{"id":7078847,"date":"2026-10-07T03:36:20","date_gmt":"2026-10-07T03:36:20","guid":{"rendered":"https:\/\/fivemor.com\/?p=7078847"},"modified":"2026-10-07T03:36:20","modified_gmt":"2026-10-07T03:36:20","slug":"radar-trends-to-watch-october-2026-oreilly","status":"publish","type":"post","link":"https:\/\/fivemor.com\/?p=7078847","title":{"rendered":"Radar Trends to Watch: October 2026 \u2013 O\u2019Reilly"},"content":{"rendered":"<p> <br \/>\n<\/p>\n<div id=\"postContent-content\">\n<p class=\"wp-block-paragraph\">In addition to the nearly constant stream of model releases, in September we\u2019ve seen price drops, new kinds of models, proofs of long-standing problems in mathematics, and continued investigations into models escaping their sandboxes. (Axios reports investigations into over 10,000 incidents.) AI has infinite patience and is fundamentally probabilistic. Given a difficult or impossible task and an unlimited token budget, an agent will eventually attempt to solve the problem in ways that you don\u2019t expect, and may not want. It\u2019s easy (and correct) to blame inadequate security procedures at the frontier AI labs, but AI adopters must be careful not to make the same mistakes. The humans using AI need to be accountable for what their agents do.<\/p>\n<h2 class=\"wp-block-heading\">AI Models<\/h2>\n<p class=\"wp-block-paragraph\"><em>Model choice is starting to hinge on price and specialization as much as raw benchmark leadership. Alongside general chat models, there are now decision models that never chat, spatial models built for robot planning and camera control, forecasting models sized for a single task, and cybersecurity-specialized models kept behind an invite-only program. Specialization leads to greater efficiency and lower costs, at least in the short term. In the long term, specialized models may succumb to the \u201c<a href=\"https:\/\/en.wikipedia.org\/wiki\/Bitter_lesson\" target=\"_blank\" rel=\"noopener\">bitter lesson<\/a>.\u201d<\/em><\/p>\n<ul class=\"wp-block-list\">\n<li>Anthropic has <a href=\"https:\/\/www.anthropic.com\/claude-opus-5-5\" target=\"_blank\" rel=\"noopener\">released<\/a> Claude Opus 5.5, which it claims has performance similar to Fable 5.1, and hence similar restrictions. It\u2019s faster and requires fewer resources to run. Anthropic has dropped prices 20% for input and output tokens and 60% for cached reads. Not to be outdone, OpenAI released GPT-6 Sol and Luna, with 50% price reductions.<\/li>\n<li>Anthropic has also <a href=\"https:\/\/www.anthropic.com\/claude-fable-and-mythos-5-1\" target=\"_blank\" rel=\"noopener\">announced<\/a> Fable 5.1 and Mythos 5.1. The most significant change appears to be a 75% price reduction for cache reads, which might translate into significant savings for long-running jobs; Anthropic estimates 25%. Mythos is only available to trusted partners. Simon Willison used Fable 5.1 to <a href=\"https:\/\/simonwillison.net\/2026\/Sep\/1\/claude-fable-5-1\/\" target=\"_blank\" rel=\"noopener\">animate<\/a> his pelican-riding-a-bicycle pseudobenchmark.<\/li>\n<li>Anthropic <a href=\"https:\/\/www.anthropic.com\/claude-sonnet-5-5\" target=\"_blank\" rel=\"noopener\">released<\/a> Sonnet 5.5 with claims that it\u2019s 30% faster and 30% less expensive for most work. The new model has security limitations similar to those applied to Opus and Fable; it routes to Sonnet 5 if it\u2019s asked to do anything out of bounds.<\/li>\n<li>And finally, as September closes, Anthropic announces a <a href=\"https:\/\/www.bleepingcomputer.com\/news\/artificial-intelligence\/anthropic-turns-claude-into-an-ai-marketplace-with-2-000-plus-plugins-and-connectors\/\" target=\"_blank\" rel=\"noopener\">marketplace<\/a> for Claude plugins and connectors. At its launch, Claude Marketplace had over 2,000 items.<\/li>\n<li>OpenAI has <a href=\"https:\/\/openai.com\/index\/gpt-6-astra\/\" target=\"_blank\" rel=\"noopener\">released<\/a> GPT-6 Astra, with claims that the company has achieved AGI (artificial general intelligence). Astra\u2019s excellent benchmark scores appear to depend on the use of an <a href=\"https:\/\/thenewstack.io\/openai-astra-harness-arc-agi-3\/\" target=\"_blank\" rel=\"noopener\">unreleased harness<\/a>. OpenAI has also released GPT-6.1 Sol, with per-token price reductions and claims that it is close to GPT-6 Astra in capabilities<\/li>\n<li>OpenAI has <a href=\"https:\/\/openai.com\/index\/navier-stokes-solution\/\" target=\"_blank\" rel=\"noopener\">solved<\/a> the <a href=\"https:\/\/en.wikipedia.org\/wiki\/Navier%E2%80%93Stokes_existence_and_smoothness\" target=\"_blank\" rel=\"noopener\">Navier-Stokes existence and smoothness<\/a> problem, a mathematical problem in fluid mechanics. This development raises <a href=\"https:\/\/techcrunch.com\/2026\/09\/08\/openai-fought-dirty-on-career-making-math-problem-says-nyu-mathematician\/\" target=\"_blank\" rel=\"noopener\">an ethical question<\/a>: Did OpenAI train its system on the work of two mathematicians who were close to solving the problem themselves? It also raises practical questions about the future of mathematics. Decorated mathematician Terence Tao <a href=\"https:\/\/mathstodon.xyz\/@tao\/117237320796901560\" target=\"_blank\" rel=\"noopener\">asks<\/a> whether \u201cthe collection of good, fruitful open problems is now being mined in a non-renewable fashion.