Why we're building Merak.

Most engineering orgs still run on rituals from 2014. Daily standups to surface what changed. Quarterly OKR docs no one re-reads after week one. Weekly status packets pushed into inboxes. Postmortems written for the doc, not the learning. PR reviews bottlenecked on a single tech lead.

None of it scaled to a team of ten engineers and a handful of ambient agents writing half the code. None of it scales to where the next two years are going.

When humans and AI share a workspace, they have to agree on what good looks like. Not in prose. Not in retros. In signals — measurements that update on their own and don’t care who’s in the room.

Humans can hold three things at once. Agents can hold thousands. The asymmetry only matters if everyone is reading from the same numbers, with the same definitions of primary and guardrail. That alignment is the product. That’s the common language of goodness.

Merak is the substrate. Goals on top — objectives, KRs, the things you’re trying. Work in the middle — tasks, artifacts, automations, the moves humans and agents make. Truth underneath — signals from your tools and from Merak itself. The agent harness runs the loop against KRs you set; guardrails fire before things break.

You stay the founder. The work compounds inside the substrate, not in seventeen Slack threads no one will read in six months.