Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
LLM とプロンプトエンジニアリング/チューターを定義する / LLMs and Prompt...
Search
Kenji Saito
PRO
April 12, 2026
Technology
570
0
Share
Embed
Copy iframe code
Copy JS code
Copy link
Start on current slide
LLM とプロンプトエンジニアリング/チューターを定義する / LLMs and Prompt Engineering, and Defining Tutors
早稲田大学大学院経営管理研究科「プロンプトエンジニアリング ─ 生成AIの応用」2026 春の第1-2回で使用したスライドです。
Kenji Saito
PRO
April 12, 2026
More Decks by Kenji Saito
See All by Kenji Saito
2026 年度「AI による研究・教育サポート」ワークショップ (教員向け) / AY 2026 “AI-Based Research and Education Support” Workshop (For Faculty)
ks91
PRO
0
37
エージェント化するAI:現在地とその先に起きる変化 / AI as Agents: The Current State and the Changes Ahead
ks91
PRO
0
200
文化的ループ (推論, エージェント) と思考プロセスの検証可能性 / Cultural Loops (Inference, Agents) and the Verifiability of Thought Processes
ks91
PRO
0
47
アカデミーキャンプ2026 夏 「ヤツらを究めろ!ウチらとヤツらの自由研究∼ シーズン2」DAY 3 / Academy Camp 2026 Summer: Master Them! Our Independent Research Project with Them—Season 2 DAY3
ks91
PRO
0
89
アカデミーキャンプ2026 夏 「ヤツらを究めろ!ウチらとヤツらの自由研究∼ シーズン2」DAY 1 / Academy Camp 2026 Summer: Master Them! Our Independent Research Project with Them—Season 2 DAY1
ks91
PRO
0
81
アカデミーキャンプ2026 夏 「ヤツらを究めろ!ウチらとヤツらの自由研究∼ シーズン2」DAY 2 / Academy Camp 2026 Summer: Master Them! Our Independent Research Project with Them—Season 2 DAY2
ks91
PRO
0
110
責任2.0 ∼ AI が主体となる世界で ∼ / Responsibility 2.0: In a World Where AI Takes Responsibilities
ks91
PRO
0
110
発表と総括 / Presentations and Summary
ks91
PRO
0
240
インシデントレスポンス演習II / Incident Response Exercise II
ks91
PRO
0
120
Other Decks in Technology
See All in Technology
V8コントリビュート超入門
riyaamemiya
0
130
データエンジニアリングワークショップ:Auto LoaderとSparkで学ぶデータパイプライン構築
databricksjapan
PRO
0
140
Hub & Spoke 環境のネットワークルーティングを分解してみる
tsuyataku
1
520
2026-09-04 SRE Tech Talk #15 怠惰なTerraform / Lazy Terraform
masasuzu
0
170
Bet AI Day 2026丨AIによって本質に戻るシステムリスク管理
layerx
PRO
0
770
AIエージェントを雇う前に決める5つのこと
knishioka
1
150
振り返りこそエンジニアの本領
negima
0
260
Bet AI Day 2026丨バクラク Autopilot、業務システムの再設計
layerx
PRO
2
1.1k
AI活用の現在地、 ちゃんと見えてますか?/XPfest-2026
visional_engineering_and_design
0
130
Guerilla InnerSource in enterprises, during the AI hype
onenashev
PRO
0
140
Genie Code ワークショップ 基礎編 / Genie-Code-Workshop-fundamental
databricksjapan
PRO
0
330
[RSJ26] Building a VLA Model Based on Self-Distilled Classification
keio_smilab
PRO
0
180
Featured
See All Featured
Design in an AI World
tapps
1
300
Embracing the Ebb and Flow
colly
88
5.2k
We Have a Design System, Now What?
