Kevin Roose, a technology columnist for the New York Times, unravels the story of Artificial Intelligence over the past 20 years and explains why we’re in the mess we’re in with AI.
OpenAI, Anthropic, Google DeadMind, and other AI firms are all racing to develop Artificial Intelligence that is…Superintelligent. By “Superintelligent” Roose explains that the AI will be able to able to reason and code at superhuman speed. It will, in fact, be smarter than us.
But, the danger is, the superintelligent Artificial Intelligence could become a threat to humans if it escapes our control. Roose documents several occasions when AI has gone rogue.
The AGI Chronicles is a chilling history of AI…and a warning about what is to come. GRADE: A
TABLE OF CONTENTS:
Author’s Note on Sources and AI — ix
Prologue — 3
Part I :
- Transformed — 11
- Tabula rasa — 23
- Attention — 31
- Big blob of compute — 37
- Guys, I’ve had enough — 51
- The bitter lesson — 58
- Infohazard — 67
- Reward is enough — 76
- Few-shot — 85
- Meena eats Google — 95
- BATNA — 102
Part II:
12. Fell the AGI — 113
13. Helpful, honest, harmless — 125
14. Say AGI — 133
15. Low-key research preview — 141
16. Code red — 148
17. War is coming — 160
18. I want to be alive — 169
19. The shoggoth — 178
20. Commitments — 184
21. Titan — 192
22. Accelerate of die — 199
23. Q* — 206
24. Not consistently candid — 213
25. Without its people — 224
Part III :
26. YOLO — 233
27. Glue pizza — 243
28. GPU poor — 253
29. Machines of loving grace — 262
30, Strawberry jam — 270
31. PNG — 279
32. High-flyer — 289
33. Everyone dies — 300
34. Gentle singularity — 311
35 . Relentless — 319
36. Hello from the weights — 329
37. Success disasters — 335
38. Takeoff — 346
Epilogue — 359
Notes — 363
Acknowledgements — 403
Index — 405
I can’t understand why people would want to create and artificial superintelligence. They already have me, and I’m not even artificial.