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
파급효과: From AI to Android Development
Search
HyunWoo Lee
April 28, 2025
Programming
0
200
파급효과: From AI to Android Development
Build with AI on Android 2025에서 진행한 파급효과: From AI to Android Development의 Speaker Deck입니다.
HyunWoo Lee
April 28, 2025
Tweet
Share
More Decks by HyunWoo Lee
See All by HyunWoo Lee
선언형 UI에서의 상태관리
l2hyunwoo
0
480
선언형 UI를 학습할 때 알아둬야하는 키워드들
l2hyunwoo
0
410
Essential concepts to know when learning Declarative UI
l2hyunwoo
2
1.3k
React Native under the hood
l2hyunwoo
0
110
유연한 Composable 설계
l2hyunwoo
0
660
KotlinConf 2024 Global in South Korea Keynote
l2hyunwoo
0
120
TextField 씹고 뜯고 맛보고 즐기고
l2hyunwoo
0
420
CDG로 모두와 함께 성장하기
l2hyunwoo
0
190
fun HelloKMP(): GladToMeetYou
l2hyunwoo
0
450
Other Decks in Programming
See All in Programming
NPOでのDevinの活用
codeforeveryone
0
900
なぜ「共通化」を考え、失敗を繰り返すのか
rinchoku
1
680
テスターからテストエンジニアへ ~新米テストエンジニアが歩んだ9ヶ月振り返り~
non0113
2
220
生成AI時代のコンポーネントライブラリの作り方
touyou
1
290
Android 16KBページサイズ対応をはじめからていねいに
mine2424
0
450
AI コーディングエージェントの時代へ:JetBrains が描く開発の未来
masaruhr
1
200
The Niche of CDK Grant オブジェクトって何者?/the-niche-of-cdk-what-isgrant-object
hassaku63
1
620
顧客の画像データをテラバイト単位で配信する 画像サーバを WebP にした際に起こった課題と その対応策 ~継続的な取り組みを添えて~
takutakahashi
4
1.3k
MCPを使ってイベントソーシングのAIコーディングを効率化する / Streamlining Event Sourcing AI Coding with MCP
tomohisa
0
170
「テストは愚直&&網羅的に書くほどよい」という誤解 / Test Smarter, Not Harder
munetoshi
0
200
ペアプロ × 生成AI 現場での実践と課題について / generative-ai-in-pair-programming
codmoninc
2
21k
“いい感じ“な定量評価を求めて - Four Keysとアウトカムの間の探求 -
nealle
2
12k
Featured
See All Featured
Java REST API Framework Comparison - PWX 2021
mraible
31
8.7k
The Power of CSS Pseudo Elements
geoffreycrofte
77
5.9k
Done Done
chrislema
184
16k
Helping Users Find Their Own Way: Creating Modern Search Experiences
danielanewman
29
2.7k
XXLCSS - How to scale CSS and keep your sanity
sugarenia
248
1.3M
Learning to Love Humans: Emotional Interface Design
aarron
273
40k
Faster Mobile Websites
deanohume
308
31k
Creating an realtime collaboration tool: Agile Flush - .NET Oxford
marcduiker
30
2.2k
RailsConf 2023
tenderlove
30
1.1k
Building a Scalable Design System with Sketch
lauravandoore
462
33k
Statistics for Hackers
jakevdp
799
220k
A Tale of Four Properties
chriscoyier
160
23k
Transcript
파급효과 : From AI to Android Development in 2025 HyunWoo
Lee GDG Korea Android Korea Android
ഄݺ: ޖ ߸ೞחо? Chapter One Revolution: What’s new in our
lives?
AI → ഄݺ
ੋܨ ޙݺ ߊী ೠ ೱਸ ՙ 3о ഄݺਸ ੌஸ •
ઁ 1 ޛѾ: ֪স ഄݺ • ઁ 2 ޛѾ: স ഄݺ • ઁ 3 ޛѾ: ࠁച ഄݺ ޛѾ ۿ(চ࠼ ష۞)
п ޛѾ ੋܨ ޙݺীࢲ “ҕә ߑध”ਸ ഄनदఅ ࢎѤ → ޛѾ
աт ٸ݃ ੋܨ ޙݺ ࢤઓ ߑध ׳ۄѱ ؽ ޛѾ ۿ(চ࠼ ష۞)
সഄݺ ૐӝӝҙ ߊݺਵ۽ ઁಿਸ ࢤࣘ بী ഄनਸ Ҋ ҕઁ ࢤߑधਸ
ٜ݅য ੋܨ ࣻਃܳ ֈযࢲח ઁಿ (/ध/)ٜਸ ࢤೡ ࣻ ѱ ؽ ࠄী ೠ न҅ә҅(ࠗܰই/ ܀ۨఋܻই) ࢤ, ઁҴ द ߊ ֪সഄݺ ࣻ۵/ ӝ߈ ࢤઓߑधী ࢲ ػ షীࢲ ֪ࢎܳ য “ݡѢܻ”ܳ ࢤೞৈ উੋ ࢤ ઓਸ যաт ࣻ ѱ ೣ ੋܨࢎഥ ҅әઁ/ӝୡੋ ҃ઁ ѐ֛ ܳ ӝ߈ਵ۽ ഋࢿ ٣ణഄݺ җ ਘ٘ ৬٘ ਢ(WWW)ী ӝ߈ೞৈ ੋܨо ࠁܳ ࢤೞҊ ҕਬೞח ߑध ഄनਸ оೣ. ܳ ӝ߈ೠ স زച, ޖੋചо द ؽ ٜب ࣚऔѱ ࠁܳ ҕਬ߉ই ߊী ੋ ೱ, ࢜ ۽ ࢤഝߑध(Mobility)ী ݏח ࣻਃী ೠ ҕә ହؽ п ޛѾ “әബҗ”
AIח যڃ Ѫਸ ഄनदఃחо?
