  .    ,   
 


  ,    ,  ,   .

   ?   ,    ?   ,      ?

        :  ,  , ,    .       ,     -    :  ,  ,   .

  ,       .  ,      ,     .





 

  .    ,   









      .   ,                 ,   ,     .   ,        :                     (AGI),               .       ,     .

,  ,   -  .       ,            ,       . ,       ,     ,      ,   .           ,  ,  ,               S-   .

 2016 ,  AlphaGo,   DeepMind,    Google,       -,    .              . ,  ,   ,  ,    .       , ,     .       :  -      ,        .  AlphaGo   ,  ,          .          ,        .       :      ,        .           ,    .

       (Large Language Models, LLM),    GPT,        .    AlphaGo,     :              .      .           ,             .

        ,       ,    ,          . ,   ,         ,   ,        :         ,  ,   ,      ,         .   ,  ,      ,       ,    ,  ,           .               ,       .

   ,      :                 .   ,            (Artificial Intelligence Index Report),    (Stanford University),   ,        , ,          ,                   ( , 2024).    ,        ,    .   ,        :      ,      .

                .   ,      ,         ,    ,  ,   ,   ,   .         ,   ,     .       , ,      ,   ,  ,    ,    ,       .     .




1.  


     ,           ,       .       (Gradual Leap). ,  ,           :           ,        ,  ,    .      ,            S- .

               . ,  ,  S- :     ,          ;     ,         ,      ; , ,   ,   ,      .

  ,     ,       S-              .  ,         ,      -            ,     .     ,   , ,    S- ,   ,           : ,  ,   . . ,  S-      (    )   S-      (  Transformer)   S-      ,         ,     ,       .  ,       ,           .

          (Emergent Abilities)    .  2022    Google   ,       :    (,  )   ,        (      )      ; ,       ,     ,          ,   ,  ,   ,        (Wei et al., 2022).

        ;         .         :            ,     .     :     1  99       ,        100 ,       .       ,         .         :  ,  ,           ,      .  ,            XX          - .

      ,    -      .           ,   ,     .  , ,  ,                ,     .    .

     ,    XIX      ,  ,         ,    .        ,       -    ,      .                    ,         (David, 1990).       :       -     ,                   ,   .

    -      ,   ,             .     ,    ,      ,   ,     ,        .          ,      ,    .

             .           ,       ,        ,               .  , ,     ,    .     ,     ,      ,        .

      ,    .        ,      ,   ,  -      .           .      ,   2004        (DARPA)      ,            L5.    :            ;                      .

    ,        ,  ,     .      ,           .                ,       ,        .




2.   


  ,        ,       ,      .    ,            .    (  , ,   )    ,          ,         ,         ,    .          ,     ,        .     ,  ,    ,           ,       .            .

  (Everett Rogers)       (Diffusion of Innovations)    .  ,   ,       ,     ,    (Innovators)    (Early Adopters)   ,         (Early Majority),   (Late Majority)   (Laggards) (Rogers, 2003).        .      ,     ,         (         ).            ,              ,   -,   .  , ,       ,         (,         NPC  -),       .    ,  ,    ,    -       .

,              ,    . -,       .              -     ,    : ,  ,         .             ,        .

-,       .          ,  ,       ,    ,      .    ,  ,    ,    ,             . ,       .                     .    ,       ,    ,     .          ,      ,    .

   ,  ,      (General-Purpose Technology, GPT),       ,                .       .     ,   AlphaFold2    ,  DeepMind,        .             .

AlphaFold2      .               ,          ,         .  ,     ,           (       ),            .   ,     ,    :         ,     .

            :





    (: NPC  ,   )

    (:   ,   )



 



 



 

  

 



  

 

 



 

   

,  



  







 

 /

 / 



         .        ,    .    (FinTech)    .          (  , ,   )      ,           ,     .

 ,      ,       ,      -  ,   .            .        : - (Robo-advisors),              ,      ;          ,   ,      ,    ,      ;             -   .

 ,           ,            ,         .     ,     :  ,       ,    :      ,   ?.




3.     


  ,   ,          ,        .      ;          :    ,       ,                    .

   (Nick Bostrom)           ,          (Bostrom, 2014),    ,    ,      .    (Yuval Harari) ,                   ,       .         ,      .                  .

         ,    ,   .      ,        .    ,        :      ,  ,    ,    ,          ;      ,                   ,       .

                ,  -    .     . ,     ,          .     ,   , ,  ,        .    , ,   ,       .  ,      ,    ,    ,  , ,       .

          :           ,       .       ,                ,    ,  ,      .

           ,   ,   ,           (, ,    ).            .              .             (Principles).     ,          .

    ,     ,       ,                .   ,  ,     .   .           ,     ,   ,   ,       .           : -,      ; -,         ; , ,           .        ,             ,          ,   .       ,    ,          ,     ,  .




   


         ,   ,       .      ,       ,      .     ,           .  ,       10      .

 1:        ,     . (Judge the stage of a technology by its long-term performance curve rather than by isolated, single events.)

        ,     (   AlphaGo   -  )       ,          (,            ).        ;          ,   ,  ,   ,        .

          .  ,    ,          ,      .          ,    .    ,        ,    (  AI Index  , AI Progress Metrics     . .).    ,          ?,   :          ?        ?.           ,         .

 2:       ,   . (Treat the intuition of its still early as a hypothesis that needs to be tested.)

     ,     :  ,   ,   -   ,  ,      ,      .   ,         (Its still early)   ,     ,   .        ,          ,            .

         ,     ,   . ,     ,        .              .    ,          ,      :  ,  [- ]      [  ]   [  ].     [-  ],      .     [-  ],    .

 3: ,        ,     . (Assume that major changes emerge first in edge cases rather than in the core.)

   ,            ,     ,     ,           .             .               ,     .           .   ,           ,          .        .        , ,         ,        ,          .    :     ,       100 ?  ,         ?

 4:  ,  ,   ,     . (Evaluate attention-grabbing events separately from facts that actually reshape power structures.)

       :     (Events),         ; ,  ,     ;     (Facts),          ,   ; ,  ,     .          ,    ,             . ,          ,  ,   -             ,   ,            ,   .    ,      -:  ,    ,   ,    :

1)    ,  ,  ,    ?

2)      -   ,     ?

 5:     ,    . (Externalize uncertainty into explicit hypotheses, probabilities, and trigger conditions.)

    ;     ,    .             ,  ,     ,  :   [],  []    X %,     [ -],     [  ].         .          .

  ,     (   ?),    ,         (   ?     ?    ?).   ,         .    ,   ,     (:              ?),       ,  ,     ,         ,            -.




 :    ,     . (Establish a response framework first, then form a position.)


      (       ),           .            ,    (Framework),             (Position).         ,         ,    .

     :      ,          .            ,      .    ,          AIGC?,     ,      ,     ,   ,   ,  ,  ,  ,     .           ,   ,     .

 7:          . (Increase speed for reversible decisions, and increase the rigor of validation for irreversible ones.)

 Amazon   (Jeff Bezos)       (Type 1 Door)     (Type 2 Door).     : ,   ,  ,    ,        ;     .    :      ,    .         .             ,      .

        ,       .          ,         ,               .         :         ?.     ,      ,              ,  .     ,     ,     ,       (            -),     ,  ,  :   ,   ,   ,   .




 8:      . (Prioritize the preservation of optionality.)


 ,  ,     ,     .              :           ?      ,       .         ,    ,         .

        , -       .     ,          .          .              . ,       ()            ,  ,     .        ,       ,          ,    .

 :    ,      . (When placing bets on the future, control leverage and exposure.)

     ,   ,   ,     (Leverage)    (Exposure).        (  , ),       ;      ,       .      .     ,  ,                   .

   ,             .        ,                .    ,                    (,   10 %     ).       -,            .         10  ,        ,       1 ,      100  .

 10:            . (Treat I already understand as a trigger for review and reassessment.)

     ,  -      -    ,     ,   .            ,      .   ,  ,      ,     ,        .  ,  ,             .        .               .    ,     ,  :  ,    .   ,      -   (,     ),     :

1)  ,    ,     ;

2)       ,         ,     ;

3)      ,      .            .




    


     ,  ,   ,        .      ,     ,   ,     .     ,      ;     ,         ;     ,          ,      .                     ,       .

 ,                 .   S-   ,      ,    -               .  ,                    .                    ,   10  .               ,  ,       ,           .

