Blog · RSS
AI news and guides
New AI models, coding tools and agent releases, explained for developers and founders, plus practical guides from my builds.
OVHcloud Launches Olala Mini 7A1B: An Open MoE Model That Runs on 8 GB GPU
OVHcloud announced its first open-source LLM, Olala mini 7A1B, a mixture-of-experts model with 7B parameters that runs on 8 GB of GPU memory. No benchmarks yet, but early deployment shows speed and cost gains.
ReadAWS Launches Kiro Bots: AI Agents That Take On Specific Development Roles
AWS has officially launched Kiro bots, long-running AI agents that handle tasks like code review, bug triage, and release preparation, letting developers focus on higher-level decisions.
ReadOpenCode Launches Browse MCP Server to Give Coding Agents Real Browser Access
OpenCode's Browse MCP server and Chromium extension let coding agents control a developer's real browser, including authenticated sessions, without sandboxing.
ReadAnthropic Launches Claude Sheets Beta: AI Spreadsheets Inside Claude
Anthropic released Claude Sheets in beta on November 10, 2026, letting users generate and collaborate on spreadsheets with natural language prompts. Available on paid plans, it targets lightweight work, not Excel replacement.
ReadI Tested Qwen 4 Max: Alibaba’s 2.4T AI Model vs Claude (10x Cheaper?)
I put Alibaba's new Qwen 4 Max through real coding tests—building games, 3D environments, and tools—and compared it directly with Claude. Here's how the 2.4T parameter model performed and what it cost.
ReadGoogle Launches Gemini Orchestrator to Coordinate Multi-Model Coding Workflows
Google's new Orchestrator in Antigravity dynamically switches between Gemini 4 Carbon and the lightweight Gemini 4 Hydrogen during coding sessions, cutting token usage by 24% and speeding up tasks by 19%.
ReadCmpnd Releases DSPy-WSP: Weight-Space Programming Optimizer for DSPy
Cmpnd has open-sourced DSPy-WSP, an experimental optimizer that lets a frontier LLM edit smaller model weights without gradients, improving decision model accuracy by up to 16 points.
ReadOpenAI Launches ChatGPT Wallet for AI Agent Payments
OpenAI announced ChatGPT Wallet on 13 November 2026, letting AI agents make purchases with user approval via biometric checks and payment tokens, initially for US Pro subscribers.
ReadBland Releases Tofu-1, a Full-Duplex Speech-to-Speech Model Built for Noisy Calls
San Francisco voice-AI company Bland has launched Tofu-1 in research preview, a full-duplex model that listens and speaks simultaneously on phone calls, with direct audio understanding and scene generation.
ReadClaude Fable 5.2: 3D Game From a Screenshot and More – My Real-World Tests
I put Claude Fable 5.2 through six real 3D and design projects, from a racing game built from a single screenshot to Blender character rigging. Here’s what worked, what didn’t, and the full spec breakdown.
ReadCloudflare open-sources Kitesurf browser engine for AI agents
Cloudflare has released Kitesurf under Apache 2.0, letting developers inspect, run locally, and deploy their own instance of the experimental AI-agent browser engine on Cloudflare Workers.
ReadAnthropic Launches Claude Opus 4.8 With Sharper Coding and a Focus on Honesty
Claude Opus 4.8 brings better agentic coding, a 4x reduction in silent code errors, and new developer controls, while running faster and cheaper than its predecessor.
ReadAnthropic releases Claude Fable 5.2 and Mythos 5.2 with lower pricing and 30% faster output
Anthropic has launched Claude Fable 5.2 and Claude Mythos 5.2 on 9 November 2026, cutting API prices and delivering up to 30% faster token generation while targeting ambitious agentic and scientific workloads.
ReadAnthropic Launches Self-Hosted Claude Managed Agents in Private Beta
Anthropic opened a private beta for self-hosted Claude Managed Agents on November 12, 2026, letting enterprises run the full agent service, sandboxes, and session history in their own infrastructure.
Readllama.cpp 1.0 Brings Stable API and Memory Optimisations
llama.cpp and ggml reach version 1.0.0 on 13 November 2026, promising semantic versioning and breaking changes only with major-version bumps. The release removes deprecated functions, adds FP8 KV caches, and hardens the server.
ReadOpenBot 1.0: Self-Hostable AI Coworker with Built-In Governance Released
CopilotKit released OpenBot 1.0 on November 9th, an MIT-licensed, self-hostable AI coworker that lets developers bring their own models and agents while enforcing policy checks on every action.
ReadGoogle Unveils Gemini 4 Hydrogen: Speed, Lower Cost, and Agentic Workflows for Developers
Google has released Gemini 4 Hydrogen, a faster, lower-cost model that replaces 3.8 Flash. It offers 2.5x speed, reduced hallucinations, and new thinking levels for everyday coding and agentic tasks.
ReadNeo4j Launches Relate: An Open-Source Embedded Graph Database for AI Agents
Neo4j has released Relate, an open-source embedded database built in Rust for AI agents. It runs in-process with optional disk persistence and includes graph-native features for agent memory, sandboxes, and edge deployments.
ReadAWS Open-Sources UAEF for Cross-Framework AI Agent Evaluation
AWS released UAEF under Apache 2.0, a framework-agnostic evaluation system with 36 metrics, hardened judge, and regression detection for AI agents.
ReadPAN 2: MBZUAI’s IFM Releases Open 29B World Model for Physical AI
IFM at MBZUAI released PAN 2, an open 29B-parameter world model that simulates video outcomes from text actions. Weights and code are under Apache 2.0.
