vidtheque.

read-only demo

Videos VrpEyglYgeU

In the Land of AI Agents, the Verifiers Are King — Tariq Shaukat, Sonar

index_state ready data_status ok

AI Engineer· published 2026-07-20· 0:18:53· en· 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 0:19, 1 of 1 keyframes kept
  5. Shot 4, 0:19 to 0:28, 1 of 1 keyframes kept
  6. Shot 5, 0:28 to 0:30, 1 of 1 keyframes kept
  7. Shot 6, 0:30 to 0:31, 1 of 1 keyframes kept
  8. Shot 7, 0:31 to 0:34, 1 of 1 keyframes kept
  9. Shot 8, 0:34 to 0:39, 1 of 1 keyframes kept
  10. Shot 9, 0:39 to 0:56, 1 of 1 keyframes kept
  11. Shot 10, 0:56 to 1:05, 1 of 1 keyframes kept
  12. Shot 11, 1:05 to 1:43, 1 of 1 keyframes kept
  13. Shot 12, 1:43 to 2:00, 1 of 1 keyframes kept
  14. Shot 13, 2:00 to 2:08, 1 of 1 keyframes kept
  15. Shot 14, 2:08 to 2:53, 1 of 1 keyframes kept
  16. Shot 15, 2:53 to 3:09, 1 of 1 keyframes kept
  17. Shot 16, 3:09 to 3:41, 1 of 1 keyframes kept
  18. Shot 17, 3:41 to 4:14, 1 of 1 keyframes kept
  19. Shot 18, 4:14 to 4:46, 1 of 1 keyframes kept
  20. Shot 19, 4:46 to 5:12, 1 of 1 keyframes kept
  21. Shot 20, 5:12 to 5:39, 1 of 1 keyframes kept
  22. Shot 21, 5:39 to 6:05, 1 of 1 keyframes kept
  23. Shot 22, 6:05 to 6:35, 1 of 1 keyframes kept
  24. Shot 23, 6:35 to 7:04, 1 of 1 keyframes kept
  25. Shot 24, 7:04 to 7:34, 0 of 1 keyframes kept
  26. Shot 25, 7:34 to 7:45, 1 of 1 keyframes kept
  27. Shot 26, 7:45 to 7:58, 1 of 1 keyframes kept
  28. Shot 27, 7:58 to 8:28, 1 of 1 keyframes kept
  29. Shot 28, 8:28 to 8:56, 1 of 1 keyframes kept
  30. Shot 29, 8:56 to 9:25, 0 of 1 keyframes kept
  31. Shot 30, 9:25 to 9:53, 1 of 1 keyframes kept
  32. Shot 31, 9:53 to 10:22, 1 of 1 keyframes kept
  33. Shot 32, 10:22 to 10:51, 1 of 1 keyframes kept
  34. Shot 33, 10:51 to 11:12, 1 of 1 keyframes kept
  35. Shot 34, 11:12 to 11:47, 1 of 1 keyframes kept
  36. Shot 35, 11:47 to 12:31, 1 of 1 keyframes kept
  37. Shot 36, 12:31 to 12:55, 1 of 1 keyframes kept
  38. Shot 37, 12:55 to 13:13, 1 of 1 keyframes kept
  39. Shot 38, 13:13 to 13:50, 1 of 1 keyframes kept
  40. Shot 39, 13:50 to 14:17, 1 of 1 keyframes kept
  41. Shot 40, 14:17 to 14:43, 1 of 1 keyframes kept
  42. Shot 41, 14:43 to 15:13, 1 of 1 keyframes kept
  43. Shot 42, 15:13 to 15:42, 1 of 1 keyframes kept
  44. Shot 43, 15:42 to 16:11, 1 of 1 keyframes kept
  45. Shot 44, 16:11 to 16:55, 1 of 1 keyframes kept
  46. Shot 45, 16:55 to 17:33, 1 of 1 keyframes kept
  47. Shot 46, 17:33 to 18:14, 1 of 1 keyframes kept
  48. Shot 47, 18:14 to 18:27, 1 of 1 keyframes kept
  49. Shot 48, 18:27 to 18:36, 1 of 1 keyframes kept
  50. Shot 49, 18:36 to 18:52, 0 of 1 keyframes kept

50 shot(s).

