uz
Feedback
Continuous Learning_Startup & Investment

Continuous Learning_Startup & Investment

Kanalga Telegramโ€™da oโ€˜tish

We journey together through the captivating realms of entrepreneurship, investment, life, and technology. This is my chronicle of exploration, where I capture and share the lessons that shape our world. Join us and let's never stop learning!

Ko'proq ko'rsatish
2 299
Obunachilar
-224 soatlar
-147 kunlar
-4130 kunlar
Postlar arxiv
์„ฑ๊ณต ํ™•๋ฅ ์„ ๋†’์ด๋Š” ๋ฐฉ๋ฒ•์€ ๋ฌด์—‡์ธ๊ฐ€์š”? ์ €์˜ ํฐ ์žฅ์  ์ค‘ ํ•˜๋‚˜๋Š” ๊ธฐ๋Œ€์น˜๊ฐ€ ๋งค์šฐ ๋‚ฎ๋‹ค๋Š” ๊ฒƒ์ž…๋‹ˆ๋‹ค. ๋Œ€๋ถ€๋ถ„์˜ ์Šคํƒ ํฌ๋“œ ์กธ์—…์ƒ๋“ค์€ ๋ช…๋ฌธ ํ•™๊ต ์ถœ์‹ ์ด๊ธฐ ๋•Œ๋ฌธ์— ๊ธฐ๋Œ€์น˜๊ฐ€ ๋งค์šฐ ๋†’์œผ๋ฉฐ, ๋‹น์—ฐํžˆ ๊ทธ๋Ÿด ๋งŒํ•ฉ๋‹ˆ๋‹ค. ๊ทธ๋“ค์€ ๋งค์šฐ ์„ฑ๊ณต์ ์ด์—ˆ๊ณ , ํ•™๊ธ‰์—์„œ ์ตœ์ƒ์œ„๊ถŒ์„ ์ฐจ์ง€ํ–ˆ์œผ๋ฉฐ, ํ•™๋น„๋ฅผ ๊ฐ๋‹นํ•  ์ˆ˜ ์žˆ์—ˆ์Šต๋‹ˆ๋‹ค. ์ง€๊ตฌ์ƒ์—์„œ ๊ฐ€์žฅ ํ›Œ๋ฅญํ•œ ๊ต์œก ๊ธฐ๊ด€ ์ค‘ ํ•˜๋‚˜๋ฅผ ์กธ์—…ํ•˜๊ณ  ๋‹ค๋ฅธ ๋›ฐ์–ด๋‚œ ํ•™์ƒ๋“ค์— ๋‘˜๋Ÿฌ์‹ธ์—ฌ ์กธ์—…ํ•˜๋ฉด ์ž์—ฐ์Šค๋Ÿฝ๊ฒŒ ๊ธฐ๋Œ€์น˜๊ฐ€ ๋†’์•„์ง‘๋‹ˆ๋‹ค. ํ•˜์ง€๋งŒ ๊ธฐ๋Œ€์น˜๊ฐ€ ๋†’์€ ์‚ฌ๋žŒ์€ ์„ฑ๊ณต์— ํ•„์ˆ˜์ ์ธ ํšŒ๋ณตํƒ„๋ ฅ์„ฑ์ด ๋‚ฎ์€ ๊ฒฝ์šฐ๊ฐ€ ๋งŽ์Šต๋‹ˆ๋‹ค. ๊ณ ํ†ต์„ ๊ฒฝํ—˜ํ•˜์ง€ ์•Š๊ณ ์„œ๋Š” ํšŒ๋ณตํƒ„๋ ฅ์„ฑ์„ ๊ฐ€๋ฅด์น˜๊ธฐ๊ฐ€ ์–ด๋ ต์Šต๋‹ˆ๋‹ค. ์ €๋Š” ์šด์ด ์ข‹๊ฒŒ๋„ ์„ฑ๊ณต์˜ ์กฐ๊ฑด์„ ์ œ๊ณตํ•œ ๋ถ€๋ชจ๋‹˜ ๋ฐ‘์—์„œ ์ž๋ž์ง€๋งŒ ๋งŽ์€ ์ขŒ์ ˆ๊ณผ ๊ณ ํ†ต์— ์ง๋ฉดํ•˜๊ธฐ๋„ ํ–ˆ์Šต๋‹ˆ๋‹ค. ์šฐ๋ฆฌ ํšŒ์‚ฌ์—์„œ๋Š” ๊ณ ํ†ต๊ณผ ์•„ํ””์ด ์ธ์„ฑ์„ ๋‹ค๋“ฌ๋Š” ๋ฐ ๋„์›€์ด ๋˜๊ธฐ ๋•Œ๋ฌธ์— ๊ณ ํ†ต๊ณผ ์•„ํ””์˜ ๊ฐœ๋…์„ ๋ฐ›์•„๋“ค์ž…๋‹ˆ๋‹ค. ์œ„๋Œ€ํ•จ์€ ์ง€๋Šฅ์ด ์•„๋‹ˆ๋ผ ์ธ์„ฑ์—์„œ ๋‚˜์˜ค๋ฉฐ, ์ธ์„ฑ์€ ๊ณ ํ†ต์„ ํ†ตํ•ด ๋งŒ๋“ค์–ด์ง‘๋‹ˆ๋‹ค. ๋ชจ๋“  ์Šคํƒ ํผ๋“œ ํ•™์ƒ ์—ฌ๋Ÿฌ๋ถ„, ํ•ด๋ฅผ ๋ผ์น˜๋Š” ๊ฒƒ์ด ์•„๋‹ˆ๋ผ ํšŒ๋ณต๋ ฅ๊ณผ ์ธ์„ฑ์„ ํ‚ค์šฐ๊ธฐ ์œ„ํ•ด ์ถฉ๋ถ„ํ•œ ๊ณ ํ†ต๊ณผ ๊ณ ํ†ต์„ ๊ฒช๊ธฐ๋ฅผ ๋ฐ”๋ž๋‹ˆ๋‹ค. ์ง์›๋“ค์˜ ๋™๊ธฐ ๋ถ€์—ฌ์™€ ๊ด€๋ จํ•˜์—ฌ ์ €๋Š” 55๋ช…์œผ๋กœ ๊ตฌ์„ฑ๋œ ๊ด€๋ฆฌํŒ€์— ๋‘˜๋Ÿฌ์‹ธ์—ฌ ์žˆ์Šต๋‹ˆ๋‹ค. ์ €๋Š” ์ง์›๋“ค์„ ์œ„ํ•ด ๋ฆฌ๋ทฐ๋ฅผ ์ž‘์„ฑํ•˜์ง€๋Š” ์•Š์ง€๋งŒ, ์ง์›๋“ค์ด ์ €์—๊ฒŒ ํ•˜๋Š” ๊ฒƒ์ฒ˜๋Ÿผ ์ง€์†์ ์ธ ํ”ผ๋“œ๋ฐฑ์„ ์ œ๊ณตํ•ฉ๋‹ˆ๋‹ค. ์ง์›๋“ค์˜ ๋ณด์ƒ์€ ๊ฐ„๋‹จํ•˜๋ฉฐ ๋†€๋ž๊ฒŒ๋„ ๋งŽ์€ ์ง์›์ด ๋˜‘๊ฐ™์€ ๋ณด์ˆ˜๋ฅผ ๋ฐ›๊ณ  ์žˆ์–ด ์šฐ๋ฆฌ์—๊ฒŒ ์ž˜ ๋งž์Šต๋‹ˆ๋‹ค. ์ €๋Š” ๊ผญ ํ•„์š”ํ•œ ๊ฒฝ์šฐ๊ฐ€ ์•„๋‹ˆ๋ฉด ์ผ๋Œ€์ผ ๋ฏธํŒ…์„ ํ•˜์ง€ ์•Š์œผ๋ฉฐ, ์ผ๋ถ€ ์†Œ์ˆ˜์—๊ฒŒ๋งŒ ๋น„๋ฐ€ ์ •๋ณด๋ฅผ ๊ณต์œ ํ•˜์ง€ ์•Š์Šต๋‹ˆ๋‹ค. ์šฐ๋ฆฌ ํšŒ์‚ฌ๋Š” ๋ฏผ์ฒฉ์„ฑ์„ ์œ„ํ•ด ์„ค๊ณ„๋˜์—ˆ์œผ๋ฉฐ, ์ •๋ณด๋Š” ๋น ๋ฅด๊ฒŒ ํ๋ฅด๊ณ  ์‚ฌ๋žŒ๋“ค์€ ์ž์‹ ์ด ์•„๋Š” ๊ฒƒ๋ฟ๋งŒ ์•„๋‹ˆ๋ผ ์ž์‹ ์˜ ๋Šฅ๋ ฅ์— ๋”ฐ๋ผ ๊ถŒํ•œ์„ ๋ถ€์—ฌ๋ฐ›์Šต๋‹ˆ๋‹ค. ์ €์˜ ํ–‰๋™์€ ์„ฑ๊ณต์„ ์ถ•ํ•˜ํ•˜๊ณ  ์‹คํŒจ๋ฅผ ์ฒ˜๋ฆฌํ•˜๋Š” ๋ฐฉ์‹์— ๋Œ€ํ•œ ๋ถ„์œ„๊ธฐ๋ฅผ ์กฐ์„ฑํ•˜๋ฉฐ ํšŒ์‚ฌ์˜ ๋ฌธํ™”์™€ ๊ฐ€์น˜๋ฅผ ์ง€์†์ ์œผ๋กœ ์ฃผ์ž…ํ•ฉ๋‹ˆ๋‹ค. ๋งค์ผ ์ค‘์š”ํ•œ ๊ฒƒ์„ ๊ฐ•ํ™”ํ•˜๊ณ  ๋ฌด์—‡์ด ์ข‹์€ ์„ฑ๊ณผ๋ฅผ ๊ตฌ์„ฑํ•˜๋Š”์ง€๋ฅผ ์ •์˜ํ•  ์ˆ˜ ์žˆ๋Š” ๊ธฐํšŒ๊ฐ€ ์ฃผ์–ด์ง‘๋‹ˆ๋‹ค.

How to improve the chances of success? One of my my great advantages is that I have very low expectations. Most Stanford graduates have very high expectations, and rightfully so, because they come from an esteemed school. Theyโ€™ve been highly successful, at the top of their class, and were able to afford tuition. Graduating from one of the finest institutions on the planet, surrounded by other incredible students, naturally sets their expectations high. However, people with high expectations often have low resilience, which is crucial for success. Itโ€™s challenging to teach resilience except through experiencing suffering. I was lucky to grow up with parents who provided conditions for success, but also faced plenty of setbacks and suffering. In our company, we embrace the concept of pain and suffering because it helps refine our character. Greatness comes from character, not intelligence, and character is forged through suffering. For all Stanford students, I wish you ample doses of pain and suffering, not to cause harm, but to build resilience and character. Regarding keeping employees motivated, Iโ€™m surrounded by a management team of 55 people. I donโ€™t write reviews for them but offer constant feedback, as they do for me. Their compensation is straightforward, and surprisingly, many are paid exactly the same, which works well for us. I donโ€™t have one-on-one meetings unless necessary, and thereโ€™s no secret information shared with only a select few. Our company is designed for agility, with information flowing quickly and people empowered by their abilities, not just what they know. My behavior sets the tone for how we celebrate success and handle failure, continually instilling the companyโ€™s culture and values. Every day presents opportunities to reinforce whatโ€™s important and define what constitutes good performance.