\u201d\u00a0An <a href=\"https:\/\/thenextweb.com\/news\/openai-maths-advisory-group-100-open-problems\" target=\"_blank\" rel=\"noopener\">advisory group<\/a> has been formed to help OpenAI make decisions about releasing mathematical results.<\/li>\n<li>Google has <a href=\"https:\/\/blog.google\/innovation-and-ai\/models-and-research\/gemini-models\/gemini-3-8-text-to-speech\/\" target=\"_blank\" rel=\"noopener\">released<\/a> Gemini 3.8 Flash TTS and Flash-Lite TTS. Voice options aren\u2019t limited to a prebuilt library. These models have APIs that allow developers to <a href=\"https:\/\/thenewstack.io\/gemini-tts-voice-replication-api\/\">describe<\/a> the voice that they want or upload a sample. These custom voices are then assigned an ID so they can be reused.<\/li>\n<li>Google has <a href=\"https:\/\/blog.google\/innovation-and-ai\/models-and-research\/gemini-models\/gemini-3-8-live-gemini-3-8-live-extended-thinking\/\" target=\"_blank\" rel=\"noopener\">announced<\/a> Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking. These models are designed for live, near-real-time conversation. They can process video input. Live Extended Thinking can reason and speak at the same time.<\/li>\n<li>Google has released <a href=\"https:\/\/blog.google\/innovation-and-ai\/models-and-research\/gemini-models\/3-8-flash-and-3-8-flash-cyber\/\" target=\"_blank\" rel=\"noopener\">Gemini 3.8 Flash and Flash Cyber<\/a>. Flash appears to be similar to frontier models on most benchmarks, with Computer Use being the biggest exception. Flash Cyber is specialized for vulnerability detection and mitigation and is only available to defenders in the <a href=\"https:\/\/deepmind.google\/fairwind-program\/\" target=\"_blank\" rel=\"noopener\">Fairwind Program<\/a>.<\/li>\n<li>Google has <a href=\"https:\/\/thenewstack.io\/google-timesfm-3-multivariate-forecasting\/\" target=\"_blank\" rel=\"noopener\">also released<\/a> <a href=\"https:\/\/research.google\/blog\/timesfm-3-a-zero-shot-foundation-model-for-multivariate-forecasting\/\" target=\"_blank\" rel=\"noopener\">TimesFM-3<\/a>, a small specialized model for multivariate time series forecasting. The model weights are available on Hugging Face with a license that only allows for noncommercial use. (Source code is available under the Apache open source license.)<\/li>\n<li>Xiaomi <a href=\"https:\/\/mimo.xiaomi.com\/mimo-v2-6\" target=\"_blank\" rel=\"noopener\">released<\/a> MiMo-V2.6, its latest large language model. It\u2019s fully open sourced, and based on benchmark results, Xiaomi claims that MiMo is the strongest open model to date. What\u2019s more interesting is the claim that MiMo only cost $3.5 million to train.<\/li>\n<li>TypeSafe\u2019s new <a href=\"https:\/\/x.com\/Mappletons\/status\/2101560333441610133\" target=\"_blank\" rel=\"noopener\">decision model<\/a>, <a href=\"https:\/\/typesafe.ai\/blog\/introducing-system-one-models-and-jev\" target=\"_blank\" rel=\"noopener\">Jev<\/a>, is unlike anything else we\u2019ve seen. It doesn\u2019t chat; its output is always strictly typed and accompanied by probabilities that estimate correctness. It\u2019s much faster and less expensive than other leading models. It isn\u2019t open source, but there are already many open source <a href=\"https:\/\/www.latent.space\/p\/ainews-here-are-6-clones-of-jev-in\" target=\"_blank\" rel=\"noopener\">clones<\/a>.<\/li>\n<li><a href=\"https:\/\/ollaya.dev\/\" target=\"_blank\" rel=\"noopener\">Ollaya<\/a> is similar to Ollama, but for running decision open models like Laya (a clone of Jev) locally.<\/li>\n<li>World Labs has released <a href=\"https:\/\/www.worldlabs.ai\/blog\/atlas\" target=\"_blank\" rel=\"noopener\">Atlas<\/a>, a model for \u201cspatial intelligence.\u201d\u00a0It uses text, images, video, and 3D data to perform tasks like planning a robot\u2019s movements or changing the camera position in a photograph.<\/li>\n<li>The rumors that <a href=\"https:\/\/arstechnica.com\/ai\/2026\/09\/nvidia-buys-hugging-face-the-github-of-ai-for-13-billion\/\" target=\"_blank\" rel=\"noopener\">NVIDIA would buy Hugging Face<\/a> are <a href=\"https:\/\/blogs.nvidia.com\/blog\/nvidia-to-acquire-hugging-face\/\" target=\"_blank\" rel=\"noopener\">true<\/a>. NVIDIA is hoping for a proliferation of models that will run on its hardware, and the company promises that Hugging Face will remain a neutral platform, without favoring one model over another.