keyframes kept every frame deduplicated

What was stored

cues
197
whisperx 197
chunks
32
from 197 cues
keyframes
47
kept of 50 captured
frames with text
47
1,699 lines read
chapters
11
from the source metadata
keyframe bytes
7.8 MB
word timings on 197 cues

Provenance

Each pipeline stage, its state and the model that produced it
stage state model started took
fetch done — 2026-08-10 00:24 1m 48s
stt done — 2026-08-10 00:25 23s
chunk done — 2026-08-10 00:26 0s
text_embed done — 2026-08-10 19:43 1s
keyframe done — 2026-08-10 00:26 2m 11s
ocr done — 2026-08-10 00:28 22s
frame_embed done — 2026-08-10 19:43 8s

Frames, and what the machine read

  • 0:02 #0 done2 line(s)

    shot 0·sharpness 456.7

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

  • 0:03 #1 done2 line(s)

    shot 1·sharpness 663.3

    1. AlEngineer0.95
    2. World's Fair0.99

  • 0:10 #2 done25 line(s)

    shot 2·sharpness 2742.5

    1. LAB & PLATINUM SPONSORS0.99
    2. Amazon AGI Lab0.97
    3. ANTHROP\C1.00
    4. Google DeepMind1.00
    5. MINIMAX0.98
    6. OpenAl0.92
    7. Akamai1.00
    8. arize1.00
    9. aws1.00
    10. Braintrust1.00
    11. bright data0.98
    12. B1.00
    13. Browserbase1.00
    14. docker1.00
    15. neo4j0.98
    16. ORACLE1.00
    17. PayPal1.00
    18. qodo1.00
    19. reducto0.99
    20. Sonar1.00
    21. Makers of0.98
    22. togetherai1.00
    23. Unblocked1.00
    24. WorkOS1.00
    25. SonarQube1.00

  • 0:16 #3 done4 line(s)

    shot 3·sharpness 214.1

    1. Sonar1.00
    2. Makers of1.00
    3. SonarQube1.00
    4. AIE0.97

  • 0:26 #4 done6 line(s)

    shot 4·sharpness 558.9

    1. TARIQ SHAUKAT0.97
    2. CHIEF EXECUTIVE OFFICER1.00
    3. Sonar1.00
    4. Makers of1.00
    5. SonarQube1.00
    6. AIE1.00

  • 0:29 #5 done17 line(s)

    shot 5·sharpness 1526.0

    1. TARIQSHAUKAT0.99
    2. ot0.58
    3. Make0.83
    4. promptql0.85
    5. ZAY0.73
    6. ineer1.00
    7. 'sFair0.99
    8. CHIEF EXECUTIVE OFFICER1.00
    9. Sonar1.00
    10. Ckeric0.85
    11. Makers of0.99
    12. SonarQube1.00
    13. AIE1.00
    14. ZAI0.98
    15. Ravenina0.94
    16. brld's Fai0.75
    17. BeNCORD0.88

  • 0:31 #6 done70 line(s)

    shot 6·sharpness 3154.4

    1. World's Fair0.97
    2. Engineer1.00
    3. World's Fa0.94
    4. air0.96
    5. THE VELOCITY ROOM0.97
    6. beigha data0.68
    7. OpenAl1.00
    8. INNGEST0.95
    9. World'sFair1.00
    10. World's Fair0.97
    11. © BAND0.89
    12. AUTOMATTIC1.00
    13. PRIOR0.98
    14. Z.AI0.98
    15. World's Fair0.96
    16. Keycard0.98
    17. :neo4j0.99
    18. World's Fair0.98
    19. Cleric1.00
    20. Meticdous0.97
    21. World's Fair0.97
    22. qodo0.96
    23. GGRAMITEE0.89
    24. World's Fair0.97
    25. Google DeepMind1.00
    26. World's Fai0.98
    27. A ATLASSIAN0.92
    28. World's Fair0.91
    29. Amazon AGI Lab1.00
    30. World's Fai0.93
    31. World's Fai0.94
    32. arize0.98
    33. PayPal0.99
    34. Gradium0.95
    35. Braintrust1.00
    36. World'sFai1.00
    37. Microsoft1.00
    38. World'sFa0.95
    39. arlucto0.83
    40. ANTHROPC0.96
    41. OpenAl0.96
    42. ORAC1.00
    43. Microsoft0.99
    44. World's Fair0.97
    45. World'sFai0.99
    46. MINIMAX1.00
    47. brighe data0.81
    48. World's Fai0.93
    49. Akamai1.00
    50. World'sFai0.95
    51. World's Fair0.87
    52. ANTHROPIC0.96
    53. World's Fair0.97
    54. DATADOG1.00
    55. World's Fai0.96
    56. Resolve.ai1.00
    57. @Arbyte0.92
    58. DATAD1.00
    59. Optiver0.99
    60. World's Fair0.98
    61. builder.io0.97
    62. Worid'sFai0.93
    63. Ravenna1.00
    64. World'sFa1.00
    65. ilotKit1.00
    66. World'sFair1.00
    67. cognee1.00
    68. World'sFa0.96
    69. BeNCORD0.91
    70. tgrs0.76

