Practice: The future of git in an AI world
Self-check questions
Section titled “Self-check questions”Answer each in your own words first, then open the answer to check.
Q1. Which parts of git’s design are likely to remain stable across the next 5-10 years?
Show answer
The snapshot data model, the distributed architecture, branches as movable pointers, content-addressing, and the accountability principle (human responsibility for what’s merged).
Q2. Name three patterns visible in 2026 that suggest where conventions and tooling are shifting.
Show answer
(Pick any three of:)
- Native AI authorship support (more direct than reused Co-Authored-By)
- Semantic-conflict detection across files and branches
- Better worktree management UI (IDE integration)
- Multi-agent orchestration as first-class tooling
- Higher-level commit primitives (stacks, change sets)
- Agent-fleet-specific branching models
Q3. List three claims about AI and git that don’t hold up against the actual mechanics.
Show answer
(Pick any three of:)
- “Git is obsolete in the AI era.”
- “AI will write code without humans.”
- “Commit message craft disappears with AI.”
- “All code will be AI-authored, so attribution doesn’t matter.”
- “Multi-agent workflows will eliminate human engineering teams.”
Q4. What are the three “stay calm” habits the lesson recommends?
Show answer
- Return to the snapshot model when evaluating new tools.
- Distinguish conventions (fast-evolving) from primitives (slow-evolving).
- Don’t chase every new tool; wait 1-2 years for adoption to signal staying power.
Q5. Why is the snapshot model still the right mental model after fifteen lessons?
Show answer
It’s the data layer underneath everything else. Tools layer on top of the snapshot model; they don’t replace it. New tools must read and write the same data git already stores. Changing the snapshot model would require changing the underlying storage of every git repo in existence; that’s not practical.
Q6. What’s the difference between a “convention” and a “primitive” in git, and why does the distinction matter?
Show answer
Primitives are the underlying mechanisms (snapshots, branches, refs, commits). They change slowly (years to decades) because changing them breaks compatibility with the entire ecosystem.
Conventions are the patterns built on top of primitives (commit message format, branch naming, trailer placement, workflow choices). They change quickly (months to years) because adopting a new convention is just a team decision.
The distinction matters because: when you see a change proposed, knowing which level it’s at tells you how much to invest in adapting to it.
Q7. The lesson offers a speculative vision of a 2030 workflow. What’s the difference between honest speculation and marketing speculation?
Show answer
Honest speculation acknowledges uncertainty, grounds claims in current patterns (“here’s what we see today; here’s how it might extrapolate”), and offers explicit caveats. The L16 vision of a 2030 workflow is offered as speculation, not prediction.
Marketing speculation makes confident claims about uncertain futures (“git is obsolete,” “AI replaces engineers”), usually without grounding in mechanics. Often connected to selling something or attracting attention.
The L1-L15 mental models help distinguish these. Marketing claims usually fail when you check them against the actual mechanics.
Q8. Why does the accountability principle persist even as AI tools become more capable?
Show answer
AI tools don’t have legal standing. They can’t be held responsible in incident reviews. They can’t be fired or promoted. The accountability has to land somewhere, and it lands on the human in the loop. This is a structural fact about how organizations and law work, not a technical limitation that better AI tools will solve.
Reflection drills
Section titled “Reflection drills”These are reflective, not operational. There are no “wrong” answers; the goal is to consolidate your own thinking.
Reflection 1: Your favorite voice anchor moment.
The voice anchor “Git stores snapshots. Every other command is just navigating those snapshots” has appeared in every lesson. Pick a moment when this anchor most helped you understand something. Write a paragraph.
Reflection 2: Your biggest shift.
Of the fifteen previous lessons, which one shifted your understanding most? What was the shift? Why did it matter?
Reflection 3: One thing you’ll change immediately.
Pick one specific workflow change you’ll make based on T7. It should be concrete: “I’ll start adding Co-Authored-By trailers.” “I’ll use git worktree for PR review.” “I’ll set up branch protection on main.” Commit to making the change this week.
Reflection 4: One question you still have.
What’s something T7 didn’t fully answer? Often the most valuable post-track outcome is a clear question. Write it down.
Reflection 5: Where will your workflow go from here?
Imagine your team’s git workflow in 2 years. What’s the same? What’s different? What’s the most likely evolution?
Identifying speculation drills
Section titled “Identifying speculation drills”For each claim below, identify whether it’s a STABLE foundation, a PLAUSIBLE EVOLUTION, or MARKETING SPECULATION:
- “Git’s content-addressable storage will be replaced by ML-based change embedding within five years.”
- “Most multi-agent workflows in 2030 will use worktree-based parallelism plus automated integration runners.”
- “Branch protection on main is a foundational guardrail that will remain best practice.”
- “AI agents will eliminate the need for code review entirely.”
- “Semantic-conflict detection tooling will improve substantially over the next 2-3 years.”
- “Co-Authored-By trailers for AI attribution may evolve into a more structured format.”
- “Commit messages will become entirely auto-generated; the human craft of WHY-focused messages will disappear.”
- “The lead orchestrator role in multi-agent workflows is real human work that requires judgment skills.”
- “Distributed version control will be replaced by centralized AI-mediated systems by 2028.”
- “The reflog as a safety net for HEAD-modifying operations is structural to git’s design and likely to persist.”
Track-wide review
Section titled “Track-wide review”This is the only “review the whole track” drill. Pick three commands or concepts from anywhere in T7 (L1-L15) that you want to make sure you remember. For each one, write a one-sentence definition and one situation where you’d use it. The goal: see what stuck with you and which areas might benefit from re-reading.
Flashcards
Section titled “Flashcards”Q. Stable foundations of git
Snapshot model, distributed nature, branches as pointers, content-addressing, accountability principle
Q. Patterns evolving in 2026
Native AI authorship support, semantic-conflict detection, worktree management UI, multi-agent orchestration frameworks, higher-level commit primitives, agent-fleet-specific branching models
Q. Marketing speculation to ignore
Git is obsolete; AI replaces humans; attribution doesn’t matter; commit message craft disappears
Q. Stay-calm habit 1
Return to the snapshot model when evaluating new tools
Q. Stay-calm habit 2
Distinguish conventions (fast-evolving) from primitives (slow-evolving)
Q. Stay-calm habit 3
Don’t chase every new tool; wait for 1-2 years of adoption
Q. Why the snapshot model persists
It’s the data layer; tools layer on top but don’t change the underlying storage
Q. Why accountability persists
AI doesn’t have legal standing; responsibility has to land on the human merger
Q. Honest vs marketing speculation
Honest acknowledges uncertainty, grounds in current patterns; marketing makes confident claims about uncertain futures
Q. T7's most durable takeaway
The voice anchor: git stores snapshots; everything else is navigation