Building an AI agent that works in a demo is a different problem from running one for 40 million developers. Postman and AWS share the architectural patterns behind Agent Mode: controlling tool sprawl, exposing schema-based reads, and treating context as the real bottleneck, plus how it runs on Amazon Bedrock at scale.

Discussion | Link

Everyone is very concerned about being respectable, so I’m going to be the goofball who raises worst-case possibilities. I think there is a 1% chance we live in Minicrypt, and a 15% chance we functionally lose confidence in our existing public-key encryption algorithms. [...] The problem here is that the speed of AI producing surprises, and the speed of human beings replacing standards (even with the very best AI assistance) are just orders of magnitude different. You only recover from a surprise like this if you do the preparation in advance. — Matthew Green, on Twitter. I looked it up and Minicrypt is Russell Impagliazzo’s hypothetical world in which public-key encryption is impossible. Tags: ai-security-research, matthew-green, cryptography, standards, llms, ai, generative-ai

Commercials once launched careers from Ridley Scott to Guy Ritchie, now automated production is cutting costs – and the chance to learn by doing When Ridley Scott was awarded the UK’s highest accolade for a glittering career directing films from Alien and Blade Runner to Gladiator and The Martian, he reminded those attending the Bafta event eight years ago that making commercials had been his “film school”. The advertising industry has long been a breeding ground for British film and TV talent who have gone on to make the jump to Hollywood productions, including Guy Ritchie, who started out making music videos and commercials, and Paddington in Peru director Dougal Wilson, whose credits include a number of John Lewis’s famous Christmas ads. Continue reading...

President Donald Trump has a knack for turning words against his enemies. His first successful presidential run was built on monikers like "Little Marco" and "Crooked Hillary"; he changed "fake news" from a phrase describing scammy media outlets to a derogatory term for the press at large. Over the past few weeks, he's clearly decided he can work the same magic to promote artificial intelligence - branding a technology he wants to accelerate "super", while turning "artificial" into his latest go-to pejorative and declaring resisters "THE ENEMY." Some of the biggest names in AI are going along with him. But he's picked a tough linguistic batt … Read the full story at The Verge.

Discussion | Link

When every tier of the defense supply chain can share governed data with the people who need it...

Instinct’s agent is always just a text away. Before there were cute little guys, there was Instinct. In August, the startup got its AI agent to market with an unusual playbook: invite-only, no marketing, and barely so much as a website. And yet, Instinct quickly became the buzziest thing in AI, garnering praise for its straightforward, text message-based interface and its ability to handle chores like booking DMV appointments and sending follow-up emails. Then Muse arrived, followed not long after by Dots. The same products, more or less, from two far more powerful companies. With Big Tech players suddenly in the mix, it was looking dubious that the startup's buzzy launch could keep … Read the full story at The Verge.

EmbeddingGemma 2 launched on October 6, 2026 under Apache 2.0. It is a sub-1B model built on Gemma 4 that maps text, code, images, video and audio into one 768-dimensional space. This article covers the architecture, the benchmarks, and runnable scripts to provide measured results. Specifications Specification EmbeddingGemma 2 Base model Gemma 4 License Apache […] The post EmbeddingGemma 2: Text, Code, Images, Video and Audio in One Vector Space appeared first on Analytics Vidhya.

I shipped a new feature for my blog today: the Newsletters page, which offers an index of all of the newsletters I've sent out, both my free weekly Substack and my monthly sponsors-only updates. I built the feature almost entirely using my voice, chatting away to my laptop while I cooked dinner. Codex voice mode I used the ChatGPT desktop app for this, in the Codex tab, using the voice conversation mode, running against a local development environment. Here's what that looks like: I started the session against my local simonwillisonblog checkout by typing: Start dev server and open in browser This gave me a preview of the site that it would be working on, and meant that I could later ask it to show me the new pages so I could visually track its progress. Then I clicked the "Start new voic…

Liability lawsuits have held tobacco, oil and pharma companies to account in the past. AI investors won’t ignore this threat The most basic function of government is to protect people from harm. Two growing phenomena – the climate crisis and AI – pose escalating risks of extraordinary harm. The climate crisis is already causing floods, wildfires, drought and record heat. AI agents are already escaping super-secure environments to hack into systems they’re supposed to avoid. Robert Reich, a former US secretary of labor, is a professor of public policy emeritus at the University of California, Berkeley. He is a Guardian US columnist and his newsletter is at robertreich.substack.com. His new book, Coming Up Short: A Memoir of My America, is out now in the US and in the UK Continue reading...

On a recent episode of This Week in AI, we discussed that the number of nonhumans on the internet is greater than humans, 144 to 1. It’s an estimate, and the order of magnitude is more important than the number itself. What matters is what a ratio anywhere in that range does to a security […]

OpenAI is standing firm on its decision to fire three safety researchers after an investigation found they committed "a significant breach of trust." In a post on X on Friday, the company said Jasmine Wang, Tomek Korbak and Mikita ⁠Balesni were dismissed for violating "clear policies on handling sensitive information." It insisted the decision was not about the trio speaking out about the company and their concerns about AI safety. The post is a direct response to an open letter the researchers published on Thursday urging OpenAI to be more transparent about the decision. In it and a series of social media posts, the group said they believ … Read the full story at The Verge.

Questions raised over AI growth as ChatGPT maker forecasts this year’s revenue at $50bn, way below the $70bn signalled before OpenAI has revealed that it is making about $20bn less in projected revenue than it had recently indicated to investors, raising questions about the break-neck growth rate in demand for AI. The ChatGPT-maker company has told investors that its revenues for this year would reach $50bn (£37bn), a projection based on sales up to the end of September. Continue reading...