  • 0:33 #7 done208 line(s)

    shot 7·sharpness 3124.5

    1. orld's Far0.91
    2. ORACLE1.00
    3. World's Fair0.95
    4. arize1.00
    5. World's Fair0.98
    6. Google DeepMind1.00
    7. World's Fair0.97
    8. :neo4j0.91
    9. World's Fair0.95
    10. Z.AI0.98
    11. World's Fair0.95
    12. bright data1.00
    13. World's Fair0.98
    14. Browserbo0.99
    15. per compute co.0.97
    16. World's Fair0.96
    17. extend1.00
    18. World's Fair0.95
    19. vast.ai1.00
    20. World's Fair0.97
    21. Ref.1.00
    22. World's Fair0.91
    23. Red Hat0.93
    24. World's Far0.98
    25. mezmo+0.91
    26. World's Fair0.98
    27. stigg1.00
    28. World's Fal0.92
    29. orld'sFair0.80
    30. RELAI0.98
    31. World's Fair0.97
    32. VAPI1.00
    33. World's Fair0.99
    34. Modal1.00
    35. World's Far0.98
    36. promptql1.00
    37. World's Fair0.94
    38. THE VELOCITY ROOM0.95
    39. World's Fair0.97
    40. fiddler1.00
    41. World's Fair0.99
    42. Supercondu1.00
    43. Surreal0.99
    44. World's Fair0.98
    45. ZERO0.99
    46. World's Fair0.96
    47. AIEngineer0.96
    48. comet1.00
    49. World's Fair0.98
    50. authe0.99
    51. World's Fal0.94
    52. orld's Fair0.91
    53. SOIOIO0.88
    54. World's Fair0.98
    55. granica1.00
    56. World's Far0.92
    57. dash00.98
    58. World's Fair0.97
    59. PRIOR1.00
    60. World's Fair1.00
    61. igitalOcean1.00
    62. World's Fair0.97
    63. Vence0.98
    64. World's Fair0.96
    65. POSTMAN1.00
    66. World's Fair0.98
    67. Composio1.00
    68. World's Fal0.92
    69. orld's Fair0.94
    70. Modular1.00
    71. World's Fair0.99
    72. MERGE1.00
    73. World's Fair0.94
    74. AUTOMATTIC1.00
    75. World's Fair0.95
    76. Buildki1.00
    77. ugabyteDB1.00
    78. World's Fair0.97
    79. Zed1.00
    80. World's Fair0.96
    81. Keycard1.00
    82. World's Fair0.94
    83. Meticulous1.00
    84. World's Fai0.97
    85. lorld's Fair0.91
    86. Daytona1.00
    87. World's Fair0.95
    88. twilio0.92
    89. World's Fair0.92
    90. GRAVITEE1.00
    91. World's Fair0.96
    92. PlanetSca0.95
    93. Temporal1.00
    94. World's Fair0.98
    95. Llamalndex0.99
    96. World's Fair0.99
    97. INNGEST0.99
    98. World's Fair0.99
    99. BAND1.00
    100. World's Fair0.96
    101. Cleric1.00
    102. World's Far0.92
    103. A ATLASSIAN0.95
    104. World's Fair0.97
