AGI (人工一般知能) と創る新しく奇妙な社会 / New and Stranger Society built with AGI

2025年2月6日(木)、某所の勉強会にて使用したスライドです。

Transcript

1. VUCA — generated by Stable Diffusion XL v1.0 AGI ( ) AGI ( ) — 2025-02-06 – p.1/34
2. https://speakerdeck.com/ks91 AGI ( ) — 2025-02-06 – p.2/34
3. ( ) ( ) 1993 ( ) 2006 ( ) SFC 24 P2P (Peer-to-Peer) 2011 ( ) 2018 2019 VR 2021.9 & VR 2022.3 2023 AI VR&RPG 2023.5 “Don’t Be So Serious” 2023 2024 AI( ) 2024 “ALOHA FROM HAWAII” 2024 2024 AI( ) 2025 (Live and Let Think) → ( ) AGI ( ) — 2025-02-06 – p.3/34
4. (2018∼2023) 2023 2018 SCIgen (+ Google ) 2023 AI AGI ( ) — 2025-02-06 – p.4/34
5. (2024∼2025) 2025 (Live and Let Think) 2024 AI( ) ( ) (Raspberry Pi 4B) “Open Interpreter” ( ) 5cm AI https://github.com/ks91/Freenove_Dog 2025 (Live and Let Think) “01” https://github.com/ks91/01 AGI ( ) — 2025-02-06 – p.5/34
6. AI ࣗಈԽ γϛϡϨʔγϣϯ ਓؒͷ֦ு ϓϩϯϓτ ΤϯδχΞϦϯά ੜ੒ "* ˞
   ͔͋ͨ΋࿑ಇऀΛ૿΍͢ ˞
   ਓؒʹ࡞༻͠޲্ͤ͞Δ ˞
   ϞσϧͰ༧ଌ͠౷ޚ͢Δ େ೴ͷ Φʔτϝʔγϣϯ খ೴ɾے೑ͷ Φʔτϝʔγϣϯ ਓମ΁ͷ࡞༻ σδλϧϨΠόʔ σδλϧπΠϯ αΠόʔϑΟδΧϧ ػցֶश ༧๷ҩྍ γϦΞεήʔϜ Ϗοάσʔλ ςϨΤάδελϯε ϦϞʔτϫʔΫ Ծ૝ɾ֦ுݱ࣮ Y3 ࠶ੜҩྍ #.* ৽ༀ ԕִҩྍ ϩϘςΟΫε ݴ༿Ͱ "* ͱର࿩͠ɺΰʔϧͷ ୡ੒ʹ޲͔͏ͨΊͷϓϩϯϓτ ΤϯδχΞϦϯά͕ɺ͢΂ͯͷ جૅʹҐஔ͚ͮΒΕ͍ͯ͘ ؀ڥٕज़ "* લఏࣾձʹ͓͍ͯɺ ؾީมಈ౳ɺਓؒͷपΓͷ ॏେͳ໰୊ʹରͯ͠΋ ࣗಈԽͱγϛϡϨʔγϣϯ ʹΑཱͬͯͪ޲͔͍ͬͯ͘ "* લఏࣾձʹ͓͍ͯɺ ΤωϧΪʔ΍৯ྉΛ͸͡Ίͱ͢Δ ੜ࢈΍ྲྀ௨ͷࣗಈԽ΍ɺ ೔ৗੜ׆ͷϩϘςΟΫεʹΑΔࢧԉ͕ ࢝·͍͖ͬͯɺ·ͨɺ ϗϫΠτΧϥʔͷ࿑ಇऀ͕ैࣄ͢Δ ࢓ࣄͷଟ͘͸σδλϧϨΠόʔͷํ͕ ͏·͘΍ΕΔΑ͏ʹͳ͍ͬͯ͘ "* લఏࣾձʹ͓͍ͯɺ ݸਓ͕Τϯύϫʔ͞Εɺ ਓͻͱΓ͕Ͱ͖Δ͜ͱ͕ ֦େ͍ͯ͘͜͠ͱʹՃ͑ɺ ݈߁͕޲্͠ɺ ߴྸԽΛՃ଎ͤͭͭ͞ ࣗಈԽʹΑΓ શһ͕࿑ಇਓޱ͔Β ୀग़Λ࢝Ί͍ͯ͘ "* લఏࣾձʹ͓͍ͯɺ ͔͋ͨ΋શਓྨͷ ༧ଌɾ෼ੳೳྗ͕޲্͢Δ AGI ( ) — 2025-02-06 – p.6/34
7. AGI ( ) AGI 5 (Amodei, 2024)( ) 5 (AGI ) ChatGPT Open Interpreter ( ) A X (Amodei, 2024) AGI ( ) — 2025-02-06 – p.7/34
8. AGI ( ) AGI 5 (Amodei, 2024)( ) 5 (AGI ) AGI ( ) — 2025-02-06 – p.8/34
9. AGI AGI : Artificial General Intelligence ( ) ( ) ( AGI ) [ AGI] “Star Trek: The Next Generation” ( ) (1) ( ) ( ) Jean-Luc Picard “Computer, tea. Earl Grey. Hot.” ( ) (2) ( ) ( ) ( ) (1) AGI ( ) — 2025-02-06 – p.9/34
10. AGI 5 (OpenAI ) 1 (ChatGPT) 2 (o1, o3) 3 ( ) (> deep research) (A ) 4 ( ) (X ) 5 (AGI ) 3 5 A X AGI ( ) — 2025-02-06 – p.10/34
11. (Amodei, 2024) 5 (AGI ) 1. AI 10 21 5∼10 2 2. AI 3. AI AI 4. 5. (AI ) AGI ( ) — 2025-02-06 – p.11/34
12. ChatGPT Open Interpreter ( ) A X AGI ( ) — 2025-02-06 – p.12/34
13. ChatGPT GPT ChatGPT GPT OpenAI (GPT-3.5, GPT-4, o1, o3) GPT Generative Pre-trained Transformer ( ) (deep learning) a GPT ( ) GPT-3.5 a : ( ) AGI ( ) — 2025-02-06 – p.13/34