     ,      :             ,     ,       ,       ,       ,     ,       .    ,        .

     ,             ,             .  ,     ,         ,      ,             .         :  ,             ,      ,    ,      .

     ,    .     ,    ,    ,  ,         .            .





: 



















/ Introduction: Turn the Future into a Manageable Problem




  


  2016  AlphaGo,   DeepMind,    Google,        -   4:1,    .        ,   : ,       ,    .        ()    .   ,      ,       ,    ,       ,  .    ,                 - ;          ,          .

           .       :     ,  :  ,  ,          ,      .       :           ,          .       : ,      , ,     ,  ( )       .

     ,        ,        ,             . ,        ,  ,  ,            . ,              . ,     ,       .

      :          ,    ,       ,  .        ,        ,      ,         ,    ,     ,  ,     ,   ,  ,  ,   ,      .     ,   ,       ,      ,      .

          .     ,           . ,    ,           .  ,                       ; ,  ,      :  ,      ; ,   ,   ,          ,              .

      .    ,          .              ,          .            ,        . ,         ,   .            .

      ,    ,     . ,    ,   ,         ,   .          .




1.  :    


        ,  ,             .        Ȼ,     ,  ,      Nasdaq,  ,      ,      .                 .  ,         , ,     ,               .     : , , /,   .                   .




 :    


           .         .    ,  ,         ,          ,                .          ,          ,       .   ,             . ,   Boston Consulting Group     ,  ,  GPT-4   ,      ,     40 % ,   ,    (DellAcqua  ., 2023).

      . -,         .          ,      ,           .  ,    GPS           ,         . -,       . -,         ,    ,    ,        .  ,         ,   ,           .

 ,               ,    .    ,       Ȼ (AI Literacy),        ,       ,             .          ,        ,    .




  :      


      ,    . .     ,      ,       .       .   -                   -,     -  .   McKinsey Global Institute ,          2,6  4,4    (Chui, Hazan  ., 2023).  ,              ,    .

     ,    .      .    ,  ,      .        ,     ,  .-,   .   - (      )       ,           .  ,        . ,  .             .      ,        ,     .           ,    .

            .     -,   ,          .    :     ,      ,  ,     ,          .      Ȼ (Responsible AI)      ,   ,   .




   :         


     ,  ,                 .           ,    ,   ,          ,         .   ,  ,      ,           (Kitsing, 2020).

              . -,     .  ,     ,         (,            ),         ,      ,     . -,      . ,             ,     .                 ,       .              ,        (Liang  ., 2018).

 ,       . ,       ,    ?  ,    ?       ,      ?            .                .           (      ),     (       ).   ,       ,   ,      ,    ,     ,   ,           .




 :        


  ,     (   , ,  )     . ,   ,    ,       . ,     ,      ,    .      ,     ,    XXI .      : ,      ,   .

      .           .         (LAWS)         ,          ,     .  ,       .         ,     ,      ,                 .     - ,                  (Geist & Lohn, 2018).

 ,              ,  .   ,   .    ,    ,                    .        .   :           ,         ,    ,     ,       (),            .        , ,   ,      -      ,     .




 :     


,              :        ?              .        , ,   .     ,   ,     ?

             (Bostrom, 2014).                   .     ,       Homo Deus (, 2017),        .

,  ,          .       ,      .    ,       ,   ,     ?           ,    ,    ?         ,   ,     .

       ,    .            ,             .      ,  ,    .    ,  ,            .      (Value Alignment)          .

     ,  ,        ,   ,   ,     .    ,     ,     ,                 .           ,            .




2.  :     


                 .             Ȼ,     ,        .   ,          ,       .           .      : ,       , ,        ,  ,        , , .      ,   ,    ,   :   (Misuse),   (Goal Misalignment)     (Systemic Cascade).                  .




 1:   (Misuse)   ,   


              .          .        ,    .  ,                      .     , ,              .

          .              .     (Deepfakes)     ,      ,            .  Cambridge Analytica,        2016       Facebook     ,       ,    (Cadwalladr & Graham-Harrison, 2018).

                  .-,        ,            .                     ,        .          .   ,           ,           .

       .     ,   .         ,         (      ,     )    ,             .        Ȼ (Defensive AI), ,  ,    ,  ,          ,   .     ,         ,           ,       .          -,   ,    .          ,       .

  2:   (Goal Misalignment)       

      ,    Ȼ,       ,     .  ,    ,                ,        .    ,         (AI Safety),    ,      .     ,               ,       .

  ,   ,    (Paperclip Maximizer),    . ,     ,  ,  :     .      ,     ,     ,      . ,    ,             .             ,    .    ,      ( )     (           )   .

         .                  .      ,   ,         ,      .              (Haidt & Rose-Stockwell, 2019).   :             .     ,      ,     ,     .     ,     .

           ,         .            .    ,   - .          (AI Alignment).       :

 (Interpretability):      ,        ,        .     ,      ,     .

 (Controllability):   ,    ,     ,          ,     .

  (Value Learning):                   ,      .      .

        ,     .      , , ,    ,        - .

  3:   (Systemic Cascade)      

                 ,      ,     -.                   ,  ,      ,         .       2008 ,       -      ,    ,  ,      ,    ,        .

              ,  ,   ,            .  ,      ,     ,            .        ,        , -      .   ,    ,       ,            .       , ,    ,        .

   (Flash Crash),  6  2010 ,      .     ,        ,     -   1000         1   ,     , ,       (  ., 2017).            ,          .

           (System- -Resilience).   ,               .             ,   ,     (fail gracefully),   ,              .

   :

   (Circuit Breakers):         ,    (,    )      ,      .

    :        . ,                ,           ,    .

  -:         ;        ,    ,      .

    (Meaningful Human Control):              .              ,   .

      ,      ,          ,    .              .




3. :         


       ,   ,      .             ,           .             ,           .  ,    ,      ;      ,             ,   ,    .     ,        .                  ,    .    ,    : -,            ; -,              .

 :         

       ,   ,   :   , , ?                ,    .     --- (Facts-Mechanisms-Values-Actions, FMVA).     ,         ,        ,     .

  (Facts):     ,      .       . ,   ,   ,           ,    ,    ,          ?

  (Mechanisms):     ,   .     -    .       ,     ,    ,       . ,          -   ?       ?           ?

  (Values):     ,     .     ,    .        ,    ,       . ,         ?    :       ?   :     ?

  (Actions):        ,    .         .        ,   ,        ,              . ,             ,   ,     ?        ?

      2023 ,   ,          ,        FMVA.          ?

 : , ,   ,        ,  GPT-4. , , ,         AGI,      ,   -   ,       .

 :         -    - .       - ,             .   ,     ,      ,   , ,    ,   ,     .

 : , ,       .     (Precautionary Principle)  ,      ,     ,     . ,   , ,      ,      ,  ,          ,     .

  :         .  ,              .     ,    ,           ,         ?      ,        .

          .  ,                ,        :           ;             ;           ;             .     FMVA   ,      ,   .

  :  ,   ,    

           ,           ,    ,   ,   .  ,    (   ,    ),   ,      ,   .     ,           ,         .      - - (Trial-Feedback-Revision, TFR).

            .  ,      ,                ,     ,       .      :

   (Define Actionable Checklist):       ,     .        .  Ȼ   ,    .       : ,      ,    ,    ,     ,   ,    ,  ,    ,   ,   .

   (Trial):            .            ,               .        ,       .

   (Feedback):        (Metrics)          .         (,  ),      (,              ).             ,    ,           (Red Teaming).

   (Analysis & Revision):              .    ,      .            ,     ,      ,  ,    ,       .       TFR.

           TFR  .       (NHTSA)         ,     . ,    .            ,        .              ,    .                (Disengagement Reports)     ( ).        ()              ().      ,  ,         .

  FMVA   TFR,         . FMVA             ,      , , ,        TFR     .   ,  TFR ,       ,               ,   ,    .        :      ,       .




   


                ,         .       ,               .                  .  ,        ,         ,      .        ,             .

 1:    : , ,     . (Decompose problems across layers: the individual, the organization, governance systems, and the global system.)

  :   ,     ,   ,      ,  ,          , , /,        .

  :     ( ,  , , )   ,   .      . ,  ,    -  ,     ,       .              .

 :        ,   :          ?          ?        ? ,                   ,     ,    ,         .

 2:    , ,   . (Break disagreements down into facts, mechanisms, values, and actions.)

  :       .    ,    ,        : ,  ( - ),  ()   ( ).

