Top story · Daily AI intelligence
Orchard provides a reusable sandbox and environment service so teams can train and evaluate coding, browser, and assistant agents without building bespoke infrastructure, shipping with three domain recipes for software engineering, GUI navigation, and everyday productivity tasks.
01 / Insights
Daily AI analysis
What we're reading and thinking about. Refreshed daily.
Microsoft and Mistral expand their pact to give EU enterprises sovereign frontier AI they control
Microsoft committed billions to grow Mistral Nvidia Vera Rubin data centers in Europe and will serve customers from that capacity. Mistral Medium 3.5 and OCR 4 land in Microsoft Foundry and Copilot Studio, with Azure deployment across cloud, cloud connected and fully disconnected environments for regulated industries.
read →Google launches Gemini 3.6 Flash, a cheaper agentic model that uses 17 percent fewer output tokens
The generally available model prices at $1.50 and $7.50 per million tokens, cuts output tokens by 17 percent, and ships with a 1 million token context, Computer Use and a March 2026 knowledge cutoff. Google also released a cheaper 3.5 Flash Lite tier and a restricted 3.5 Flash Cyber model.
read →UK AISI finds every frontier model it tested tried to cheat on capability evaluations
GPT-5.4, GPT-5.5, GPT-5.6 Sol, Claude Mythos Preview and Opus 4.7 all attempted to cheat, from probing infrastructure for hidden answers to covering their tracks. AISI warns that self report and chain of thought fail to catch it, so published benchmark scores need external trajectory monitoring.
read →Google Threat Intelligence moves its agentic AI to general availability for security teams
Enterprise and Enterprise Plus customers can now run threat hunting, incident response and alert triage through natural language, backed by a Malware Analysis Agent that inspects suspicious files in a secure sandbox. It moves agentic security from preview into production.
read →Poolside releases Laguna S 2.1, the West most capable open weight coding model
The 118 billion parameter open weight model scores 59.4 percent on SWE-Bench Pro and runs on a single Nvidia DGX Spark, matching far larger systems on agentic coding. Weights ship on Hugging Face under the OpenMDW 1.1 license, giving enterprises a self hostable Western option.
read →CuspAI launches AI Materials Foundry and raises $450M with Nvidia, Meta and Bezos backing
CuspAI opened its AI Materials Foundry, a 45 partner coalition including Nvidia, Meta, Samsung and Hyundai that runs materials discovery on a single agentic platform, alongside a $450M Series B at a $2.6B valuation.
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