Andrew Yeung didnโ€™t know ANYONE when he moved to the U.S. 3 years ago, so he started organizing "tech parties". Today, he's built a community of 25,000 founders and investors via his events. Hereโ€™s exactly how he built his network from scratch - and how you can too ๐Ÿ‘‡ Before we dive in... Quick backstory: - Andrew Yeung was born and raised in China - When he turned 18, he moved to Canada for college - He landed his first job at the "AT&T of Canada" - Worked in telecommunications for 3 years - He felt stuck in his career when the pandemic hit - He sent 1000s of DMs to interesting people - 150 of them agreed to a "virtual coffee chat" - Within 6 months, he landed his first job in tech - He fulfills his lifelong dream of moving to NYC - Andrew becomes well-known for his monthly "tech parties" - 3 years later, he has connected 25,000 people in tech 3 steps to grow your network: Step 1: Don't be afraid to send a "cold DM" Andrew always knew he wanted to live in New York City one day, but he had no idea how to "break into tech". He sent thousands of direct messages to strangers on Linkedin, Reddit, and Fishbowl. 150 of them agreed to a "virtual coffee chat". 6 months later, he landed his first job in tech!!! Want to do the same but don't know how? Here's his framework for cold DMs: โœ”๏ธ Make it relevant โœ”๏ธ Introduce yourself โœ”๏ธ Demonstrate value โœ”๏ธ Gratitude & praise โœ”๏ธ Make a clear ask Step 2: Organize your own "meet-up" Today, Andrew organizes 1,000-person tech events in every large city in the United States, but he didn't start out that way. His advice? Start small. Identify 6-8 people with a common interest. Ask them if they'd be interested in meeting up. Organize a small meet-up at a bar, restaurant, or your home. Ask everyone if they'd be interested in another event in the future. Repeat. Step 3: Attend networking events - even if you feel awkward at first 3 tips to avoid feeling "cringe" ๐Ÿ‘‡ 1) Go on Twitter, Reddit, or Linkedin to find people who are also going to the event. Try to organize a smaller meet-up before the actual event. 