    105. FACTORY1.00
    106. World's Fai0.94
    107. orld's Fair0.94
    108. baseten1.00
    109. World's Fair0.99
    110. Snorkel0.91
    111. World's Fair0.97
    112. Z.AI0.99
    113. World's Fair0.97
    114. qodo0.97
    115. World's Fair0.96
    116. PayPal1.00
    117. World's Fair0.98
    118. Gradium1.00
    119. World's Fair0.91
    120. ANTHROP1.00
    121. zon AGI Lab0.99
    122. World's Fair0.98
    123. Browserbase1.00
    124. Word's Fair0.88
    125. :neo4j0.92
    126. World's Fair0.93
    127. Google DeepMind1.00
    128. World's Fair0.95
    129. arize1.00
    130. World's Fair0.93
    131. reducto1.00
    132. World's Fair0.98
    133. Microsoft0.98
    134. World's Fal0.92
    135. orld's Fair0.88
    136. OpenAl0.98
    137. World's Fair0.96
    138. WorkOS0.88
    139. World's Fair0.96
    140. Amazon AGI Lab0.97
    141. World's Fair0.96
    142. Microsoft1.00
    143. World's Fair0.98
    144. ORACLE1.00
    145. World's Fair0.94
    146. bright data1.00
    147. World's Fair0.95
    148. Google Deeph0.98
    149. Microsoft0.96
    150. World's Fair0.95
    151. docker1.00
    152. World's Fair0.99
    153. Braintrust1.00
    154. World's Far0.98
    155. Oper1.00
    156. World's Fair0.99
    157. MINIMAX0.99
    158. World's Fair0.97
    159. Z.AI0.98
    160. World's Fair0.96
    161. ANTHROPIC1.00
    162. World's Fa0.99
    163. orld's Fair0.90
    164. snyk1.00
    165. World's Fair0.98
    166. aws0.99
    167. World's Fair0.96
    168. togetherai0.96
    169. Wor0.97
    170. Akamai1.00
    171. Word's Fair0.90
    172. Unblocked0.97
    173. World0.97
    174. Lightrun1.00
    175. ain0.98
    176. World's Fair0.97
    177. World's Fair0.96
    178. DA0.79
    179. World's Fair0.97
    180. Resolve.ai0.99
    181. World's Far0.92
    182. World's Fa0.94
    183. Op0.96
    184. WorldsFa0.99
    185. orld's Fair0.89
    186. World's Far0.97
    187. LanceDB1.00
    188. World's Fair0.95
    189. builderio0.96
    190. Worl0.97
    191. Ravenna1.00
    192. World's Fair0.98
    193. CopilotKit1.00
    194. VERIS1.00
    195. Wor0.99
    196. escupe0.93
    197. Microsoft1.00
    198. t1.00
    199. Word's Far0.84
    200. TOPK1.00
    201. World's Fair0.98
    202. cr0.64
    203. e0.73
    204. World's Fair0.92
    205. ENCORD0.92
    206. World's Fair0.98
    207. rid's Fa0.83
    208. Worid's Fa0.92

