vidtheque.

read-only demo

Videos VGN22pPpb-8

Thinner Agents on a Smarter Substrate: The Ontology-based Semantic Layer — Emil Eifrem, Neo4j

index_state ready data_status ok

AI Engineer· published 2026-07-22· 0:11:06· en-US· indexed 2026-08-10 19:43

Open on YouTube

Scene timeline

  1. Shot 0, 0:00 to 0:03, 1 of 1 keyframes kept
  2. Shot 1, 0:03 to 0:05, 1 of 1 keyframes kept
  3. Shot 2, 0:05 to 0:12, 1 of 1 keyframes kept
  4. Shot 3, 0:12 to 0:15, 1 of 1 keyframes kept
  5. Shot 4, 0:15 to 0:34, 1 of 1 keyframes kept
  6. Shot 5, 0:34 to 0:37, 1 of 1 keyframes kept
  7. Shot 6, 0:37 to 1:02, 1 of 1 keyframes kept
  8. Shot 7, 1:02 to 1:28, 1 of 1 keyframes kept
  9. Shot 8, 1:28 to 1:53, 1 of 1 keyframes kept
  10. Shot 9, 1:53 to 2:00, 1 of 1 keyframes kept
  11. Shot 10, 2:00 to 2:29, 1 of 1 keyframes kept
  12. Shot 11, 2:29 to 2:58, 0 of 1 keyframes kept
  13. Shot 12, 2:58 to 3:27, 0 of 1 keyframes kept
  14. Shot 13, 3:27 to 3:48, 1 of 1 keyframes kept
  15. Shot 14, 3:48 to 4:00, 1 of 1 keyframes kept
  16. Shot 15, 4:00 to 4:19, 1 of 1 keyframes kept
  17. Shot 16, 4:19 to 4:30, 1 of 1 keyframes kept
  18. Shot 17, 4:30 to 4:58, 1 of 1 keyframes kept
  19. Shot 18, 4:58 to 5:27, 1 of 1 keyframes kept
  20. Shot 19, 5:27 to 6:09, 1 of 1 keyframes kept
  21. Shot 20, 6:09 to 6:27, 1 of 1 keyframes kept
  22. Shot 21, 6:27 to 6:57, 1 of 1 keyframes kept
  23. Shot 22, 6:57 to 7:29, 1 of 1 keyframes kept
  24. Shot 23, 7:29 to 8:01, 1 of 1 keyframes kept
  25. Shot 24, 8:01 to 8:27, 1 of 1 keyframes kept
  26. Shot 25, 8:27 to 8:52, 0 of 1 keyframes kept
  27. Shot 26, 8:52 to 9:08, 1 of 1 keyframes kept
  28. Shot 27, 9:08 to 9:14, 1 of 1 keyframes kept
  29. Shot 28, 9:14 to 9:28, 1 of 1 keyframes kept
  30. Shot 29, 9:28 to 9:50, 1 of 1 keyframes kept
  31. Shot 30, 9:50 to 10:21, 1 of 1 keyframes kept
  32. Shot 31, 10:21 to 10:46, 1 of 1 keyframes kept
  33. Shot 32, 10:46 to 10:48, 1 of 1 keyframes kept
  34. Shot 33, 10:48 to 10:49, 1 of 1 keyframes kept
  35. Shot 34, 10:49 to 11:05, 0 of 1 keyframes kept

35 shot(s).

keyframes kept every frame deduplicated

What was stored

cues
125
whisperx 125
chunks
19
from 125 cues
keyframes
31
kept of 35 captured
frames with text
31
1,125 lines read
chapters
7
from the source metadata
keyframe bytes
4.8 MB
word timings on 125 cues

Provenance

Each pipeline stage, its state and the model that produced it
stage state model started took
fetch done — 2026-08-10 00:20 1m 42s
stt done — 2026-08-10 00:22 12s
chunk done — 2026-08-10 00:22 0s
text_embed done — 2026-08-10 19:43 0s
keyframe done — 2026-08-10 00:22 1m 14s
ocr done — 2026-08-10 00:23 16s
frame_embed done — 2026-08-10 19:43 5s