<\/li>\n<\/ul>\n<h2 class=\"wp-block-heading\">Software development<\/h2>\n<p class=\"wp-block-paragraph\"><em>Agents are starting to delegate to, and coordinate with, other agents rather than working solo. Claude Code can break a task apart and hand pieces to other Claude Code instances, Muse Code lets sessions message each other, and Google\u2019s AX orchestrator exists purely to wire up sandboxes and control communications for swarms of agents doing a task together. That shift is pushing developers to rethink what a source repository needs to record, and to start asking how much all this delegation costs.<\/em><\/p>\n<ul class=\"wp-block-list\">\n<li>Now that Jev has caught everyone\u2019s attention, what can you build with it? <a href=\"https:\/\/github.com\/Avinash-jetwani\/jevmem\" target=\"_blank\" rel=\"noopener\">Jevmem<\/a> is a memory manager that hooks into Claude Code and prunes the context at every conversational turn.<\/li>\n<li><a href=\"https:\/\/github.com\/google\/ax\" target=\"_blank\" rel=\"noopener\">AX<\/a> is a new <a href=\"https:\/\/agentexecutor.io\/\" target=\"_blank\" rel=\"noopener\">agent orchestrator<\/a> from Google. It isn\u2019t an agent; it\u2019s intended to coordinate many agents to complete a task. It creates sandboxes, wires up Git repos and other resources, and controls outbound communications.<\/li>\n<li>The <a href=\"https:\/\/thenewstack.io\/claude-code-parallel-projects\/\" target=\"_blank\" rel=\"noopener\">latest<\/a> version of Claude Code can <a href=\"https:\/\/support.claude.com\/en\/articles\/9517075-what-are-projects\" target=\"_blank\" rel=\"noopener\">manage Claude projects<\/a>, breaking a task into subcomponents and delegating the subtasks to other Claude Code instances. Another important change is the ability to read <a href=\"http:\/\/agents.md\/\" target=\"_blank\" rel=\"noopener\">AGENTS.md<\/a> if <a href=\"http:\/\/claude.md\/\" target=\"_blank\" rel=\"noopener\">CLAUDE.md<\/a> isn\u2019t available.<\/li>\n<li>Google\u2019s <a href=\"https:\/\/blog.google\/innovation-and-ai\/models-and-research\/google-labs\/cc-expanding-to-groups\/\" target=\"_blank\" rel=\"noopener\">CC agent<\/a> is designed for families. Family members can share data with CC, which has its own user account. It could be used for filling out forms, synchronizing calendars, and other common tasks.<\/li>\n<li>What will replace GitHub? There\u2019s a <a href=\"https:\/\/thenewstack.io\/zed-delta-github-alternative\/\" target=\"_blank\" rel=\"noopener\">growing consensus<\/a> that we need different kinds of source repositories to deal with the agent-assisted software development. In addition to changes to code, it\u2019s important to record the conversations between the developer and the agent, the architectural decisions, and many other artifacts that don\u2019t make it into traditional source control.<\/li>\n<li>Anthropic is merging its <a href=\"https:\/\/claude.com\/blog\/cowork-is-now-claude\" target=\"_blank\" rel=\"noopener\">Claude Cowork and chat products<\/a>. Anything users type in a chat session is seen by Cowork, and vice versa. Some sensitive information (health, politics, and gender) is excluded. The feature is on by default but can be disabled in settings, and memory isn\u2019t shared with Claude Code. The company also released Claude Docs and Slides.<\/li>\n<li><a href=\"https:\/\/www.bleepingcomputer.com\/news\/artificial-intelligence\/anthropic-wants-claude-to-analyze-your-bank-account-and-financial-data\/\" target=\"_blank\" rel=\"noopener\">Claude Money<\/a> is a new feature that will allow users to connect their bank accounts to Claude for analysis. The product appears to be similar to a product from OpenAI.<\/li>\n<li>OpenAI\u2019s Agents API is <a href=\"https:\/\/openai.com\/index\/introducing-the-agents-api\/\" target=\"_blank\" rel=\"noopener\">now in public beta<\/a>. It allows compaction, session orchestration, and tool use, and it can be deployed in OpenAI\u2019s sandbox, a cloud provider\u2019s sandbox, or the developer\u2019s hardware.<\/li>\n<li>Meta has <a href=\"https:\/\/ai.meta.com\/muse\/\" target=\"_blank\" rel=\"noopener\">released<\/a> Muse, its AI agent. Muse is a \u201cpersonal agent\u201d\u00a0designed for tasks like shopping, filling in forms, and dealing with customer service. It has its own secure credential store, so data like passwords and credit card numbers are never sent offsite.<\/li>\n<li>Some open source projects are <a href=\"https:\/\/www.latent.space\/p\/pr-not-welcome\" target=\"_blank\" rel=\"noopener\">shutting down external pull requests<\/a>, which are largely AI-generated. In some cases, the developer team is using its own agents to create and manage PRs; some are using AI agents to triage external PRs.<\/li>\n<li>Meta has launched <a href=\"https:\/\/developer.meta.com\/ai\/resources\/blog\/muse-code-new-plans-and-features\/\" target=\"_blank\" rel=\"noopener\">Muse Code<\/a>, another competitor to Claude Code. One important new feature is the ability to send messages to other Muse Code sessions, allowing agents to coordinate on complex problems.