History · Trends · Tools
From glorified autocomplete to the orchestration of agent fleets
Markus Klug
Glorified autocomplete
Copy/paste programming
AI bolted onto the IDE it already had
Rebuilt from first principles around the model
Cursor Mar 2023 · Composer mid-2024 · Windsurf Nov 2024 · VS Code agent mode Feb 2025
The agent moves into your terminal
Research preview Feb 2025 · GA May 2025 · Codex CLI, Gemini CLI and Amp followed
Parallelism
Adding workers doesn't remove a bottleneck, it relocates one
The building blocks arrived one by one
One session fans out research, review, and search to specialized workers
Every agent gets its own git worktree — parallel edits and merge queues
Agents run long jobs detached and report back — no more spinner-watching
Agents keep working in the cloud while your laptop sleeps
From conversation to orchestration
The new craft
Stop prompting every step — design the loop that prompts the agent
Every coding agent is a while-loop: call the model, run tools, feed results back. Engineering that loop — not the prompt — is the leverage
A testable goal and explicit exit criteria — defined before it runs
Real feedback: tests, compilers, logs — not vibes
Context that gets summarized, not accumulated
Failure paths: what retries, what escalates to you
The new craft
Loops made agent behavior programmable — graphs make agent organizations programmable
Nodes: agents with a role, a domain, memory
Edges: data flow and delegation paths
Advisor–orchestrator: a planner routes to workers
Zone defense: specialists own stable domains
Council: rival views deliberate, then synthesize
Role boundaries, handoffs, failure isolation — the org chart becomes code
Five years, six form factors
Audience check
Steve Yegge's 8 stages of the AI-assisted coding journey
Stage 1 · Near-Zero AI
Stage 2 · IDE Agent
Stage 3 · IDE, YOLO
Stage 4 · Wide Agent
Stage 5 · CLI, Single
Stage 6 · CLI, Multi
Stage 7 · 10+ Agents
Stage 8 · Own Orchestrator
Steve Yegge, "Welcome to Gas Town" (2026) · code surface vs agent surface · click a stage to zoom
read the postNew blog post: "Software 2.0"
↳ the essay arguing neural nets, not hand-written code, would become the software
The hottest new programming language is English
"There will be no [human] programmers in five years."*
It is our job to create computing technologies that nobody has to program… the programming language is human: everybody in the world is now a programmer.
There's a new kind of coding I call "vibe coding", where you fully give in to the vibes, embrace exponentials, and forget that the code even exists…
In three to six months, AI is writing 90% of the code… in 12 months, essentially all of the code.
For 25% of the Winter 2025 batch, 95% of lines of code are LLM generated. That's not a typo.
I've never felt this much behind as a programmer. The profession is being dramatically refactored…
The era of humans writing code is over. Disturbing for those of us who identify as SWEs, but no less true.
I don't prompt Claude anymore. I have loops that are running… My job is to write loops.
You shouldn't be prompting coding agents anymore. You should be designing loops that prompt your agents.
Are we still talking loops or did we shift to graphs yet?
Outlook
A new set of problems — and a new paradigm to solve them in
The landscape
Four layers, each rebuilt from scratch in under three years — and already shedding casualties
the thing that reads, edits and runs
many agents, one human
where the agent lives in the editor
protocols, sandboxes, review, plumbing
click any name to open it · struck = already gone
Assemble your own
What my setup is made of — two open-source tools, one rendering library, one web companion, and a handful of hooks
Terminal workspace manager for agents. Mine adds declarative per-project layouts, label-addressed panes and popup launchers.
hh <project> · h-send / h-read · prefix+o ops · prefix+c claude
Review-first terminal diff viewer, paired to each Claude pane: it shows exactly what that session changed, live.
baseline at SessionStart · touched-set per edit · notes → next prompt
My very own local companion — live hook-event feed, session and worktree diffs with per-hunk staging, inline notes back to the agent. Live demo in a minute.
built on Claude Code hooks, herdr and Pierre components
The rendering layer under both hunk and otomo — Shiki-highlighted split or unified diffs and file trees, framework-agnostic.
shadow DOM · split / stacked · line selection
The glue. Three lifecycle events drive the whole review loop — no agent cooperation required, and every hook exits 0.
SessionStart · PostToolUse · UserPromptSubmit
One key in the files panel: staged diff + your last 15 commits → a single Haiku call → pick, edit, commit. Key from the Keychain.
C → git diff --cached → POST /v1/messages → gum