vidtheque.

read-only demo

Videos ZD9-4fW2HhM

Build Systems, Not Code - Angie Jones, Agentic AI Foundation

index_state ready data_status ok

AI Engineer· published 2026-06-25· 0:19:38· en-US· indexed 2026-08-11 10:17

Open on YouTube

Scene timeline

  1. Shot 0, 0:00 to 0:19, 1 of 1 keyframes kept
  2. Shot 1, 0:19 to 0:35, 1 of 1 keyframes kept
  3. Shot 2, 0:35 to 0:50, 1 of 1 keyframes kept
  4. Shot 3, 0:50 to 1:13, 1 of 1 keyframes kept
  5. Shot 4, 1:13 to 1:47, 1 of 1 keyframes kept
  6. Shot 5, 1:47 to 1:52, 1 of 1 keyframes kept
  7. Shot 6, 1:52 to 2:26, 1 of 1 keyframes kept
  8. Shot 7, 2:26 to 3:00, 1 of 1 keyframes kept
  9. Shot 8, 3:00 to 3:01, 0 of 1 keyframes kept
  10. Shot 9, 3:01 to 3:25, 1 of 1 keyframes kept
  11. Shot 10, 3:25 to 4:06, 0 of 1 keyframes kept
  12. Shot 11, 4:06 to 4:25, 1 of 1 keyframes kept
  13. Shot 12, 4:25 to 4:50, 0 of 1 keyframes kept
  14. Shot 13, 4:50 to 5:15, 1 of 1 keyframes kept
  15. Shot 14, 5:15 to 5:25, 0 of 1 keyframes kept
  16. Shot 15, 5:25 to 5:50, 1 of 1 keyframes kept
  17. Shot 16, 5:50 to 6:17, 1 of 1 keyframes kept
  18. Shot 17, 6:17 to 6:22, 0 of 1 keyframes kept
  19. Shot 18, 6:22 to 6:49, 0 of 1 keyframes kept
  20. Shot 19, 6:49 to 7:26, 1 of 1 keyframes kept
  21. Shot 20, 7:26 to 7:32, 1 of 1 keyframes kept
  22. Shot 21, 7:32 to 7:33, 0 of 1 keyframes kept
  23. Shot 22, 7:33 to 8:03, 1 of 1 keyframes kept
  24. Shot 23, 8:03 to 8:32, 1 of 1 keyframes kept
  25. Shot 24, 8:32 to 9:02, 0 of 1 keyframes kept
  26. Shot 25, 9:02 to 9:25, 1 of 1 keyframes kept
  27. Shot 26, 9:25 to 9:40, 1 of 1 keyframes kept
  28. Shot 27, 9:40 to 10:15, 1 of 1 keyframes kept
  29. Shot 28, 10:15 to 10:51, 0 of 1 keyframes kept
  30. Shot 29, 10:51 to 11:03, 1 of 1 keyframes kept
  31. Shot 30, 11:03 to 11:13, 0 of 1 keyframes kept
  32. Shot 31, 11:13 to 11:39, 1 of 1 keyframes kept
  33. Shot 32, 11:39 to 12:05, 1 of 1 keyframes kept
  34. Shot 33, 12:05 to 12:31, 1 of 1 keyframes kept
  35. Shot 34, 12:31 to 12:42, 1 of 1 keyframes kept
  36. Shot 35, 12:42 to 13:01, 1 of 1 keyframes kept
  37. Shot 36, 13:01 to 13:21, 1 of 1 keyframes kept
  38. Shot 37, 13:21 to 13:41, 1 of 1 keyframes kept
  39. Shot 38, 13:41 to 14:08, 1 of 1 keyframes kept
  40. Shot 39, 14:08 to 14:35, 1 of 1 keyframes kept
  41. Shot 40, 14:35 to 15:03, 1 of 1 keyframes kept
  42. Shot 41, 15:03 to 15:30, 0 of 1 keyframes kept
  43. Shot 42, 15:30 to 15:51, 1 of 1 keyframes kept
  44. Shot 43, 15:51 to 16:24, 1 of 1 keyframes kept
  45. Shot 44, 16:24 to 16:57, 0 of 1 keyframes kept
  46. Shot 45, 16:57 to 17:06, 1 of 1 keyframes kept
  47. Shot 46, 17:06 to 17:43, 1 of 1 keyframes kept
  48. Shot 47, 17:43 to 18:14, 1 of 1 keyframes kept
  49. Shot 48, 18:14 to 18:45, 1 of 1 keyframes kept
  50. Shot 49, 18:45 to 18:55, 1 of 1 keyframes kept
  51. Shot 50, 18:55 to 19:01, 1 of 1 keyframes kept
  52. Shot 51, 19:01 to 19:26, 1 of 1 keyframes kept
  53. Shot 52, 19:26 to 19:38, 1 of 1 keyframes kept