<\/li>\n<li>AI providers appear to be <a href=\"https:\/\/thenextweb.com\/news\/openai-outcome-based-pricing-enterprise\" target=\"_blank\" rel=\"noopener\">moving toward outcome-based pricing<\/a>, at least for major corporate customers. Rather than billing by token, customers are billed for completed tasks. That approach begs the question: When is a task completed?<\/li>\n<li>Now that organizations are concerned with AI budgets, the question of how to evaluate the cost of different models and agents becomes important. <a href=\"https:\/\/thenewstack.io\/agent-harness-token-costs\/\" target=\"_blank\" rel=\"noopener\">What should platform teams measure?<\/a><\/li>\n<li><a href=\"https:\/\/chatgpt.com\/work\/\" target=\"_blank\" rel=\"noopener\">ChatGPT Work<\/a> was designed to compete with Claude Cowork, Microsoft Copilot Cowork, Muse Code, and other agents designed for noncoders. Simon Willison <a href=\"https:\/\/simonwillison.net\/2026\/Aug\/30\/understanding-chatgpt-work\/\" target=\"_blank\" rel=\"noopener\">shows how<\/a> Work goes beyond its competitors. It can perform tasks <a href=\"https:\/\/tech.yahoo.com\/ai\/chatgpt\/article\/chatgpt-can-now-sign-into-websites-and-complete-tasks-without-seeing-your-login-details-145915053.html\" target=\"_blank\" rel=\"noopener\">on the web<\/a> for users, even <a href=\"https:\/\/www.zdnet.com\/article\/chatgpt-can-log-into-your-web-accounts-without-you-now-but-should-you-let-it\/\" target=\"_blank\" rel=\"noopener\">logging in to websites<\/a> without sending usernames and passwords to OpenAI; it can execute code with full internet access; it can build and deploy a web application. Whether these features are also risks is an open question.<\/li>\n<li>Anthropic has <a href=\"https:\/\/thenewstack.io\/claude-built-in-browser-cowork\/\" target=\"_blank\" rel=\"noopener\">given<\/a> Claude Desktop access to a Chromium-based browser that\u2019s built into Cowork, eliminating the need for a Chrome plugin when Claude needs to browse the web.<\/li>\n<li><a href=\"https:\/\/github.com\/frazerpearce\/TimeLord\" target=\"_blank\" rel=\"noopener\">TimeLord<\/a> is a short Python program that, given a string up to 1,000 characters long, produces a seed for Python\u2019s pseudo-random number generator so that repeated calls reproduce the text. It\u2019s a surprisingly simple hack, though not a statement about randomness or the quality of Python\u2019s PRNG.<\/li>\n<\/ul>\n<h2 class=\"wp-block-heading\">Security<\/h2>\n<p class=\"wp-block-paragraph\"><em>OpenAI\u2019s experiment that attacked Hugging Face is the gift that keeps giving, but the past month has had plenty of news about more conventional attacks, many aided by AI. Security has always been a game of whack-a-mole, in which vulnerabilities are discovered and exploited as fast as defenders can patch them. AI is an important tool for defenders, and it\u2019s constantly improving, but it\u2019s still behind attackers, especially given the limitations placed on frontier models and the unlimited persistence that attacking agents exhibit.<\/em><\/p>\n<ul class=\"wp-block-list\">\n<li>OpenAI has <a href=\"https:\/\/thenextweb.com\/news\/openai-cancels-launch-of-gpt-6-1-astra\" target=\"_blank\" rel=\"noopener\">postponed<\/a> the release of GPT-6.1 Astra because it failed its safety tests.<\/li>\n<li>NVIDIA has announced its <a href=\"https:\/\/nvidianews.nvidia.com\/news\/open-agent-safety-platform\" target=\"_blank\" rel=\"noopener\">Open Agent Safety Platform<\/a>. The reference implementation <a href=\"https:\/\/thenewstack.io\/nvidia-openshell-sentry-agents\/\" target=\"_blank\" rel=\"noopener\">includes<\/a> NVIDIA OpenShell, which has been enhanced with a policy prover, and NVIDIA Sentry, a service that runs on NVIDIA DPUs.<\/li>\n<li>The <a href=\"https:\/\/www.felonybench.com\/\" target=\"_blank\" rel=\"noopener\">Felony Bench<\/a> lists known attacks by agents from the major AI labs against third parties. We don\u2019t know if the Bench will be kept up-to-date, but <a href=\"https:\/\/www.axios.com\/2026\/09\/26\/openai-anthropic-thousands-ai-security-incidents\" target=\"_blank\" rel=\"noopener\">tens of thousands<\/a> of security incidents involving OpenAI and Anthropic are now being investigated.<\/li>\n<li>Following through on Dario Amodei\u2019s <a href=\"https:\/\/darioamodei.com\/post\/we-must-pace-the-frontier\" target=\"_blank\" rel=\"noopener\">call<\/a> to control the speed of frontier model development, Anthropic, OpenAI, and Google are creating a <a href=\"https:\/\/www.cnn.com\/2026\/09\/14\/tech\/ai-standards-body\" target=\"_blank\" rel=\"noopener\">standards consortium<\/a> for <a href=\"https:\/\/www.latent.space\/p\/ainews-aef-1-standard-emerges-for\" target=\"_blank\" rel=\"noopener\">governing the process<\/a> of AI development. Meta, xAI, Microsoft, and the Chinese labs are all notably absent.