14. GPT ( ( ← DX )) GPT Alec Radford, Karthik Narasimhan, Tim Salimans, and Ilya Sutskever. 2018. “Improving Language Understanding by Generative Pre-Training”. Available at: https://cdn.openai.com/research-covers/language-unsupervised/language_understanding_paper.pdf. GPT-2 Alec Radford, Jeff Wu, Rewon Child, David Luan, Dario Amodei, and Ilya Sutskever. 2019. “Language Models are Unsupervised Multitask Learners”. Available at: https://paperswithcode.com/paper/language-models-are-unsupervised-multitask. GPT-3 Tom B. Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, Sandhini Agarwal, Ariel Herbert-Voss, Gretchen Krueger, Tom Henighan, Rewon Child, Aditya Ramesh, Daniel M. Ziegler, Jeffrey Wu, Clemens Winter, Christopher Hesse, Mark Chen, Eric Sigler, Mateusz Litwin, Scott Gray, Benjamin Chess, Jack Clark, Christopher Berner, Sam McCandlish, Alec Radford, Ilya Sutskever, and Dario Amodei. 2020. “Language Models are Few-Shot Learners”. Available at: https://doi.org/10.48550/arXiv.2005.14165. GPT-4 OpenAI. 2023. “GPT-4 Technical Report”. Available at: https://doi.org/10.48550/arXiv.2303.08774. AGI ( ) — 2025-02-06 – p.14/34
15. Generative Pre-Training, Language Models (GPT, GPT-2, GPT-3) : : (or ) Improving Language Understanding (GPT) ( ) ( ) Unsupervised Multitask Learners (GPT-2) Few-Shot Learners (GPT-3) → GPT ( ) AGI ( ) — 2025-02-06 – p.15/34
16. — Open Interpreter Code Interpreter a (CLI ) CLI : Command Line Interface UNIX Python R Javascript ( ) ( )( Linux ) : a Code Interpreter : GPT Python AGI ( ) — 2025-02-06 – p.16/34
17. Open Interpreter $ python3 -m venv oi [Python ] $ source oi/bin/activate [Python ] (oi) $ pip install -U pip [ ] (oi) $ pip install open-interpreter [Open Interpreter ] (oi) $ interpreter --model gpt-4o [Open Interpreter ] [ 1] > 1 100 [ 2] > 1 : Python 2 : Raspberry Pi “Pi” Python ( ) AGI ( ) — 2025-02-06 – p.17/34
18. A X — A M. A /X MIT 2013 LEGO X A A X A X X X (2013 ) ⇒ 10 A Open Interpreter AGI ( ) — 2025-02-06 – p.18/34
19. A X A A A A ↓ × ↑ X X X AGI ( ) — 2025-02-06 – p.19/34
20. Nobel Turing Challenge https://www.nobelturingchallenge.org 2050 AI 2050 . . . ( ) X A AGI X X X A AGI ( ) — 2025-02-06 – p.20/34
21. ͜Ε·Ͱ ͜Ε͔Β ਓؒʹΑΔ ൃݟɾൃ໌ ਓྨͷ஌ࣝͷશൣғ ਓྨͷ஌ࣝͷશൣғ ࿦จ ࿦จ ࣗಈ ੜ੒ ࿦จ ࣗಈ ੜ੒ ࿦จ ࣗಈ ੜ੒ ࿦จ ͢Ͱʹ ಡΈ੾Εͳ͍ ֶҐɾ໊੠ͳͲͷ ΠϯηϯςΟϒ΍ ޷ح৺Ͱۦಈ શࣗಈ͕ͩ ௐ੔͞ΕΔ ·ͬͨ͘ ಡΈ੾Εͳ͍ ͕ɺਓྨશମ ͷ஌ͷ֦ுͷ ଍৔ͱͳΔ DG
    ;1% ཧ࿦ ϝλࣗવબ୒ AGI ( ) — 2025-02-06 – p.21/34
22. ( GPT AGI) ( OpenAI) ( ) AGI → AGI ( > > ) ( ) X A ( ) A (A ) A A AGI ( ) — 2025-02-06 – p.22/34
23. ( ) . . . (Amodei, 2024) 1 : 2 : AGI ( ) — 2025-02-06 – p.23/34
24. SAY NO TO BULLSHIT JOBS David Graeber, Bullshit Jobs: A Theory, Penguin, 2018. ( , , , — , , 2020) Bullshit Jobs ( ) Shit Jobs ( ) ( ) . . . AGI ( ) — 2025-02-06 – p.24/34
25. Alexa, ত࿨࣌୅
    
