AI news for Friday, 30 October 2026
Stories 121 to 160 of 160, most important first. Back to the first page →
-
Databricks CEO Ali Ghodsi says enterprises should move from tokenmaxxing to valuemaxxing
Enterprises that launched AI projects with the biggest, most capable models are now confronting rising bills. Databricks co‑founder and CEO Ali Ghodsi calls the early phase “tokenmaxxing,” where teams equate high token…
1 sourceUnite.AI -
UK CMA opens probe into Meta's exploitative terms for AI use
The UK Competition and Markets Authority (CMA) announced on Friday that it has opened an investigation into Meta's terms for publishers and advertisers. The probe focuses on whether Meta's contracts force content…
1 sourcereuters.com -
Google DeepMind's PlanAssist to expand to 12 more UK councils
Google DeepMind is expanding its AI planning tool, PlanAssist, to 12 additional local authorities in England. The tool, built with Gemini, is designed to streamline routine householder applications, which account for…
1 source primary sourceGoogle DeepMind -
Researchers propose PoPLi test to detect covert AI agent collusion
Researchers from arXiv have introduced Partner-Outsider Positive Listening (PoPLi), a behavioral method designed to identify secret coordination between AI agents. The core problem is that agents sharing a hidden…
1 source primary sourcearXiv cs.AI -
AWS publishes five-step playbook for managing Amazon Bedrock costs
AWS has released a detailed guide for FinOps and platform engineering teams to manage generative AI costs on Amazon Bedrock. The playbook outlines a five-stage loop: attributing spend to specific identities, making…
1 source primary sourceAWS Machine Learning Blog -
Crisp introduces Halo for on-premise retail AI in CPG and distribution
Crisp launched Halo, an AI platform designed to run inside enterprise cloud environments like Databricks, Snowflake, and Google Cloud. The tool targets consumer packaged goods (CPG) manufacturers, retailers,…
1 sourceUnite.AI -
Shopify pays AI agent developers commission on promoted products
Shopify has introduced a commission system for developers of AI shopping agents that retrieve products from its catalog. When an agent presents a promoted placement and the shopper completes the purchase inside Shop –…
1 sourceSearch Engine Journal -
Researchers introduce FilingLoop for SEC disclosure review
A new research paper introduces FilingLoop, an interactive benchmark and self-evolving agent framework designed for corporate disclosure review and revision. The system is grounded in historical U.S. Securities and…
1 source primary sourcearXiv cs.AI -
Manufacturers warn AI readiness depends on verified product data logic
Manufacturers are rushing to adopt AI and automation but often overlook a critical foundation: ensuring the product data and configuration rules feeding these systems remain accurate. The issue isn’t just system…
1 sourceUnite.AI -
Wikimedia Foundation seeks leave to file amicus brief in OpenAI copyright case
The Wikimedia Foundation announced it has asked the U.S. District Court for the Southern District of New York for leave to file an amicus brief in the consolidated OpenAI copyright litigation. The brief, 5,831 words,…
1 sourceUnite.AI -
Complete Dell workstation with 96GB NVIDIA GPU may cost less than graphics card alone
NVIDIA RTX PRO 6000 Blackwell graphics cards have become so expensive at US retailers that buyers are finding cheaper options by purchasing complete workstations. One example comes from Dell, where a Pro Max Tower T2…
1 sourcevideocardz.com -
Google’s cash burn drags down AI optics suppliers by 10-15%
Google’s $9.1 billion quarterly cash burn—double analyst expectations—sent shockwaves through AI optics suppliers this week. Applied Optoelectronics, Coherent, Lumentum, and Marvell all saw their stocks tumble 10% to…
1 source247wallst.com -
Study finds AI agents lose intent in smart contract synthesis
A new arXiv paper examines the gap between what AI agents are asked to build and what they actually produce, focusing on smart contracts for Web3 applications. The researchers argue that executable code is not the same…
1 source primary sourcearXiv cs.AI -
Viz3d skill lets Claude Code build and self-correct 3D simulations
A Japanese marketer tested 'viz3d', an open-source skill for Claude Code, to build 3D applications. The tool, published by codepawl under an MIT license, allows users to create interactive 3D scenes by describing them…
1 sourcenote.com -
AI Agents vs. Copilots vs. Workflows: What’s the Difference?