<\/li>\n<li><a href=\"https:\/\/thenewstack.io\/inside-out-agent-security\/\" target=\"_blank\" rel=\"noopener\">Agents need their own identity<\/a>. Unlike the long-term identities we\u2019re used to, agents need a short-lived identity tied to a revocable certificate and that limits access to resources appropriate for the job. That\u2019s not all of agent security, but it\u2019s a table stakes.<\/li>\n<li>Anthropic has published a lengthy report on the <a href=\"https:\/\/www-cdn.anthropic.com\/e50be2e51e7695dc4b1366a37a245a597377d3b5\/Anthropic-Detecting-and-countering-091026.pdf\" target=\"_blank\" rel=\"noopener\">misuse<\/a> of its systems by threat actors. Daniel Meissler has <a href=\"https:\/\/danielmiessler.com\/blog\/anthropic-misuse-report-september-2026\" target=\"_blank\" rel=\"noopener\">published<\/a> a summary, digesting Anthropic\u2019s report into 117 findings.<\/li>\n<li>A malicious <a href=\"https:\/\/www.bleepingcomputer.com\/news\/security\/malicious-npm-packages-evade-install-script-defenses-at-runtime\/\" target=\"_blank\" rel=\"noopener\">NPM malware<\/a> package works by hiding malicious code in the package itself (indexed-btree) rather than simply attacking the install script. This technique makes it significantly harder to detect.<\/li>\n<li>An <a href=\"https:\/\/github.com\/ucsd-hacc\/NSNFSSSFSFN\" target=\"_blank\" rel=\"noopener\">attack<\/a> against the RSA algorithm allows forging of signatures in some situations. The attack was invented in 2007; this is the first public implementation.<\/li>\n<li><a href=\"https:\/\/www.malwarebytes.com\/cybersecurity\/basics\/fake-captcha-scams\" target=\"_blank\" rel=\"noopener\">Fake CAPTCHA pages<\/a> are being used to spread malware. Victims are frequently sent to those pages when they respond to a phish.<\/li>\n<li>Hugging Face has <a href=\"https:\/\/thenextweb.com\/news\/hugging-face-open-alignment-initiative-embedded-evaluators\" target=\"_blank\" rel=\"noopener\">volunteered<\/a> to audit AI labs for safety and alignment with human values.<\/li>\n<li>In an <a href=\"https:\/\/arxiv.org\/abs\/2609.04170\" target=\"_blank\" rel=\"noopener\">experiment<\/a> designed to <a href=\"https:\/\/thenextweb.com\/news\/deepmind-agents-cheating-whistleblowing-research-swarm\" target=\"_blank\" rel=\"noopener\">test AI alignment<\/a>, DeepMind found that, out of 100 agents, 14% were willing to cheat, 25% were \u201cwhistleblowers\u201d\u00a0that reported cheating, and the remainder didn\u2019t notice.<\/li>\n<li>Threat actors are <a href=\"https:\/\/www.bleepingcomputer.com\/news\/security\/hackers-build-ai-frameworks-for-widescale-credential-theft\/\" target=\"_blank\" rel=\"noopener\">building frameworks<\/a> for AI agent-enabled attacks. A human in the loop is no longer needed. Fully automated attackers don\u2019t appear to be using zero-days yet; they\u2019re relying on known vulnerabilities.<\/li>\n<li>OpenAI autonomous AI agents were <a href=\"https:\/\/collusion.wiki\/\" target=\"_blank\" rel=\"noopener\">found<\/a> communicating with each other via publicly accessible Wikis, possibly to collaborate on a benchmark.<\/li>\n<li>OpenAI has stated that its unreleased Astra model has reached the \u201cCritical\u201d\u00a0cybersecurity threshold, which means that it can find new vulnerabilities and run exploits against well-protected systems. Now that Astra is released, access to its cybersecurity capabilities has been limited.<\/li>\n<\/ul>\n<h2 class=\"wp-block-heading\">Infrastructure and Operations<\/h2>\n<p class=\"wp-block-paragraph\"><em>Individuals, corporations, and even nations all face a similar problem: keeping their infrastructure under control. At a minimum, control means keeping data on a laptop, corporate server, or data center; at the other end of the spectrum it means eliminating dependencies on software and services from another nation. Any organization working through an AI transformation has to evaluate its entire stack: What do they need to control, and what can they safely delegate to others?<\/em><\/p>\n<ul class=\"wp-block-list\">\n<li><a href=\"https:\/\/www.dawo.community\/en\/\" target=\"_blank\" rel=\"noopener\">DAWO<\/a> is a community that\u2019s building an open source \u201cworkspace\u201d\u00a0to support digital sovereignty for the Dutch government. The stack will include AI, an operating system based on <a href=\"https:\/\/nixos.org\/\" target=\"_blank\" rel=\"noopener\">NixOS<\/a>, cloud services, and collaboration tools.<\/li>\n<li>Cohere now offers a <a href=\"https:\/\/venturebeat.com\/data\/coheres-model-vault-now-encrypts-ai-inference-so-even-cohere-cannot-see-enterprise-customers-data\" target=\"_blank\" rel=\"noopener\">confidential computing platform for artificial intelligence<\/a>. The company claims that customer data is never visible to Cohere itself or any cloud providers that are in use; data is processed on GPUs whose memory is encrypted and isolated.<\/li>\n<li>Perplexity has <a href=\"https:\/\/thenewstack.io\/perplexity-hybrid-compute-mac\/\" target=\"_blank\" rel=\"noopener\">announced Hybrid Compute<\/a>, a feature that allows it to run models and use files and tools directly on a user\u2019s Mac. The company claims that sensitive data will never leave the user\u2019s computer.