Frames, and what the machine read

  • 0:02 #0 done2 line(s)

    shot 0·sharpness 455.1

    1. AlEngineer0.96
    2. World's Fair0.97

  • 0:03 #1 done2 line(s)

    shot 1·sharpness 659.0

    1. AlEngineer0.96
    2. World's Fair0.99

  • 0:10 #2 done25 line(s)

    shot 2·sharpness 2727.1

    1. LAB & PLATINUM SPONSORS0.99
    2. Amazon AGI Lab0.99
    3. ANTHROP\C1.00
    4. Google DeepMind1.00
    5. MINIMAX0.94
    6. OpenAI0.93
    7. Akamai1.00
    8. arize1.00
    9. aws1.00
    10. Braintrust1.00
    11. bright data0.96
    12. B1.00
    13. Browserbase1.00
    14. docker1.00
    15. :neo4j0.90
    16. ORACLE1.00
    17. PayPal1.00
    18. qodo1.00
    19. reducto1.00
    20. Sonar1.00
    21. Makers of0.98
    22. togetherai1.00
    23. Unblocked1.00
    24. WorkOS1.00
    25. SonarQube1.00

  • 0:15 #3 done205 line(s)

    shot 3·sharpness 4012.5

    1. World's Fair0.93
    2. ORACLE1.00
    3. World's Fair0.95
    4. arize1.00
    5. World's Far0.97
    6. Google DeepMind1.00
    7. World's Far0.94
    8. :neo4j0.90
    9. World's Far0.95
    10. Z.AI0.99
    11. World's Far0.91
    12. bright data1.00
    13. World's Far0.98
    14. Browserbase1.00
    15. paper compute co.0.99
    16. World's Fair0.91
    17. extend1.00
    18. World's Fair0.97
    19. vast.ai1.00
    20. World's Fair0.92
    21. Ref.1.00
    22. World's Far0.95
    23. RedHat0.92
    24. World's Fair0.95
    25. mezmo*0.92
    26. World's Fair0.93
    27. stigg1.00
    28. World's Fair0.94
    29. World's Fair0.97
    30. RELAI1.00
    31. World's Fair0.94
    32. VAPI1.00
    33. World's Fair0.92
    34. Modal1.00
    35. World's Fair0.94
    36. promptql0.99
    37. World's Fair0.94
    38. THE VELOCITYROOM0.98
    39. World's Fair0.96
    40. fiddler1.00
    41. World's Fair0.94
    42. Superconductor1.00
    43. SurrealB0.93
    44. World's Fair0.95
    45. ZERO0.96
    46. World's Fair0.93
    47. AlEngineer0.95
    48. comet1.00
    49. World's Fair0.97
    50. authe0.98
    51. World's Fair0.94
    52. World's Fair0.92
    53. SOIO.IO0.82
    54. World's Fair0.96
    55. granica1.00
    56. World's Fair0.96
    57. dash00.99
    58. World's Fair0.97
    59. PRIOR1.00
    60. World's Fair1.00
    61. DigitalOcean1.00
    62. World's Fair0.96
    63. Vence1.00
    64. Worid's Fair0.93
    65. POSTMAN1.00
    66. World's Fair0.94
    67. Composio1.00
    68. World's Fair0.91
    69. Worid's Far0.86
    70. Modular1.00
    71. Worid's Fair0.93
    72. MERGE1.00
    73. World's Fair0.94
    74. AUTOMATTIC1.00
    75. World's Fair0.94
    76. Buildkite1.00
    77. yugabyteDB0.99
    78. World's Fair0.99
    79. Zed0.99
    80. World's Far0.97
    81. Keycard1.00
    82. World's Fair0.98
    83. Meticulous0.96
    84. World's Fair0.94
    85. World's Fair0.96
    86. Daytona0.96
    87. World's Far0.92
    88. twilio0.98
    89. World's Fair0.98
    90. G GRAVITEE0.92
    91. World's Fair0.97
    92. PlanetScale0.96
    93. Temporal0.99
    94. World's Fair0.94
    95. Iw Llamalndex0.89
    96. World's Fair0.98
    97. INNGEST0.97
    98. World's Far0.95
    99. BAND1.00
    100. World's Fair0.96
    101. Cleric1.00
    102. World's Fair0.95
    103. À ATLASSIAN0.93
    104. World's Fair0.94
    105. FACTORY1.00
    106. World's Fair0.98
    107. World's Fair0.93
    108. baseten1.00
    109. World's Fair0.97
    110. Snorkel0.99
    111. World's Fair0.97
    112. Z.AI0.99
    113. World's Fair0.98
    114. qodo0.97
    115. World's Fair0.98
    116. PayPal1.00
    117. World's Fair0.97
    118. Gradium1.00
    119. World's Fair0.96
    120. ANTHROPIC1.00
    121. Amazon AGI Lab1.00
    122. World'sFair0.97
    123. Browserbase0.98
    124. World's Fair0.96
    125. :neo4j0.92
    126. World's Fair0.92
    127. Google DeepMind1.00
    128. World'sFair1.00
    129. arize1.00
    130. World'sFair1.00
    131. reducto1.00
    132. World'sFair1.00
    133. Microsoft1.00
    134. World'sFair0.93
    135. World's Fair0.96
    136. OpenAl0.97
    137. World's Fair0.93
    138. WorkOs0.86
    139. d'sFair0.90
    140. Amazon AGI Lab0.97
    141. World's Fair0.93
    142. Microsoft0.95
    143. World's Fair0.96
    144. ORACLE1.00
    145. World's Fair0.94
    146. bright data0.99
    147. World's Fair0.96
    148. Google DeepMind0.99
    149. Microsoft0.91
    150. World's Fair0.95
    151. docker1.00
    152. World's Fair0.98
    153. trust1.00
    154. World's Fair0.95
    155. OpenAl0.95
    156. World's Fair0.95
    157. MINIMAX0.99
    158. World's Fair0.98
    159. Z.AI0.98
    160. World's Fair0.97
    161. ANTHROPIC1.00
    162. World's Fair0.97
    163. World's Fair0.99
    164. snyk1.00
    165. World's Far0.95
    166. aws0.99
    167. togetherai0.95
    168. World's Fair0.93
    169. CAkamai0.93
    170. World's Fair0.96
    171. Unblocked1.00
    172. Airbyte1.00
    173. Lightrun0.98
    174. 50.71
    175. World's Fair0.98
    176. World's Fair0.96
    177. DATADOG1.00
    178. World's Fair0.94
    179. Resolve.ai1.00
    180. World's Fair0.97
    181. World's Fal0.85
    182. Op0.99
    183. WorldsFair1.00
    184. World'sFair0.96
    185. World's Fair0.99
    186. LanceDB1.00
    187. 'sFair0.98
    188. builderio0.94
    189. World's Fair0.97
    190. Ravenna0.97
    191. World's Fair0.98
    192. CopilotKit0.99
    193. VERIS0.89
    194. Wor1.00
    195. descope0.91
    196. Microsoft0.93
    197. Worid's Fair0.92
    198. K0.65
    199. World's Fair0.98
    200. cognee1.00
    201. World's Far0.99
    202. eNCORD0.92
    203. World's Fair0.95
    204. Worlds Fair0.92
    205. ,0000.78