  • 0:38 #8 done11 line(s)

    shot 8·sharpness 346.0

    1. AlEngineer0.96
    2. AI0.94
    3. World's1.00
    4. AlEngir0.96
    5. Fair1.00
    6. World'1.00
    7. Akan0.99
    8. AlFnai0.72
    9. Reso0.98
    10. DOG1.00
    11. Worlu1.00

  • 0:44 #9 done9 line(s)

    shot 9·sharpness 418.6

    1. AlEngineer0.99
    2. OpenAl0.92
    3. World's Fair0.96
    4. AlEngineer1.00
    5. World's Fair0.97
    6. Akamai1.00
    7. AlEngineer0.99
    8. DATADOG1.00
    9. Jorld's Fair0.99

  • 0:57 #10 done75 line(s)

    shot 10·sharpness 4067.5

    1. World's Fair0.96
    2. World's Fal0.97
    3. World's Fair0.99
    4. ted Hat0.88
    5. MERGE1.00
    6. World's Fair0.99
    7. THIE VELOCITY ROOM0.93
    8. mezmo{0.93
    9. ©fiddler0.91
    10. World's Fair0.97
    11. INNGEST1.00
    12. PRIOR1.00
    13. SROK0.63
    14. World's Fair0.95
    15. AUTOMATTIC1.00
    16. World's Fair0.98
    17. ©BAND0.91
    18. Keycard1.00
    19. Bterupx0.59
    20. Worid's Fair0.92
    21. World's Fai0.96
    22. World's Fair0.98
    23. Cleric1.00
    24. GRAVITEE0.91
    25. Meticulous0.97
    26. Kexa0.84
    27. World's Fair0.98
    28. Amazon AGI Lab0.99
    29. World'sFai1.00
    30. World's Fa0.96
    31. qodo0.96
    32. arize1.00
    33. PayPal1.00
    34. À ATLASSIAN0.98
    35. Gradium0.99
    36. FACTORY1.00
    37. PareSode0.59
    38. dnid0.58
    39. diy0.57
    40. orld's Fair0.99
    41. World's Fair0.98
    42. Braintrust1.00
    43. World's Fair0.96
    44. World's Fa0.96
    45. World'sFai1.00
    46. OpenAl0.96
    47. World's Fal0.91
    48. Microsoft0.99
    49. World's Fal0.89
    50. ACLE1.00
    51. reducto1.00
    52. Z.AI0.98
    53. bright data0.99
    54. ANTHROPIC1.00
    55. Microsoft0.99
    56. ANTHROPIC0.99
    57. Google DepMind0.88
    58. Crusce0.85
    59. OpenAl0.91
    60. rld'sFair0.99
    61. World's Fair0.97
    62. @Artbyte0.87
    63. gruptie0.65
    64. Res0.96
    65. CODERI0.97
    66. Optiver0.99
    67. builder.io1.00
    68. Worid'sFa0.94
    69. rlid'sFair0.78
    70. d's Fair0.93
    71. World'sFair1.00
    72. Worid'sFai0.93
    73. World'sFa0.97
    74. Workd's Fai0.91
    75. tigus0.70

  • 1:31 #11 done9 line(s)

    shot 11·sharpness 415.3

    1. AlEngineer0.96
    2. OpenAI0.92
    3. World's Fair1.00
    4. AlEngineer0.99
    5. together.ai0.99
    6. orld's Fair0.98
    7. AlEngineer0.97
    8. World's Fair0.97
    9. DATADOG1.00

  • 1:58 #12 done9 line(s)

    shot 12·sharpness 411.5

    1. AlEngineer0.98
    2. OpenAI0.92
    3. World's Fair0.97
    4. AlEngineer0.99
    5. Wor'sF:0.92
    6. Akamai1.00
    7. AlEngineer0.98
    8. DATADO1.00
    9. World's Fair0.97

  • 2:06 #13 done30 line(s)

    shot 13·sharpness 2950.5

    1. AlEngineer0.98
    2. World'sFair1.00
    3. CNN Business0.93
    4. Markets0.98
    5. Tech Media0.99
    6. Calculators1.00
    7. Videos1.00
    8. • Watch0.89
    9. Listen0.90
    10. Another 'hallucinated' court filing highlights the0.99
    11. difference between Silicon Valley and the rest of0.99
    12. the world0.97
    13. APR 23, 20260.85
    14. SULLIVAN1.00
    15. CROMWELL1.00
    16. 80.89
    17. d's Fair0.91
    18. ngineer1.00
    19. Ope1.00
    20. Sonar e20260.96
    21. - AlEngi0.96
    22. gether0.99
    23. rld1.00
    24. In the Land of Al Agents, the Verifiers Are King0.99
    25. ngineer1.00
    26. d'sFair1.00
    27. DAT1.00
    28. Tariq Shaukat / Chief Executive Officer0.97
    29. Sonar1.00
    30. Makers of0.93

  • 2:26 #14 done18 line(s)

    shot 14·sharpness 2512.9

    1. AlEngineer0.99
    2. World's Fair0.98
    3. Every industry is1.00
    4. struggling with AI slop.1.00
    5. AIEn0.93
    6. enAl0.99
    7. Worlc1.00
    8. Sonar e20260.97
    9. Engineer1.00
    10. Id's0.95
    11. Ak1.00
    12. In the Land of Al Agents, the Verifiers Are King1.00
    13. AIEn0.87
    14. ATADOr0.97
    15. Worlc0.97
    16. Tariq Shaukat / Chief Executive Officer0.99
    17. Sonar1.00
    18. Makers of0.99

