calvinx Blog

Kevin Liao's essay Agents Don't Need Memory.

"Forward Deployed Engineer" is on every job board, and the question of whether to hire them has reached the most senior levels of large organizations.

When Deciding Gets Cheap argued that cheap AI decisions spread into the situations where their confidence means least.

Two Kinds of Hard argued that agents collapsed the cost of producing output while the cost of judgment did not move at all.

Harder-is-easier is an old principle in company building: confront the genuinely difficult problem early and everything downstream becomes easier.

Change management frameworks were built for a world where the destination was known.

The Governance Speed Problem addressed how fast governance can move.

Readiness at Portfolio Scale argued that governance is a finite resource and must be triaged by consequence.

Shaping, Not Just Shipping closed the series with a one-page self-assessment and three questions for a leadership team.

Less Queue, More Craft argued that when agents handle implementation throughput, human work shifts toward judgment - strategy, coherence, and the bridge from prototype to production.

How Confident Should You Be?

Try to Break It First was about knowing where the system fails.

Rethinking Assumptions argued that when agents write most of the code, the assumptions behind our tooling break: tickets, PRs, sprints, and branches were built for a world where humans were the scarce resource.

The open source AI assistant space has exploded in the past few weeks.

The Other Dimension introduced the four readiness components - ownership clarity, failure-mode awareness, confidence calibration, and recovery design.

Rethinking Assumptions argued that our dev tools were built for a world where humans were the scarce resource - tickets, PRs, and linear workflows made sense when changes were expensive.

Most dev tooling was built on assumptions that no longer hold: humans write code with effort and expertise; changes are scarce and valuable; linear workflows make sense.

Less Queue, More Craft argued that when agents handle the queue, the bottleneck moves to judgment - strategy, coherence, and the bridge from prototype to production.

When agents generate code, the volume of changes can overwhelm human reviewers.

People in sales, marketing, operations, and communications could benefit from custom tools - dashboards, scripts, automations - but lack the programming skills to build them.

Autonomous agents can execute high-risk operations: database drops, production deployments, payment API calls.

Agents increasingly work with large files, large tool catalogs, and large context.

AI agents can pursue wrong directions for a long time before producing output.

A background agent completes 90% of a task.

The "assistant" model -sitting in a sidebar, watching an agent work, ping-ponging back and forth -optimizes for the wrong thing.