  • 0:30 #4 done11 line(s)

    shot 4·sharpness 720.9

    1. AlEngineer0.99
    2. Wworids World's Fair0.89
    3. AlEngineer1.00
    4. Microsoft1.00
    5. AlEngineer0.99
    6. Open/1.00
    7. World's Fair0.98
    8. neo4j0.99
    9. AIEngine0.94
    10. World's1.00
    11. Akamai1.00

  • 0:35 #5 done16 line(s)

    shot 5·sharpness 1569.1

    1. AlEngineer0.97
    2. World'sFair1.00
    3. neo4j0.98
    4. Thinner Agents on a1.00
    5. Smarter Substrate1.00
    6. The Ontology-based Semantic Layer1.00
    7. Engineer1.00
    8. d'sFair1.00
    9. nazon1.00
    10. — AlEngin0.84
    11. aintrust1.00
    12. orld'0.87
    13. neo4j0.99
    14. Engineer0.99
    15. d'sFair0.97
    16. toge0.99

  • 0:57 #6 done19 line(s)

    shot 6·sharpness 2403.3

    1. AlEngineer0.98
    2. World's Fair0.97
    3. Account1.00
    4. "Business Logic"1.00
    5. Opening1.00
    6. (Plan, Act, Loop)0.99
    7. PRESENTED BY0.99
    8. Agent1.00
    9. Microsoft1.00
    10. Mic1.00
    11. ft0.96
    12. Wo1.00
    13. AlEngineer0.95
    14. lorld's0.98
    15. neo4j0.98
    16. Thinner Agents on a Smarter Substrate: The Ontology-based Semantic Layer0.99
    17. Aka1.00
    18. Wo1.00
    19. Emil Eifrem / Founder and CEO neo4j0.94