2) Be curious to be interesting. Some people believe that by namedropping and talking about themselves, theyโ€™ll convince others that they are interesting people. That rarely works. Instead, be interested in others. Ask questions and listen more than you talk. 3) Stand in areas that help spark a conversation - introduce yourself during the elevator ride up, stand by the bar, or join an existing circle. It can be nerve-racking to talk to a stranger, but remember that everyone is there for the same goal: to meet new people. Building a community? You can now book Andrew Yeung on Intro.

Today we are thrilled to share that weโ€™ve raised $106M in a new round led by Salesforce Ventures with participation from Coatue and our existing investors. Our vision is to rapidly bring innovations from research to production and to ultimately build the best platform we can for developers, startups, and enterprises to run generative AI applications built on open-source models at production scale. Existing investors participating in the round included Kleiner Perkins, Lux Capital, Emergence Capital, Prosperity7 Ventures, New Enterprise Associates (NEA), Greycroft, Definition, Long Journey Ventures, Factory, Scott Banister, and SVA. We are also thrilled to have participation from industry luminaries including Clem Delangue ๐Ÿค—, CEO of HuggingFace, Soumith Chintala, the creator of PyTorch, and Manu Sharma, CEO of Labelbox. https://lnkd.in/g5UAm-R8

<๋‹น์‹ ์€ ์˜์›…์˜ ์—ฌ์ •์„ ๊ฑท๊ณ  ์žˆ์Šต๋‹ˆ๊นŒ : 30๋ถ„๋งŒ์— ๋‚ด ์ปค๋ฆฌ์–ด๊ฐ€ ๋” ์†Œ์ค‘ํ•ด์ง€๋Š” ๋ฐฉ๋ฒ•> ์ด๋ฒˆ์— ์˜ฌ๋ฆฌ๋Š” ๊ธ€์€ 30๋ถ„๋งŒ์— ์ปค๋ฆฌ์–ด ์˜๋ฏธ๋ฅผ ๋” ๋А๋ผ๊ณ  ์ง๋ฌด ๋งŒ์กฑ๋„๊ฐ€ ์˜ฌ๋ผ๊ฐ€๋Š” ๋น„๋ฒ•์ž…๋‹ˆ๋‹ค. ์šฐ์„  ๋‹ค์Œ์˜ ์งˆ๋ฌธ๋“ค์— ์†”์งํ•˜๊ฒŒ ๋‹ต์„ ํ•ด๋ณด์„ธ์š”. 