  • 2:59 #15 done18 line(s)

    shot 15·sharpness 2383.9

    1. AlEngineer0.99
    2. World's Fair0.96
    3. But, isn't0.99
    4. software development1.00
    5. different?1.00
    6. AIE1.00
    7. penAl0.91
    8. Worl1.00
    9. Sonar e20260.97
    10. AlEngineer0.95
    11. ld'sFai0.93
    12. In the Land of Al Agents, the Verifiers Are King1.00
    13. AIE0.99
    14. ATADOG1.00
    15. Worl1.00
    16. Tariq Shaukat / Chief Executive Officer0.99
    17. Sonar1.00
    18. Makers of0.99

  • 3:34 #16 done49 line(s)

    shot 16·sharpness 3481.9

    1. Coding agents are quickly getting a lot better0.99
    2. AlEngineer0.99
    3. World'sFair1.00
    4. Time horizon of software tasks Al can complete0.97
    5. at 50% accuracy1.00
    6. 50% success rate1.00
    7. Claude Mythos0.98
    8. Preview (early)0.99
    9. 16 hours1.00
    10. Tas u (ns)0.76
    11. Claude Opus 4.61.00
    12. 12 hours1.00
    13. GPT-5.2 (high)1.00
    14. 8 hours1.00
    15. Gemini 3.1 Pro1.00
    16. 6 hours0.99
    17. Claude Opus 4.51.00
    18. GPT-51.00
    19. 4 hours1.00
    20. 030.97
    21. 2 hours1.00
    22. 1 hour1.00
    23. GPT-21.00
    24. GPT-31.00
    25. GPT-3.51.00
    26. GPT-41.00
    27. OpenAI0.93
    28. W0.99
    29. 01.00
    30. 20191.00
    31. 20201.00
    32. 20211.00
    33. 20221.00
    34. 20231.00
    35. 20241.00
    36. 20250.98
    37. 20261.00
    38. 20271.00
    39. )Sonar e20280.88
    40. LLM release date1.00
    41. Source: METR: Task-Compietion Time Horizons of Frontier AI Models0.98
    42. AlEngineer1.00
    43. Vorld'sl0.92
    44. In the Land of Al Agents, the Verifiers Are King1.00
    45. DATAD1.00
    46. W0.99
    47. Tariq Shaukat / Chief Executive Officer0.98
    48. Sonar1.00
    49. Makers of0.99

  • 4:07 #17 done38 line(s)

    shot 17·sharpness 3628.8

    1. Coding agents are quickly getting a lot better1.00
    2. AlEngineer1.00
    3. World'sFair1.00
    4. at 50% success rate0.99
    5. Time horizon of software tasks Al can complete0.98
    6. at 50% accuracy0.98
    7. 50% success rate1.00
    8. PRESENTED BY0.99
    9. Preview (early)1.00
    10. Claude Mythos0.99
    11. 16 hours1.00
    12. Microsoft1.00
    13. Tas u (tns)0.76
    14. Claude Opus 4.60.97
    15. 12 hours1.00
    16. GPT-5.2 (high)1.00
    17. 8 hours1.00
    18. Gemini 3.1 Pro1.00
    19. 6 hours0.99
    20. Claude Opus 4.51.00
    21. GPT-51.00
    22. 4 hours1.00
    23. 030.98
    24. 2 hours1.00
    25. 1 hour1.00
    26. GPT-21.00
    27. GPT-31.00
    28. GPT-3.51.00
    29. GPT-41.00
    30. :air0.90
    31. UMINIM0.97
    32. 01.00
    33. 20191.00
    34. 20201.00
    35. 20211.00
    36. 20221.00
    37. 20231.00
    38. 20241.00