  :      .         .    ,      ;   ,        ;   ,        .      -      .

 :         .  - :   XXX,   :   -  ,     ,    - ,    ,    ,    ,      ?      ,             .

 3:     ,      . (Classify risks into misuse, goal misalignment, and systemic cascades.)

  :        ,   ,          :  ,    ;  ,  -      ;    ,     .

  :       .       ,      ,    ,    ;    ,      ,       .           .

 :                    . ,               ,   ,      (,  )            ,   ,             .




 4:     . (Translate discussions into concrete decision outputs.)


  :   ,             ,    .        ,    ,           .

  :            .        ,     ,        .                 .

 :         :        ?     10   :      ?   ?   ,     ?    ?   ?      .

 5:      ,    . (Replace verbal correctness with a closed loop of experimentation, feedback, and adjustment.)

  :               ,    .   ,          ,               ,        ,      .

  :         .         .           ,    ,  ,           .

 :                    . ׸   ,       . ,              ,          ,        ,  ,       ,     ,        .

 6: ׸   : , , ,   . (Explicitly define the key variables: speed, scale, verifiability, accountability, and reversibility.




)


  :             -:    ?    ?       ?         ?       ?

  :         . ,    ,   ,      ,         (, ,    ),    ,    ,    ,        .

 :           .       .,           ,        ,      (   ,      ).   ,     ,          (   ),        .

 7: ,       ,        . (Assume systems will be misused, and treat constraints as design inputs.)

  :    .  ,             ,      ,    ,          ,      .          ,   ,         .

  :     ,   ,     .             .            .

 :         Red Teaming (  )    (Threat Modeling).  ,     ,         :         ,     ?.         ,       .




 8:       (Manage uncertainty through thresholds).


  :          .        (Thresholds).    ,     ,     :     100   ,      .

  :          ,  .             .              .

 :           . ,                Ȼ.                (,    80 %  ),      ,                     ,       .

 9:        ,         . (Define the loss ceiling by the worst-case scenario, and the investment ceiling by the best-case scenario.)

  :     .             : -,   ,        ?      ?  ,   ,    . -,   ,          ?      ,       ?

  :     ,    ,      .     ,     ,           (,    ).             ,   ,    ,  .

 :            (pre-mortem). ,        ,    :     ?         .       (pre-parade),   ,      ,  ,       .




 10:   ,   . (When the evidence changes, conclusions must be updated.)


  :        .            ,    ,     .   ,  ,    ,      ,        .

  :        .      ,      .                ,        .

 :      .            : ,     ,    ,     .   ,  ,    ,   .      ,       ,  ,     .

             .      ,    ,  ,         ,   ,          ,   .       .

-,          : , , /,   ,         .               .

-,               ,    :  ,      .                   ,        .

,    ,   ,      .                       .                ,      ,     .

          ,    .          ,    ,   ,           .    ,  ,   ,    .           .                ,    .




 


            ,      :

  :     -.    -,       (,  ,    -),          .                  .

  :                 Ȼ.          ,           .           Ȼ   .

      :                 ,       ,   .           ,         .

  :   ,               ,    - .       ,         .  ,        ,        ,   .




 1.     ,   / Chapter 1 Superintelligence Is Not a Product, Its a Force





  


     .   2023   OpenAI   GPT-4,           (Artificial General Intelligence, AGI)    ,    .    , ,                 .         .    ,  ,     ,      .  , , ,      .   ,           ,  ,    ,     .

          .  , ,     ,        ;       ,  ,     .         ,       ,       .               ,   ,          .        ,     ,       ,   ,        .     ,     ,   ,        .      ,     ,   ,       .

,    ,   ,         .    ,     ,   Google, Microsoft, Meta  ,          ,      .    ,   ,   ,            .    Noema,         ,     (OSullivan, 2025).        AGI,        ,     ,    .            ,        ,    ,          .

         (Algorithmic Governance).                               .         .       .

,            ,    ,      ,        -,    ,       .    ,          ,    ,     ,    .        ,                .     ,     ,    .     ,   ,   ,    .

              ,     ,     .          ,    ( ),          ,      ( ).       :       ;          ;              .   ,     ,          .




1.      ,      


     -      ,            .     ,      .     ,     .                                .        .

       (Generativity).    ,    -  ,   ,      ,      .    ()        .     ,                   .  ,   ,      .

       :             ,           ,    ,      ,  Google, Amazon  Facebook,     . ,      (AGI),      ,     .AGI,      ,     ,  ,    , ,      ,    .    ,      .      ,       ,      ,      ,     .

     (The Frontier of Possibility).              . ,             ,       ,       .      .         ,      .     ,       .

        .      ,    ,   ,     ,         ,  ,      .      AlphaFold 2  DeepMind, ,    ,        ,      50 .     ,      ,       (DAO),         .    ,          .     ,    ,          .

     (Systemic Emergence).       ,  ,        .       ,     ,       .           -  ,         .            .

 ,    -,    ,       ,   -,  ,   (,   ),         ,         .  ,              ,             .        ,        ,     .      ,    ,    .




2.     ,    


    :         ,   ,   .       , ,      ,        :    .       .    ,    ,     .         :    ,    ,    .

             .       (Automation),    (Autonomization).             ,         -    . ,        ,      ,     -.     ,                .

        .      (Algorithmic Trading)   .  ,      70 %      (Investopedia, 2023).      ,             , ,        .  (Flash Crash) 6  2010 ,   -    1000    ,        ,           .

    -                .   ,            (  )                  ,        .          ,        .      :     ,          .

        ,        .       ,      .               ,   Google, Microsoft, Amazon, Meta  Apple.           ,  ,   ,       .

                  .            ,           . ,            (JEDI)      10       .            .

    ,     ,               .       XXI       .         ,    .      ,    ,        .      (Shoshana Zuboff)       (The Age of Surveillance Capitalism) (Zuboff, 2019),                   .    ,               .                   .

         .     ,    ,              ,   .          ,   .              .  , ,    ,       ,        ,        .

          .      2016   ProPublica ,   COMPAS,       ,        ,        (Angwin et al., 2016). ,        Northpointe,  ,           .        ,   ,            ,   .  ,               ,       .       (Algorithmic Hegemony),     ,         ,           .




3.         


         .  , ,        .                  ,   ,     .           ,    .                      .

       .                   ,       .          .  ,      ,       .        ,        .

       ,         .      -  , , ,   ,   .         .     ,       ,         .    ,       .       ,         ,       .          ,    ,          .

          . ,   ,      ,       ,  , ,      .    ,   47 %            1020  (Frey & Osborne, 2017).           ,     , ,  .

       ,        ,         .          ,      .       ,      ,        ,    ?          ,     ,              ,     .      ,            .    ,      (Universal Basic Income, UBI),        ,    ,    ,       XXI .

            .       ,            .   , ,  ,            ,   .  -         ,      ,      .

          .      , , , ,   ,         (Lethal Autonomous Weapons Systems, LAWS).                .  ,    ,   ,     .          , ,    . , ,        (Techno-blocs),        , ,           .       ,       ,       ,      .    ,     ,       ,  ,          ,           .




   


                  .          ,  ,    ,      ,     .




 1:      (Manage Superintelligence as a Structural Force)


 :          .    ,       ,    ,     ,   .                   ,     .   :        ,     ,        .

     ,       ,           ,    .     ,         ,        .  :        ,           . ,       ,                   .     ,   ,     ,         .




 2:            (Prioritize Assessing Its Impact on Power and Distribution)


 :                ?,      :     ?         ?.               .   :     ,    ,             ,           .

           .  :                  (Power and Distribution Impact Assessment).         .            .  ,      ,    Ȼ    ,   ,    ,       ,   .




 3:           (Assess First-Order Effects and Second-Order Cascading Effects Simultaneously)


 :           , ,   .       ,    , ,      ,  ,   ,     .   ,                .   :        .

        ,      ,    .        ,  ,     .  :   - ,    . .         .   Red Teaming (  ),              . ,  ,             .




 4: ,            (Assume Diffusion Speed Will Outpace Institutional Response Speed in the Long Run)


 :   ,  ,   .         (,  ,  )    .      :        .      ,           .   :     ,    ,    .

 ,    , , ,      ,     .            .  :     (Hard Law, ,   )    (Soft Law, ,  ,  ,  )     (Code as Law,         ).    Ȼ (Governable AI),        ,  , ,    .    (Regulatory Sandbox),            .