7์ ์€ ๋งค์šฐ ๊ทธ๋ ‡๋‹ค, 1์ ์€ ์ „ํ˜€ ์•„๋‹ˆ๋‹ค. 1. ๋‚ด ์ปค๋ฆฌ์–ด๋ฅผ ์ƒ๊ฐํ•  ๋•Œ, ๋‚˜๋Š” ๋งค์šฐ ๋ช…ํ™•ํ•œ ๋ชฉํ‘œ์™€ ๋ชฉ์ ์ด ์žˆ๋‹ค. 2. ๋‚ด ์ปค๋ฆฌ์–ด์—์„œ ๋‚˜์˜ ์กด์žฌ๋Š” ๊ทธ ์ž์ฒด๋กœ ๋ชฉ์ ๊ณผ ์˜๋ฏธ๊ฐ€ ์ถฉ๋ถ„ํžˆ ์žˆ๋‹ค. 3. ๋‚˜๋Š” ์ปค๋ฆฌ์–ด์—์„œ ๋ช…ํ™•ํ•œ ๋ชฉํ‘œ์™€ ๋งŒ์กฑํ•  ๋งŒํ•œ ๋ชฉ์ ์„ ๊ฐ–๊ณ  ์žˆ๋‹ค. 4. ๋‚˜๋Š” ๋‚ด ์ปค๋ฆฌ์–ด์—์„œ ์˜๋ฏธ๋‚˜ ๋ชฉ์  ํ˜น์€ ์‚ฌ๋ช…์„ ์ฐพ๋Š” ๋Šฅ๋ ฅ์ด ๋งค์šฐ ์šฐ์ˆ˜ํ•˜๋‹ค๊ณ  ์ž๋ถ€ํ•œ๋‹ค. ํ•œ ์—ฐ๊ตฌ์—์„œ ํ”ผ์‹คํ—˜์ž 450๋ช…์˜ ํ‰๊ท  ์ ์ˆ˜๋Š” ์•ฝ 5์ ์ด์—ˆ์Šต๋‹ˆ๋‹ค. ๋ณธ์ธ์˜ ์ ์ˆ˜๋Š” ์–ด๋–ค๊ฐ€์š”? ๋งŒ์•ฝ ๋ณธ์ธ์˜ ์ ์ˆ˜๊ฐ€ 5์ ์ด ์•ˆ๋œ๋‹ค๋ฉด ์•„๋ž˜ ๊ธ€์„ ๊ผญ ์ง„์ง€ํ•˜๊ฒŒ ์ฝ์–ด๋ณด์‹ค ๊ฑฐ๋ฅผ ๊ถŒํ•ฉ๋‹ˆ๋‹ค. ์ด ์ธก์ •์€ ์Šค์Šค๋กœ ์‚ถ๊ณผ ์ปค๋ฆฌ์–ด์˜ ์˜๋ฏธ๋ฅผ ์–ผ๋งˆ๋‚˜ ๋А๋ผ๋‚˜๋ฅผ ํ‰๊ฐ€ํ•ฉ๋‹ˆ๋‹ค. ์ด ์ ์ˆ˜๋Š” ์ง์ž‘์„ ํ•˜์‹œ๊ฒ ์ง€๋งŒ ๋ณธ์ธ์˜ ์›ฐ๋น™, ์‚ถ์˜ ๋งŒ์กฑ๋„์™€ ํฐ ๊ด€๋ จ์ด ์žˆ์Šต๋‹ˆ๋‹ค. ํฅ๋ฏธ๋กœ์šด ์—ฐ๊ตฌ๊ฐ€ ์žˆ์Šต๋‹ˆ๋‹ค[1]. ๋ณธ์ธ์˜ ์‚ถ์ด ์†Œ์œ„ ์˜ํ™”๋‚˜ ์†Œ์„ค์—์„œ ๋ณด๋Š” ์˜์›…์˜ ์—ฌ์ •๊ณผ ์œ ์‚ฌํ•˜๋‹ค๊ณ  ๋А๋‚„์ˆ˜๋ก ์ด ์‚ถ์˜ ์˜๋ฏธ๊ฐ€ ๋†’์€ ๊ฒ๋‹ˆ๋‹ค. ์•„๋ž˜ ์งˆ๋ฌธ๋“ค์— ๋†’์€ ์ ์ˆ˜๋ฅผ ์ฃผ๋ฉด ์ž์‹ ์˜ ์‚ถ์„ ์˜์›…์˜ ์—ฌ์ •์œผ๋กœ ๋ณธ๋‹ค๋Š” ๋œป์ด ๋ฉ๋‹ˆ๋‹ค. ๋‚˜๋Š” ๋‚ด ์‚ถ์„ ์–ด๋–ค ์Šคํ† ๋ฆฌ๋กœ ๋ณด๋Š” ๊ฒฝ์šฐ๊ฐ€ ์ž์ฃผ ์žˆ๋‹ค. ๋‚ด ์‚ถ์—๋Š” ๋ช…๋ฃŒํ•œ ์„œ์‚ฌ์  ๊ถค์ ์ด ์žˆ๋‹ค. ๋‚ด ์‚ถ์€ ๋ชจํ—˜์œผ๋กœ ๊ฐ€๋“ ์ฐจ ์žˆ๋‹ค. ๋‚˜๋Š” ๋‹ค๋ฅธ ์‚ฌ๋žŒ๋“ค์˜ ์ง€์ง€๋ฅผ ๋ฐ›๋Š”๋‹ค. ๋‚˜๋Š” ์‚ด์•„์˜ค๋ฉด์„œ ์žฅ์• ๋ฌผ์„ ๊ทน๋ณตํ–ˆ์–ด์•ผ ํ–ˆ๋‹ค. ๋‚˜๋Š” ์‹œ๊ฐ„์ด ์ง€๋‚˜๋ฉด์„œ ํ•œ ์‚ฌ๋žŒ์œผ๋กœ์„œ ์„ฑ์žฅํ•ด ์™”๋‹ค. ๋‚˜๋Š” ๋‚จ๋“ค์—๊ฒŒ ์ง€์†์ ์ธ ์˜ํ–ฅ์„ ๋‚จ๊ธธ ๊ฒƒ์ด๋‹ค. ์กฐ์…‰ ์บ ๋ฒจ์ด๋ผ๋Š” ์‹ ํ™”ํ•™์ž๋Š” <์ฒœ์˜ ์–ผ๊ตด์„ ๊ฐ€์ง„ ์˜์›…>์ด๋ผ๋Š” ์ฑ… ๋“ฑ์„ ํ†ตํ•ด ์—ฌ๋Ÿฌ ๋ฌธํ™”๊ถŒ์˜ ์˜์›… ์‹ ํ™”๊ฐ€ ๊ณตํ†ต๋œ ์›ํ˜•์„ ๊ฐ–๊ณ  ์žˆ๋‹ค๊ณ  ์ฃผ์žฅํ–ˆ์Šต๋‹ˆ๋‹ค. ์˜ˆ๋ฅผ ๋“ค์–ด ์ดˆ๊ธฐ์— ์ฃผ์ธ๊ณต์ด ์†Œ๋ช…์„ ๊ฑฐ๋ถ€ํ•˜๊ฑฐ๋‚˜ ํ˜น์€ ์Šค์Šน์„ ๋งŒ๋‚˜๊ฑฐ๋‚˜, ๋˜๋Š” ์‹œ๋ จ์„ ๊ฒช๊ณ  ๊ฐ์„ฑ์„ ํ•œ๋‹ค๋“ ๊ฐ€ ํ•˜๋Š”. ์šฐ๋ฆฌ๊ฐ€ ์ตํžˆ ์•Œ๊ณ  ์žˆ๋Š” ์Šคํƒ€ ์›Œ์ฆˆ๋‚˜ ๋ฐ˜์ง€์˜ ์ œ์™• ๊ฐ™์€ ์˜ํ™”๋“ค๋„ ์ด ์˜์›…์‹ ํ™”์˜ ๊ตฌ์กฐ๋ฅผ ๊ฐ–๊ณ  ์žˆ์ฃ . ์ด ์—ฐ๊ตฌ๋Š” ์ด ์กฐ์…ˆ ์บ ๋ฒจ์˜ ์˜์›… ์‹ ํ™” ๊ตฌ์กฐ๋ฅผ ์ด์šฉํ–ˆ์Šต๋‹ˆ๋‹ค. ๊ทธ๊ฑธ 7๊ฐœ ๋‹จ๊ณ„๋กœ ํ†ต๊ณ„ ๋ถ„์„์„ ํ†ตํ•ด ๋” ๋‹จ์ˆœํ™” ํ–ˆ์Šต๋‹ˆ๋‹ค. ๊ทธ๋ฆฌ๊ณ  ์ž์‹ ์˜ ์‚ถ์ด ์˜์›… ์‹ ํ™” ๊ตฌ์กฐ์™€ ์–ผ๋งˆ๋‚˜ ์œ ์‚ฌํ•˜๊ฒŒ ๋А๋ผ๋Š”๊ฐ€๊ฐ€ ๊ทธ ์‚ฌ๋žŒ์˜ ์‚ถ์˜ ์˜๋ฏธ๋ฅผ ์˜ˆ์ธกํ•œ๋‹ค๋Š” ๊ฑธ ๋ฐœ๊ฒฌํ–ˆ๊ณ , ๋” ๋‚˜์•„๊ฐ€ ์‹คํ—˜์„ ํ†ตํ•ด, ์‚ฌ๋žŒ๋“ค์ด ์ž์‹ ์˜ ์‚ถ์„ ์˜์›… ์‹ ํ™” ๊ตฌ์กฐ์— ๋งž๊ฒŒ ํ’€์–ด ์“ฐ๋„๋ก ์œ ๋„๋ฅผ ํ•˜๋ฉด, ์‹ค์ œ ์‚ถ์˜ ์˜๋ฏธ๊ฐ€ ์ฆ๊ฐ€ํ•˜๋Š” ๊ฑธ ์ฐพ์•„๋ƒˆ์Šต๋‹ˆ๋‹ค. ๊ทธ๋ฆฌ๊ณ  ๋‹จ์ˆœํžˆ ์‚ถ์˜ ์˜๋ฏธ ์ฆ๊ฐ€๋กœ ๋๋‚˜๋Š” ๊ฒŒ ์•„๋‹ˆ๋ผ, ํฅ๋ฏธ๋กญ๊ฒŒ๋„ ์ง๋ฌด ๋งŒ์กฑ๋„๋„ ๋ฐ”๋กœ ํ–ฅ์ƒ๋˜๊ณ  ์šฐ์šธ ์ฆ์ƒ๋„ ์ค„์–ด๋“ค์—ˆ์Šต๋‹ˆ๋‹ค. ์ค‘์š”ํ•œ ๋ถ€๋ถ„์€ ์ด๋Š” ๋‹จ์ˆœํžˆ ์ƒ๊ด€๊ด€๊ณ„ ์—ฐ๊ตฌ๊ฐ€ ์•„๋‹ˆ๋ผ ์ธ๊ณผ๊ด€๊ณ„๋ฅผ ๋ณผ ์ˆ˜ ์žˆ๋Š” ์—ฐ๊ตฌ์˜€๋‹ค๋Š” ์ ์ž…๋‹ˆ๋‹ค. ์—ฌ๋Ÿฌ๋ถ„๋“ค๋„ ์ง์ ‘ ํ•ด๋ณผ ์ˆ˜ ์žˆ์Šต๋‹ˆ๋‹ค. ์‹ค์ œ๋กœ ์—ฐ๊ตฌ์—์„œ ์‚ฌ์šฉํ•œ ๊ฐœ์ž…์˜ ์ถ•์•ฝํŒ์ž…๋‹ˆ๋‹ค. ์ž์‹ ์˜ ์‚ถ์—์„œ ์ฃผ์ธ๊ณต์€ ๋ˆ„๊ตฌ์˜€๋Š”์ง€, ๋‚ด๊ฐ€ ์ฒ˜์Œ ์–ด๋–ค ์ „ํ™˜/์ƒˆ๋กœ์šด ๊ฒฝํ—˜์„ ๋งž์ด ํ–ˆ๊ณ , ๊ทธ๋ฅผ ํ†ตํ•ด ์–ด๋–ค ํ€˜์ŠคํŠธ๋ฅผ ์ˆ˜ํ–‰ํ•ด ์™”์œผ๋ฉฐ, ๋‚ด๊ฐ€ ๋งŒ๋‚œ ํ˜‘๋ ฅ์ž/๋ฉ˜ํ† ๋Š” ๋ˆ„๊ตฌ์˜€๊ณ , ๋Œ€๋งˆ์™•์€ ๋ˆ„๊ตฌ์ด๊ณ (ํฐ ๋„์ „), ๊ทธ๋ฅผ ํ†ตํ•ด ๋‚˜๋Š” ์–ด๋–ป๊ฒŒ ๋ณ€์‹ ํ–ˆ๊ณ , ์œ ์‚ฐ์€ ๋ฌด์—‡์ธ์ง€. ๊ทธ๋ฆฌ๊ณ  ๋งˆ์ง€๋ง‰์œผ๋กœ ์ด ๋ชจ๋“  ๊ฑธ ๊ณ ๋ คํ•ด์„œ ๋‚ด ์‚ถ์„ ์˜์›…์˜ ์„œ์‚ฌ์‹œ์  ์—ฌ์ •์œผ๋กœ ์–ด๋–ป๊ฒŒ ๋ณผ ์ˆ˜ ์žˆ๋Š”์ง€ ์“ฐ๊ธฐ. ์ˆœ์„œ๊ฐ€ ์ค‘์š”ํ•ฉ๋‹ˆ๋‹ค. ์ด๊ฑธ ๊ทธ๋ƒฅ ๊ธ€๋กœ ๋”ฑ 30๋ถ„๊ฐ„ ์“ฐ๋Š” ๊ฒƒ๋งŒ์œผ๋กœ๋„ ์—ฌ๋Ÿฌ๋ถ„์˜ ์‚ถ๊ณผ ์ปค๋ฆฌ์–ด์˜ ์˜๋ฏธ๊ฐ€ ๋” ์ƒ๊ธฐ๊ฒŒ ๋ฉ๋‹ˆ๋‹ค. ์‚ถ์˜ ์˜๋ฏธ๊ฐ€ ์ด๋ ‡๊ฒŒ ๊ฐ„๋‹จํ•œ ๊ฐœ์ž…๋งŒ์œผ๋กœ๋„ ์œ ์˜๋ฏธํ•˜๊ฒŒ ์ฆ๊ฐ€ํ•  ์ˆ˜ ์žˆ๋‹ค๋Š” ์ ๋„ ๋†€๋ž์ง€๋งŒ, ์‹ฌ์ง€์–ด ์ž์‹ ์˜ ์ตœ๊ทผ ๋ช‡๋…„๊ฐ„ ์‚ถ์—์„œ ํž˜๋“ค์—ˆ๋˜ ๊ณ ๋‚œ๊ณผ ๋ฌธ์ œ(์˜ˆ์ปจ๋Œ€ ์‹ค์ง, ๊ฑด๊ฐ•๋ฌธ์ œ, ๊ด€๊ณ„ ๋“ฑ)์— ๋Œ€ํ•œ โ€œํšŒ๋ณต๋ ฅโ€œ(resilience)์ด ํ•˜๋ฃจ ์‚ฌ์ด์— ์œ ์˜๋ฏธํ•˜๊ฒŒ ์ฆ๊ฐ€ํ–ˆ์Šต๋‹ˆ๋‹ค. ์ž์‹ ์˜ ๋ฌธ์ œ๋ฅผ ๊ธ์ •์ ์ธ ์‹œ๊ฐ์—์„œ ์ƒˆ๋กญ๊ฒŒ ๋ณด๋Š” ๊ธ์ •์  ์žฌํ‰๊ฐ€(positive reappraisal)๊ฐ€ ์ฆ๊ฐ€ํ–ˆ๊ณ , ํšŒ๋ณต์  ๋Œ€์‘ ์ „๋žต(Resilient Coping)์„ ๋” ์ ๊ทน์ ์œผ๋กœ ์‚ฌ์šฉํ•œ๋‹ค๊ณ  ๋ณด๊ณ ํ–ˆ์Šต๋‹ˆ๋‹ค. ์ €๋Š” ์ด ์—ฐ๊ตฌ๋ฅผ ๋ณด๊ณ  ์˜๊ฐ์„ ๋ฐ›์•„ ์›Œํฌ์ˆ์„ ๋งŒ๋“ค์–ด๋ณด๊ธฐ๋กœ ํ–ˆ์Šต๋‹ˆ๋‹ค. ๊ทธ๋ž˜์„œ ํƒ„์ƒํ•œ ๊ฒƒ์ด ๋ฐ”๋กœ, <์Šคํ† ๋ฆฌ ๊ฒŒ์ž„์„ ํ†ตํ•œ ์ปค๋ฆฌ์–ด ์žฌ์„ค๊ณ„ ์›Œํฌ์ˆ>์ž…๋‹ˆ๋‹ค. ์‚ฌ๋žŒ๋“ค์˜ ์ธ์ง€์  ๊ณ ์ฐฉ ๋•Œ๋ฌธ์— ์ž์‹ ์˜ ์‚ถ์„ ๋‹ค๋ฅธ ๊ฐ๋„๋กœ ๋ณด๋Š” ๊ฒƒ์ด ์–ด๋ ต๋‹ค๋Š” ๊ฒƒ์—์„œ ์ฐฉ์•ˆํ•ด์„œ ์ž์‹ ์˜ ์‚ถ์„ โ€œ์‚ฌ์ด๋ฒ„ํŽ‘ํฌโ€ ์‹œ๋Œ€์˜ ์บ๋ฆญํ„ฐ๋กœ ๋Œ€์‘์„ ์‹œ์ผœ์„œ ์Šคํ† ๋ฆฌ ๊ฒŒ์ž„์„ ํ•˜๊ฒŒ ๋ฉ๋‹ˆ๋‹ค. ์ด๊ฑด TRPG(ํ…Œ์ด๋ธ”ํ†ฑ ๋กคํ”Œ๋ ˆ์ž‰ ๊ฒŒ์ž„์ด๋ผ๊ณ  ํ•˜๊ณ  ๋˜์ „ ์•ค ๋“œ๋ž˜๊ณค ๊ฐ™์€ ๊ฒŒ์ž„์ด ๋Œ€ํ‘œ์ )์—์„œ ๋ช‡๊ฐ€์ง€ ํ˜•์‹์„ ๋นŒ๋ ค์˜จ ๊ฒ๋‹ˆ๋‹ค. ์ €ํฌ๋Š” ์ด๊ฑธ โ€œ๋ฏธ๋Ÿฌ ์›”๋“œโ€œ๋ผ๊ณ  ๋ถ€๋ฆ…๋‹ˆ๋‹ค. ์ฐธ๊ฐ€์ž๋Š” โ€œ๋ฏธ๋Ÿฌ ์›”๋“œโ€œ์™€ โ€œ๋ฆฌ์–ผ ์›”๋“œโ€œ๋ฅผ ์˜ค๊ฐ€๋ฉด์„œ ์ฐธ๊ฐ€์ž์™€ ์บ๋ฆญํ„ฐ๊ฐ„์˜ ํ˜‘๋ ฅ์„ ํ†ตํ•ด ์˜์›…์˜ ์—ฌ์ •์„ ๋งŒ๋“ค์–ด ๊ฐ€๊ฒŒ ๋ฉ๋‹ˆ๋‹ค. ์ด๊ฑธ ํ•˜๊ณ  ๋‚˜์„œ๋Š” ์ฐธ๊ฐ€์ž ์ค‘ ํ•œ๋ถ„์€ โ€˜๋‚ด๊ฐ€ ํ˜„์‹ค์—์„œ ํ•˜๋Š” ์ผ์„ ๋Œ€์ˆ˜๋กญ์ง€ ์•Š๋‹ค๊ณ  ์ƒ๊ฐํ–ˆ๋Š”๋ฐ, ์Šคํ† ๋ฆฌ ๊ฒŒ์ž„์„ ํ•˜๋ฉด์„œ ๊ทธ ์ผ๋“ค์ด ์—„์ฒญ ์˜๋ฏธ์žˆ๋Š” ์ผ์ด์—ˆ๊ตฌ๋‚˜ ํ•˜๋Š” ๊นจ๋‹ฌ์Œ์„ ์–ป์—ˆ๋‹คโ€˜๊ณ  ๋ง์”€ํ•ด์ฃผ์‹œ๊ธฐ๋„ ํ–ˆ์Šต๋‹ˆ๋‹ค. ๊ทธ๋ฆฌ๊ณ  ์—ฌ๊ธฐ์—์„œ ํ•œ ๊ฑธ์Œ ๋” ๋‚˜์•„๊ฐ‘๋‹ˆ๋‹ค. ์œ„ ๋…ผ๋ฌธ์˜ ์—ฐ๊ตฌ์—์„œ๋Š” ์ž๊ธฐ ๊ณผ๊ฑฐ ์‚ถ์„ ์ƒˆ๋กญ๊ฒŒ ํ•ด์„ํ•˜๋Š” ๊ฑฐ์—์„œ ๋ฉˆ์ถ”๋Š”๋ฐ ์ €ํฌ๋Š” ์ด๊ฑธ ๋ฏธ๋ž˜๋กœ ํ™•์žฅํ•ด์„œ ํ–ฅํ›„ ์ปค๋ฆฌ์–ด๋ฅผ ๋””์ž์ธํ•˜๊ฒŒ ๋•์Šต๋‹ˆ๋‹ค. ์ „ํ˜€ ์ƒˆ๋กœ์šด ์‹œ๊ฐ์—์„œ์š”. ์ด ๋ถ€๋ถ„์€ ์žก ํฌ๋ž˜ํ”„ํŒ…์˜ ์—ฐ๊ตฌ๋ฅผ ์ฐธ๊ณ ํ•ด์„œ โ€œ์ปค๋ฆฌ์–ดโ€ ํฌ๋ž˜ํ”„ํŒ…์œผ๋กœ ํ™•์žฅํ–ˆ์Šต๋‹ˆ๋‹ค. ์ด๊ฑธ ํ•œ ํ›„์—๋Š” โ€˜์ž์‹ ์ด ๊ฟˆ์— ๊ทธ๋ฆฌ๋Š” ์ปค๋ฆฌ์–ด ๋ฐœ์ „์„ ์œ„ํ•ด ์˜ค๋Š˜ ๋‹น์žฅ ์‹œ์ž‘ํ•ด๋ณผ ์ˆ˜ ์žˆ๋Š” ์ผ๋“ค์ด ๋ถ„๋ช…ํ•ด์ ธ์„œ ์ข‹์•˜๋‹คโ€™๋Š” ๋ถ„๋„ ๊ณ„์…จ์–ด์š”. ์‹ฌ์ง€์–ด ์—ฌ๊ธฐ์—์„œ ํ•œ๊ฑธ์Œ ๋” ๋‚˜์•„๊ฐ€์„œ, ์›Œํฌ์ˆ ์ข…๋ฃŒ ํ›„ 3์ฃผ ํ›„์—๋Š” ์˜จ๋ผ์ธ ์„ธ์…˜์„ ํ†ตํ•ด ํ–‰๋™๊ณ„ํš์˜ ์ง„ํ–‰์ƒํ™ฉ์„ ํ•จ๊ป˜ ์ฒดํฌํ•˜๊ณ  ํ”ผ๋“œ๋ฐฑ ๋“œ๋ฆฌ๋Š” ์‹œ๊ฐ„๋„ ์ค€๋น„๋˜์–ด ์žˆ์Šต๋‹ˆ๋‹ค. ๊ทธ๋ž˜์„œ ๊ฒฐ๊ณผ์ ์œผ๋กœ ์ €ํฌ ํ”„๋กœ๊ทธ๋žจ์—์„œ๋„ ์‚ถ์˜ ์˜๋ฏธ ํ–ฅ์ƒ์ด ์ผ์–ด๋‚ฌ์„๊นŒ์š”? ์•ž์„œ ์‚ถ์˜ ์˜๋ฏธ๋ฅผ ์ธก์ •ํ•˜๋Š” ์งˆ๋ฌธ์„ ํ–ˆ์„ ๊ฒฝ์šฐ, ์›Œํฌ์ˆ ์ „๊ณผ ํ›„๋ฅผ ๋น„๊ตํ–ˆ์„ ๋•Œ 7์  ์ฒ™๋„์—์„œ ์•ฝ 1.9์ ์˜ ํ–ฅ์ƒ์ด ์žˆ์–ด์„œ ์ƒ๋‹นํžˆ ํฐ ํญ์˜ ํ–ฅ์ƒ์ด ์žˆ์—ˆ์Šต๋‹ˆ๋‹ค. ์ฐธ๊ณ ๋กœ ์–ธ๊ธ‰๋œ ์—ฐ๊ตฌ์—์„œ๋Š” ์•ฝ 0.5์ ์˜ ํ–ฅ์ƒ์ด ์žˆ์—ˆ๊ณ , ์ €ํฌ๊ฐ€ ์„ค๊ณ„ํ•œ ์ธํ„ฐ๋ฒค์…˜์—์„œ๋Š” ์ด ํ–ฅ์ƒ์˜ 4๋ฐฐ์— ๋‹ฌํ•˜๋Š” ํ–ฅ์ƒ์ด ์ผ์–ด๋‚œ ๊ฒ๋‹ˆ๋‹ค. ๊ทผ๋ฐ ์ด๊ฒƒ๋„ ์‚ฌ์‹ค ์˜จ๋ผ์ธ ์„ธ์…˜์„ ํ•˜๊ธฐ ์ „์˜ ํ–ฅ์ƒ์น˜๋ผ๋Š” ๊ฑธ ์—ผ๋‘์— ๋‘ฌ์•ผ ํ•ฉ๋‹ˆ๋‹ค.

