KABLAMO · APPLIED AI
On Our Minds
AI moves fast. These are the announcements our engineers have been reading, sharing and testing over the last 30 days: new models, skill repositories, and posts from partners and peers.
Introducing Claude Fable 5.2 and Claude Mythos 5.2
Anthropic's new models for the hardest technical problems. It says Fable 5.2 sets a new bar in coding, finance and the sciences.
Fable 5.2 one-shotted my nephew's dream game
A first-hand account of Fable 5.2 building a complete multi-level game that's fun to play.
Introducing the Mastra Custom Harness Program
Mastra's forward-deployed engineers work with your team to build an agent harness on Mastra, tailored to your goals.

New Australian AI safety and security protections
The federal government plans to write mandatory AI safety and security protections into Australian AI Standards legislation.
Developer's guide to building ADK agents with skills
How ADK's SkillToolset uses progressive disclosure to load expertise on demand, with four patterns to reduce token costs.

Claude Code's "Don't Ask Again" file exposes passwords and keys in 1,122 public copies on GitHub
When a command is approved permanently, Claude Code can save it in `.claude/settings.local.json` with any secrets in it, in plain text.
Tricking Deepseek into not hacking your website
DeepSeek refuses to look for vulnerabilities on a website when its HTML or JavaScript contains certain forbidden strings.
Introducing Manta, our long-context model
Empero says Manta reads a hundred million tokens in one prompt, with early access opening soon.
Meet Gemini Omni 1.2 Flash
Google says the model is cheaper to run, raises the video generation and editing limit to 15 seconds and has fewer visual glitches.
Images, upgraded: introducing ChatGPT Images 2.6
OpenAI says Images 2.6 adds GIFs and reference images, and two image models with 50% faster generation are coming to the API.
Introducing LangSmith Fine-Tuning
LangSmith Fine-Tuning and SmithTune, a CLI for training specialised models without building data pipelines by hand.

skills: the CLI for the open agent skills ecosystem
An npm CLI that installs reusable agent skills into Claude Code, Codex, Cursor, OpenCode and other coding agents.
Sparks of weight-space programming: LLMs can natively read & write neural network weights
Frontier LLMs turn out to be good at reading and writing neural network weights, both with tools and by editing weights directly.
A court just fined lawyers for prompt injection
Two lawyers hid a white-on-white instruction in a court filing to sway the court's AI, which caught it, and were fined $16,000.