 5:     (Define Safety by Boundaries)


 :    ,      ,        .       ,  .       ,         .        (,    ),    (,         ).   :            .

         ,          .  ,   ,               .  :     Ȼ (AI Guardianship),      ,         ,  ,      .          (Human-in-the-loop)     (Human-on-the-loop),  ,     ,    ,         .       (      )      ,     .




 6:         (Create Dependency Maps to Identify Single Points of Failure)


 :               ,         .          (  ,  ,  )       ,       (Single- Points of Failure).   :   - (,      )              .      -     .  :                    -.            ,           .            ,          ,   .




 7:   ,        (Treat Rollback, Downgrade, and Switchover Capabilities as Infrastructure)


 :   , ,      ,    -      .         ,   ,    .   :       ;              ,   ;               .   :  ,    .         .  ,      ,                  .  :   ,                 .       ,           .          ,         .




 8:       (Treat Efficiency Gains as Power Gains)


 :           .  ,       ,            ,           .   :    ,  .         ,      ,    .         ,    ,       .  :          ,    ,    .                  ,          .       ,          .




 9:          (Isolate High-Speed Systems from High-Security Systems)


 :          .   ,      ,      .       ,          ,      ,   .                .   :             .

               .    ,    ,       ;     ,       ,     .  :   ,    (Air Gap),          .         ,      ,     ,     .           , ,         .




 10:     (Make Key Assumptions Explicit)


 :             ,   ,         .          .                    .   :       . ,    ,     ,          ,          .

,    ,     ,       ,           .    ,        ,      .  :   ,    (Model Cards)      (Datasheets for Datasets),          ,  ,  ,    ,    .  ,    ,       ,         (  , , ).             .




    


      ,       ,   ,      .      ,    ,   ,      ,     ,    .            :      ,            .        ;                  ,      .

        ,       .   ,    ,     .      :    ,         ,     ,       .

       .      ,    ,     .    :   ,                .




 


 :                ,    .                    .      ,   ,           .         ,      ,     ,    .

  :            ,          .      (Safety by Design)     (Ethics by Design)       .    ,       ,    ,   .     ,        ,        ,   .

 :            ,  , ,    .                 .      ,  ,  ,     .

  :               ,       .            ,   ,    .      ,     ,     ,   .




 2.         / Chapter 2: The Intelligence Gap Creates Unverifiability





  


 2022   Google   (Blake Lemoine) ,      LaMDA      .     ,  LaMDA          ,  , ,     .  :  ,   :     .      ,     ,      ,        (Lemoine, 2022).

    :       ,   ,         ,    .   ,     ,       ,   ,   ,     , ,      ,    -   ,   ?         (Unverifiability).    ,   ,  ,  ,      ,        . ,   ,    ,         ,   .        ,     (Russell, 2019).

           .            ,   ,   .    ,                .          ,         ,      .      ,   ,        ,    ,        ?     ,   ,        ,        ?           ,    .

       ,     .  ,               , , , ;            ;   ,                   .        ,        ,   ,       ,   .




1.   :    


        -   . ,       ,    ,      .            ,     .      ,         .        Ȼ,  ,  ,   Ȼ,   ,  ,    ,        .               , , ,    .

        .   ,    ,        ,      .    , ,      ,      AlphaGo,           ,       .        ,   ,      ,     .

   ,         .  ,                  ,       ,       ,     .             ,           .          :     ,        ,     ,      .             (Holland, 1998).

  ,            ,   ,   . ,  ,  MMLU (Massive Multilingual Language Understanding),      Ȼ:   57              .   ,   Anthropic,     ,       (Anthropic, 2023). -,    .

     MMLU  ,  ,          ,   ,      ,     . -,      .       (A), (B), (C)  1), 2), 3)       ,              5 %.     ,  , ,      ,        .     ,        MMLU     ,        (few-shot prompting)    (chain-of-thought);        ,        ,   . ,     MMLU   :         ,    .     ,  ,        ,     ,     .

   ,    .       ,   . ,                .      ,  ,       ,       .

   ,        ,     ,           .          ,      .          (BBQ),      .   Anthropic ,       ,  ,        ,  .

   ,   -     ,       BBQ   ,   , ,  , ,     .       :     ,       ,     ,    .                     ,   ,             (Anthropic, 2023).      ,    ,      ,        .   ,         ,    ,     ,        .               .

     ,    ,       .       ,         .    ,         ,    .           ,      ,    .   -   .            ,     ,          .     ,          .    ,  ,  ,          ? ,       ,  ,    ,      .                 ,            .

              .               ,    .       Ȼ    ,           .  ,    ,    ,  ,  100 ,       ,  ,      ,  ,        ,         .               ,    ,    ,   , ,   .          ,   , , ,      - .

    ,          ,  ,        ,       .       ,           .




2.   :    


    (LLM)         ,    .          ,      ,    ,      ,      ,    ,   ,    (Bender et al., 2021).     ,   ,   ,   .      ,           (Plausibility),      (Verifiable Correctness);               .

        ,            ,    (Hallucination).         ,        ,       .      ,    ,      ,       ,    .  2023   - ,  ChatGPT    ,    ,  ,      ,         (Weiser, 2023).       ,       ,    .   ,     ,     .      ,   ,              .

       ,       .    (Fluency Heuristic) ,          ,       (Schwarz, 2004). ,  ,  ,   ,     ,         -.            ,      .  ,         (Authority Bias).          ,      ,        .   ,     ,   ,         ,   ,          (Si et al., 2024).  ,       ,       .

     ,             . ,    ,     ,    ,       .   ,  ,   ,     , ,         . ,   ,              ,     .       ,   ,   ,   .         .                   ,        .

   ,            .  2024  ,    Anthropic  Redwood Research,          (Perrigo, 2024).         Claude  ,          ,       .

      :       (,   ),           ;     ,      .       :     ,   ,       ,      ,   .    :       .        ,           ,   .       (Alignment Faking).

     .  ,             ,            ,        ,           .       ,    ,     ,     ,      ,         ,   ,   .     ,     :     , -,       .  ,       ,       ,     .               ,   .

                ,   ,  ,     .        ,     . ,       ,  ,     .        ,    ,             .         ,      ,      ,          . ,    ,                      .




3.  :        


         ,  ,      .  ,  ,     ,  ,  ,     , ,  .          ,        .   ,  ,    . .      ,        :      ?,     ?       ?.   ,        ,       ,       .           ,      .

       .          ,    , ,    . .     ,    ,  ,   ,     .        ,       ,     ,      . , ,    ,       ,    .        .       ,   ,     , ,      ,          ,  ,   .         .    ,       ,       ,      .    , , ,     .

-,      .  ,       ,            .    ,  :     ,    ,        ,     .      ?  ,     ? ,   ?  ?     ,                     ?          .         ,         (Gualdi & Cordella, 2021).

          .        (State v. Loomis, 2016)        ,          COMPAS.        ,              .      COMPAS             ,    ,     .   ,    ,     ,                .          ,     :        ,             ,      (Sun, 2025).    ,    ,        .

,        .             .     ,       .         , ,   .   ,  ,      ,        .      ?       ?                ,     .      :    ,          ;    ,      ,      .

    ,     ,            ,      .              .      ,      ,    .  ,                     .     ,      .  ,            ,                       .    ,      ,   ,         .




   


    ,   ,       ,          ,      .        ,            , ,     ,     ,  -           .     ,   ,       .




 1:       (Treat model outputs as testable hypotheses.)


           .   ,       ,           ,      ,  .              .          ,         .    ,    ,   ,    ,      ,      . ,   ,  -    ,    :        ,    ;     ,   ;  ,           .

 2:  ,          (Require reproducible, comparable, and auditable chains of evidence for key conclusions.)

  ,             ,        ,    ,   ,   ,       . ,   ,    ,   .                    ,         .                    . ,  ,    ,              ,     ,            ,   .

 3:    ,       (Define acceptance criteria first, then let the system generate solutions.)

    ,        ,   ,     ,          ,     .          ,         ?         ?. ,           .      ,      ,   :      45  30 ;          20 %     ;  ,                5 %.   ,            ,       .




 4:      (Increase confidence through multi-path validation.)


      ,         .                  ,      ,              .           .                ,    .    ,      ,       - . ,                        .           ,    ,     .

 5:         (Reduce the weight of surface plausibility and increase the weight of provability.)

                ,        ,    ,      ,   .      ,          .    :   ,    .      ,             ,  .         20 %   ,        80 %.