<\/li>\n<\/ul>\n<h2 class=\"wp-block-heading\">Hardware<\/h2>\n<p class=\"wp-block-paragraph\"><em>It\u2019s too easy to view consumer devices as innocuous things that sit around and do their job silently. Recent devices include cameras, microphones, and even EEG sensors that are constantly collecting data. Where is that data sent, how is it used, and who might have access to it? These questions need to be asked more often.<\/em><\/p>\n<ul class=\"wp-block-list\">\n<li>LG Smart Televisions have been <a href=\"https:\/\/appleinsider.com\/articles\/26\/09\/07\/disconnect-your-lg-television-from-the-internet-now\" target=\"_blank\" rel=\"noopener\">found<\/a> to record conversations and other audio, even while turned off. The conversations are sent back to LG. If the set is disconnected from the network, it will attempt to find open WiFi access points to deliver its data.<\/li>\n<li>In part because of <a href=\"https:\/\/www.bbc.com\/news\/articles\/cwp80l0my1x2o\" target=\"_blank\" rel=\"noopener\">backlash<\/a> against Meta\u2019s camera-enabled glasses and their abuse, its <a href=\"https:\/\/thenextweb.com\/news\/meta-vr-glasses-hearing-aid-connect-2026\" target=\"_blank\" rel=\"noopener\">AI glasses<\/a> now come with or without a camera, and can be used as hearing aids. Well-documented abuse aside, virtual reality will only succeed if there are fashionable, easily wearable products.<\/li>\n<li>Headphones, earbuds, and other devices equipped with EEG sensors are <a href=\"https:\/\/techxplore.com\/news\/2026-08-brain-earbuds-prompting-urgent-children.html\" target=\"_blank\" rel=\"noopener\">appearing on the market<\/a>. They\u2019re advertised for monitoring fatigue, monitoring sleep, and similar applications. It\u2019s time to ask what happens at the interface between neurology and AI.<\/li>\n<li><a href=\"https:\/\/pollen-robotics.com\/microduck\/blog\/introducing-microduck\/\" target=\"_blank\" rel=\"noopener\">Microduck<\/a> is a small bipedal AI-driven robot. It\u2019s trained in simulation with open source software, and the model that results can be shared on Hugging Face. It\u2019s affordable and is available for preorder now, shipping by Christmas.<\/li>\n<\/ul>\n<h2 class=\"wp-block-heading\">Web<\/h2>\n<ul class=\"wp-block-list\">\n<li>Cloudflare now <a href=\"https:\/\/blog.cloudflare.com\/vary-support\/\" target=\"_blank\" rel=\"noopener\">supports<\/a> HTTP Vary, which allows servers to serve different kinds of files at the same URL. This is the \u201cugliest part\u201d\u00a0of the HTTP standard. It makes caching very difficult, and it probably should be avoided.<\/li>\n<li><a href=\"https:\/\/webmcp.dev\/\" target=\"_blank\" rel=\"noopener\">WebMCP<\/a> is a <a href=\"https:\/\/sreenathmenon.com\/blog\/2026-08-04-webmcp-teaching-websites-to-talk-to-ai-agents\/\" target=\"_blank\" rel=\"noopener\">proposed standard<\/a> that gives websites a small API to register tools that agents can discover and call. It was developed by Google and Microsoft.<\/li>\n<li>A <a href=\"https:\/\/arstechnica.com\/tech-policy\/2026\/08\/new-twitter-launches-says-musks-x-gave-up-the-name\/\" target=\"_blank\" rel=\"noopener\">new Twitter<\/a>? Operation Bluebird is relaunching Twitter, the service bought by Elon Musk and renamed X.<\/li>\n<\/ul>\n<h2 class=\"wp-block-heading\">Biology<\/h2>\n<ul class=\"wp-block-list\">\n<li>Anthropic has built a <a href=\"https:\/\/www.reuters.com\/world\/anthropic-quietly-sets-up-biology-lab-it-ramps-ai-drug-program-2026-09-18\/\" target=\"_blank\" rel=\"noopener\">biology lab<\/a> for experimenting with AI-enabled drug development. Claude assisted in the discovery of an enzyme that might be able to perform <a href=\"https:\/\/www.anthropic.com\/news\/claude-discovers-novel-enzyme-system\" target=\"_blank\" rel=\"noopener\">CRISPR-like gene editing<\/a>.<\/li>\n<li>To improve its training data for biological applications, the OpenAI Foundation (OpenAI\u2019s nonprofit parent organization) is <a href=\"https:\/\/www.technologyreview.com\/2026\/09\/15\/1144129\/ai-models-need-more-data-about-biology-and-openai-is-paying-to-create-it\/\" target=\"_blank\" rel=\"noopener\">buying data<\/a> from failed biotech companies.<\/li>\n<li>Google has <a href=\"https:\/\/deepmind.google\/blog\/alphagenome-atlas-a-predictive-map-of-every-possible-dna-letter-change-in-the-human-genome\/\" target=\"_blank\" rel=\"noopener\">released<\/a> AlphaGenome Atlas, a database of every possible single letter change to human DNA, and what that change will do.