     ੈلޙ൒ ࣾձγεςϜ ࣾձ؀ڥ ػց ٕज़ ਓؒ ࣗಈγεςϜ ྩ࿨࣌୅
    
     ੈلલ൒ ిؾ͚ͭͯ ͸͍ ͸͍ ిؾ͚ͭͯ ͜ͷ৚݅ͷͱ͖ ٕज़Ͱ͋Γࣾձ؀ڥ
    ˠ
    ࣗવ؀ڥ ਓ ⇒ ( ) ⇒ ( ) AGI ( ) — 2025-02-06 – p.25/34
26. by ⇒ ⇒ AGI ( ) — 2025-02-06 – p.26/34
27. by AI SNS AI AI / ⇒ ( ) ⇒ AGI ( ) — 2025-02-06 – p.27/34
28. (4 ) : M. McLuhan and E. McLuhan, “Laws of Media: The New Science”, University of Toronto Press, 1988 ( , , , NTT , 2002) AGI ( ) — 2025-02-06 – p.28/34
29. (tetrad) = ( ) 4 : ? : ? : ? : ? — AGI ( ) — 2025-02-06 – p.29/34
30. (4 ) ↑ AGI ( ) — 2025-02-06 – p.30/34
31. Honda UNI-CUB ( ) AGI ( ) — 2025-02-06 – p.31/34
32. (4 ) ↑ ( ) AGI ( ) — 2025-02-06 – p.32/34
33. ( ) (4 ) ( ) ( ) ↑ ( ) . . . AGI ( ) — 2025-02-06 – p.33/34
34. AGI ⇒ ( ) ( ) 24 21 24 AGI ( ) — 2025-02-06 – p.34/34