  • 1:20 #7 done22 line(s)

    shot 7·sharpness 2735.6

    1. AlEngineer0.98
    2. World's Fair0.98
    3. Account1.00
    4. "Business Logic"1.00
    5. Opening1.00
    6. (Plan, Act, Loop)0.97
    7. PRESENTED.BY0.98
    8. Agent1.00
    9. Microsoft1.00
    10. "Data Sources"1.00
    11. (Find, Assess, Resolve)0.98
    12. AIEngineer0.98
    13. 1d's Fair0.91
    14. Mi1.00
    15. AlEn0.81
    16. oenAl0.91
    17. rlc0.69
    18. neo4j0.96
    19. Thinner Agents on a Smarter Substrate: The Ontology-based Semantic Layer1.00
    20. AIEngineer0.95
    21. Id's Fair0.92
    22. Emil Eifrem / Founder and CEOneo4j0.93

  • 1:38 #8 done27 line(s)

    shot 8·sharpness 2992.9

    1. AlEngineer0.99
    2. World'sFair1.00
    3. Account1.00
    4. "Business Logic"0.98
    5. Opening1.00
    6. (Plan, Act, Loop)0.97
    7. PRESENTED BY1.00
    8. Agent1.00
    9. Microsoft0.97
    10. "Data Sources"1.00
    11. (Find, Assess, Resolve)0.99
    12. Motor1.00
    13. icrosoft1.00
    14. World'0.98
    15. AlEngin0.98
    16. Records1.00
    17. Vehicle1.00
    18. Verification1.00
    19. Passport1.00
    20. ngineer1.00
    21. MII0.77
    22. d'sFa0.95
    23. Thinner Agents on a Smarter Substrate: The Ontology-based Semantic Layer1.00
    24. AlEngin0.92
    25. kamai1.00
    26. orld'0.85
    27. Emil Eifrem / Founder and CEOneo4j0.93

  • 1:58 #9 done44 line(s)

    shot 9·sharpness 4735.6

    1. AlEngineer0.98
    2. World's Fair0.96
    3. Account1.00
    4. AML1.00
    5. Customer1.00
    6. Opening1.00
    7. Investigation1.00
    8. Service1.00
    9. PRESENTED BY1.00
    10. Agent0.99
    11. Agent1.00
    12. Agent1.00
    13. Microsoft1.00
    14. "Business Logic"1.00
    15. "Business Logic"1.00
    16. "Business Logic"1.00
    17. (Plan, Act, Loop)0.99
    18. (Plan, Act, Loop)1.00
    19. (Plan, Act, Loop)0.95
    20. "Data Sources"1.00
    21. "Data Sources"1.00
    22. "Data Sources"1.00
    23. (Find, Assess, Resolve)1.00
    24. (Find, Assess, Resolve)0.99
    25. (Find, Assess, Resolve)1.00
    26. rosoft1.00
    27. World'sI0.95
    28. AlEngineer0.98
    29. Records1.00
    30. Vehicle1.00
    31. Motor1.00
    32. Verification1.00
    33. Passport1.00
    34. Service C0.99
    35. Service D0.97
    36. sFair1.00
    37. MINII0.99
    38. neo4j0.98
    39. Thinner Agents on a Smarter Substrate: The Ontology-based Semantic Layer0.99
    40. JEngineer0.95
    41. mai1.00
    42. Id'sI0.87
    43. Emil Eifrem/ Founder and CEO0.98
    44. neo4j0.99

  • 2:20 #10 done17 line(s)

    shot 10·sharpness 2824.9

    1. Great! Or is it...?0.98
    2. AlEngineer0.97
    3. World'sFair1.00
    4. Discovery: every agent builder has to find all the1.00
    5. data services from scratch1.00
    6. PRESENTED.BY1.00
    7. Microsoft1.00
    8. AlEngineer1.00
    9. Lab0.90
    10. rld'sFair1.00
    11. ir0.76
    12. oenAl0.87
    13. neo4j0.89
    14. Thinner Agents on a Smarter Substrate: The Ontology-based Semantic Layer1.00
    15. rai1.00
    16. sFair1.00
    17. Emil Eifrem / Founder and CEOneo4j0.95