  • 4:42 #18 done

    shot 18·sharpness 3752.5

  • 4:57 #19 done

    shot 19·sharpness 3807.1

  • 5:18 #20 done

    shot 20·sharpness 3809.8

  • 5:52 #21 done

    shot 21·sharpness 3843.8

  • 6:15 #22 done

    shot 22·sharpness 3526.5

  • 6:52 #23 done

    shot 23·sharpness 3509.0

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

197 cues· 2,917 words· 16,153 chars

  1. 0:12 Please join me in welcoming the Chief Executive Officer at Sonar, Tariq Shawkat.
  2. 0:34 Morning, everyone.
  3. 0:36 Did you enjoy that last talk?
  4. 0:37 That was amazing.
  5. 0:39 You particularly loved the end, the being unreasonable part.
  6. 0:42 I thought that was awesome.
  7. 0:44 I also wanna just, I'm trying to calculate the odds of Tarek following Tarek as the first two sessions in the morning.
  8. 0:51 I think the odds are pretty low on this one.
  9. 0:54 But thrilled to be here today.
  10. 0:56 As we just mentioned, I am with Sonar.
  11. 0:59 We are in the code verification space, and I'm here today to talk about verification.
  12. 1:04 And I think we're all here in large part because we believe to some extent that AGI is here, it's coming.
  13. 1:12 The models we just heard about, Fable, it's really incredible what is going on in the world today.
  14. 1:18 And yet we work almost exclusively with enterprises around the world and the conversation that we have more is the question mark version.
  15. 1:27 Is AGI here and why are they asking these questions?
  16. 1:31 It's because you can,
  17. 1:33 read the news every day, and I'm not trying to name and shame here, but if you look at KPMG putting out reports that they have to retract because of hallucinations, EY doing the same thing, law firms getting into lots and lots of trouble because of made up citations, made up case law, things like this.
  18. 1:56 I think we can really start to question, how do we get value out of AI?
  19. 2:01 The models are amazing, as we just heard, but the hard part, as the other Tarek just said, is getting value out of it.
  20. 2:09 The struggle is that AI slop is everywhere.
  21. 2:14 I'm sure you all see this inside of your organizations.
  22. 2:16 I'm sure you see this in your everyday life, that AI is amazing.
  23. 2:21 The models are incredible at generating very plausible output.
  24. 2:25 They're incredible at generating things that sound correct, but are they correct?
  25. 2:30 And how do you know that they're correct is a big problem.
  26. 2:33 And it's a big problem in professional services, as we saw.
  27. 2:36 It's a big problem in legal.
  28. 2:38 But really, I think if we're honest, it's a big problem in every sector, in every field, whether it's marketing or finance or you name it.
  29. 2:46 You have this question of how do you actually know if it's true?
  30. 2:50 How do you know if it's good or if it is slop?
  31. 2:53 And the question that we deal in the coding space in particular, we deal with software development.
  32. 3:00 And the question we get as we talk to I'm sure many of the people here in the room and a lot of our customers is isn't software development different?
  33. 3:09 And we can look at the data on this and the mythos models.
  34. 3:15 This is data from meter.
  35. 3:18 You may have seen this METR.
  36. 3:20 The coding agents are getting better very quickly.
  37. 3:23 They're getting a lot better very quickly.
  38. 3:25 And you can see the progression, the exponential curve here.
  39. 3:28 What this shows on this chart is how capable are the models at completing tasks
  40. 3:34 that humans would take.
  41. 3:36 So can they complete a task that takes one hour, two hours, whatever it is?
  42. 3:40 The latest mythos model, at least per the benchmarking, which was done a month or so ago in the preview mode, was you're getting to 16 to 18 hours.
  43. 3:49 So the agents are able to complete long-running tasks.
  44. 3:54 And it really is starting to transform how work is happening.
  45. 3:59 But the critical caveat when you read the data is this is at a 50% success rate.
  46. 4:05 Okay, so it is again able to complete tasks, but is it able to complete tasks correctly is the question.
  47. 4:13 So if you start looking at, all right, let's dial up.
  48. 4:16 the accuracy rate, you dial it up to 80%, and there's still progress, but it is much slower progress.
  49. 4:22 Instead of 18 hours, you're at about three and a half hours or something along these lines.
  50. 4:27 And by the way, this is still at 80% accuracy.

Chapters

  1. 0:00 Introduction and the current state of AI adoption
  2. 1:30 The challenge: Distinguishing AI utility from "AI slop"
  3. 3:09 Analyzing the performance data of AI coding agents
  4. 6:17 The productivity paradox: Why gains dissipate after three months
  5. 8:28 Introducing the AC/DC (Agent-Centric Development Cycle) framework
  6. 9:31 Stage 1: Guide (Providing context and constraints)
  7. 11:22 Stage 2: Verify (Zero-trust, multi-layered verification)
  8. 13:00 Stage 3: Solve (Maintenance loops and technical debt control)
  9. 14:32 The necessity of systems-level thinking for AI agents
  10. 16:56 Real-world impact: 92% reduction in issues with disciplined verification
  11. 17:42 Conclusion and final thoughts on enterprise AI

Open at this second