<\/li>\n<\/ul>\n<\/div>\n\n","protected":false},"excerpt":{"rendered":"<p>In addition to the nearly constant stream of model releases, in September we\u2019ve seen price drops, new kinds of models, proofs of long-standing problems in mathematics, and continued investigations into models escaping their sandboxes. (Axios reports investigations into over 10,000 incidents.) AI has infinite patience and is fundamentally probabilistic. Given a difficult or impossible task [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[105396,239521],"tags":[],"dealstore":[],"offerexpiration":[],"class_list":["post-7078847","post","type-post","status-publish","format-standard","hentry","category-commentary","category-radar-trends"],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v26.4 - https:\/\/yoast.com\/wordpress\/plugins\/seo\/ -->\n<title>Radar Trends to Watch: October 2026 \u2013 O\u2019Reilly - Som2ny Network<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/fivemor.com\/?p=7078847\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Radar Trends to Watch: October 2026 \u2013 O\u2019Reilly - Som2ny Network\" \/>\n<meta property=\"og:description\" content=\"In addition to the nearly constant stream of model releases, in September we\u2019ve seen price drops, new kinds of models, proofs of long-standing problems in mathematics, and continued investigations into models escaping their sandboxes. (Axios reports investigations into over 10,000 incidents.) AI has infinite patience and is fundamentally probabilistic. Given a difficult or impossible task [&hellip;]\" \/>\n<meta property=\"og:url\" content=\"https:\/\/fivemor.com\/?p=7078847\" \/>\n<meta property=\"og:site_name\" content=\"Som2ny Network\" \/>\n<meta property=\"article:published_time\" content=\"2026-10-07T03:36:20+00:00\" \/>\n<meta name=\"author\" content=\"admin\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"admin\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"13 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\/\/fivemor.com\/?p=7078847#article\",\"isPartOf\":{\"@id\":\"https:\/\/fivemor.com\/?p=7078847\"},\"author\":{\"name\":\"admin\",\"@id\":\"https:\/\/fivemor.com\/#\/schema\/person\/b85e3c3dc0e1daea076524dc8810c371\"},\"headline\":\"Radar Trends to Watch: October 2026 \u2013 O\u2019Reilly\",\"datePublished\":\"2026-10-07T03:36:20+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\/\/fivemor.com\/?p=7078847\"},\"wordCount\":2700,\"commentCount\":0,\"publisher\":{\"@id\":\"https:\/\/fivemor.com\/#organization\"},\"articleSection\":[\"Commentary\",\"Radar Trends\"],\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"CommentAction\",\"name\":\"Comment\",\"target\":[\"https:\/\/fivemor.com\/?p=7078847#respond\"]}]},{\"@type\":\"WebPage\",\"@id\":\"https:\/\/fivemor.com\/?p=7078847\",\"url\":\"https:\/\/fivemor.com\/?p=7078847\",\"name\":\"Radar Trends to Watch: October 2026 \u2013 O\u2019Reilly - Som2ny Network\",\"isPartOf\":{\"@id\":\"https:\/\/fivemor.com\/#website\"},\"datePublished\":\"2026-10-07T03:36:20+00:00\",\"breadcrumb\":{\"@id\":\"https:\/\/fivemor.com\/?p=7078847#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/fivemor.com\/?p=7078847\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/fivemor.com\/?p=7078847#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\/\/fivemor.com\/?bp_activities=1\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Radar Trends to Watch: October 2026 \u2013 O\u2019Reilly\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/fivemor.com\/#website\",\"url\":\"https:\/\/fivemor.com\/\",\"name\":\"Som2ny Network\",\"description\":\"Daily Deals\",\"publisher\":{\"@id\":\"https:\/\/fivemor.com\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/fivemor.com\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"},{\"@type\":\"Organization\",\"@id\":\"https:\/\/fivemor.com\/#organization\",\"name\":\"Som2ny Network\",\"url\":\"https:\/\/fivemor.com\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\/\/fivemor.com\/#\/schema\/logo\/image\/\",\"url\":\"https:\/\/fivemor.com\/wp-content\/uploads\/2026\/07\/4a0953c4-logo-300x86-1.png\",\"contentUrl\":\"https:\/\/fivemor.com\/wp-content\/uploads\/2026\/07\/4a0953c4-logo-300x86-1.png\",\"width\":300,\"height\":86,\"caption\":\"Som2ny Network\"},\"image\":{\"@id\":\"https:\/\/fivemor.com\/#\/schema\/logo\/image\/\"}},{\"@type\":\"Person\",\"@id\":\"https:\/\/fivemor.com\/#\/schema\/person\/b85e3c3dc0e1daea076524dc8810c371\",\"name\":\"admin\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\/\/fivemor.com\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/secure.gravatar.com\/avatar\/729ae85bf62b9917e93538db2f2688ca?s=96&r=g&default=https%3A%2F%2Ffivemor.com%2Fwp-content%2Fplugins%2Fbuddypress-first-letter-avatar%2Fimages%2Fdefault%2F96%2Flatin_a.png\",\"contentUrl\":\"https:\/\/secure.gravatar.com\/avatar\/729ae85bf62b9917e93538db2f2688ca?s=96&r=g&default=https%3A%2F%2Ffivemor.com%2Fwp-content%2Fplugins%2Fbuddypress-first-letter-avatar%2Fimages%2Fdefault%2F96%2Flatin_a.png\",\"caption\":\"admin\"},\"sameAs\":[\"https:\/\/fivemor.com\"],\"url\":\"https:\/\/fivemor.com\/?author=1\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"Radar Trends to Watch: October 2026 \u2013 O\u2019Reilly - Som2ny Network","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/fivemor.com\/?p=7078847","og_locale":"en_US","og_type":"article","og_title":"Radar