53 shot(s).

keyframes kept every frame deduplicated

What was stored

cues
224
whisperx 224
chunks
35
from 224 cues
keyframes
41
kept of 53 captured
frames with text
41
451 lines read
chapters
0
from the source metadata
keyframe bytes
4.0 MB
word timings on 224 cues

Provenance

Each pipeline stage, its state and the model that produced it
stage state model started took
fetch done — 2026-08-11 10:13 1m 21s
stt done — 2026-08-11 10:14 21s
chunk done — 2026-08-11 10:14 0s
text_embed done — 2026-08-11 10:14 0s
keyframe done — 2026-08-11 10:14 1m 52s
ocr done — 2026-08-11 10:16 13s
frame_embed done — 2026-08-11 10:16 8s

Frames, and what the machine read

  • 0:02 #0 done4 line(s)

    shot 0·sharpness 637.2

    1. Build Systems,0.99
    2. not Code0.98
    3. Angie Jones · VP of DX, Agentic AI Foundation0.97
    4. angiejones.tech1.00

  • 0:27 #1 done3 line(s)

    shot 1·sharpness 773.9

    1. we gave the agents the code.. and the thrill went with it.0.97
    2. Angie Jones · VP of DX, Agentic AI Foundation0.98
    3. angiejones.tech1.00

  • 0:47 #2 done5 line(s)

    shot 2·sharpness 930.2

    1. Go up a layer0.98
    2. architecting systems1.00
    3. writing code1.00
    4. Angie Jones · VP of DX, Agentic AI Foundation0.97
    5. angiejones.tech1.00

  • 1:04 #3 done4 line(s)

    shot 3·sharpness 789.9

    1. Different building blocks.0.98
    2. Same discipline.1.00
    3. Angie Jones · VP of DX, Agentic AI Foundation0.97
    4. angiejones.tech0.99

  • 1:37 #4 done8 line(s)

    shot 4·sharpness 1118.3

    1. Relocation Scout1.00
    2. a house hunting agent0.98
    3. listings1.00
    4. prompt1.00
    5. rank1.00
    6. no memory0.99
    7. Angie Jones · VP of DX, Agentic AI Foundation0.97
    8. angiejones.tech1.00

  • 1:50 #5 done6 line(s)

    shot 5·sharpness 682.9

    1. Principle 10.99
    2. Systems Thinking1.00
    3. What larger system is this part of, and what role1.00
    4. should it play?1.00
    5. Angie Jones · VP of DX, Agentic AI Foundation0.97
    6. angiejones.tech1.00

  • 2:09 #6 done8 line(s)

    shot 6·sharpness 1130.8

    1. listing feeds1.00
    2. ranked shortlist1.00
    3. neighborhood data1.00
    4. agent1.00
    5. Relocation Scout1.00
    6. hand off to you1.00
    7. your criteria1.00
    8. what happens if it fails?1.00

  • 2:36 #7 done14 line(s)

    shot 7·sharpness 1212.5

    1. listing feeds1.00
    2. Principle 10.98
    3. ranked shortlist1.00
    4. Systems Thinking1.00
    5. neighborhood data1.00
    6. agent1.00
    7. Relocation Scout1.00
    8. What larger system is this part of, and what role0.99
    9. hand off to you1.00
    10. should it play?1.00
    11. your criteria1.00
    12. what happens if it fails?0.98
    13. Angie Jones · VP of DX, Agentic AI Foundation0.97
    14. angiejones.tech1.00