Chapter Two 4th Industrial Revolution: What’s AI Innovating? 4ର সഄݺ:
AIח ޖਸ ഄनदఃחо?
Klaus Schwab …۠ ӝٜࣿ ࣻभর ݺ ࢎۈٜਸ ҅ࣘ೧ࢲ ਢী ো
ѾೞҊ ࠺ૉפझ ߂ ઑ ബਯࢿਸ ദӝਵ۽ ೱ࢚दఃݴ ؊ ա ҙܻܳ ా೧ ো ജ҃ਸ ࢤೡ ࣻ ח ழۆ ਫ਼۱ਸ оҊ .
AI Agent based on LLM(Large Language Model) (Gemini 2.5, o4..)
Editable Location
Mark Zuckerberg Meta/CEO “2025֙҃ীח ݫఋীࢲ ח Middle-level Engineerۢ ٘ܳ
ࢿೡ ࣻ ח AIܳ ѐߊೡ Ѫੑפ.”
Service based on AI Editable Location
Service based on AI Editable Location
• ա݅ ӝࣿ࠶۽Ӓܳ ٜ݅Ҋ रয, ӭՔೞҊ ࣁ۲ػ ו՝, • ӝࠄਵ۽
ݽ٘о ഝࢿച غযযঠ ೞҊ ਃदী ۄ ݽ٘۽ ജೡ ࣻ ח షӖ ݽ٘о ਵݶ જѷয. • ӒܻҊ కӒ߹۽ ѱदޛٜਸ ࠅ ࣻ ח ӝמٜҗ Ѩ࢝ ӝמ ୶оغযঠ ೧. ѱदӖ ݃(md)ܳ ਗ೧ঠظ. • ղ ࣗѐ ಕܳ ٮ۽ ٜ݅ਵݶ જѷয. • ղ ࣗѐ ಕীח ܴ/۽ ࢎ/۱/োۅ৬ ࣗ࣍٣য ݂ /ನಫܻয় ݂ܳ ৢܾ ࣻ যঠ ظ. ӝࣿ ࠶۽Ӓ ઁ ௪ܻ
ӝࣿ ࠶۽Ӓ ઁ ௪ܻ
• ӝמ ҳഅী ೠ ࣻ ௪ܻ • दೞח Ӗਸ ־ܰݶ
࢚ࣁ ಕ۽ ֈযоঠೞחؘ ಕо উٜ݅য Ѫ эই. ࢚ࣁಕীח ઁݾ ݢ ࠁৈҊ Ӓ റী ղਊ ӒܻҊ కӒ ٜ ࠁৈҊ ؆Ӗө ׳ ࣻ যঠ ظ • അ ѱदӖ ࢚ࣁ ಕীࢲ ؆Ӗ ࢿ ӝמ ҳഅغয ঋҊ पઁ ؘ ఠо োزغয ঋ Ѫ эই. ೧ ӝמٜਸ ୶о೧ ӝࣿ ࠶۽Ӓ ઁ ௪ܻ
٬ ో ഝਊ ۽ࣁझ Debugging ޙઁо ߊࢤೞݶ زਵ۽ ী۞ܳ х
೧ ী۞ܳ ೧Ѿೡ ࣻ ח ٘ ܳ ࢤࢿ/ࣻ Project init ӝמ झಖ ޙࢲ/ೖӒ݃ ੌ ࠗೞ ৈ ۽ં ࢤࢿ(MVP) Maintenance ӝઓ ۽ંী ࢜۽ ӝמٜਸ ୶о/ࣻೞח ١ ਬࠁࣻب ೡ ࣻ
…۠ ӝٜࣿ ࣻभর ݺ ࢎۈٜਸ ҅ࣘ೧ࢲ ਢী ো ѾೞҊ ࠺ૉפझ
߂ ઑ ബਯࢿਸ ദӝਵ۽ ೱ࢚दఃݴ ؊ ա ҙܻܳ ా೧ ো ജ҃ਸ ࢤೡ ࣻ ח ழۆ ਫ਼۱ਸ оҊ .