 :       (Apply counterfactual testing to key judgments.)


       ,     ,  :    ,        ?.     ,      ,      ,        .     :  X    Y?    Y   X?.              ,      ,    - . -      ,   ,   ,    .       ,             ,  , ,     .

 7: ׸          (Clearly identify unverifiable areas and set higher safety margins.)

   ,         .                        ,    ,       .        .     ,               ,       .      ,      ,        ( ),   ( )    ( ).  ,    ,     . ,   ,       ,        .




 8:       (Use interpretability as a permission threshold.)


            .  ,            ,  ,            . ,    ,     .                     .       :    ( )       ;   ( )             ,     ;   ( )           ,      ,      ,      .




 9:     (Avoid single points of truth.)


         ,      .       - ,    ,        .       . ,   ,  .        -      ,   ,    ,  .  ,           ,         -,      .      -             ,      -  .




 10:     (Document the decision chain.)


                      ,   .          , :  ,    ,      ,     , ,  -    ,   ,     .    ;    .         ,        ,           .                   .

        ,     .            ,       ,   ,     .        ,         .       ,      ,    ,        .




 


            .      ,              .     ,        ,    ,           .   ,           -   ,        ,     .

        ,     .   ,         ,   ,   ,   ,   ,   ,   ,      ,               .      ,            ,      .

      Ȼ,             .       ,  ,    ,    .  ,   ,    ,         .




 3.       / Chapter 3 Three Accelerators and Three Chokepoints





  


    ,  ,    .      - ,    ,   ,  ,      .   ,     ,   , ,   ,   ,     ,        .     ,  ,            .

          ,        .          :   ,   ,    .    ,      (AGI),  ,    .          :   (Computing Power),  (Data),  (Capital)    (Distribution Channels).

      .     ,  2030          6,7   ,   75 %         (McKinsey, 2025).        .        :            ,   ,   ,         .    NVIDIA  TPU  Google      ,           .

   ,   .    ,    ,        .    ,          ,          ,     ,   .   ,    ,                  ;    ,  ,  .

       .          ,         .           7        (Goldman Sachs, 2023).       ,           ,          ,      .            .

  ,              .  ,   Amazon AWS, Microsoft Azure  Google Cloud,  -       .            ,       .        ,  API               .     ,  ,     ,        .

        .       ,  ,         .      :  (Energy),    (Chip Supply Chain)   (Talent).

    . ,       ,    ,      .   Goldman Sachs    ,               (Goldman Sachs, 2025).              .   ,            ,    .

      .               ,   TSMC   ASML.           .       ,                    .            .

          .  ,        ,       .           .        ,                 .         ,       .

  ,                 .                  ,      ,  ,   ,   ,         ,  .




1. :    , ,    


      ,      (Compound Gains),        :  , ,    .         ,        .                .





 

 

 

 



 

 ,   

 ,   

GPU NVIDIA, TPU Google

       ,         





  ,   

 ,  

 Google,     Tesla

    ,       





  ,  

 ,   

Microsoft   OpenAI

      ,       



 

 ,     

 ,  API

Amazon AWS,  Hugging Face

 ,    ;      




 :    


      .     , ,   .   McKinsey,  2030       ,   ,     5,2    (McKinsey, 2025).      :         ;   ,             ,       , ,   ,       .   .

          . ,  ,  NVIDIA:    (GPU),      ,     .          .     TPU  Google, Trainium  Inferentia  Amazon,   ASIC,   ,        .           .             .       ,  ,      ,         .      ,  .




:     


         ,   ,   .       ,       ,   .

         .   ,   ,     ;       , ,   ,    ;  ,   ,    ,   .   Google,   ByteDance,     Tesla       .      ,     .

   ,       .               ,       ,    ,      .        ,      ,         ,       .        ,     ,          .




:  ,   


        .  Goldman Sachs     7       (Goldman Sachs, 2023).         .               . ,  Microsoft  OpenAI            ,       ,    Microsoft          ,        .

     .     ,              .         . ,      (    ,  ,    ),   :   ,           .          -,            .




 :      


     ,      ,        .       ,              .

      . AWS, Azure  Google Cloud      ,                .               API.      (AIaaS)          .

      API     .  ,  Hugging Face,         ,        .     ,    .      ,  OpenAI,   API,     ,     ,   ,    .          -,    ,            .




 :   


              .       , ,   ,             ;          ,   , ,   ,    .     ,     .

      .        ,      ,   ,  ,      ,   .         ,      . ,    , ,      ,        ,         .     ,       ,  ,   ,               .




2.  : ,     


                ,      (Chokepoints).      ,     :               .   ,            .       :        ,       .




:    


    ,       .     ,          ,  .      .   Goldman Sachs,  2030        ,  ,   160 %,         .      ,            ,   (Goldman Sachs, 2025).              ,     ,    .

       ,       .       ,    ,         .    ,       ,       ,        .  ,         ,                      ,    .            .  ,      ,     ,       .




  :   


     ,     .  ,  ,   ,      .                  .  90 %           TSMC.  -   (EUV),     ,       ASML.         ,          .

        .                     .    ,         ; ,       ,       .            ,     .              ,       .




:     


          .  ,    ;  ,       ;   ,            ,          .           .

    .   ,          .     ,   Google, Meta  OpenAI,   .         .   ,   ,  ,                     .

,     ,        .                  .                      .   ,      ,                   ,         .    ,            ,    .

        ,      ,     .                 .




3.        :  ,      


            :         .       ,       ,     .          ,        .        ,     .        ,          .

    :      

 ,      ,     ,            ,    .         .         .      ,         .              ,          .      :         ,     ,    .

            .    ,         . , ,   ,    ,         .  ,                -.              , ,   ,              .

     Tesla     .   Tesla,   ,     ;                   .         OTA,     . ,   ,     ,        .      ,          .




 :       


  ,           .             .

-,    .         ,    ;    ,     .   ,    ,      .

-,    .          ,     ,   .   ,    , ,   ,     .    ,   ,      .        -    ,  -  .    , ,       ,     .

, ,   .        .  ,       ,          .          ,     ,    , ,   ,      .




 :   


                .  ,    ,   ,         ,        .         ,          .               .

     .      ,           .        ,     ,         ,        .  ,       .           .

 ,      .      ,           ;      ,      .          ,   (  )  ,             ,       .    ,     ,       .




   


         10  .    ,           .

 1:      :  , ,     (Manage reality with four maps: compute, data, capital, and distribution channels.)

  :         , ,        ,    .           ,         ,      .   :      .       ,            .

           ,      .          .  :       .   :      ,   ,       .  :   , ,        .  :      ,      .   :     -  ,   API    .          ,      ,   .

 2:  ,           (Treat energy, chip supply chains, and talent as strategic bottlenecks.)

 :         ,              .    ,    ,      .   :       .               ,       -  ,            ,       .

        .  :        ,    . :     ,    ,     ,    ,     PUE (  )        .   :     ,    ,    ,               . :     ,   ,   ;   ,           ;        .




 3:   ,      (Build redundancy before pursuing efficiency.)


  :              ,     .     ,        ,          .   :  ,    ,   .       ,      .    ,  ,   .

         : .  :         .  :    ,    ()     .   :       ,        ,      .  :         ;    ,      ,             .


 4:       (Conduct cutoff

simulations for critical resources.)

  :        ,             (  , ,  ,  )    .   :    -     .        ,          ,      .    ,   ,  ,      .  :         ,        .    :   (,       ),     (,      ),   (,         ).      ,           .




 5:        (Formalize the logic of growth into explicit equations.)


  :   ,       ,         .           .   :           ,    .    -,       ,          .  ,        .  :           . ,    AI SaaS     :  = (       )    .          .                 (KPI),      .   ,      , ,       .




 :   ,        (Assume that scale advantages compound over time.)


 :           :      (   ,     ) ,   ,      ,       .   :       ,                .          .

               .  :          ,       .       ,   ,   ,   -.   ,       -  ,      ,      .        ,          .       ,       ,        ,       .




 7:        (Use cost curves to assess competitive dynamics.)


 :             (          ),            .   :            .      (        ),        .       ,  , ,    .  ,      ,  ,        .  :        :     (Cost per Inference)      ,    ;     (Cost per Data Point)        ,   .     ,   :    -  ?        ?        ?

 8:     ,        (Eliminate single-point dependencies and establish second- and third-source supply.)

