Skip to content

Instantly share code, notes, and snippets.

@bparanj
Created May 22, 2026 15:17
Show Gist options
  • Select an option

  • Save bparanj/ac5d518dc4ce37b74831d650348a33cb to your computer and use it in GitHub Desktop.

Select an option

Save bparanj/ac5d518dc4ce37b74831d650348a33cb to your computer and use it in GitHub Desktop.

Stave fits very well.

The article’s core point is:

AI is only valuable if it reduces long-term maintenance cost, not just short-term output speed.

James Shore argues that every new line of code creates future maintenance work, and AI coding agents can make this worse if they increase output without reducing the cost of understanding, fixing, and changing that output. He says the math only works when AI reduces maintenance cost in proportion to the productivity gain. (James Shore)

Stave’s fit

Stave should be positioned as:

AI-assisted security that reduces maintenance cost by turning cloud security knowledge into durable invariants.

Not:

AI generates more security checks.

That creates maintenance burden.

Better:

Stave helps convert security judgment into stable, reviewable rules that can be reused across snapshots, incidents, reports, and customer assessments.

The key distinction

AI can produce more security work.

Stave should reduce the cost of maintaining security work.

AI-only security workflow Stave workflow
AI generates findings Human defines invariant
More alerts Fewer, higher-value unsafe states
Hard to trust Deterministic output
Hard to reproduce Snapshot-based repeatability
Prompt-dependent Rule-based evaluation
Expensive to review again Reusable invariant
Point-in-time answer Time-aware evaluation

This is the strongest angle:

AI increases output. Stave preserves judgment.

How Stave reduces maintenance cost

1. Security knowledge becomes reusable

Without Stave, each S3 exposure review is manual.

The engineer asks:

  • Is the bucket public?
  • Is it public through policy?
  • Is it public through ACL?
  • Is Public Access Block masking a latent exposure?
  • Did this unsafe state persist over time?
  • Is the evidence clear enough for an executive report?

With Stave, that knowledge becomes an invariant.

The next review is cheaper.

The same rule runs again.

2. AI output becomes bounded

AI can help write:

  • explanations
  • draft invariants
  • test cases
  • incident summaries
  • remediation text

But Stave gives the hard boundary:

The final answer must pass deterministic evaluation.

That reduces review cost.

The engineer does not have to trust a generated paragraph. They review a rule, a snapshot, and an output.

3. Maintenance moves from prose to contracts

A prompt is soft.

A document is passive.

A scanner finding is temporary.

An invariant is active.

It becomes a contract:

This unsafe state must not exist.

That is easier to maintain than repeated expert analysis.

4. Stave avoids AI lock-in

The article warns that if AI creates hard-to-maintain code, the cost remains even after the AI is removed. (James Shore)

Stave avoids that trap because its durable asset is not the AI conversation.

The durable asset is:

  • snapshot JSON
  • invariant YAML
  • deterministic output
  • golden tests
  • report evidence

AI can assist, but the system does not depend on AI to keep working.

Best positioning line

Use this:

Stave uses AI where it helps, but keeps security decisions in deterministic invariants.

Or sharper:

Stave turns AI-assisted cloud security work into maintainable security rules.

Homepage angle

Headline

Reduce the maintenance cost of cloud security reviews.

Subheadline

Stave turns real cloud security failures into offline, deterministic invariants, so AI-assisted security work becomes repeatable, reviewable, and cheaper to maintain.

Article angle

Write a post titled:

AI Security Tools Should Reduce Maintenance Cost, Not Create More Findings

Structure:

  1. AI can generate more findings.
  2. More findings increase review burden.
  3. Security teams do not need more output.
  4. They need reusable judgment.
  5. Stave captures that judgment as invariants.
  6. Example: S3 public exposure.
  7. Run once, reuse forever.

The Stave thesis

The thesis is:

The future of AI security is not more generated findings. It is maintainable security knowledge.

Stave fits because it makes security knowledge executable.

Simple final positioning

Stave is not an AI scanner.

It is a maintenance-cost reducer for cloud security work.

It helps teams go from:

“Ask AI again and review another answer.”

to:

“Run the invariant and inspect deterministic evidence.”

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment