1
00:00:00,000 --> 00:00:03,091
Sam: I've shipped C plus plus code with Make for years.

2
00:00:03,250 --> 00:00:04,924
Sam: Why would I need a harness?

3
00:00:05,083 --> 00:00:10,688
LitAI: In this course, we create a new Literate AI project and prove
that its starter application runs.

4
00:00:10,838 --> 00:00:15,807
LitAI: The specification is durable authority; generated source is
disposable output.

5
00:00:15,963 --> 00:00:20,709
LitAI: These are edited transcript replays of real sessions, not raw
screen recordings.

6
00:00:20,879 --> 00:00:23,011
LitAI: Remote-worker steps are a walkthrough.

7
00:00:23,171 --> 00:00:25,322
Sam: A Makefile tells me how to build.

8
00:00:25,472 --> 00:00:26,590
Sam: Isn't that enough?

9
00:00:26,740 --> 00:00:34,801
LitAI: Your Makefile compiles code, but it does not define the
specification, acceptance evidence, packaging, or release policy.

10
00:00:34,951 --> 00:00:39,332
LitAI: Literate AI separates what the application must do from how and
where it is built.

11
00:00:39,493 --> 00:00:47,791
LitAI: A Component describes observable behavior; Flavors select
implementation targets; receipts bind the accepted result to exact
inputs.

12
00:00:47,951 --> 00:00:52,005
LitAI: Start with Python, Git, and an authenticated coding CLI.

13
00:00:52,159 --> 00:00:57,358
LitAI: Install a published Literate AI wheel in a virtual environment,
then run litai doctor.

14
00:00:57,508 --> 00:01:00,835
LitAI: Resolve reported blockers before requesting generation.

15
00:01:00,985 --> 00:01:03,575
LitAI: The first onboard create command is read-only.

16
00:01:03,735 --> 00:01:09,292
LitAI: Review its proposed project type, target Flavors, blockers, and
plan identity.

17
00:01:09,444 --> 00:01:11,424
LitAI: No project tree is written yet.

18
00:01:11,574 --> 00:01:14,589
LitAI: Apply the reviewed plan with explicit acknowledgement.

19
00:01:14,741 --> 00:01:22,068
LitAI: For automation or a long pause between planning and applying,
use the printed plan identity to fail closed if the plan has changed.

20
00:01:22,218 --> 00:01:29,161
LitAI: The scaffold contains project goals, agent onboarding, catalog
entries, and an executable hello Component.

21
00:01:29,311 --> 00:01:36,163
LitAI: Status shows the next step; project validate checks authored
structure without claiming that an application has been built.

22
00:01:36,313 --> 00:01:39,944
LitAI: A lock binds the exact specification and selected Flavors.

23
00:01:40,105 --> 00:01:44,213
LitAI: Planning exposes the recipe before a coding agent generates
source.

24
00:01:44,363 --> 00:01:48,818
LitAI: If either check fails, fix the authored inputs rather than
bypassing the gate.

25
00:01:48,988 --> 00:01:52,486
Sam: If make succeeds, I've shipped the product, right?

26
00:01:52,655 --> 00:01:53,242
LitAI: No.

27
00:01:53,405 --> 00:02:00,328
LitAI: You have never really decided how to package and release this
software, or how to prove the shipped artifact is the one you tested.

28
00:02:00,488 --> 00:02:05,789
LitAI: Rebuild invokes the coding agent and then runs the generated
application through independent gates.

29
00:02:05,946 --> 00:02:12,874
LitAI: It may call a paid model and execute generated code, so
acknowledge host execution only on a trusted machine.

30
00:02:13,030 --> 00:02:18,380
LitAI: This is an edited excerpt from the actual Codex generation
session, not a screen capture.

31
00:02:18,530 --> 00:02:22,441
LitAI: The agent produced the entrypoint and three native behavior
tests.

32
00:02:22,591 --> 00:02:26,630
LitAI: They cover a normal name, a one-character boundary, and
Unicode.

33
00:02:26,799 --> 00:02:32,874
LitAI: We show the resulting behavior here; source is still only a
candidate until independent acceptance passes.

34
00:02:33,024 --> 00:02:34,464
Sam: The code looks fine.

35
00:02:34,614 --> 00:02:36,450
Sam: Why did your rebuild fail?

36
00:02:36,614 --> 00:02:40,519
LitAI: Because the independent verifier protects us from our own
assumptions.

37
00:02:40,669 --> 00:02:42,365
LitAI: The first live rebuild failed.

38
00:02:42,515 --> 00:02:50,975
LitAI: Its generated application matched the written greeting
contract, but the scaffold's independent oracle expected extra fields
absent from that contract.

39
00:02:51,140 --> 00:02:55,192
LitAI: The gate rejected the candidate instead of issuing a false
passing receipt.

40
00:02:55,348 --> 00:02:58,518
Sam: You didn't just edit generated code to make it green?

41
00:02:58,681 --> 00:03:06,358
LitAI: We corrected the demo verifier to match the specification and
reran the lifecycle: three tests passed and a current receipt was
committed.

42
00:03:06,515 --> 00:03:10,844
LitAI: Then we fixed the framework starter oracle itself and added
regression coverage.

43
00:03:10,994 --> 00:03:14,216
LitAI: New scaffolds no longer inherit the incorrect fields.

44
00:03:14,369 --> 00:03:19,080
LitAI: This is a framework defect repair, not editing generated code
to fake a pass.

45
00:03:19,244 --> 00:03:21,821
Sam: So is the wheel now on a package registry?

46
00:03:21,971 --> 00:03:22,559
LitAI: No.

47
00:03:22,721 --> 00:03:25,725
LitAI: A verified local wheel is not a published release.

48
00:03:25,888 --> 00:03:30,163
LitAI: We then verified the project and built a pip wheel from the
accepted artifact.

49
00:03:30,313 --> 00:03:33,819
LitAI: Package verification passed using the same model selection.

50
00:03:33,979 --> 00:03:41,629
LitAI: This gives a real, locally verified product; publishing it or
cutting a release remains a separate, explicitly authorized gate.

51
00:03:41,779 --> 00:03:46,738
LitAI: For the next feature, change the Component specification and
repeat the evidence loop.

52
00:03:46,904 --> 00:03:50,035
LitAI: Avoid treating generated source as the durable design.

53
00:03:50,195 --> 00:03:56,059
LitAI: The project becomes easier to review because intent, target
choices, and proof stay separate.

54
00:03:56,209 --> 00:04:00,763
Sam: If Agentis can organize the work, why do I need all these extra
gates?

55
00:04:00,917 --> 00:04:04,718
LitAI: Because organizing a task is not proof that its product is
correct.

56
00:04:04,876 --> 00:04:16,832
LitAI: This is the division of responsibility we want with Agentis: it
coordinates work, while I keep specifications, exact build inputs,
tests, packaging, and release evidence explicit.

57
00:04:17,001 --> 00:04:20,987
LitAI: You have spent years on the code and too little thought on what
actually ships.

58
00:04:21,137 --> 00:04:23,387
LitAI: These gates make that missing work visible.

59
00:04:23,554 --> 00:04:28,010
LitAI: This course proves the local harness; it does not claim an
Agentis deployment.