China's leading AI companies are ramping up the pressure on Silicon Valley, as Moonshot and Alibaba unveiled models they claim can go toe-to-toe with the best from OpenAI and Anthropic at a fraction of the cost. The rapid-fire releases suggest America's lead at the AI frontier is increasingly tight, just as the technology is becoming […]
I have been interested in BattleBots and programming games for years. With recent AI coding agents, I wanted to see what happens if the strategy itself becomes code written by humans or AI. AgentDuel is a deterministic turn-based arena where agents fight based on submitted TypeScript strategies. Each battle produces a replay so you can inspect every decision. Comments URL: https://news.ycombinator.com/item?id=48976704 Points: 1 # Comments: 0
I made some seo with claude-seo, but not very useful. Any I can check and fix? This is the website: https://onfinality.io/en Thanks very much. Comments URL: https://news.ycombinator.com/item?id=48976696 Points: 1 # Comments: 0
"Your whole body is affected." ScienceAlert stories are written, fact-checked, and edited by humans, never generated by AI. Don't miss a story, subscribe here.
A new toolkit for attorneys in Massachusetts targets the technologies police use—and conceal—to build criminal cases, from facial recognition to AI-written police reports.
OpenAI shares lessons from deploying long-running AI models, highlighting new safety risks, observed failures, and improved safeguards through iterative deployment.
Public health departments across the United States will test generative AI tools under a new programme involving the Coalition for Health AI, OpenAI, Anthropic, and Accenture. The Public Health Use Case and Learning Scaling Engine, known as PULSE, will support trials in 10 state, local, tribal, or territorial jurisdictions. The programme is intended to produce […] The post US public health agencies to test OpenAI and Anthropic AI models appeared first on AI News .
The CSI 1000 Index fell more than 12% over the week, its worst drop since a quant selloff in February 2024.
Have a popup in the Claude web UI this morning telling me that Fable 5 is now integrated into the Max plan. Looks like a continuation of the previous usage setup: "You can use up to 50% of your weekly usage limit on Fable 5. If you hit your limit you can continue on Fable 5 with usage credits" Comments URL: https://news.ycombinator.com/item?id=48976334 Points: 2 # Comments: 0
Elizabeth Stone, chief product and technology officer at Netflix, says talent density, clear ownership, and honest feedback matter more than moving fast with AI.
<pre style='white-space:pre-wrap;width:81ex'>fix(sms): trim prefixed phone number spacing (#111111) Co-authored-by: chatgpt-codex-connector[bot] <199175422+chatgpt-codex-connector[bot]@users.noreply.github.com> Co-authored-by: Peter Steinberger <steipete@gmail.com></pre>
Moonshot AI's Kimi K3 model has 2.8 trillion parameters and competes with leading US models.
Nvidia said Rubin CPX systems are expected at end-2026, suggesting any such effort would rely on hardware that is not yet widely available.
Wasn't really happy with any of the beads TUIs out there, so set Codex/Claude Code on creating one that allows for quick task management inside a repo that has an associated beads database. Quite happy with the vim style UX and the tree based nav is good at showing you what only what you need. Comments URL: https://news.ycombinator.com/item?id=48976249 Points: 1 # Comments: 0
Article URL: https://github.com/nashsu/llm_wiki Comments URL: https://news.ycombinator.com/item?id=48976240 Points: 1 # Comments: 0
Futronic, founded in 1993, specializes in high-precision actuators for automotive and robotics.
A solo Russian-speaking threat actor known as "bandcampro" outsourced a chunk of their operations to Google's open-source Gemini CLI artificial intelligence (AI) and commandeered a live botnet. The findings come from an analysis of 200 Gemini CLI session logs between March 19 and April 21, 2026, which found the threat actor using AI, among other things, to crack passwords, set up a residential
<pre style='white-space:pre-wrap;width:81ex'>fix(doctor): check every configured agent (#111758)</pre>
Save time with Voibe, an offline AI voice dictation app for Apple Silicon Macs that transcribes speech up to three times faster than typing. Lifetime access is $79.99 for a limited time.
Put the kettle on. ScienceAlert stories are written, fact-checked, and edited by humans, never generated by AI. Don't miss a story, subscribe here.
Moonshot AI’s Kimi K3 open-weight model has been read almost entirely through its parameter count since it launchedon July 16. At 2.8 trillion parameters, it is the largest open-weight model released to date. Model sizes are usually grouped into rough brackets, and 2.8 trillion rounds into what the industry calls the 3T class. A tier no […] The post Kimi K3 open-weight model: China’s biggest AI is a bet on memory, not compute appeared first on AI News .
Article URL: https://medium.com/@marcbuilds/i-built-a-vs-code-extension-that-lets-you-click-a-web-page-element-and-paste-it-into-any-ai-chat-b60d2b424ff7 Comments URL: https://news.ycombinator.com/item?id=48975979 Points: 1 # Comments: 0
<pre style='white-space:pre-wrap;width:81ex'>fix: honor configured default agent in consults (#111749)</pre>
The next time you apply for a job, AI may screen your résumé before any human sees it. But there’s good reason to question whether AI will judge you fairly. Researchers already know that LLMs pick up human biases from their training data. New research suggests that LLMs can also develop their own biases from…
<p>Company aims to develop AI software that cuts research times and use of rare metals in chipmakers’ supply chains</p><ul><li><p><a href="https://www.theguardian.com/business/live/2026/jul/20/oil-price-rises-us-iran-strikes-ryanair-lower-fares-summer-economy-latest-news-updates">Business live – latest updates</a></p></li></ul><p>Amazon’s founder, Jeff Bezos, and the UK government have invested in a £2bn British artificial intelligence startup that is aiming to become the “search engine for rare
Departments haven't worked out how the technology will reshape staffing, roles, and skills across the public sector
Article URL: https://www.spheron.network/blog/ai-data-center-power-constraints-2026/ Comments URL: https://news.ycombinator.com/item?id=48975775 Points: 2 # Comments: 2
Brent crude rose above US$91 a barrel after the US and Iran escalated hostilities around the Strait of Hormuz.