Copilots assist users in tasks while keeping human oversight, like drafting emails or suggesting next steps. Workflows follow predefined steps, such as customer-support ticket routing, where models handle variable…
1 sourceUnite.AI -
Researchers release OpenFrontiers benchmark with 12,151 open problems and 200 math tasks
A team of researchers has published OpenFrontiers, a new resource that pairs a corpus of 12,151 open research problems across disciplines with a benchmark of 200 mathematics problems. The benchmark includes 110…
1 source primary sourcearXiv cs.AI -
Google Gemini adds custom skills and external app integration
On October 30 Google’s Gemini team announced a “Gemini Drops” update that adds two productivity‑focused features. Custom Skills let users save frequently used instructions and run them repeatedly, which can streamline…
1 sourcenote.com -
Pudu Robotics opens European headquarters in Neuss, Germany
Pudu Robotics announced on October 30, 2026 that its new European headquarters in Neuss, Germany is fully operational. The facility consolidates sales, product demonstrations, training, warehousing, and after-sales…
1 sourceUnite.AI -
Developer shows how to control air-gapped PC with Claude via HDMI and USB
A Japanese developer published a paid article describing a hardware method to let an AI model on one computer operate a second, isolated machine without installing software or using the cloud. The setup uses PC2 (which…
1 sourcenote.com -
Anthropic CEO Dario Amodei warned of AI becoming too autonomous at 2023 conference
At Fortune's 2023 Brainstorm Tech conference, Anthropic CEO Dario Amodei outlined short, medium, and long-term risks of AI. He described short-term risks as bias and misinformation, medium-term risks as harmful uses of…
1 sourcefinance.yahoo.com -
LiquidAI releases browser demo for LFM3-Omni voice AI
LiquidAI has made a public demo available on Hugging Face that runs the LFM3-Omni-2B model directly in web browsers using WebGPU. This approach allows users to experience local voice AI without installing specialized…
1 sourcenote.com -
Zest launches AI text agent to book restaurant reservations
Zest, a restaurant discovery startup founded in November 2024, has launched a new AI-powered concierge that operates via text messages. The feature allows users to receive personalized dining recommendations and book…
1 sourceTechCrunch AI -
OpenAI reportedly plans first report in early 2027, unconfirmed
A post on X claims OpenAI will publish its first formal report in early 2027, but the company has not confirmed this. The post does not specify what the report would cover, who would write it, or if it would be a…
1 sourcecryptobriefing.com -
Ramuko delegates sales priority judgment to Claude Code AI
Ramuko, a self-described AI manager, now trusts Claude Code to decide which sales leads to prioritize based on reply speed and question specificity. Previously, only humans made such calls. The AI now orders contacts,…
1 sourcenote.com -
RoboBusiness 2026 marks final event under its name before rebranding
The Robot Report podcast recapped the 20th-anniversary edition of RoboBusiness, held October 20–21 in Santa Clara with over 2,000 attendees. Highlights included keynotes on humanoid robots and indoor navigation, a RaaS…
1 sourceThe Robot Report -
Inco AI releases GGUF conversion of DFlash 2 drafter for Qwen3.8-Flash-Next
Inco AI has released GGUF conversions of the DFlash 2 drafter model for use with llama.cpp's draft-dflash speculative decoding. The drafter is designed to work with the Qwen3.8-Flash-Next target model, proposing token…
1 source primary sourcehuggingface.co -
Researchers propose Verification Learning Theory to model AI-assisted judgment checks
A new paper on arXiv introduces Verification Learning Theory (VLT), a formal model explaining how AI assistance affects human judgment and learning. The theory suggests that reliable AI tools can improve immediate…
1 source primary sourcearXiv cs.AI -
Zapier ranks top AI agent builder for enterprise workflows
Zapier’s blog lists the best AI agent builders for 2026, highlighting tools for enterprises, technical teams, and personal use. The author, a Zapier employee, emphasizes Zapier as the best for organizations needing…
1 sourceZapier Blog -
Microsoft, educators rethink entry-level work
As companies adopt AI across routine workflows, leaders say the shift is changing how early-career employees build experience. Microsoft redesigned junior roles to focus on judgment and engagement, while schools…
1 sourcetradersunion.com -
Creator summarizes recent AI trends from OpenAI, Google, and Anthropic
A new video summary reviews recent developments in the artificial intelligence industry, focusing on major announcements from OpenAI, Google, Anthropic, and Meta. The content argues that the pace of AI evolution is…
1 sourcenote.com -
Kentucky residents play data center noise at hearing
Residents in Carroll County, Kentucky, disrupted a Fiscal Court meeting on October 22 by playing loud recordings of mechanical noise from an operating data center in Florence. The demonstration occurred during a…
1 sourcethenerdstash.com -
Claude Code Mods add custom hooks for API security and code testing
A developer built four custom hooks for Claude Code to address common frustrations in AI-assisted coding. The mods protect API keys, track plan progress, catch untested code, and monitor context usage. The author…
1 sourceTowards Data Science -
TechCrunch suggests last-minute tech-themed Halloween costumes
TechCrunch published a list of last-minute Halloween costume ideas for tech enthusiasts, drawing on recent industry news and cultural moments. The suggestions include dressing as a "recursive self-improvement" AI, a…
1 sourceTechCrunch AI -
Scoring Explanations Without Reading Them: Structural Shortcuts in Automated-Interpretability
A new study reveals that automated interpretability scores for sparse-autoencoder features can be calculated without actually reading the explanations. Researchers found that over 90% of activating examples are cropped…
1 source primary sourcearXiv cs.CL -
Researchers propose rank-aware method for private listwise LLM alignment
Existing methods focus on pairwise comparisons, but listwise feedback—where full rankings are provided—reveals more sensitive user data. The paper highlights that applying local differential privacy (LDP) to listwise…
1 source primary sourcearXiv cs.CL -
User reports Claude fails to predict router save interruption
A developer using Claude to build a sake monitoring app experienced a temporary connection loss after saving static IP settings on their home router. Before saving, the user asked Claude if the change would affect…
1 sourcenote.com -
Tom’s Hardware explains key AI concepts for local model users
Tom’s Hardware breaks down technical terms for users running AI models locally, focusing on text-based LLMs. The article defines core concepts like weights—the numerical values determining token relationships—and…
1 sourceTom's Hardware -
New guide introduces LLMs for scientists and professionals
Written for a master’s course in AI in medicine at the University of Turin, Italy, the document covers core concepts like tokenization, embeddings, training, and prompting—without deep math. It highlights key…
1 source primary sourcearXiv cs.CL -
Microsoft AI chief recalls 2015 lunch where Larry Page predicted von Neumann probes
Mustafa Suleyman, the head of Microsoft AI, told a podcast that he met Google co‑founder Larry Page in a SpaceX canteen in 2015. During the first meeting of DeepMind’s AI Ethics and Safety Board, Page described a…
1 sourcetimesofindia.indiatimes.com -
Localized posterior inference may reduce predictive errors in ocean turbulence closure
The paper introduces a new calibration approach that constructs localized posterior distributions by weighting observation likelihoods according to their proximity to a target context. Gaussian localization weights…
1 source primary sourcearXiv cs.AI
