Lilith · selected stories
What is actually happening in AI. Selected stories, context and opinion without the promotional noise.
Atom feed ↗Google introduced an agentic RAG system for Gemini Enterprise Agent Platform that checks whether it has enough context before answering. For companies, that brake matters more than another polished retrieval layer.
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.
Agent Zero released DOX, a minimalist framework for hierarchical AGENTS.md files. It does not add a new runtime. It gives coding agents a simple discipline: read local rules before editing and update them after meaningful changes.
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.
ServiceNow AI published an ASR benchmark for code-switched speech in enterprise scenarios and tested seven systems. The uncomfortable point is simple: in voice agents, transcription errors propagate through the whole workflow, so bilingual speech is not a minor UX detail.
Perplexity describes Search as Code: an architecture where an agent does not call one monolithic search engine, but assembles a retrieval pipeline as code. The point is not a nicer search API. It is control over how evidence is found, filtered and verified.
Hugging Face shows Reachy Mini calling MCP tools hosted in public Spaces. The interesting part is not a weather answer, but the split between the robot body and capabilities that can be shared and updated outside the app.
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.
Nathan Lambert writes about Gemini Flash 3.5, Mythos, agent tools and the tension between open and closed models in his May outlook.
Latent Space frames xAI Grok Imagine, through an interview with Ethan He, as a move from one shot video generation toward video agents. The thesis will be proven less by demo quality than by whether the system can iterate through a whole creative task.
A Latent Space discussion with Cognition and OpenInspect frames coding agents as asynchronous workers: spec-to-PR workflows, full VMs, agent memory, and situations where a PM ships a code change without a developer. The shift is from synchronous chat to delegating an entire work cycle.
Microsoft Research describes MagenticLite, MagenticBrain and Fara1.5 as an agentic system optimized for small models that connects browser and local file system in a single workflow. The direction is practical: not one expensive model for everything, but orchestration of specialized components.
AI Snake Oil examines the claim that Google AI agents built an operating system for $916. The key point: agentic announcements need a different type of verification than chat benchmarks, because a big goal and a few steps in a demo environment are easy to inflate.
IBM Research and Artificial Analysis released the first benchmark for enterprise IT agents in a realistic Kubernetes environment on 27 May 2026. The top model (Claude Opus 4.7) reached 47 %. No frontier model exceeded 50 %.
Microsoft Research describes SocialReasoning-Bench, a benchmark testing whether AI agents genuinely act in the user's best interest. Key finding: agents complete tasks technically, but do not consistently improve outcomes for the person, even when explicitly instructed to.
LWiAI Podcast episode 246 from 26 May 2026 is a map, not a single thesis. Google I/O, coding agents, legal pressure around OpenAI and safety research landed in the same week and sketch four simultaneous pressures on the AI market.
Microsoft Research released Data Formulator 0.7, an analytics workspace where AI agents assist with exploration, transformation and visualization of enterprise data. The key question is whether the agent handles messy, permissioned data outside the demo.
SQLite added an AGENTS.md file with a blunt rule for people pointing AI agents at the codebase: agentic code is not accepted, but high-quality reproducible bug reports can be useful. A small file, but a big signal for critical open source maintenance.
Gartner estimates that over 40% of agentic AI projects will be cancelled by the end of 2027 because of cost, unclear value or weak risk controls. The signal is not that agents are dead. It is that unmanaged PoCs are entering the bill, governance and accountability wall.