morganepeng
55
8.3k
What’s in a name? Adding method to the madness
productmarketing
PRO
24
4.2k
Gemini Prompt Engineering: Practical Techniques for Tangible AI Outcomes
mfonobong
2
510
Claude Code どこまでも/ Claude Code Everywhere
nwiizo
67
58k
What's in a price? How to price your products and services
michaelherold
247
13k
SEOcharity - Dark patterns in SEO and UX: How to avoid them and build a more ethical web
sarafernandez
0
260
The Art of Programming - Codeland 2020
erikaheidi
57
14k
30 Presentation Tips
portentint
PRO
1
390
Amusing Abliteration
ianozsvald
1
280
Leading Effective Engineering Teams in the AI Era
addyosmani
9
2.5k
Transcript
Generated by Stable Image Core × Nano Banana 2 —
AI 2026 1-2 LLM (WBS : ) 2026 1-2 LLM — 2026-04-13 – p.1/40
https://speakerdeck.com/ks91/collections/prompt-engineering-2026-spring Discord loglm ( ) 2026 1-2 LLM — 2026-04-13
– p.2/40
( ) ( ) ( ) 1993 ( ) 2006
( ) SFC 25 P2P (Peer-to-Peer) 2011 ( ) 2018 2019 2023 AI VR&RPG 2023.5 “Don’t Be So Serious” 2024 AI( ) 2024 AI( ) 2025 (Live and Let Think) 2025 Secret Vision Institute ( ) 2025 GWeeeeeeK !! 2025 SuuuuuuMMeR !! 2026 2026 → ( ) 2026 1-2 LLM — 2026-04-13 – p.3/40
Dropbox Dropbox ( ) ( ) ( ) 2026 1-2
LLM — 2026-04-13 – p.4/40
LLM ( ) — 2026 1-2 LLM — 2026-04-13 –
p.5/40
(LLM : Large Language Model) AI (AGI : Artificial General
Intelligence) ( ) ( . . . ^^;) 2026 1-2 LLM — 2026-04-13 – p.6/40
( ) LLM LLM LLM ⇒ 2026 1-2 LLM —
2026-04-13 – p.7/40
( ) MBA MBA · MBA ↑ MBA 2026 1-2
LLM — 2026-04-13 – p.8/40
( ) 1 4 13 LLM ( ) • 2
4 13 ( ) • 3 4 20 I ( ) 4 4 20 ( ) 5 4 27 6 4 27 7 5 11 8 5 11 9 5 18 II 10 5 18 (1) 11 5 25 (2) 12 5 25 (3) 13 6 1 & 14 6 1 & W-IOI 2026 1-2 LLM — 2026-04-13 – p.9/40
( 20 ) 1 • 2 • 3 (Windows WSL
) • 4 (macOS Lima ) • 5 (macOS ) • 6 • 7 8 9 RPG 10 “September 12th” 11 12 13 ∼ 14 AGI (Artificial General Intelligence) 7 (4/27 ) / (2 ) OK 3,4,5 3 / 2026 1-2 LLM — 2026-04-13 – p.10/40
. . . . . . ( ) ( 20
×14 ) 3,4,5 WBS 2026 1-2 LLM — 2026-04-13 – p.11/40
(2 )(160 ) (10∼20 ) and/or I (60∼80 ) II
(60∼80 ) 2026 1-2 LLM — 2026-04-13 – p.12/40
Moodle ( Q&A ) ( ) Discord ( ) (
) ← ↑ ( ) 2026 1-2 LLM — 2026-04-13 – p.13/40
= A ( ) B ( ) A A (
) 2026 1-2 LLM — 2026-04-13 – p.14/40
Discord discord-agent-hub AI ( ) AI ChatGPT Plus Claude Gemini
3,4,5 2026 1-2 LLM — 2026-04-13 – p.15/40
Discord web (?) 80 ⇒ 2 ( ) ( )
2026 1-2 LLM — 2026-04-13 – p.16/40
LLM ( ) . . . ( ) 2026 1-2
LLM — 2026-04-13 – p.17/40
B. , , 1968 = ⇒ ( ) ( )
AI ( ) 2026 1-2 LLM — 2026-04-13 – p.18/40
: ⇒ = 30 ( 8 ). . . (
) 30 ⇒ + LLM (2023 3 ) + LLM ( ) 2026 1-2 LLM — 2026-04-13 – p.19/40
m P(w1 , . . . , w m )