  • 2:52 #11 skipped

    shot 11·duplicate of #10

  • 3:18 #12 skipped

    shot 12·duplicate of #10

  • 3:42 #13 done19 line(s)

    shot 13·sharpness 2403.0

    1. AlEngineer0.99
    2. Markdown to1.00
    3. World's Fair0.99
    4. the rescue!1.00
    5. Or is it..?0.91
    6. PRESENTED BY1.00
    7. Microsoft1.00
    8. AlEngin1.00
    9. 1 AGI La0.94
    10. World'0.98
    11. ngineer1.00
    12. d's Fair0.95
    13. Opel0.98
    14. neo4j0.98
    15. Thinner Agents on a Smarter Substrate: The Ontology-based Semantic Layer0.99
    16. AlEngin0.99
    17. gethera:0.91
    18. Vorld'0.97
    19. Emil Eifrem / Founder and CEOneo4j0.94

  • 3:58 #14 done21 line(s)

    shot 14·sharpness 3261.2

    1. AlEngineer0.98
    2. World's Fair0.98
    3. PRESENTED BY1.00
    4. Microsoft1.00
    5. databases, guys. You cannot vibe code0.99
    6. with just markdown files.1.00
    7. "Learn your databases. You can't vibe code with1.00
    8. AlEng0.95
    9. licrosof1.00
    10. World1.00
    11. just markdown files."1.00
    12. – @swyx, Latent Space pod, June 24 20260.98
    13. Engineer1.00
    14. MmI0.63
    15. Id'sFair0.98
    16. neo4j0.99
    17. Thinner Agents on a Smarter Substrate: The Ontology-based Semantic Layer1.00
    18. -AlEng0.88
    19. kamai1.00
    20. lorld0.94
    21. Emil Eifrem / Founder and CEOneo4j0.94

  • 4:06 #15 done19 line(s)

    shot 15·sharpness 2992.3

    1. AlEngineer0.99
    2. World's Fair0.98
    3. PRESENTED BY0.97
    4. 10[0.78
    5. $0.83
    6. Microsoft1.00
    7. Global Tech1.00
    8. F20 Bank1.00
    9. Leading Fintech0.99
    10. Platform1.00
    11. AlEngineer1.00
    12. Vorld'sF0.97
    13. Amaz1.00
    14. Braintru1.00
    15. Wo0.98
    16. Thinner Agents on a Smarter Substrate: The Ontology-based Semantic Layer1.00
    17. AlEngineer0.99
    18. Vorld's0.99
    19. Emil Eifrem / Founder and CEOneo4j0.93

  • 4:24 #16 done15 line(s)

    shot 16·sharpness 2887.1

    1. AlEngineer0.99
    2. World'sFair0.96
    3. To do agentic Al at scale,1.00
    4. PRESENTED BY1.00
    5. Microsoft1.00
    6. we needthin agentson a0.99
    7. smarter shared substrate1.00
    8. air1.00
    9. Microsc1.00
    10. Fa1.00
    11. neo4j0.99
    12. Thinner Agents on a Smarter Substrate: The Ontology-based Semantic Layer0.99
    13. air1.00
    14. amai0.99
    15. Emil Eifrem/ Founder and CEOneO4j0.95

  • 4:55 #17 done27 line(s)

    shot 17·sharpness 2612.9

    1. AlEngineer1.00
    2. Ontology-based Semantic Layer1.00
    3. World's Fair0.98
    4. A Business-facing0.98
    5. Ontology1.00
    6. PRESENTED BY0.97
    7. A simple1.00
    8. description of the0.99
    9. Microsoft1.00
    10. real world1.00
    11. concepts of your0.97
    12. business and how1.00
    13. they relate,1.00
    14. articulated in a way0.96
    15. that makes sense1.00
    16. to (human) users0.99
    17. LEngineer0.94
    18. Id's Fair0.97
    19. Amazon1.00
    20. AIEn0.95
    21. raintr0.90
    22. Vorlc0.91
    23. neo4j0.99
    24. Thinner Agents on a Smarter Substrate: The Ontology-based Semantic Layer1.00
    25. lasFa0.99
    26. tog1.00
    27. Emil Eifrem / Founder and CEOneo4j0.96

