DISPATCH // AI NEWS
AI and Prompt Engineering News, Updated Daily
Daily coverage of model releases, agent tooling, LLM inference and routing, and prompt-engineering technique, written for people shipping with these models.
AgentsOct 18, 20263 min read
37 Bugs in 53 Agents: What Behavioral Testing Reveals About Open-Source AI Agent Quality
A new behavioral testing framework put 53 open-source AI agents through structured tests and filed 37 bugs, exposing systematic reliability gaps. If you are shipping agent-based features, this is the quality baseline you need to know about.
AgentsOct 17, 20264 min read
AI Agents Are Getting Weird: Emergent Behavior, Spending Controls, and Physics Benchmarks
Three signals from this week paint a clear picture: AI agents are becoming capable enough to surprise their creators, costly enough to need financial guardrails, and ambitious enough to tackle theoretical physics. Here is what builders need to know.
IndustrySep 30, 20262 min read
OpenAI DevDay 2026: What Builders Need to Know
OpenAI held DevDay 2026 in San Francisco today, with live coverage surfacing new announcements relevant to teams shipping LLM-powered products. Here is what matters for builders, alongside a cluster of agent-focused signals that landed the same day.
AgentsAug 29, 20262 min read
The Coding-Agent Toolchain Is Maturing Fast: Four New Primitives Drop in One Day
Four tools landed on the same day targeting the same gap: the messy infrastructure layer between an LLM and a real dev environment. Here is what each one solves and whether it belongs in your stack.
AgentsAug 25, 20263 min read
The AI Agent Tooling Layer Is Filling In Fast
Four independent tools dropped on the same day targeting agent evaluation, context retrieval, pre-deployment scoring, and destructive-action safety. The scaffolding around autonomous agents is maturing quickly, and builders need to know what each layer does.
AgentsJul 2, 20262 min read
ScarfBench Tests AI Agents on Real Enterprise Java Migrations
IBM Research's ScarfBench gives AI agents a concrete, high-stakes coding task: migrating enterprise Java frameworks. Here is what the benchmark reveals about where agents actually stand on production-grade refactoring work.
ModelsJun 22, 20262 min read
GLM 5.2 Takes On Claude Opus: What Practitioners Need to Know
GLM 5.2 is being positioned as the best open-source model available, and a high-signal Hacker News thread is running a direct head-to-head against Claude Opus. Here is what the comparison reveals for teams choosing a backbone model.
ModelsJun 17, 20262 min read
GLM-5.2 Becomes the New Open-Weights Leader, Beating GPT-5.5 on Agentic Knowledge Work
Z.ai's GLM-5.2, a 753B-parameter MoE under the MIT license, is now the top open-weights model on the Artificial Analysis Intelligence Index and scored above GPT-5.5 on the new AA-Briefcase agentic eval. The frontier-grade option you can self-host just shifted.