Top story · Daily AI intelligence
Attackers used AI generated voice clones of executives to try to trick employees at major hedge funds into granting system access. Two Sigma blocked the attempt with no data compromised, while Point72 told investors it had been targeted.
01 / Insights
Daily AI analysis
What we're reading and thinking about. Refreshed daily.
UN and ITU launch an AI for Good Global Commission led by Kagame and Benioff
More than 40 founding members including Jensen Huang, Andy Jassy and Brad Smith will push responsible AI and access for developing nations, meeting first at the Geneva AI for Good Summit on July 7.
read →White House nears a voluntary release framework for frontier AI models
The US is in advanced talks with OpenAI, Google and Anthropic on benchmarks, testing timelines and access rules for the next model tier, with an announcement possible the week of July 7.
read →Microsoft launches Frontier Company with 6,000 engineers and $2.5B to fix failing enterprise AI pilots
Microsoft is embedding forward deployed AI engineers directly inside enterprise customers to design, deploy and run their AI systems, its answer to research showing 95 percent of gen AI pilots deliver no measurable ROI. Amazon, OpenAI and Anthropic have all stood up similar units.
read →Mistral releases Leanstral 1.5, an Apache 2.0 Lean 4 proof agent solving 587 of 672 PutnamBench problems
The open weight 119B model with 6.5B active params saturates miniF2F and sets a new state of the art on formal math benchmarks at about 4 dollars per problem, roughly 75 times cheaper than rivals. It ships free on Hugging Face and via API.
read →Microsoft plans a unified Copilot super app in August with new AutoPilot background agents
An internal memo says Microsoft will merge its consumer and enterprise Copilot into one paid app built around coding and AutoPilot agents that handle scheduling and email, while stripping out underused features. It follows Anthropic and OpenAI into the AI super app race.
read →UK AI Security Institute finds standard benchmarks systematically understate what AI agents can do
Fixed compute budgets cut evaluations short, and giving agents more compute lifted success rates by up to 25 percent on cyber and software tasks. Capability is a curve over compute, not a fixed score, which reshapes how enterprises should assess agent risk.
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