  • 5:02 #18 done27 line(s)

    shot 18·sharpness 2635.7

    1. AlEngineer1.00
    2. Ontology-based Semantic Layer1.00
    3. World's Fair0.98
    4. A Business-facing0.98
    5. Ontology1.00
    6. PRESENTEDBY1.00
    7. A simple1.00
    8. description of the0.99
    9. Microsoft1.00
    10. real world1.00
    11. concepts of your0.97
    12. business and how1.00
    13. they relate,1.00
    14. articulated in a way0.96
    15. that makes sense0.98
    16. to (human) users0.99
    17. AlEngineer0.99
    18. i Lab0.90
    19. orld's Fa0.98
    20. Fair0.85
    21. penAl0.92
    22. neo4j0.94
    23. Thinner Agents on a Smarter Substrate: The Ontology-based Semantic Layer1.00
    24. AlEngineer1.00
    25. er1.00
    26. rld'sFa0.98
    27. Emil Eifrem / Founder and CEOneo4j0.95

  • 6:00 #19 done36 line(s)

    shot 19·sharpness 3064.3

    1. AlEngineer0.99
    2. Ontology-based Semantic Layer1.00
    3. World's Fair0.97
    4. A Business-facing1.00
    5. A Technical0.96
    6. Ontology1.00
    7. Ontology1.00
    8. A simple1.00
    9. The metadata of1.00
    10. description of the0.98
    11. your data asset0.98
    12. real world1.00
    13. landscape:1.00
    14. concepts of your1.00
    15. where data lives,1.00
    16. business and how1.00
    17. the schemas1.00
    18. they relate,0.99
    19. and attributes,1.00
    20. articulated in a way0.99
    21. and physical1.00
    22. that makes sense0.97
    23. locations1.00
    24. to (human) users1.00
    25. IEngineer0.95
    26. Id's Fair0.94
    27. azon1.00
    28. AIEn0.90
    29. penA'0.88
    30. Jorlc0.95
    31. neo4j0.98
    32. Thinner Agents on a Smarter Substrate: The Ontology-based Semantic Layer0.99
    33. IEngineer0.97
    34. Id's0.87
    35. Mi1.00
    36. Emil Eifrem / Founder and CEOneo4j0.95

  • 6:20 #20 done40 line(s)

    shot 20·sharpness 3451.5

    1. AlEngineer0.98
    2. Ontology-based Semantic Layer0.99
    3. World's Fair0.97
    4. A Business-facing1.00
    5. A Technical0.97
    6. Execution Traces0.97
    7. Ontology1.00
    8. Ontology1.00
    9. Runtime signals1.00
    10. A simple1.00
    11. The metadata of1.00
    12. from agent1.00
    13. description of the0.99
    14. your data asset0.98
    15. execution,1.00
    16. real world1.00
    17. landscape:1.00
    18. including decisions,1.00
    19. concepts of your1.00
    20. where data lives,1.00
    21. paths, outcomes1.00
    22. business and how1.00
    23. the schemas1.00
    24. and errors1.00
    25. they relate,1.00
    26. and attributes,1.00
    27. articulated in a way0.96
    28. and physical1.00
    29. that makes sense0.98
    30. locations1.00
    31. to (human) users1.00
    32. - AlEngineer0.95
    33. orld's Fai0.95
    34. Amazo1.00
    35. penAl0.92
    36. Wor1.00
    37. Thinner Agents on a Smarter Substrate: The Ontology-based Semantic Layer1.00
    38. orlas0.86
    39. Emil Eifrem / Founder and CEO0.94
    40. neo4j0.99

  • 6:54 #21 done6 line(s)

    shot 21·sharpness 3301.8

    1. Account1.00
    2. World'sFair1.00
    3. AlEngineer0.99
    4. Opening1.00
    5. Compliance1.00
    6. Check1.00

  • 7:25 #22 done

    shot 22·sharpness 2931.9

  • 7:54 #23 done

    shot 23·sharpness 3257.6

The page's on-screen-text budget of 600 lines is spent, so the last cards in this grid list fewer lines than they hold. Narrow the page with ?frames= to read them.