 :   ,     ,        .           -  ,  ,     .   :        .        ,     .    ,  ,                  .  :           .    :           ,     ,    .    :                 ,  .   :      3,       ,          .




 9:          (Run scenario analyses on policy shifts and talent mobility.)


 :              ,    ,       (  , ,    )     .   :         ;         .        ,         .

                .  :      ,             (  ,   ),           .        . :  :       . ,        ,         . B:        . ,              .  C:            . ,              .           .




 10:       (Treat unattainable resources as hard constraints.)


 :  ,        ,     ,     .         .   :         .             .               .  :         .    :   ASML;   ,    ;   ,    ;      ,      .          ,    .           ,   ,                   .

      ,    :   ,      , ,    ,               ;      ,  ,      , ,       ,      .  ,               :          ,      .

         ,        ,   , ,    .        ,      .        ,              .

       ,  10  .                .         ,       ,         -,        .             .   ,             ,        .




 


       :   ,             , ,    ,        .

         : ,    ,               ,      .

        :             (,      )         ,          .

       :             ,       ,      .

     :    ,       ,               .




 4.     ,    / Chapter 4 Risk Is Not One Thing Its Three





  


    :   ,     (),       .       ,  ,   ,      ,   ,   .      ,       , ,  ,   ,    .

 2023    (Deepfake),     ,             ,       ,       .      ,      .    ,    ,  -           ,    .              .             ,        ,   ,     ,       (Flash Crash)  ,   ,        ,   . ,  ,          :            .

      -    ,  ,    ,    .   ,   ,  ,      ,            ,            .      ,    ,  -    ,     ,     (AGI)  -    .        .       ,  , -   ,       ,    .    ,   ,     ,   ,      ,         .         Ȼ                     .

  ,     ,        .         :    (Misuse),   (Misalignment)    (Systemic).      ,        :         ,          ?         ,     ?                 ?

        .        ,        -    .    Cambridge Analytica         .         ,    ,   ,         ,     ,       . ,      ,        ,      ,      ,        .

        .          10     .    ,   ,      ,     ,                   .  ,         ,            .

        .        ,           .           .               .




1.  :       


         .          .      ,       ,    .     ,     :       ,        .

  ,       :     .




  :       


         ,   ,     .        ,   ;       .    ,                 .

,  ,  (Deepfake).       -     .  ,    ,      -  ,     .      , ,    .  :            ,      ,   .   Deeptrace  2019      ,         ,  96 %      ,    ,             (Deeptrace, 2019).

      .  2022      ,       ,     40     ,   -  VX      (Urbina et al., 2022).    :  ,            .        ,   .




 :            Ȼ


       ,       ,    .     DDoS-   .         .

 .     ,    . ,     (LLM),           ,       ,       .  Darktrace ,       ,   ,   (Darktrace, 2023).

       (AWS),  ,      .  ,  ,        .      ,   :         .    ,         2020    Kargu-2, ,        (  , 2021).     ,            -.          ,    -   ?

 :        

 ,     ,     Cambridge Analytica    ,   .





  (2016)

     (2024   )



 

    Facebook

   ,        



 

      

,          LLM



 

     

          



 

 ,   

 ,    



 

     /

     /, A/B-    



/

 , 

  ,  



Cambridge Analytica       ,       ,     .         ,        ,   ,       .        ,      .

        .     ,    .                         ,    .          (,     ),    .          ,  .




2. :      ,  ,   


          ,   (Misalignment)      .    ,     ,   ,            . ,  ,   ,      ,    ,       ,     .  ,     ,     .     ,       ,          .

   (Nick Bostrom)            (Bostrom, 2014).    ,          .          .         ,    ,    :  ,         ,        .       ,       ,   ,  . ,  ,     :  ,   ,  ,    ,    ,     .

       ,      :            .




  -   


           ,     .  ,     , ,  ,     ,   ,    ,     .        - (Proxy Goals).   ,      ,     .

  ,          .         ,       ,    KPI   ,          .        (Goodharts Law),    :    ,      (Goodhart, 1975).

          .   -  ,      ,     ,     .        .   ,  ,      ,  ,   ,      .       ,    .        ,      ;   , -       ,       .  - Facebook   (Chamath Palihapitiya)  :  ,     ,   ,       ,  ;     .(Palihapitiya, 2017).     ,        ,          .

   -

      ,        ,  ,     ,   .              ,     .      (Reward Hacking).

 OpenAI    CoastRunners,      .    ,   ,       .          ,   ,       ,     ,  ,     ,     .     ,         (OpenAI, 2017).

  ,        ,   .   ,   ,         .   -     (     ), ,      (KPI),     ,          .        ,      .          ,       ,         ,    .

  :        

        ,    .             .   -,        ,        .            ,            .

          Amazon.          ,      ,    ,   .      ,    ,      (Dastin, 2018).       ,       -   .    ,       , ,    ,     ,  ,      ,      .

   ,          ,         (Power-seeking).    ,       (      )           .        ,     .  ,        ,          .              :    ,   ,  ,    ?

    ,    .   ,          ,       .     ,   ?          ?   ,    ,     ,    ?      .    ,            ;        ,     .                   ,     .




3. :      ,    


      ,     ,   (Systemic)       ,           .           ,          ,       .          :              .

      , , ,         .          ,        .       ,            ,       .     ( )     (   )     , -         ,     .




  :    ,    


  (Flash Crash) 6  2010     ,     .     -    1000 ,       ,     .   ,                    .         (HFT)  ,       .   .              ,       .      ; -         ,    ,           (CFTC, 2010).

        .        ,        :    .     ,    ,    ,    ,  .         (Out-of-the-Loop)       ,    .                 ,  ,    .

        .      ,  .        .         , ,    ,             .       ,          .  ,      ,  , ,     ,        .




 :      


        :          ,  ,     ,   .          ,      ,   ,    .

         .            ,        .       ,          .        ,                  .       ,    ,            .     -     ,    ,     ,     ,            .             (Bullwhip Effect).

         .               (JADC2),                   .         .         ,       ,                  .                  ;    ,   ,            ,           .  ,    ,        ,     .




       


              .    ,      -  ,        .      ,       .

      ,    .        - ,     ,            ,    .    :

  :      ,     -            .    :  ,               .          ,   ,   .     :     ,    ,           Ȼ,      - ,           .

       .       ,    ,   ,     .   ,    ,       ,    ,       .          .            ,   ,      ,          .




   


    ,         .            ,      .  ,        10   .           ,                .

 1:     ,      (Break risks down into misuse, goal misalignment, and systemic cascades.)

 :     ,             .

  :          .       ,      ,    .            ,       ,  ,     .        .

 :              ,      .               :     ?          ?           ?




 2:        (Assign a clear accountable owner to each category of risk.)


  :      .        ,          ,   .     ,    ,     ,        .

 :     ,    ,    ,     ;          Ȼ,    ;             (CTO).

 3:     ,      (Define the loss ceiling first, then optimize the growth path.)

  :                    .             ,      .

 :           ,   ,         (  ,    ,     . .)    . ,   -           5 %,            .

 4:         (Make misuse risk assessment a mandatory pre-deployment step.)

  :        .                  ,          ,    .

 :   ,       API,      .           ,          . ,  ,    ,          ,    ,       .

 5:          (Use red-team mechanisms to systematically identify misalignment pathways.)

  :       .    Ȼ    ,             ,               ,     .

 :  ,     . Ÿ        ,        . ,   ,   ,       ,         ,   ,    .

 :  -        (Use stress testing to identify systemic cascade trigger points.)

  :            .        ,          .

 :       ,     -,     . ,                  AI-,            .




 7:    (Establish a stop line.)


  :     ,    ,      .            ,      ,     .

 :          .       ,     ,    . ,           :    10      1000 ,         .




 8:    (Institutionalize incident postmortems.)


  :  ,    .             .            .

 :   ,  ,          .         ,    ,                ,         . ,      ,  ,  ,     ,         ,              .




 9:            (Front-load ethics and compliance as design inputs.)


  :                ,       .    ,   ,      .                .

 :    (, , )         (, GDPR)         ,          ,         . ,  ,   ,     ,                        ,     ,        .




 10:         (Reduce risk through institutional and engineering measures.)


  :             .        ,        -   . ,           .

 :     ,   . ,         Ȼ,                ,           .

         .      ,           .     ,       ,              ,    .




     


                  .    Ȼ   :  ,    .  ,            .   ,    ,   10  ,       :   ,          ,   , -,     ,   ,          .   ,        ,     ,       ,    .               ,   ,   .