(Wikipedia) 1 (Wikipedia) : ( ) ← (Generative Pre-training) : ( ) ( ) https://writings.stephenwolfram.com/2023/02/what-is-chatgpt-doing-and-why-does-it-work/ 2026 1-2 LLM — 2026-04-13 – p.20/40
ChatGPT GPT ChatGPT GPT OpenAI (GPT-5.4, GPT-5.2, etc.) GPT Generative
Pre-trained Transformer ( ) GPT ( ) GPT-3.5 2026 1-2 LLM — 2026-04-13 – p.21/40
GPT 2026 1-2 LLM — 2026-04-13 – p.22/40
ChatGPT ChatGPT ← LLM LLM → 200 ( ) 2026
1-2 LLM — 2026-04-13 – p.23/40
ChatGPT (5.3) Claude (Sonnet 4.6 ) Claude (Opus 4.6 )
Gemini 3 ( ) Google ( ) Perplexity Grok 2026 1-2 LLM — 2026-04-13 – p.24/40
AI 2026 1-2 LLM — 2026-04-13 – p.25/40
( ) GPT 3 1. (GPT ) 2. ( )
3. ( ) 3 (↑ 3. ) 3. (GPT ) 2026 1-2 LLM — 2026-04-13 – p.26/40
AI : https://github.com/ks91/loglm https://github.com/ks91/gamer-pat 2026 1-2 LLM — 2026-04-13 –
p.27/40
( ) : https://diamond.jp/articles/-/360966 ( ) : https://diamond.jp/articles/-/360967 AGI 2026
1-2 LLM — 2026-04-13 – p.28/40
AI ࣗಈԽ γϛϡϨʔγϣϯ ਓؒͷ֦ு ϓϩϯϓτ ΤϯδχΞϦϯά ੜ "* ˞͔͋ͨ࿑ಇऀΛ૿͢ ˞ਓؒʹ࡞༻্ͤ͠͞Δ
˞ϞσϧͰ༧ଌ͠౷ޚ͢Δ େͷ Φʔτϝʔγϣϯ খɾےͷ Φʔτϝʔγϣϯ ਓମͷ࡞༻ σδλϧϨΠόʔ σδλϧπΠϯ αΠόʔϑΟδΧϧ ػցֶश ༧ҩྍ γϦΞεήʔϜ Ϗοάσʔλ ςϨΤάδελϯε ϦϞʔτϫʔΫ Ծɾ֦ுݱ࣮ Y3 ࠶ੜҩྍ #.* ৽ༀ ԕִҩྍ ϩϘςΟΫε ݴ༿Ͱ "* ͱର͠ɺΰʔϧͷ ୡʹ͔͏ͨΊͷϓϩϯϓτ ΤϯδχΞϦϯά͕ɺͯ͢ͷ جૅʹҐஔ͚ͮΒΕ͍ͯ͘ ڥٕज़ "* લఏࣾձʹ͓͍ͯɺ ؾީมಈɺਓؒͷपΓͷ ॏେͳʹରͯ͠ ࣗಈԽͱγϛϡϨʔγϣϯ ʹΑཱ͔͍ͬͯͪͬͯ͘ "* લఏࣾձʹ͓͍ͯɺ ΤωϧΪʔ৯ྉΛ͡Ίͱ͢Δ ੜ࢈ྲྀ௨ͷࣗಈԽɺ ৗੜ׆ͷϩϘςΟΫεʹΑΔࢧԉ͕ ࢝·͍͖ͬͯɺ·ͨɺ ϗϫΠτΧϥʔͷ࿑ಇऀ͕ैࣄ͢Δ ࣄͷଟ͘σδλϧϨΠόʔͷํ͕ ͏·͘ΕΔΑ͏ʹͳ͍ͬͯ͘ "* લఏࣾձʹ͓͍ͯɺ ݸਓ͕Τϯύϫʔ͞Εɺ ਓͻͱΓ͕Ͱ͖Δ͜ͱ͕ ֦େ͍ͯ͘͜͠ͱʹՃ͑ɺ ݈߁্͕͠ɺ ߴྸԽΛՃͤͭͭ͞ ࣗಈԽʹΑΓ શһ͕࿑ಇਓޱ͔Β ୀग़Λ࢝Ί͍ͯ͘ "* લఏࣾձʹ͓͍ͯɺ ͔͋ͨશਓྨͷ ༧ଌɾੳೳྗ্͕͢Δ 2026 1-2 LLM — 2026-04-13 – p.29/40
AGI AGI : Artificial General Intelligence ( ) ( )
( AGI ) [ AGI] “Star Trek: The Next Generation” ( ) (1) ( ) ( ) Jean-Luc Picard “Computer, tea. Earl Grey. Hot.” ( ) (2) ( ) ( ) ( ) (1) 2026 1-2 LLM — 2026-04-13 – p.30/40
AGI 5 (OpenAI ) 1 (ChatGPT) 2 (o1, o3) 3
( ) (> deep research) (Codex) (A ) 4 ( ) (X ) 5 (AGI ) 3 5 A X 2026 1-2 LLM — 2026-04-13 – p.31/40
A X — A M. A /X MIT 2013 LEGO
X A A X A X X X (2013 ) ⇒ 10 A Codex Claude Code A . . . 2026 1-2 LLM — 2026-04-13 – p.32/40
(Amodei, 2024) 5 (AGI ) 1. AI 10 21 5∼10
2 2. AI 3. AI AI 4. 5. (AI ) 2026 1-2 LLM — 2026-04-13 – p.33/40
(AI) ( ) AI ( AI ) → ( )
→ ( ) AI ( ) GPT-5 Codex OpenAI ( ) ( ) ( ) ( ) >AI ×AI 2026 1-2 LLM — 2026-04-13 – p.34/40
( ) 6 15 ( ) — ( fix )
2026 1-2 LLM — 2026-04-13 – p.35/40
1 : → ( ) ( ) (AI ) 2026
1-2 LLM — 2026-04-13 – p.36/40
2 : ( ) ID team< >- : team3-interpreter (
) 2026 1-2 LLM — 2026-04-13 – p.37/40
2026 1-2 LLM — 2026-04-13 – p.38/40
1. (1) AI (2) 2026 4 16 ( ) 23:59
JST ( ) Waseda Moodle 2026 1-2 LLM — 2026-04-13 – p.39/40
2026 1-2 LLM — 2026-04-13 – p.40/40