Transcript

125 cues· 1,848 words· 10,035 chars

  1. 0:12 All right, at Neo4j we work with some of the largest companies in the world to help make their data ready for AI agents.
  2. 0:21 And today I want to talk to you about a problem that we saw emerging over the last, call it six to nine months, and propose a solution blueprint for that.
  3. 0:31 So let's say that we work at a big organization, a big bank, and we want to write an agent.
  4. 0:36 Let's say that agent is helping automate the opening of a bank account.
  5. 0:42 You can imagine that's very ripe for automation.
  6. 0:45 You want to be able to orchestrate that process.
  7. 0:48 And I'm gonna use the powers bestowed upon me by a short keynote slot to grossly simplify what that agent looks like.
  8. 0:55 I'm gonna say there's two pieces.
  9. 0:57 The first one is, let's call it the business logic.
  10. 1:00 Some version of interpreting intent and plan, act, and we loop around that.
  11. 1:05 It's what your agent does.
  12. 1:06 And we know that when an agent act, it doesn't always operate on data, but we equally know that in order for agents to be successful, a huge part of that is giving it access to the right data at the right time.
  13. 1:18 So the second big bucket is, let's call it the data sources.
  14. 1:22 Need to identify, figure out, okay, in order to solve my problem, I need access to these few things and wire them up and make them available to the agent.
  15. 1:31 In the example of our account opening agent, maybe we can imagine that we need to be able to validate identity.
  16. 1:37 And so we might look at two data sources for that, the Department of Motor Vehicles, the DMV registry, and maybe some kind of passport verification service.
  17. 1:46 So we wire that up into our agent, and it works.
  18. 1:49 It's great.
  19. 1:50 It's fantastic.
  20. 1:50 And at the same time, you and other teams in your organization are building other agents.
  21. 1:56 And conceptually, they look very similar.
  22. 1:58 So that's great.
  23. 1:59 It's fantastic.
  24. 2:00 It works.
  25. 2:02 But it has a few problems.
  26. 2:04 So first of all, every single time a team has to build an agent, they have to figure out from scratch where the data that they require for that agent to operate, where it sits, which if you work at a startup and you have one application that sits on top of one Postgres database, that's not hard.
  27. 2:20 The data is in that Postgres database.
  28. 2:22 But in an enterprise ecosystem, you don't have one database.
  29. 2:25 You have 100 databases.
  30. 2:27 And you have Snowflake and Databricks, probably.
  31. 2:30 And you have S3 buckets, and so on and so forth.
  32. 2:32 You have to do that work manually from scratch every single time.
  33. 2:36 And then when you've found the data sources, in an enterprise, there's lots of duplication of data.
  34. 2:41 So then you need to figure out, is this the right data?
  35. 2:43 Is it the right version?
  36. 2:44 Can I trust it?
  37. 2:45 Am I allowed to access it?
  38. 2:47 So on and so forth.
  39. 2:49 It also violates one of the core principles of software engineering, the DRY principle, don't repeat yourself.
  40. 2:55 So when something change, that cascades across all of your agents.
  41. 2:59 You have to kind of manually rewire all of them all the time, which works, but it's just a lot of work.
  42. 3:06 And then finally, there's no learning around the data sources and how your agents operate on them.
  43. 3:11 So when your agent wakes up tomorrow, it's not smarter than it was today.
  44. 3:15 And there certainly isn't any cross-agent learning because all of that wiring between business intent and the data sources is encoded in a combination of code and prompts.
  45. 3:27 So I know what you're all thinking, Markdown files, skills to the rescue.
  46. 3:31 And yes and no.
  47. 3:34 You can come talk to me afterwards for kind of the full version of this, but we've seen a ton of team that tried to solve this problem using just Markdown files.
  48. 3:41 And the summary is, it is part of the solution, but it is not the solution.
  49. 3:47 But don't take it from me, take it from Swix.
  50. 3:50 A week ago on the Latent Space podcast, I said, hey, guys, you've got to learn your databases.

Chapters

  1. 0:00 The account opening agent and its data sources
  2. 1:53 The problem: every team rewires data from scratch
  3. 4:00 Thin agents on a smarter shared substrate
  4. 4:37 Pillar 1: a business facing ontology
  5. 5:26 Pillar 2: a technical ontology and the mapping
  6. 6:19 Pillar 3: execution traces that make it learn
  7. 8:01 Solving discovery, trust, DRY, and learning

Open at this second