AI News Today

<!-- SC_OFF --><div class="md"><p>I'm working through a free series on GPU kernel engineering (my own channel; I posted the roadmap here earlier: <a href="https://www.reddit.com/r/MachineLearning/comments/1xa4h2g/open_gpu_kernel_engineering_series_naive_gemm_to/">https://www.reddit.com/r/MachineLearning/comments/1xa4h2g/open_gpu_kernel_engineering_series_naive_gemm_to/</a> ) . Session 1 is a setup session. While preparing it I collected a few things that cost me time and that I rarely see in one

Promises to build ‘IBM Z Atlas’ tool that will tidy up messy big iron and enable AIOps

<p> Computers and sandboxes for your agents </p> <p> <a href="https://www.producthunt.com/products/paddock-2?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/1294009?app_id=339">Link</a> </p>

OpenAI released GPT-6.1 Terra, a fast computer-use and vision model priced at $1/$5 per million tokens. It matches GPT-6 Astra’s medium-effort Agents’ Last Exam score at $0.29 per task versus $4.82. The post OpenAI Launches GPT-6.1 Terra, Halving GPT-6.1 Sol’s API Price for Agent Work appeared first on MarkTechPost .

<!-- SC_OFF --><div class="md"><p>Recently I've been getting bombarded with reach outs from companies that build RL environments and sell to frontier labs like Anthropic, OAI, etc. I'm not really understanding how and why these companies exist. </p> <p>Lets say you somehow create an environment that no other model is able to perform well on. Your second criteria would be that this environment is suitable for RLVR, perhaps you have some metric of gradient information per example in the dataset. S

"May be a surprisingly effective route to HIV cure." ScienceAlert stories are written, fact-checked, and edited by humans, never generated by AI. Don't miss a story, subscribe here.

<p>Tasks from determining asylum seekers’ ages to border surveillance now involve AI, says Open Rights Group </p><p>AI has been embedded across the Home Office in near-total secrecy, reinforcing the so-called hostile environment for immigration, according to research.</p><p>Tasks from the detection of potential sham marriages to determining the age of asylum seekers and checking through the contents of phones of new arrivals now involve use of the technology, according to a report by Open Rights

The Vercel Plugin is now available in OpenCode . The plugin provides on-demand guidance on Vercel APIs and recommended patterns, with skills for AI SDK, Vercel Functions, Workflow, and more. It also includes specialist agents for deployment, performance optimization, and AI application design. Use /vercel:deploy to deploy projects and /vercel:env to manage environment variables. The plugin requires OpenCode 2. To install it, run this command in your terminal: Then run /mcps in OpenCode, select v

You can now query AI traffic in Vercel Observability to understand how AI bots are accessing your site. AI traffic insights show: Requests by verified AI bot: how many requests each bot makes Most-crawled pages: pages most frequently requested by AI crawlers for indexing or model training Assistant retrievals: pages AI assistants fetch in response to user actions or search queries Markdown requests: pages AI bots request in Markdown format Errors: pages where requests from AI bots return errors

Vercel Agent sessions in the dashboard and Slack now cost 67% less on average. We rebuilt how Agent manages context, reducing the tokens needed per session and improving prompt-cache reuse. These improvements are live and require no action. What changed: Workflow-based CLI skills: Updated Vercel CLI skills give agents clearer guidance for common workflows and error recovery. Other agents using these skills also benefit. Incremental context management: Agent manages context more efficiently as co

The AI SDK HarnessAgent layer now accepts image inputs across every harness adapter. Pass images alongside text in the standard AI SDK message format: a content array mixing { type: 'text' } and { type: 'file' } parts. Image input is processed by the harnesses' underlying vision models and supported across all available adapters: Claude Code, Cline, Codex, Cursor, Deep Agents, fx, GitHub Copilot, Grok Build, OpenCode, Pi. Read the HarnessAgent image inputs documentation to get started. Read more

With OpenAI Daybreak, Check Point speeds up the research behind new security protections, enabling its analysts to turn vulnerability intelligence into customer defenses faster.

<p><strong><a href="https://openai.com/index/introducing-gpt-6-1-terra/">GPT-6.1 Terra</a></strong></p> It's always interesting to note which aspects of a new model are emphasized in the release announcement.</p> <p>In the case of GPT-6.1 Terra, OpenAI went with visual reasoning and computer use:</p> <blockquote> <p>We’re introducing <strong>GPT-6.1 Terra</strong>, a fast, efficient model that combines detailed visual reasoning and precise computer use alongside clearer, more natural writing. A

It's more important than chips, 'leccy and the net, after all

Less than a month after its mass release of hundreds of manuscripts that it said contained solutions to long-standing mathematics questions, OpenAI is back with three more "major" discoveries in math, computer science, and physics. What's different this time is the choice to have external experts that OpenAI describes as "shepherds" guiding the results in […]