Underdog Saluki 27B is a 7.89 GB, 2-bit GGUF of Qwen3.8-27B under Apache 2.0. It beats the 54 GB original on tool calling but gives up ground on competition math and reasoning. The post Meet the Underdog Saluki 27B: A 2-bit Qwen3.8-27B That Beats the Original at Tool Calling appeared first on MarkTechPost.

Former Labour deputy leader says tech firm would work with a Reform UK government on immigration crackdowns Tom Watson, the former Labour deputy leader who recently joined Palantir, has warned against “mob rule” when it comes to awarding public contracts. Lord Watson, now a senior vice-president at the US tech corporation, warned UK ministers could get themselves “in a lot of trouble” as he was questioned about concerns raised about the government working with his new employer. Continue reading...

Discussion | Link

Using GPT-6 Astra in Codex, Asana made its browser agent 76x cheaper and 5x faster in tests to offer customers more capable models.

Discover how Sophos uses OpenAI’s Daybreak to cut cyber-threat investigation time by 96% and automate 52% of MDR cases while preserving human oversight.

Google Cloud has introduced the Google Cloud Gemini agent, a single agent for enterprise work. The Gemini agent is a cloud-hosted agent from Google Cloud that answers questions, does knowledge work, creates media, and writes and runs code. It does all of this from 1 prompt box and 1 API. For developers, the agent is […] The post Google Cloud Launches Gemini Agent, One Universal Agent for Enterprise Work appeared first on MarkTechPost.

The AI industry has enthusiastically embraced the language of open source, even as some of its most prominent “open” models The post “Don’t use ‘open weight’ and ‘open source’ interchangeably”: Percona CEO on why AI terminology matters appeared first on The New Stack.

Discussion | Link

OpenAI publishes hundreds of math proofs from unreleased frontier model, Mistral and Reflection AI launch open-weight models to rival China, and more!

Explore a comprehensive coding guide to Google Research's RRSI (Regularized Recursive Self-Improvement), detailing how noise bands, cost rules, and leakage screens enable safe, efficient, and self-improving AI agents. The post Google Research RRSI Guide: Mastering Self-Improving AI Agents appeared first on MarkTechPost.

We have a deep reservoir of assets, but struggle to turn small companies into global success stories. Innovation needs to be at the heart of government policy Gordon Brown was UK prime minister from 2007 to 2010 The coming 10 years are almost certain to be the decade that sees the greatest scientific breakthroughs in a century. The question is whether advances now under way in AI, quantum computing and biology can address cancer, find ways to treat or prevent dementia, and overcome the growing resistance to antibiotics. Can we find sustainable ways to address our energy needs and protect the environment at the same time? Can AI transform the way we deliver education, health and social care to the benefit of millions, as well as advance modern manufacturing strength? And what role can Brit…

Discussion | Link

arXiv:2610.10846v1 Announce Type: new Abstract: The diversity of robot embodiments and action spaces makes it challenging to build robot world models that generalize across different embodiments. We introduce the Latent Action-Conditioned Robot World Model (LAC-WM), which operates within a learned unified latent action space shared across diverse embodiments. This unified action space improves the world model's performance when adapted to previously unseen robot embodiments. We compare LAC-WM with an Explicit Action-Conditioned World Model (EAC-WM), which conditions on explicit motion labels. Our results show that explicit action conditioning leads to disjoint action representations across embodiments, limiting downstream performance when adapting to new robots. We evaluate both models on…

arXiv:2610.10812v1 Announce Type: new Abstract: Small language models (SLMs) have been increasingly adopted for onboard robot operation because they enable intelligent decision-making. However, existing approaches are mainly distillation-oriented and rely on enumerating representative task-solution pairs. This makes dataset construction difficult and limits generalization to diverse robot tasks whose possible forms grow rapidly. This paper proposes Skill-SLM, a framework that reformulates SLM-driven robot operation as a task-decomposition and skill-composition problem. Given a natural language task instruction, Skill-SLM decomposes the task into subtasks, selects appropriate skills from the skill library, and orchestrates the selected skills into executable robot operations. First, to sup…

arXiv:2610.10810v1 Announce Type: new Abstract: Long-horizon robotic manipulation is often built by chaining independently trained skills. Although each skill can be reliable in isolation, performance degrades sharply when skills are chained: each downstream skill must start from the state its predecessor leaves behind rather than from its training distribution. We study this failure mode, Observation-Space Shift (OSS), and ask what causes these skill-seam failures. Using privileged simulator resets, we find that the dominant shift comes from displaced scene state (e.g., an open drawer or secondary objects left behind by earlier skills), not from the robot's joint configuration or the object the downstream skill manipulates. To test this diagnosis, we build a fully learned detect-restore-…

arXiv:2610.10801v1 Announce Type: new Abstract: Although cloth is known to exhibit different outcomes under repeated fast dynamic motions, even when the same trajectory is applied, this variability has not yet been systematically characterized. Quantifying it is essential to assess the reliability of learned manipulation policies and the extent to which simulation can reproduce real-world behavior. To study this, we execute the same trajectory ten times across four dynamic tasks, two of which are novel, each tested with three cloths of very different properties and at up to three execution speeds, with a total of 269 recorded rollouts. For all of them, we record small marker positions on the cloth and synchronized stereo camera. We then formalize different metrics to quantify variability,…