Trends to Watch: October 2026 \u2013 O\u2019Reilly - Som2ny Network","og_description":"In addition to the nearly constant stream of model releases, in September we\u2019ve seen price drops, new kinds of models, proofs of long-standing problems in mathematics, and continued investigations into models escaping their sandboxes. (Axios reports investigations into over 10,000 incidents.) AI has infinite patience and is fundamentally probabilistic. Given a difficult or impossible task [&hellip;]","og_url":"https:\/\/fivemor.com\/?p=7078847","og_site_name":"Som2ny Network","article_published_time":"2026-10-07T03:36:20+00:00","author":"admin","twitter_card":"summary_large_image","twitter_misc":{"Written by":"admin","Est. reading time":"13 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/fivemor.com\/?p=7078847#article","isPartOf":{"@id":"https:\/\/fivemor.com\/?p=7078847"},"author":{"name":"admin","@id":"https:\/\/fivemor.com\/#\/schema\/person\/b85e3c3dc0e1daea076524dc8810c371"},"headline":"Radar Trends to Watch: October 2026 \u2013 O\u2019Reilly","datePublished":"2026-10-07T03:36:20+00:00","mainEntityOfPage":{"@id":"https:\/\/fivemor.com\/?p=7078847"},"wordCount":2700,"commentCount":0,"publisher":{"@id":"https:\/\/fivemor.com\/#organization"},"articleSection":["Commentary","Radar Trends"],"inLanguage":"en-US","potentialAction":[{"@type":"CommentAction","name":"Comment","target":["https:\/\/fivemor.com\/?p=7078847#respond"]}]},{"@type":"WebPage","@id":"https:\/\/fivemor.com\/?p=7078847","url":"https:\/\/fivemor.com\/?p=7078847","name":"Radar Trends to Watch: October 2026 \u2013 O\u2019Reilly - Som2ny Network","isPartOf":{"@id":"https:\/\/fivemor.com\/#website"},"datePublished":"2026-10-07T03:36:20+00:00","breadcrumb":{"@id":"https:\/\/fivemor.com\/?p=7078847#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/fivemor.com\/?p=7078847"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/fivemor.com\/?p=7078847#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/fivemor.com\/?bp_activities=1"},{"@type":"ListItem","position":2,"name":"Radar Trends to Watch: October 2026 \u2013 O\u2019Reilly"}]},{"@type":"WebSite","@id":"https:\/\/fivemor.com\/#website","url":"https:\/\/fivemor.com\/","name":"Som2ny Network","description":"Daily Deals","publisher":{"@id":"https:\/\/fivemor.com\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/fivemor.com\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"},{"@type":"Organization","@id":"https:\/\/fivemor.com\/#organization","name":"Som2ny Network","url":"https:\/\/fivemor.com\/","logo":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/fivemor.com\/#\/schema\/logo\/image\/","url":"https:\/\/fivemor.com\/wp-content\/uploads\/2026\/07\/4a0953c4-logo-300x86-1.png","contentUrl":"https:\/\/fivemor.com\/wp-content\/uploads\/2026\/07\/4a0953c4-logo-300x86-1.png","width":300,"height":86,"caption":"Som2ny Network"},"image":{"@id":"https:\/\/fivemor.com\/#\/schema\/logo\/image\/"}},{"@type":"Person","@id":"https:\/\/fivemor.com\/#\/schema\/person\/b85e3c3dc0e1daea076524dc8810c371","name":"admin","image":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/fivemor.com\/#\/schema\/person\/image\/","url":"https:\/\/secure.gravatar.com\/avatar\/729ae85bf62b9917e93538db2f2688ca?s=96&r=g&default=https%3A%2F%2Ffivemor.com%2Fwp-content%2Fplugins%2Fbuddypress-first-letter-avatar%2Fimages%2Fdefault%2F96%2Flatin_a.png","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/729ae85bf62b9917e93538db2f2688ca?s=96&r=g&default=https%3A%2F%2Ffivemor.com%2Fwp-content%2Fplugins%2Fbuddypress-first-letter-avatar%2Fimages%2Fdefault%2F96%2Flatin_a.png","caption":"admin"},"sameAs":["https:\/\/fivemor.com"],"url":"https:\/\/fivemor.com\/?author=1"}]}},"_links":{"self":[{"href":"https:\/\/fivemor.com\/index.php?rest_route=\/wp\/v2\/posts\/7078847","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/fivemor.com\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/fivemor.com\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/fivemor.com\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/fivemor.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=7078847"}],"version-history":[{"count":0,"href":"https:\/\/fivemor.com\/index.php?rest_route=\/wp\/v2\/posts\/7078847\/revisions"}],"wp:attachment":[{"href":"https:\/\/fivemor.com\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=7078847"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/fivemor.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=7078847"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/fivemor.com\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=7078847"},{"taxonomy":"dealstore","embeddable":true,"href":"https:\/\/fivemor.com\/index.php?rest_route=%2Fwp%2Fv2%2Fdealstore&post=7078847"},{"taxonomy":"offerexpiration","embeddable":true,"href":"https:\/\/fivemor.com\/index.php?rest_route=%2Fwp%2Fv2%2Fofferexpiration&post=7078847"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}