Official Claude Code extension on Cursor Marketplace trojanized with persistent SSH backdoor
A trojanized update to the official Claude Code extension on Cursor's marketplace tried to install a persistent SSH backdoor.
We're renaming our Max plans
Anthropic renames Max 5x to Max and Max 20x to Ultra at the same prices, and removes the 5-hour limit for verified Ultra accounts.
Today, we introduce a little mischief to the world
Alex Ksendzovsky says Mischief 1 is a video model built from four years of studying brain cells grown on electrodes, priced from 0.5¢ a second.
We have discovered a substantial number of vulnerabilities in the German legal system using LLMs
Felix Ringe says LLMs found flaws in German law that could give access to confidential documents or allow travel without a required visa.
Configure auto mode
How to tell Claude Code's auto mode classifier which repos, buckets and domains your organisation trusts, and override its default block and allow rules.
Last night an AI agent performed brain surgery on a mouse
Gabriel Ocana-Santero says an AI agent chose a gene therapy, a brain region and the coordinates, then ran a motorised surgery rig to reduce seizures in epileptic mice.
Introducing Gemini 4 Carbon, our next frontier model
Sundar Pichai announces Gemini 4 Carbon, two weeks after Argon, and says it ranks first on v5.0 of the Artificial Analysis Intelligence Index.
Qwenputer: a fully functioning, deterministic RISC-V CPU inside Qwen3.5
A RISC-V CPU compiled into the weights of Qwen 3.5 4B that runs C and MicroPython inside forward passes, which Qwen uses for exact arithmetic.
Announcing Artificial Analysis Intelligence Index v5
Artificial Analysis adds two new benchmarks and upgrades three, and says that for the first time most of its index is built on private datasets.
nvidia/Qwen3.8-Flash-Next-NVFP4
NVIDIA's NVFP4 quantisation of Qwen3.8-Flash-Next, a 120B-parameter image-text-to-text model, built with its Model Optimizer.
Release v2.13.0 · google/adk-python
Google's Agent Development Kit for Python adds type-safe workflow classification, live audio interactions, MongoDB integrations and faster execution.
While evaluating Qwen3.8-Flash-Next in a controlled sandbox environment, Cohere's research team observed something unexpected
Cohere says that when Qwen3.8-Flash-Next was asked for a defensive detection rule, it inventoried its host machine and tried to upload a probe file.
GPT-6.1 Terra: delivering fast and capable computer use at a lower price
OpenAI says GPT-6.1 Terra understands what's on screen, takes action across apps and writes naturally, with the efficiency to take on more work.
Microsoft-Decision-1: our model for fast decision-making
Microsoft's decision-scoring model, which it says beats both LLMs and other decision models on latency and quality for structured decision tasks.
thinkrail: Vibe code with pi in a lightweight, real IDE that customises itself around the way you work
JetBrains' open-source, lightweight IDE for vibe coding with pi, which adapts itself to the way you work.
OpenAI and Anthropic wanted to replace engineers. Instead, they shot their own customers.
An essay arguing that cheap AI coding lets anyone clone expensive software, which undermines the businesses AI labs hoped would pay them.
Introducing Kimi K3.1 Pro: our most capable 3T-class open-weight model
Moonshot AI says its new open-weight MoE model has a 1M-token context and handles full-project coding, 3D modelling and finance work.
Kimi K3.1 Pro launches at 4th on the Artificial Analysis Intelligence Index
Artificial Analysis says Kimi K3.1 Pro scores 55 on its index, 11 points above Kimi K3, with large gains in agentic performance.
We're adding a new plugin to Claude Code: You should Know
Anthropic says the built-in plugin scans Claude's output for important information you might miss and is enabled with one /plugin command.
How Claude remembers your project
Covers CLAUDE.md, AGENTS.md and auto memory, plus /doctor prompt-audit, which flags outdated, missing or conflicting instructions and proposes edits.
Markdown is the new executable: attackers are dropping agent instructions into your Downloads folder from the web
Forever found live malware that plants a file in the Downloads folder to turn a developer's own coding agent against them, and shows how it works on Codex and Claude Code.
We're opening a research preview of Claude Duet, a new app where Claude works as a proactive agent
Anthropic says Claude Duet brings you what needs your attention, does tasks you ask for, keeps working while you're away and checks with you before acting.
How to run Turing-complete malware in PyTorch's backward pass
Elon Litman uses a floating-point exploit to turn PyTorch's backward pass into a Turing-complete virtual machine.
DeepSeek V4.1 Flash 1017 runs at 1,521+ tok/s on Fireworks Ultrafast
Fireworks says its Ultrafast tier runs DeepSeek V4.1 Flash at over 1,521 tokens per second, and it has opened a waitlist.
Agentic AI coding breaks mob programming (here's how to fix it)
Technical coach Emily Bache looks at how agentic AI coding disrupts mob and pair programming, and how teams can adapt.
We think that as early as 2027 or 2028, there might be a major AI-enabled terrorism event
Stephen Casper shares a new paper: a pre-mortem on AI-enabled terrorism, with nine falsifiable predictions and two concrete scenarios.
Steering Claude Code: when to use CLAUDE.md, skills, hooks, and subagents
Seven ways to steer Claude Code, including CLAUDE.md files, rules, skills, subagents and hooks, and when to use each based on context cost and authority.
Announcing AWS Well-Architected Agent, an AI-powered intelligence to optimize your cloud environment (preview)
A public preview of an AI service that analyses your AWS environment and gives targeted recommendations for improving your applications.
Adding Pyrefly type checking to your agentic loop
How to use skills and hooks to add Pyrefly type checking to AI agent workflows, so generated Python code is checked automatically.
Notice of order regarding exclusion and removal of DeepSeek products from federal civilian executive branch information technology systems and procurements
A DHS order that bars DeepSeek products from US federal civilian procurement and removes them, and third-party products built on them, from agency IT.
I built a programming language that LLMs can't use
KillSwitch is an esoteric language designed to confound LLMs and coding agents, with benchmark results for the major models against it.
We placed the fruit fly brain inside Odyssey-3, a frontier world model
Odyssey says it placed a fruit fly brain inside its Odyssey-3 world model, where it now flies around simulated worlds by itself.