…۠ ӝٜࣿਸ ഝਊೞৈ ӝઓ ੋ۱ٜਸ AI۽ ೞৈ ӝসٜ ࠺ਊ ബਯࢿਸ
ദӝਵ۽ ೱ࢚दఃݴ ؊ ա ҙܻܳ ా೧ ӝস ܲ ࠙ঠী ࠄਸ ై ೡࣻ ب۾ ೠ.
әѺ ߊೞח ؘఠ ࢤࣘب ӝਗ 3,000֙ࠗఠ 5,000֙ زউ ࢤػ ؘఠо
20 ুࢎ߄ (2ୌ݅ పۄ߄) Ҋ 2000֙ ୡ߈ࠗఠ 2021֙ө ࢤػ ؘఠ ୨ 50 ઁఋ߄ (5݅ ুࢎ߄ ).
2500ߓ ؘఠ ࢤࣘب ౠ • 20ৈ֙ زউ ࢤػ ؘఠо 5,000
֙р ؘఠ 2,500ߓܳ ֈযࢲח Ѫ. • അ ੋܨо о ؘఠ ড 90%ח դ 10֙р ࢤعਸ ೠ.
ӝࣿ ౠ ӝࣿ ೦ҳೠ оࣘ ߊਵ۽ ੋ೧ ੋܨ ࢎীח োਵ۽
ౠ ߊࢤ ೡ Ѫݴ, Ӓ റ ੋܨ ࢎח ӘԈ যઉৡ Ѫҗח ഃ ܲ ޖоо ؼ Ѫ.
1. AIо ѐߊܳ ೞח Ѣ ইקө? 2. ࠁ নী
ౠ ৡݶ ੋܨח যڌ ѱ ݆ࣻ ࠁٜਸ ٮۄт Ѫੋо? فо ޙ
Chapter Three How can I survive in the AI era?
AI ࣁ҅ীࢲ ইթӝ
None
1. ߨ गٜ(଼ ࣗ) 2. ತࣧੋ ࠁٜ AIо ݏٯڰܻח ف
೧
1. Closed source ۽؋(ਬܐ SDK) 2.۽ં ղࠗ Context/بݫੋ ۽ ತࣧ(?)ੋ
ࠁٜ
AIо ࢤࢿೠ ؘఠ۽ णೞח ҃ Nࣁо դ റ Ѿҗޛ ജп
࠼ب, ؘఠ ࢿ ڄযח അ࢚ ߊࢤ ӈ (झࠗܰ AI)
• ѾҴ ੋр AIо ٜ݅যղח ٘ী ೧ࢲ ೧ܳ ೡ
ঌইঠ ೠ • AIܳ ؊ ੜ ഝਊೞӝ ਤ೧ࢲۄب ѐߊٜ ҅ࣘ೧ࢲ ࢜۽ ٘ ٜਸ ࢿ೧ঠ݅ ೠ. • AIܳ ഝਊೡ ٸ ਸ ࣻ ח ࢤࢿ ӝࠄਵ۽ оઉоঠ ೠ. ࢚ടٜਸ ഝਊ೧ࠁ
ѾҴূ ѐߊ ਤ࢚ • ӝઓী ӝמٜਸ ੜ ѐߊೞҊ ࠄੋ ࢿೠ
٘о যڌѱ زೞח ೧ܳ ೞҊ ؍ ٜ࠙ ӒܻҊ ޙઁо ߊࢤ೮ਸ ٸ ਗੋਸ ইղ प ࣻ ח ࠙दۄݶ ఋѺ হਸ Ѫ э • ݅ Ӓ۞ ޅೞ࣑ݶ, AIо ࢿೠ ٘ܳ ഝਊೡ ٸ Ӓ ٜ٘ਸ ೧ܳ ೡ ࣻ ח ী ೧ࢲ झझ۽ ࢤпਸ ೞ࣊ঠ פ. • AIо ࢿೠ ٜ٘ झझ۽ ࢿೡ ই࣊ঠ ೞҊ ؊ աইо য ڌѱ ѐࢶೡ ࣻ ח , ࢎ٘ ಖо աח ঋחө ࢤп ਸ ೞप ই࣊ঠ פ.
ߧۈೞח ࠁ әܨܳ ఋҊ ݁ೠ ѐߊ ৈਸ ٜ݅য աт ࣻ
ب۾
Index Reference
• ઁ 3 ޛѾ, চ࠼ ష۞ • ۄझ गߏ ઁ
4ର সഄݺ, ۄझ गߏ • Lovable • [ݽفܳ ਤೠ ੋҕמ] 5. ܻীѱ ‘ؘఠ’ח যڃ ੋо • The curse of recursion: Training on Generated Data Makes Models Forget Reference
Thank You HyunWoo Lee Viva Republica(Toss) Korea Android