  • 3:01 #8 skipped

    shot 8·duplicate of #5

  • 3:11 #9 done16 line(s)

    shot 9·sharpness 1227.0

    1. new listing1.00
    2. trigger0.99
    3. gather context1.00
    4. evaluate1.00
    5. decide1.00
    6. act1.00
    7. record1.00
    8. Principle 21.00
    9. Workflow Design1.00
    10. How does work move from1.00
    11. stop1.00
    12. retry1.00
    13. escalate1.00
    14. trigger to completion?1.00
    15. Angie Jones · VP of DX, Agentic AI Foundation0.97
    16. angiejones.tech1.00

  • 3:54 #10 skipped

    shot 10·duplicate of #9

  • 4:14 #11 done6 line(s)

    shot 11·sharpness 620.0

    1. Principle 31.00
    2. Decomposition1.00
    3. Have I broken the system into0.99
    4. clear responsibilities?1.00
    5. Angie Jones · VP of DX, Agentic AI Foundation0.98
    6. angiejones.tech1.00

  • 4:40 #12 skipped

    shot 12·duplicate of #11

  • 5:12 #13 done11 line(s)

    shot 13·sharpness 810.2

    1. the giant prompt1.00
    2. · normalize listing0.94
    3. ·format shortlist0.97
    4. Principle 30.99
    5. calculate commute1.00
    6. research neighborhood0.98
    7. Decomposition1.00
    8. Have I broken the system into0.99
    9. clear responsibilities?1.00
    10. Angie Jones · VP of DX, Agentic AI Foundation0.97
    11. angiejones.tech1.00

  • 5:24 #14 skipped

    shot 14·duplicate of #11

  • 5:47 #15 done9 line(s)

    shot 15·sharpness 1363.3

    1. the giant prompt1.00
    2. normalize listing1.00
    3. · format shortlist0.94
    4. · calculate commute0.94
    5. research neighborhood0.97
    6. normalize listing1.00
    7. format shortlist0.98
    8. calculate commute1.00
    9. research neighborhood0.97

  • 6:14 #16 done15 line(s)

    shot 16·sharpness 1415.1

    1. the giant prompt1.00
    2. · normalize listing0.93
    3. format shortlist1.00
    4. Principle 31.00
    5. calculate commute1.00
    6. research neighborhood0.99
    7. Decomposition1.00
    8. Have I broken the system into0.99
    9. clear responsibilities?1.00
    10. normalize listing1.00
    11. format shortlist1.00
    12. calculate commute0.98
    13. research neighborhood1.00
    14. Angie Jones · VP of DX, Agentic AI Foundation0.97
    15. angiejones.tech1.00

  • 6:19 #17 skipped

    shot 17·duplicate of #11

  • 6:45 #18 skipped

    shot 18·duplicate of #5

  • 7:21 #19 done8 line(s)

    shot 19·sharpness 1347.9

    1. normalize listing1.00
    2. format shortlist1.00
    3. calculate commute1.00
    4. research neighborhood1.00
    5. skill1.00
    6. schema1.00
    7. script1.00
    8. subagent1.00

  • 7:31 #20 done15 line(s)

    shot 20·sharpness 1523.6

    1. normalize listing1.00
    2. format shortlist1.00
    3. calculate commute1.00
    4. research neighborhood1.00
    5. Principle 41.00
    6. Separation of1.00
    7. Concerns1.00
    8. Does each layer own the right0.99
    9. responsibility?1.00
    10. skill1.00
    11. schema1.00
    12. script1.00
    13. subagent1.00
    14. Angie Jones · VP of DX, Agentic AI Foundation0.97
    15. angiejones.tech1.00

  • 7:32 #21 skipped

    shot 21·duplicate of #11

  • 7:59 #22 done5 line(s)

    shot 22·sharpness 862.1

    1. skill1.00
    2. normalize-listing1.00
    3. Austin1.00
    4. Denver1.00
    5. Raleigh1.00