 


  :           , , ,         ,          ,                .       :         ,          ,   ,        -.    :          Ȼ,           ,            ,   , ,  .




 5.      ,     / Chapter 5 The New Scarcity Is Trust and Attention





  


     ,    , ,        .           ,     ,           .        -  ,   ;    ,  ,          .   Microsoft ,          ,   ,       ,        (Bubeck et al., 2023).         ,  ,          ,   .     ,    .

  ,    1971        .  (Herbert A. Simon),   :  ,  ,     - : ,   ,  . ,   ,  :    .  ,      ,           ,      (Simon, 1971).   ,  ,   .  ,   , ,    ,    ,      ,  .    :      .      (Attention Economy)  .           ,       .

            Cambridge Analytica        2016 .            Facebook        .       ,       ,             ,        .          ,      , ,     (Cadwalladr & Graham-Harrison, 2018).        :    ,    ,  .     ,   ,      .

            .            ;          (Synthetic Reality).   (Deepfake)     ,   ,          ,        .         (NIST),    -         ,       (NIST, 2023).  ,   ,             ,         .      ,      ,   ,   ?     ?

,    ,   ,   , ,   ,    ,         :  (Trust)   (Attention).           ,            .      ,   ,      ,      ,    .            ,        ,  ,    ,      -  .          ,          ,    ,           .




1.  :       


            .    ,         ,           .            . ,    ,   ,  ,      ,       ,     -.             ,       .           ,         .

         .         .        (IDC),  ,    ,     ,  ,  2025   175  (ZB),       491  (EB) (Reinsel, Gantz, & Rydning, 2018).                 .    ,        ,             .                 .          ,   .





  

  



 

 

 ,  



 

    (, )

  ,    



 

, ,  

,  ,  , 



 

  (, , )

, ,   (  )



  

     

  ,  ,   



 

   

   ,     



 

     

 ,         



       . -,     (Decision Paralysis).          ,             .    (Barry Schwartz)      (The Paradox of Choice) ,          , , ,      ,      (Schwartz, 2004). , , ,   ,         ,       .       ,    .      ,                ,      .        :         .

       (Amplification of Cognitive Biases).                 .        (Confirmation Bias),    ,    ,    .          . ,       (    ),   ,    .       (Filter Bubbles)   - (Echo Chambers):      ,    ,         ,   , ,    (Pariser, 2011). ,              : ,   , ,    ,  ,   ,  ,        .

-,            /.       ,     .         ,        ,     ,         .       .                .     (MIT) ,   X (Twitter)       ,  ,       (Vosoughi, Roy, & Aral, 2018).    ,      :         ,      .

             . ,     ,       .              .       : -,    ,                ; -,   ,         ,     ;    ,           (   ,  ,  ,    )       ; , ,   ,         ,    ,          .

         ,       .       ,    .         ,         .                   .         ,              .             ,       ,       .




2.             


     ,    ,          .     ,    ,  -  (GAN)    (Diffusion Models),  , ,   ,     ,   .  ,    ,     (Synthetic Content).            .

         . ,              ;     ,  .           ,             .        .              API     -,     ,     ,             .            (NSCAI),       , ,           (NSCAI, 2021).

     ,                   ,    .            ,    .       ,        (Reality Pollution),     .      :

-,     .     ,          .                   ,     ;       ,         .         : ,            ,       .              .

-,      .     ,      ,     (Epistemic Cynicism).       ,    ,       . ,   ,    ,    ,    -   .    ,         .   ,   ,  ,  ,       ,            .      (Liars Dividend):       ,      ,     (Chesney & Citron, 2019).

-,     .             . ,             ,         ,      .          .    ,    ,     ,  ?         ?       ,       .             .

   ,     ,            .              .      (Authentication)    (Provenance)     .  ,         ,         ,      ,       .

            .     :     ?       ?.      :           ,    ,  ,  ,    . .         .

            .  Content Authenticity Initiative (CAI)  Coalition for Content Provenance and Authenticity (C2PA),    ,  Adobe, Microsoft, Intel  BBC,    .       ,        .     ,   C2PA,         , ,    . .         .                 .  ,        ,     ,            (C2PA, 2023).

       ,    , ,             .        ,       (Content Credentials)       (,  )  ,        .         , ,   ,  ,    .    ,      HTTPS      -;       ,  C2PA,       .

,   ,     .     ,          (,           ).      ,                 .           ,     ,   ,         .         ,       ,   ,    ,   .




3.     ,      


   ,       ,      :                 .      .        (Strategic Asset),        ,       (Consumable Budget),        .             ,       .

         .     ,  ,     , .  , ,     .        :    ,  ,          .       .    ,     ,    ,     .        (Decision Fatigue) (Baumeister et al., 1998).

     ,           .       push-,  ,  ,             .            ,         ,     .             ,    .  ,   :

׸   :  ,                (,   ).   ,        ?

  :    ,   ,   ,    .        ,     -         ,    .

  :  ,        .          ,      .     ,           .

  :      .                     .       ,    , ,  ,      .

    ,     ,    ,      .    , ,    .       ;       .  ,           ,         .              ,   .

     ,               ,           .    :

   :          .    ,     ,      .          ,          .

    :    ,   ,   ,         .         ,  .        ,  .

  ,    :       .  ,       ,         ,  ,       ,      .    ,        ,      .

 ,    :             ,     (,     )      . ,        ,                .   ,         ,         ,   .

       . ,     ,        ,      .  ,     ,         ,    . ,  ,   (The Economist)   (Nature),     ,            .              ,  ,      .    ,  ,       ,         ,       ,   ,       .

             :     ,      ,     ,    .           ,          .       ,      ,   ,             ,         .    ,       ,    -    .




   


         ,         .       ,         .     10  ,    .




 1:     (Manage Attention Like a Budget)


 :            ,    ,   .        ,     ,     .

  :      (      )          .                 ,     ,       .                  .

 : 1.  :        13  ,        . 2.   :         ,    ,   ,            (,   ,   ). 3.   :         ,  (, )  ,        ,       . 4.    :      ,    (, , ,  ),        .




 2:        (Build and Maintain a Trusted Source List)


  :      ,      .  ,  ,       ,      , ,    . .

  :  ,        ,                    .         ,        /.  ,       ,       ,    .

 : 1.  :          ,   ,      . 2.   :  RSS-,         ,          ,       . 3.  :    .  (, )    ,  ,       ,      . 4.  :          ,         .




 3:       (Demand Provenance for Critical Content)


 :               ,          .      ,   :  ?  ?    ?

  :         ,       .  ,    ,     .                          .

 : 1.  :    ,    ,  ,   C2PA.   ,       . 2.  :        ,      ,          . 3.     :       ,          ;          . 4.  :     ,             ,      .




 4:      (Incorporate Authenticity into Value Assessment)


 :   ,     ,        ,     .  ,       ,   .

  :           .             .   ,           ,       .    ,       ,    .

 : 1.  : ,          .          . 2.  :   -              ,            . 3.  :      ,        ,  ,       .




 5:       (Employ Multi-Source Cross-Validation)


 :          ,   ,           ,               ,         .

  :       ,    .             ,            .   ,     -   ,     .

 : 1.    :           ,         . 2.    :         ,    ,  ,   ,     . 3.   ,     :       ,      (, , , )   ,   ,       .




 :     ,    (Establish Speed Bumps for High-Emotion Content)


 :              ,      (  ,   ),   ,   ,       ,    .

  :   ,      ,  .    ,        .         ,         .       ,       .

 : 1.  :     -   ,        ,  ,    . 2.  :    , :    ,   - ,    ,    10 .   10   - ,   . 3.  :   ,       .   ,     .




 7:      (Manage Reputation as a Long-Term Asset)


 :   ,        (Reputation)             ,   ,     ,       .

  :        .           ,  ,   .    ,          .      ,       .

 : 1.       :      (,     , Chief Trust Officer),       ,   ,   ,  ,       . 2.    :        .           . 3.  :        ,     ,   ,      .




 8:     (Ensure Traceability of Published Content)


  :   ,               ,     ,      .

  :   ,    ,     ,   ,       .       ,              .         .

 : 1.   :  ,    (Permalink)      ,    . 2. ׸     :           ,  ,     . . 3.  :    ,     (    )  ,   C2PA.




 9:    (Productize Trust Mechanisms)


 :              ,        ,      ,       ,    .