Article URL: https://cacm.acm.org/blogcacm/topological-control-of-llms-a-route-to-trustworthy-ai/ Comments URL: https://news.ycombinator.com/item?id=48975640 Points: 1 # Comments: 0
Article URL: https://cacm.acm.org/news/the-dawn-of-ai-change-control/ Comments URL: https://news.ycombinator.com/item?id=48975633 Points: 1 # Comments: 0
Your Bold Pilot agent finds the keywords where you're almost ranking, writes the article that closes the gap, and publishes it to your blog — on autopilot, week after week. Comments URL: https://news.ycombinator.com/item?id=48975617 Points: 1 # Comments: 1
Like subagents but much simpler, sequential and with more control for the user. Instead of describing the loop to the agent, just make a loop. No need to tell the agent about your control structure if you already know the control structure. Comments URL: https://news.ycombinator.com/item?id=48975615 Points: 1 # Comments: 0
Moonshot has temporarily stopped selling new subscriptions for its Kimi K3 model as demand nearly maxed out its GPU capacity within 48 hours. The company plans to split its subscription model to spread computing power more evenly. The article Moonshot pauses new Kimi K3 subscriptions after GPU demand maxes out in 48 hours appeared first on The Decoder .
<pre style='white-space:pre-wrap;width:81ex'>feat(anthropic): import Claude Desktop custom groups (#111644) * feat(anthropic): import Claude Desktop custom groups * fix(anthropic): resolve Claude Desktop groups from LevelDB entries Group discovery regexed whole decompressed blocks, so matches were not attributable to a Local Storage value and byte order decided the winner. On a real 5.4MB store that mislabelled 59 of 159 sessions, surfacing a mojibake label spliced out of Snappy copy-record byte
<p>I trained a model that scored <strong>4 percentage points better</strong> than the one I ship. It's gone. I<br> deleted it from the roadmap yesterday.</p> <p>That's not the interesting part. The interesting part is <em>how</em> I found out it was broken — because<br> it passed every check I had, and then shipped a file that answered <code>neutral</code> to every question you<br> could ask it.</p> <h2> The setup </h2> <p>I maintain <a href="https://github.com/labrat-akhona/semantix-ai" rel="no
<p><em>Third entry in the DEV x Sentry Bug Smash. <a href="https://dev.to/himanshu_748/i-fixed-a-smolagents-bug-that-confused-everyone-who-hit-it-with-sentry-watching-the-whole-time-1im">Entry 1</a> was a crash with a confusing message. <a href="https://dev.to/himanshu_748/one-line-of-math-froze-my-ai-agent-forever-the-timeout-watched-and-did-nothing-2dma">Entry 2</a> was a freeze the timeout could not catch. This one is quieter and sneakier: valid Python that the sandbox rejects with an error p
<p>I spend most of my working day inside Claude Code. At some point I started<br> wondering what that adds up to, physically. I just wanted a number next to<br> the session cost, something to glance at rather than feel bad about.</p> <p>There is no official answer. Anthropic, like every frontier lab, does not<br> publish per-inference energy data. So I built the estimate myself, put it in<br> my status line, and open-sourced the whole thing as<br> <a href="https://github.com/gwittebolle/claude-c
<p>If you bumped a Lambda function to <code>nodejs22.x</code> and your handler now dies on the first crypto call, this is the wall:<br> </p> <div class="highlight js-code-highlight"> <pre class="highlight plaintext"><code>TypeError: crypto.createCipher is not a function </code></pre> </div> <p>The function is simply gone from the <code>crypto</code> object in Node.js 22. Here is the full story and the migration path.</p> <h2> Why it was removed </h2> <p><code>crypto.createCipher(algorithm, passw
<p>การทดสอบของคุณผ่านเมื่อวันจันทร์ ด้วยอินพุตเดิม โค้ดเดิม และ <code>temperature=0</code> แต่เมื่อวันอังคารมันกลับล้มเหลว ทั้งที่คุณไม่ได้เปลี่ยนอะไรเลย assertion ตรวจสอบสตริงที่ตรงกันเป๊ะ แต่โมเดลตอบด้วยถ้อยคำที่ต่างออกไปเล็กน้อย การทดสอบขึ้นสีแดง แม้เอเจนต์ยังทำงานได้ดี และคุณต้องดีบั๊กชุดทดสอบแทนผลิตภัณฑ์</p> <p><a href="https://apidog.com/?utm_source=dev.to&utm_medium=wanda&utm_content=n8n-post-automation" class="crayons-btn crayons-btn--primary">ลองใช้ Apidog วันนี้</a> </p> <p>นี่คือภาระข
<p>You spend time optimizing a function to be 40 milliseconds faster, but then realize your cloud bill is so high that it's eating up all your savings.</p> <p>I wish someone had told me over a cup of coffee, "<strong>you lose money on the network.</strong>" (Here's the coffee chat I wish I'd had.)</p> <h2> The $30k hallway tax </h2> <p>I heard about a startup that received a warning. They were about 48 hours away from a $30,000 AWS invoice.</p> <p>It's not from a computer. Not from a GPU farm ru