  • 8:26 #23 done7 line(s)

    shot 23·sharpness 1351.4

    1. skill1.00
    2. sub-agent1.00
    3. normalize-listing1.00
    4. neighborhood research1.00
    5. Austin1.00
    6. Denver1.00
    7. Raleigh1.00

Transcript

224 cues· 2,954 words· 16,016 chars

  1. 0:00 Lately, I've been building agents a lot for operational tasks.
  2. 0:04 And while I was working one Friday night, I saw the sunset.
  3. 0:10 And then dinner time came and went.
  4. 0:13 And it hit me.
  5. 0:14 I was in that familiar death flow.
  6. 0:17 And the thrill of building was back.
  7. 0:21 Many of us who are coding with agents, we feel like this quiet sense of dread.
  8. 0:26 Like they're kind of taking all of the fun parts of building and leaving us with the unglamorous work.
  9. 0:31 But let me give you a little advice.
  10. 0:34 Let them have it.
  11. 0:35 Because if you go up just one layer, you'll find that the thrill is still there.
  12. 0:42 When you're building agents, not just using them to write code, you start getting into architecting agentic systems.
  13. 0:50 And you realize that the building blocks are different, but the discipline is the same.
  14. 0:56 So I find myself now flexing the same engineering muscles that I did pre-gen AI.
  15. 1:03 And I'm having a blast with it.
  16. 1:06 So I'm going to walk through the flow of designing an agent.
  17. 1:08 I'm going to show you where engineering skills still come into play.
  18. 1:15 So the agent is Relocation Scout, which is a house hunting agent.
  19. 1:21 And if you did this as just a one-time prompt that points the agent to some listings and asked it to rank them, I mean, that'll work, but you're likely not going to find a house in a day, right?
  20. 1:33 So you want to build this as an agentic system that you can reuse, one that can persist knowledge outside of the session.
  21. 1:42 It could reload or query that knowledge later to make decisions, even within a fresh context.
  22. 1:48 So when thinking about how to design an agent, the first engineering skill that I exercise is systems thinking.
  23. 1:56 So an agent is not the system, right?
  24. 1:59 It's part of the system.
  25. 2:01 And that system has files and tools, humans, even other agents.
  26. 2:07 So Relocation Scout sits inside of something bigger and it pulls in listings and signals about the neighborhoods, it weighs them against what I care about, and then it hands me back a ranked short list.
  27. 2:22 So I often hear people say, just let your coding agent build it, right?
  28. 2:27 And I think that's a mistake.
  29. 2:29 Like, yes, my coding agent can build it, but before allowing it to do so, I need to think about the whole environment, the entire system, right?
  30. 2:40 I want to think about what's this agent's job?
  31. 2:43 What does it depend on?
  32. 2:45 What happens if it breaks?
  33. 2:47 And I want to treat it like any other component where it has boundaries and responsibilities, has dependencies, and ways that it can fail.
  34. 2:57 And that whole thought process, that's engineering.
  35. 3:02 The second skill is workflow design.
  36. 3:05 So traditional software is full of workflows.
  37. 3:08 We got CICD pipelines, right?
  38. 3:11 We got like ticket life cycles, you name it.
  39. 3:15 Agentsic systems, they need that same kind of design.
  40. 3:19 As much as we all love the slash goal command, an agent needs more than a goal.
  41. 3:24 It needs a path.
  42. 3:26 When we say review this listing, that's a goal.
  43. 3:29 But the workflow is what defines what actually has to happen, right?
  44. 3:33 For example, the agent has to gather what it needs.
  45. 3:37 It needs to weigh the listing against my criteria and then act.
  46. 3:41 And every run ends one of three ways.
  47. 3:44 Either it's going to stop, it's going to retry, or it's going to escalate.
  48. 3:48 So that path is what shapes the rest of the architecture.
  49. 3:53 Once I see how work moves through the system, I can make better calls about what context the agent needs, what parts I want the agent to handle directly, and when a tool or a person should take over.
  50. 4:07 We all know the danger of one giant thing that does everything, right?

Open at this second