ReadAlibaba Opens Qwen4-Max Preview, Targeting Complex Coding and Agentic Workflows
The Qwen4-Max preview is now available across Alibaba’s products, with a focus on complex programming, front-end design, and autonomous tool use. Pricing includes promotional discounts and a ¥690/month Qwen Studio plan.
Read0sec Labs Open-Sources 0 AI Security Harness With Web UI, CLI and MCP
Zurich-based 0sec Labs has released 0, an open-source AI security harness that runs vulnerability scans, pentesting and reverse engineering via a browser console, CLI or MCP server.
ReadGoogle Open-Sources Elephant-Goldfish Model Plugin for Spec-First AI Coding
Google released the Elephant-Goldfish Model plugin under Apache 2.0, a workflow that forces AI coding agents to draft a specification and have it reviewed by fresh, zero-context subagents before any code is written.
ReadTesting OwnVenture: Can AI Take You From Idea to First Paying Customer?
I put OwnVenture to the test with a bedtime story app idea. Here's how the AI handled market research, product previews, and marketing strategy on autopilot.
ReadGoogle Donates Agent Substrate, gVisor, and OpenRL to CNCF
Google has donated Agent Substrate, gVisor, and OpenRL to the CNCF Sandbox. Agent Substrate 1.0 is now GA on GKE, providing an open stack for running AI agents at scale.
ReadClaude Code Cloud Now Automatically Sets Up Repo Dependencies and API Keys
Anthropic announced on 6 November 2026 that Claude Code cloud sessions on Pro, Max and Ultra plans can automatically detect dependencies, network access and API keys, saving the environment for future sessions.
ReadBuoyant Releases Kinference: Open Source Control Plane for vLLM on Kubernetes
Buoyant has released Kinference, an Apache 2.0 control plane that adds policy-driven routing, KV-cache-aware load balancing, and per-tenant metrics to vLLM inference on Kubernetes.
ReadOpenAI Launches GPT-6 Cyber: A Cybersecurity Model That Stays in Scope
OpenAI released GPT-6 Cyber on 5 November 2026, a model trained for authorised vulnerability research that scores 48.5% on ExploitGym and stays within scope 99.93% of the time without safety prompts.
ReadGoogle Cloud launches Agent Compute Engine for isolated AI agents
Google Cloud has introduced Agent Compute Engine, a microVM service for running AI agents with hardware isolation and idle optimisations, alongside open-source tools and governance controls.
ReadThinking Machines releases Inkling 2 Small, a western open-weight model that challenges Chinese AI
The ex-OpenAI CTO's startup has launched a 594B-parameter model with strong coding and agent capabilities. Weights will be released next week, making it the highest-scoring open model built outside China.
ReadBuild an iOS App with AI and Then Let It Market Itself – Superapp Selfdrive
I show how to build a native iOS app from plain English using Superapp, then hand the entire App Store growth plan, ads, SEO and UGC to its Selfdrive agent.
ReadGemini Omni 1.2 Flash Now GA: 15-Second Clips and 30% Lower Price
Google has made Gemini Omni 1.2 Flash generally available, bringing 15-second video generation, conversational editing, and a 30% price cut to $0.07 per second for 720p.
ReadBaidu open-sources ERNIE 5.5 Flash Base, a 280B-parameter MoE model for developers to fine-tune
On 5 November 2026 Baidu released ERNIE 5.5 Flash Base, a pre-training checkpoint with 280B total parameters and 14B active per token. Weights are on Hugging Face under a custom licence that restricts commercial use and geography.
ReadHow Framia's Node-Based Canvas Simplifies AI Video Creation
In this walkthrough, I build a sports drink commercial using Framia’s node canvas and AI agents, showing how you can create consistent, cinematic videos without switching between tools.
ReadCognition Launches SWE-2 Flash, a Faster and Cheaper Coding Model for Routine Agent Work
Cognition released SWE-2 Flash on 4 November, a 276B-parameter mixture-of-experts model with 12B active parameters that runs 3x faster and costs one-fifth as much as SWE-2 inside Devin. It is free through 30 November.
ReadClaude Duet: Anthropic's Waitlisted Background Agent Arrives in Research Preview
Anthropic has launched Claude Duet as a U.S.-only research preview for Pro, Max and Ultra subscribers, with background tasks, waitlist access, and new Shop Pay checkout support.
ReadPentagon Threatens Anthropic with Defence Production Act Over Military AI Use
The US Defence Department has given Anthropic a Friday deadline to agree to its terms for military use of its AI, or face forced compliance under the Defence Production Act, according to sources.
ReadGPUWorks Goes Open Source: A 40 MB Platform to Manage 1,000+ GPU Servers
ZStack open-sourced GPUWorks under Apache 2.0 on November 4, 2026. The 40 MB platform deploys in 60 seconds, manages heterogeneous GPU clusters, and supports time-shared scheduling with fast wake-up.
Readllm-d 1.0 Released to Cut AI Inference Costs with Multi-Tenant Controls
The CNCF-governed project llm-d reaches 1.0, adding multi-tenant routing, high availability, and session-aware caching to help developers run large language model inference more efficiently on shared infrastructure.
ReadTesting Synthesia Roleplay Sessions: Can AI Help You Handle Difficult Conversations?
I put Synthesia's new AI Roleplay Sessions to the test, simulating a frustrated customer to see if an avatar can teach real-world communication skills under pressure.
ReadWhy I’m Attending Alibaba.com CoCreate London 2026 (And You Should Too)
I’m heading to ExCeL London to see if Alibaba.com’s new AI sourcing engine can fix my supplier headaches. Here’s what founders, investors and sellers can expect from the event.
Readfal Agent Review: A Creative Workspace That Remembers Your Project
I test fal Agent’s project memory, side-by-side model comparisons, and Auto mode to see if it solves the consistency problem for AI filmmakers.
Read