Top story · Daily AI intelligence
OpenAI published Agent Plugins 1.0, an open format that bundles Agent Skills and MCP servers into portable plugins any compatible client can discover and load. A steering committee of OpenAI, Amazon, Microsoft, Cursor, and Vercel backs it, Google joined as a core maintainer, and ChatGPT, Codex, Cursor, GitHub Copilot, and VS Code support it at launch.
01 / Insights
Daily AI analysis
What we're reading and thinking about. Refreshed daily.
Sight Machine releases an agentic manufacturing platform
Built on a Semantic Model that maps sensor and operations data so agents can reason over assets, processes and KPIs, letting process experts deploy improvements in days.
read →Research: agent memory systems can degrade accuracy via sycophancy
Popular memory/personalization systems can bias models toward stored preferences even when irrelevant; the suggested mitigation is A/B-checking outputs with and without retrieved memory.
read →OpenClaw v2026.6.6-beta.2 tightens security and delivery
Tighter security boundaries with exec approvals failing closed on timeout, safer Telegram and iMessage delivery through tool calls, lower Control-UI latency, and default compaction timeout dropped to 180s.
read →OpenAI and Visa let ChatGPT agents shop and pay autonomously
OpenAI partnered with Visa so ChatGPT agents can complete real purchases on a user's behalf via a linked Visa card — agentic commerce on a major payment rail.
read →Anthropic: Claude now authors 80%+ of merged code, warns on self-improvement
Anthropic published a recursive self-improvement note saying Claude writes most of its own merged code; Amodei is pushing for mandatory third-party audits of frontier models.
read →Google DeepMind releases DiffusionGemma, a text-diffusion open model
DiffusionGemma 26B-A4B (Apache 2.0) uses text diffusion on the Gemma 4 backbone, hitting 1000+ tok/s on an H100 and fitting in ~18GB VRAM quantized.
read →02 / Weekly Deep Dive
How agents actually ship
One researched edition a week: real deployments, the stacks behind them, and the honest counter-facts.
Caylent builds its own agent orchestrator and cuts a 2,600-hour migration to 220
Instead of buying a migration tool, Caylent built a multi-agent orchestrator on the Claude Agent SDK that runs inside the customer's own cloud and cut one engagement from 2,600 engineering hours to about 220.
Eve makes the buy-a-vertical-vendor case for small plaintiff law firms
Plaintiff firms with no in-house ML team buy their agents off the shelf, and Eve customers report large jumps in lead conversion, case value, and drafting speed.
Suzano puts a natural-language-to-SQL agent in front of 50,000 staff
A Brazilian pulp-and-paper maker built a Gemini agent that turns plain-language questions into SQL, which Google Cloud says cut query time by about 95 percent for 50,000 employees.
03 / Latest project
MomMe, a lifestyle community, web & iOS
A full lifestyle community platform, web portal and native iOS app, taken end to end, with an AI moderation flow keeping the community safe at scale.
- MomMe web portal + iOS application
- Full lifestyle community with 3,000+ registered users
- AI moderation flow
04 / Contact
Get in touch
Questions, ideas, or a project in mind? Send us a note and we'll get back to you, usually within a day.
hello@kmmt.ai