Latest
CLIs are hands, skills are knowledge, MCP is reach
A simple model for deciding when an agent needs a command-line tool, when it needs a skill, and when it needs an MCP server.
Weekly AI Roundup
What caught our eye with AI this week, in five minutes.
- Issue 4
What data from hundreds of organisations says about AI in software delivery, why refactoring cuts token costs, benchmarking tech stacks for agents, and a governed chain from conversation to code.
- Issue 3
OpenAI’s DevDay announcements, a million agent skills, why context compounds, and agents moving into Slack.
- Issue 2
Small decision models versus generative LLMs, how much context is the right amount, and a repository-native take on agent task tracking.
More from the crew.
To review, or not to review?
Thoughts on how to approach reviewing code changes in the era of AI development agents.
The pipeline is the only reviewer that scales
Why AI-native delivery means automatic deployments to every environment, including production, gated by tests instead of people.
Best place for your AI context (for now)
Linear says issue tracking is dead. For enterprises, the boring answer is still: put your context where the whole business already reads it.
Kaiku at Vibe Code Finland
Graham Daw and Akseli Manninen on why shared context is what makes AI-assisted delivery scale.
We moved to Kaikukatu after one epic year
Kaiku is moving to our own office on Kaikukatu.
Skills: everyone's job, no one's responsibility
Shared ownership make skills hard to get right, but gold when you do.
Kaiku at Context is King
Graham Daw delivered a keynote at Wolt's office on how context changes everything in AI-augmented delivery.
The technology partner for forward-looking, AI-powered engineers and companies.
Kaiku Crew Oy / 3520495-1
Kaikukatu 4, 00530 Helsinki