeScript state-machine framework, the ability to record non-obvious learnings that would otherwise require significant context to reconstruct across sessions.</p> <p>I turned to the classic note-taking methodology developed by the German sociologist <a href="https://en.wikipedia.org/wiki/Niklas_Luhmann" rel="noopener noreferrer">Niklas Luhmann</a>: the <a href="https://en.wikipe

<p><a href="https://github.com/GiulioDER/cca-audit" rel="noopener noreferrer">CCA-Audit — open source (MIT)</a></p> <p>AI code review has a trust problem, and it's not that it misses bugs. It's that it invents them.</p> <p>If you've run an LLM over a diff, you've seen it: a "possible null dereference" on a value that's guarded three lines up. A "SQL injection" your ORM already parameterizes. A "race condition" that can't happen. And then — worse — it confidently rewrites working code to "fix" th

made a fun little website with AI to explain how the Oceans in FFIX are authored after days of slamming my head against the keyboard trying to deconstruct it Comments URL: https://news.ycombinator.com/item?id=48808276 Points: 2 # Comments: 1

This story originally appeared in The Algorithm, our weekly newsletter on AI. To get stories like this in your inbox first, sign up here. OpenAI CEO Sam Altman’s oft-discussed promise that Americans will share in the wealth AI creates was in the news again last week. On Thursday, the Financial Times reported that Altman is in…

<pre style='white-space:pre-wrap;width:81ex'>fix(test): skip live auth browser caches (#91907) * fix(test): skip live auth browser caches * fix(test): scope Gemini live auth exclusions --------- Co-authored-by: Peter Steinberger <steipete@gmail.com></pre>

Article URL: https://gamowlabs.com/sota-genome-interpretation-with-agentic-ai.html Comments URL: https://news.ycombinator.com/item?id=48808099 Points: 9 # Comments: 1

Article URL: https://www.botfluencerz.com/ Comments URL: https://news.ycombinator.com/item?id=48808061 Points: 2 # Comments: 0

Article URL: https://www.anthropic.com/research/global-workspace Comments URL: https://news.ycombinator.com/item?id=48808002 Points: 81 # Comments: 17

Is SpaceX working on a mobile device? It seems like it. Here's everything we know so far.

Article URL: https://link.springer.com/book/10.1007/978-3-032-26782-5 Comments URL: https://news.ycombinator.com/item?id=48807971 Points: 2 # Comments: 0

<table> <tr><td> <a href="https://www.reddit.com/r/MachineLearning/comments/1up4cjh/lingbotvision_masked_boundary_modeling_for/"> <img src="https://preview.redd.it/ha08vg49bnbh1.png?width=140&height=78&auto=webp&s=cbd1e4aed6c0571b7f0acee245c21543fe356719" alt="LingBot-Vision: masked boundary modeling for self-supervised pretraining (0.296 NYUv2 linear-probe RMSE at 1.1B vs 0.309 for DINOv3-7B, trails on ImageNet); weights in 4 sizes[R]" title="LingBot-Vision: masked boundary modeling for self-su

Article URL: https://github.com/forgedlogicdev/pyforge-memory Comments URL: https://news.ycombinator.com/item?id=48807937 Points: 3 # Comments: 0