<p>Tech firm announces two-part fundraising after Elon Musk warns hundreds of AI startups face collapse</p><p>Mark Zuckerberg’s Meta has announced plans for $95bn (£71bn) of fundraising to cover its AI expenditure, even as Elon Musk warned that funding for smaller players in the sector was drying up.</p><p>The owner of Facebook, Instagram and WhatsApp is a leading investor in AI and is ploughing multibillion-dollar sums into datacentres, the central nervous systems of the technology.</p> <a href

Article URL: https://github.com/jdavid465/JDA-ZIP-PROTOCOL Comments URL: https://news.ycombinator.com/item?id=50252604 Points: 2 # Comments: 0

A hospital staffing gap often becomes most expensive when there is little time left to solve it. Managers scramble for coverage, nurses field last-minute requests, and overtime or premium labor fills the remaining holes. symplr is trying to move those decisions earlier with the launch of symplr Time and Scheduling, an integrated workforce management solution that combines AI-powered scheduling with timekeeping. Announced October 26, the product brings anticipated staffing demand, available…

Like no other amphibian we've ever heard of. ScienceAlert stories are written, fact-checked, and edited by humans, never generated by AI. Don't miss a story, subscribe here.

Not a bombshell, but it rivals pricier models on some tasks at a tenth of Astra's API price

Cloudflare on October 26, 2026 published an incident report disclosing a cross-request KV cache exposure in the GLM-5.3 model on its Workers AI platform, under which 155 responses to very long prompts contained data from other customers' sessions. Security researcher Gadi Evron of Knostic reported the tenant-isolation issue to Cloudflare on September 8, 2026 through the company's HackerOne program. The exposure affected the model, served as @cf/zai-org/glm-5.3, starting August 27, 2026, the day…

Article URL: https://www.brangus.io/survival-llm Comments URL: https://news.ycombinator.com/item?id=50252314 Points: 2 # Comments: 0

<p> Click any UI bug on localhost and your AI agent fixes it </p> <p> <a href="https://www.producthunt.com/products/pulsedev?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/1293889?app_id=339">Link</a> </p>

It's widely believed AI's inner workings are largely inscrutable. A former OpenAI researcher now says this narrative could be wrong. The post Is AI Really a Black Box? Neural Networks May Be Easier to Understand Than We Thought appeared first on SingularityHub .

Months after Elon Musk declared that the xAI model would outclass Christopher Nolan’s mythological epic, the bot itself admitted defeat.

<p>If proposed copyright reforms proceed, Australia will become a global IP-laundering paradise. Why would a UK business license its TV shows here if they can be scraped at will?</p><p>Australia and the UK have always had a shared love of creativity. We consume each other’s TV, news, music and books. Our chefs appear on each other’s TV channels and quite frankly I was bereft when the ABC’s The Newsreader ended its run in the UK.</p><p>That shared love though, appears not to be reciprocated by th

OpenAI said in an announcement on October 26, 2026, that its internal models produced three research results in mathematics, computer science, and physics, including a disproof of the Exponential Time Hypothesis, a longstanding conjecture in computational complexity theory. In the company's view, the AI-generated manuscripts amount to raw research data that must still be converted into scientific understanding, and it is experimenting with a new release approach toward that end. OpenAI asked…

Article URL: https://knell.digital/alan-watts-predicted-llm-usage-in-1951/ Comments URL: https://news.ycombinator.com/item?id=50251716 Points: 2 # Comments: 1

A configuration mismatch in GLM 5.3 on Workers AI triggered an SGLang bug that exposed KV cache pages across requests. Here is how we addressed the issue.

A step toward bigger machines. ScienceAlert stories are written, fact-checked, and edited by humans, never generated by AI. Don't miss a story, subscribe here.

How Greptile uses GPT-6 Astra to test user workflows, Daybreak to strengthen security, and Codex to reduce rework.

As it prepares to go public, Anthropic's S-1 filing makes generative AI's methane dependency glaringly obvious. "Competition for access to electrical supply and data center locations with adequate power availability is intense." In documents filed with the US Securities and Exchange Commission Monday, Anthropic acknowledges that data centers are becoming increasingly reliant on "natural gas" […]

<p> AI code review that learns your repo's own rules </p> <p> <a href="https://www.producthunt.com/products/revix-ai-2?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/1293843?app_id=339">Link</a> </p>

Vertiv Holdings Co introduced the Vertiv SmartIT solution on October 26, 2026, a pre-engineered, factory-validated infrastructure platform for high-density AI deployments based on NVIDIA MGX and Open Compute Project ORV3 architectures. The platform is available now for AI deployments globally and is designed for enterprises, research institutions and colocation providers across existing facilities and new data centers. Vertiv said the scalable platform supports deployment across existing…

But there's not much evidence this is anything more than echoes of unsanitized training data

Anthropic’s S-1 filing is hundreds of pages long—and it leaves out quite a few financial details. “Anthropic owes investors more information,” one analyst tells WIRED.

Bedrock Robotics has secured $600 million to scale autonomous heavy equipment and combat construction labor shortages. The post Bedrock Robotics nears $1B in total funding as demand for autonomous heavy equipment surges appeared first on The Robot Report .