Skip to content
techspike.tv

techspike.tv

Vibeslop before you vibeship

A techspike is a small, time-boxed coding experiment that answers one question.

Click and drag the globe to glitch it.

01 · Definition

What is a techspike?

A techspike is the smallest thing you can build to answer one risky question. Not a feature, not a roadmap — a probe that turns a guess into evidence before you bet real time on the build.

  • One question, one probe — scoped to whatever you're unsure about.
  • Time-boxed to hours or a weekend, never a quarter.
  • Disposable by design — the answer matters, the code doesn't.
  • Built to be deleted the moment you've learned.

02 · Value

Why spike first?

Estimates are fiction until someone builds the risky part. A spike burns that risk down in a day — so you commit to the real build on evidence, not optimism.

Evidence beats opinion
A running probe settles arguments a whiteboard can't. Spend a day to learn the API is wrong, not a quarter.
A decision, not a deliverable
The output is a yes/no on your hypothesis. Ship the decision; throw the code away without a second thought.
Loops at agent speed
Small and single-question is exactly the shape an AI agent can take a first pass at while you review. Bets that took weeks now take minutes.

03 · Method

How to ship one

Run the loop at machine speed. Each pass turns one question into evidence and seeds the next bet.

  1. 01

    Name the question

    State the one risky assumption. If you're not unsure, you don't need a spike.

  2. 02

    Build the smallest probe

    The simplest thing that could answer it — no abstractions, no gold-plating.

  3. 03

    Run it, read the result

    Verify against reality, not your intent. The result is the deliverable.

  4. 04

    Keep or kill

    Commit to the build or walk away — then re-bet. Survivors get archived here.