Slide contents

1. 1. AI4SDLC: what I would do differently

   If I knew then what I know now

2. 2. First, I would draw the ownership boundary

   Rent · adapt · own

3. 3. Rent what the market improves faster

   Models — Frontier capability portfolio · providers directly or through OpenRouter

   Agent loop — Leading harnesses: Claude Code; open Codex and OpenCode

   Infrastructure — Standard compute and serving · cloud first, vLLM when needed

4. 4. Adapt the moving boundary to your environment

   Routing — Smart task-to-model routing with budget optimization

   Context — Packaging, memory, and compaction tuned to platform-specific constraints

   Adapters — Semantics across models, tools, and runtime

5. 5. Own authority and switch providers

   Authority — Tool contracts, access, and policies—where architecture meets security

   Evidence — Owned evals and outcome traces

   Switchability — Portable replay and a provider-switch drill

6. 6. The model does not define the system

   One checkpoint produces different production systems

7. 7. Capability without authority cannot deliver outcomes

   Model capability becomes an outcome through authorized action and independent verification

8. 8. Advantage comes from combining the layers correctly

   Changes become improvements as they pass through the entire stack

9. 9. Not every improvement requires a new model

   Fast and slow loops improve different layers

10. 10. Most improvements ship through the fast loop

    Route to another model

    Repackage the context

    Change a tool description

    Tighten policy or gate

11. 11. Slow-loop infrastructure changes the feasible architecture

    Precision · memory · fabric · serving

12. 12. At least four ways close the loop

    Vertical · big labs · product-first · open-weight ecosystem

13. 13. A vertical closes the loop internally

    Product → evidence → behavior → economics → next release

14. 14. Big Labs own the model loop

    Research → model → harness → evals inside; compute with AWS

15. 15. Cursor closed its loop; SpaceX added compute

    SpaceX acquired Cursor on August 14, 2026

16. 16. Open weights move the assembly point

17. 17. A model learns a particular action world

    A compatible schema does not guarantee compatible behavior (well documented by Cursor)

18. 18. The harness covers the next weakness

    A mature capability becomes a tool

19. 19. The contract bounds action consequences

    Identity · effect · provenance

20. 20. Observing does not imply training rights

    Telemetry · evals · training are different modes

21. 21. A convincing trace cannot prove the outcome

    Trace

    Actions and tools

    Errors and recovery

    Cost and policy events

    Outcome

    Environment end state

    Hidden executable checks

    Repeated-run reliability

22. 22. An improvement must survive replay

    Capture task, trace, and final state

    Replay in a frozen environment

    Change one controlled layer

    Run the same independent checks

    Release only a measured improvement

23. 23. Turn the SDLC into verifiable episodes

    Keep the frontier model where search is required

24. 24. The workload profile determines the serving architecture

    Colocate · split phases · disaggregate KV storage

25. 25. Repeatable episodes fit smaller models

    The same gate · escalation under uncertainty

26. 26. A frontier model is not needed everywhere

    Your own learning loop is always required

    Buy capability, not dependency

    Own verifiable episodes

    Remove repeatable variability

    Uncertain → frontier · repeatable → system