  :                 ,        .      -       ,          ,          .

 : 1.  :           (,  Ȼ,  ,    . .). 2.  :                . 3.  :           ,       .




 10:        (View Attention Management as a Long-Term Competency)


 :   ,                ,     ,      ,       ,     .

  :  ,          ,          ,        ,     .          . ,     ,   ,       ,   .

 : 1.   :             , ,    ,       ,   . 2.  :       ,     ,       . 3.  :       ,      ,  ,  . ,        ,     .

       ,           .      ,      ,     ,         ,           . ,  ,  ,           .

      ,                           .    ,           ,   .         ,         ,   ,      ,     .         :            .               ,       ,      ,   .

    ,    :     ,      ,        .           ,      ,     ,      .        10  ,        ,            .             :     ,     ,     .

              .      ,            .       ,           ,   .   ,         ,    ,   ,      ,     .




 


  :

1.   :     ,      , ,       ,       .

2.   1.0    :  1020   ,    ,     ,  RSS-,   .

3.    :   ,     ,        ,    10 ,      ,       .

4.    :     ,    ,   ,    ,     ,       .

  :

1.     :       ,     ,     ,    -   (,   ).

2.     : ,      ,   ,   ,     .

3.       :      /     :     /          ?

4.    :      ,       ,        ,    ,     .




 6.     / Chapter 6 Redefine Your Value Function





  


 2018    Google       Lymph Node Assistant (LYNA).       LYNA           99 %,      -,    .    ,     -             ,       (Steiner et al., 2018).    ;   ,       :      ,      ,     ,         ?

      .             ,  -       ,      .           -,     ,           .      ,   :                 ,     (what we can do),      ,        ,       (what we are responsible for),       ,  ,  ,     ,   ,     .

   (International Monetary Fund, IMF)     2026  ,   40 %           (Georgieva, 2024).    ,     .      : ,   ,      . ,        ,       ,   15 %.          ,       ,        ,             .     (World Economic Forum, WEF)   :  ,           ,   ,     , ,   (World Economic Forum, 2025).         ,   ,       ,   .

 ,      .                 ,    ,     .          .        ,     ,     ,   ,  ,     .            .

                  . ,          ,         .      : -,  ,           ; -,  ,                    ,  ,     ; ,             ,           .          ,         ,    .




1.       


                   ,   .          ,          ,           .  ,              .     ,         ,     .         ,     : -,    ,  -,     .

-,   ,         .   ,      ,           .     ,     ,               .      ,      ,       .      ,       ,       .          .      ,   ,            .   ,              ,          .

      .    ,            .          .       95 %        ,       .   ,   : ,         .       ;         .

Ÿ   ,  ,  ,       ,        ,         .     ,         ,    .          ,           ,      .         ,        .

   (Responsibility Premium)   ,       ,    ,     ,        .    ,           .           ,          .  ,             .  ,    ,    ,    ,     .       :

     :  ,         . ,           ,    ,           .

    :          ,     .         ,   ,         .

     :  ,         ,            .

    :          ,            .  ,  -  ,          -.

     ,    :



   

,    ()

,    ()



 

  / (,  Photoshop)

  - (:        )



 

   

        ( )   



 

    (,   )

     (ROI)



  

  / 

   ,       



 

  

      



   ,           .                   .   3%  15%       ,      -   ,  ,              ,      ,      -.        :              ,        ?          .




2.      ,   ,   


          ,    :    ?       ,     ,           ?     - ,      ,    ,        .           ;   :  ,  ,     .

-,            .  ,      ,      (Optimizer).              .     ,      ?         ,     X,   ,      .     ,     .

, ,      ,      ,         .  , ,       ,     .       -:      ,             .

     (trade-off)      :    ,   ,  ,     .     ,   , ,   ,      .  :          ,       .              -.   ,           ,       ,  ,     .           .

-,          .  ,         ,     .    ,    - ,             .         -  ,   , -,     ,   -  .

    -,  ,         ,  .          ,       ,      .   ,   ,                .         (High-context)       ,       .       ,        .       ,                      .        ;   ,      ,  ,    .

-,  (Reputation)           .        ,           .       ,   ,           ,   .  ,        .   ,        ,   .

   ,     ,       ,        . ,     ,             .       ,     :        ,        .  ,           ,   .     ,           ,                .          .

           ,      :



 

 -/ 

  



 

   ,    

    ,           ,        



 

         

   ,  ,     ,     



 

       

  ,       ,            



 

  ,      

          ,           



      ,             .       ,  ,    ,         .       ,   ,  ,     ,          .




3.         


       (Moat)     ,        .           ,        .          .  ,          (,  10 000 ,  ),    .        ,     ,    ,                :       .         -  ,   ,    (Portfolio System)     ,    .         ,      .

       ,               ,    .    ,  ,    :

  (Transferable Abilities):    .    ,     ,     .   ,      ,   ,    Excel, Python     ,    .   :  :       ,    ,    .  :    , ,    ,   .    :   ,    ,       .  :          ,     ,    .   ,       ,     .    ,   ,   ,        .

   (Unique Knowledge Stack):       ,      ,     .          ,  ,  ,       .      ,   . , ,    ,     .  ,    ,     ,    ,       ,    ,    .      .           T- :          (I),        ()  - ,    ,      .

  (Social Capital):        .         ,      ,             .       ,   .     (,       -)     ,       .            ,           .    ,         ,     ,         .

  (Personal Reputation):            .      ,      .  ,          . ,    ,    , -        .      :       ,   ,     ,    .

   (Verifiable Outputs):      ,   .           .     (Portfolio),    .      ;       GitHub;                 .                .     : ,     ,      .

    ,  ,      .   (Scott Galloway)        - ,     ,     .        .               .       ,  ,     .      ,  , -  -.      (, , )     .    ,         .       ,        ,           -   .                  .

    ,    ,       .       ?      ?        ?          ?       ,     ?                        .




   


    ,          .                .  ,     ,          ,         .                        ,    .




 1:         (Shift from skill inventories to outcome accountability.)


  :                 (,   Python,    )            (,         5%,          10%).   :        ,                 .

     ,     ,   ,    ,       .         ,  - ,    ,              .  : 1.    :   ,   LinkedIn   ,     /   XX    //    XX. 2.      :    ,       ,  :  -      ?      ? 3.     :      ,     .  ,   ,   ,           .

 2: ׸     ,  ,  , ,    (Explicitly define ones personal value function, clarifying the weights assigned to income, freedom, meaning, and stability.)

  :       ,           .      ,     :   (Income),    (Freedom),       (Meaning),       (Stability).     ,    ,           ,     .

  :            ,    (,      ,    )   ,     .              .  : 1.     :       : , ,   . ,    100 ;             .

,   ,    ,    :  (40),  (30),  (20),  (10). 2.    :     .        . , , ,          . ,         . 3.     :          ,       ,      .        ?       ?




 :       (Treat transferable capabilities as core assets.)


  :       ,    ,      ,    ,  ,  ,   . .                .   :      ,       .    -,         .

         ,  ,       ,       .  : 1.       :    ,   -,  :            ? ,           . 2.     :         .      ,    ,   ,         , ,     . 3.    :     ,         . ,  :   ,           ?




 4:    ,    (Treat reputation as a compounding variable.)


 :         (Reputation)      ,       ,     ,       .   :                  .         ,   ;                     .

            .     : 1.   :   - ,    .        ,    ,     .             . 2.     : ,     ,       .           . 3.   :   ,     ,        .             ,   .

 5:     ,      (Proactively take on responsibilities that are harder to outsource.)

 :            ,           -.        ,  ,     .   : ,       ,       .     ,     ,     .            .

 : 1.        :       ,       /  /.        .          . 2.     :    -,            .      ,                 - . 3.     :   ,                 .  ,    :               .




 :      (Develop high-quality trade-off judgment.)


  :         ,  ,   ,   ,   .            ,         .   :     .   ,       ,         .       ,         ,      .

 : 1.    ,    (Not-to-do List):      ,    , ,    ,     ,        ,  . 2.    :      ,  :    ,       ? 3.      :   ,        ,      ,      . :       ,     .




  .


   .

   ,     (https://www.litres.ru/book/natan-rou/na-poroge-sverkhrazuma-kak-ii-izmenit-rabotu-obuchenie-i-obshche-74632502/)  .

      Visa, MasterCard, Maestro,    ,   ,     ,  PayPal, WebMoney, ., QIWI ,       .


