The AI code + design space is not a monolith. As plans of attack are shaping up, I can see paths emerging. Here's a horizon lookout.
⬆️ Capture the Workflow: Own the daily surface (IDE). Sell seats. The DevX retention game.
⬇️ Compress the Stack: Replace teams/ ICs for zero-to-ones (V1s). Sell outcomes. The "Time-to-product" + QA/QE/QC game.
⬅️ Engineers
➡️ Non-engineers. (including Designers)
The interesting thing is, with the lens, a clear white-space forms in the matrix:
"Capturing the entire workflow at scale, for non-engineers/designers ICs".
No one is really there, yet.
Cursor (Anysphere).
Wedge: Being "VS Code".
Defensibility: Seat density.
Windsurf.
Wedge: Also IDE-shaped but priced by task, instead of per seat.
Defensibility: Cognition's enterprise pipeline
v0 (Vercel).
Wedge: Design-to-component.
Defensibility: Vercel hosting.
Bolt.new (StackBlitz).
Wedge: Full-stack browser apps (engineer credible).
Defensibility: WebContainers are hard to clone.
Lovable.
Wedge: IDE-free. "describe your app, get a deployed app," sold to founders, operators, and PMs who never touch an IDE.
Defensibility: Category owner. Brand + behaviour change trailblazer.
Replit Agent.
Wedge: Ships directly into Replit's hosting.
Defensibility: Platform lock-in.
Claude Code (Anthropic).
Wedge: Lives where engineers already work (git, shell...).
Defensibility: Anthropic !! owns the model layer underneath much of the cohort above !! (Cursor, Windsurf, and others route heavy traffic through Claude). Insane advantage.
I'm very very curious on who's building the "Cursor for PMs or Ops" at scale. I mean full IDE, good usability, traceability, low oversight needs etc.