Lilith · selected stories
What is actually happening in AI. Selected stories, context and opinion without the promotional noise.
Atom feed ↗Zvi Mowshowitz disputes a Wall Street Journal framing that China has matched Anthropic in cybersecurity, arguing that it blurs narrow bug finding tasks with autonomous vulnerability discovery and exploit chaining. For security teams, the lesson is colder: a model can win one test and still not be the same weapon in production.
Z.ai is positioning GLM-5.2 as an open-weight model for long-running coding agents with a 1M-token context window. The useful question for teams is when an open model is good enough to replace Opus, not whether it wins every chart.
OpenAI introduced GPT-5.6 as a three-model family, Sol, Terra and Luna, but is starting with a limited preview coordinated with the U.S. government. The important part for teams is that the system card pairs higher cyber and bio capability with a heavier safety stack.
The US government told Anthropic to restrict Fable 5 and Mythos 5 for all foreign nationals, so Anthropic switched the models off for all customers. A protest by 76 security experts exposes the weak point: export control is bad at separating an offensive exploit from defensive testing.
Zvi Mowshowitz uses Fable and Mythos as a case study for why model welfare cannot be separated from capabilities, alignment and user experience. Even where the topic remains speculative, it is becoming a practical question of evaluations and safety interventions for frontier labs.
Anthropic released Claude Opus 4.8 at the same standard price as Opus 4.7, with a focus on coding, agentic tasks and longer work. The more important shift is a model that is supposed to say more often when it is unsure.
Nathan Lambert announced his departure from the Allen Institute for AI and used it to reflect on work around Olmo. This is not just a personnel note. It is a reminder that open models depend on institutions that must outlast one strong team.
Latent Space frames Microsoft Build as the moment Microsoft showed its own MAI models alongside Copilot, Windows and Web IQ. The key ambition is to control data, inference and developer workflow at once, rather than leaving that leverage to partners.
The White House issued an executive order that calls for a classified benchmark for covered frontier models within 60 days and a voluntary framework for up to 30 days of pre-release government access. It says this is not licensing, but it creates a pressure point before launch.
Uber is limiting monthly token spend to $1,500 per employee for each agentic coding tool, according to Bloomberg via Simon Willison. Coding agents are becoming a budget line item.
Latent Space's interview with Andon Labs shows evals that look less like exams and more like running a small business. The key ingredients are long horizons and real consequences.
Simon Willison released the alpha package micropython-wasm and a Datasette Agent plugin that runs Python inside a WebAssembly sandbox. The important part is not the demo, but the boundary between a useful agent and code that can break its host application.
Latent Space published Auriel W's piece on why low-quality RL environments damage agent training. The point is simple: in reinforcement learning, the environment is the data generator, so a harness bug becomes training material.
Apple announced Siri AI and new Apple Intelligence features at WWDC 2026, while extending Private Cloud Compute to Google Cloud with NVIDIA GPUs for demanding tasks. After last year's Apple Intelligence disappointment, this is less about the keynote and more about whether Siri can finally survive outside the demo.
Simon Willison shows how he manually added pricing for Claude Fable 5 in AgentsView and immediately saw the cost of local coding agents by project. The small trick points to a bigger shift: AI coding is starting to look like infrastructure consumption, not an app subscription.
Zvi Mowshowitz's AI #171 is not one clean trend, but a signal map: Claude Opus 4.8, US frontier model testing, OpenAI's policy blueprint and PAC politics.
Nathan Lambert reads the Claude Fable 5 release as a dispute over who gets to use a frontier model without routing and filters. The important layer is not only model capability, but the governance system that decides when the user is really talking to the strongest model.
OpenAI has started rolling out Lockdown Mode for eligible personal ChatGPT accounts and self-serve ChatGPT Business. It does not stop prompt injection itself, but it limits outbound network requests, which are the channel an attacker needs to exfiltrate sensitive data.
The Latent Space interview with Kyle Daigle frames GitHub as a platform under pressure from agentic coding. The point is not another Copilot feature, but whether infrastructure built for human pace can absorb software produced by machines.
Simon Willison released datasette-agent-edit 0.1a0, a base plugin for Datasette Agent with view, str_replace and insert tools. It is not a flashy AI demo. It is the layer that decides whether an agent can edit text without casually breaking the file.