We are an applied AI lab focused on reasoning, and today we're excited to introduce Devin, the first AI software engineer. โ€ Weโ€™re building AI teammates with capabilities far beyond todayโ€™s existing AI tools. By solving reasoning, we can unlock new possibilities in a wide range of disciplinesโ€”code is just the beginning. We are well funded, including a $21 million Series A led by Founders Fund And weโ€™re grateful for the support of industry leaders including Patrick Collison, John Collison, Elad Gil, Sarah Guo, Chris Re, Eric Glyman, Karim Atiyeh, Erik Bernhardsson, Tony Xu, Fred Ehrsam and so many more. More about Devin: - Devin is the new state-of-the-art on the SWE-Bench coding benchmark, has successfully passed practical engineering interviews from leading AI companies, and has even completed real jobs on Upwork. - Devin is an autonomous agent that solves engineering tasks through the use of its own shell, code editor, and web browser. - When evaluated on the SWE-Bench benchmark, which asks an AI to resolve GitHub issues found in real-world open-source projects, Devin correctly resolves 13.86% of the issues unassisted, far exceeding the previous state-of-the-art model performance of 1.96% unassisted and 4.80% assisted. Our team is small and talent-dense. Among our founding team, we have world-class competitive programmers, former founders, and leaders from companies at the cutting edge of AI including Cursor, Scale AI, Lunchclub, Modal, Google DeepMind, Waymo, and Nuro. โ€ Building Devin is just the first stepโ€”our hardest challenges still lie ahead. If youโ€™re excited to solve some of the worldโ€™s biggest problems and build AI that can reason, learn more about our team and apply to join us here: https://lnkd.in/eZq3s_KQ

https://bluedskim.github.io/posts/tech/%ED%95%A8%EA%BB%98-%EC%9E%90%EB%9D%BC%EA%B8%B0_%EA%B9%80%EC%B0%BD%EC%A4%80_%EB%A9%94%EB%AA%A8/ ์ž๋ผ๊ธฐ(ํ•™์Šต) 2์ข…๋ฅ˜์˜ ํ•™์Šต, ์•ผ์ƒํ•™์Šต โ‡” ํ•™๊ต ํ•™์Šต ์•ผ์ƒํ•™์Šต์˜ ํŠน์ง• ํ˜‘๋ ฅ์  ๋น„์ˆœ์ฐจ์  ์ž๋ฃŒ์— ํ•œ์ •์ด ์—†์Œ ํ‰๊ฐ€๊ธฐ์ค€์ด ์—†์Œ ์ •๋‹ต์ด ์—†์Œ ๋ชฉํ‘œ๊ฐ€ ๋ถˆ๋ช…ํ™•ํ•˜๊ณ  ๋ณ€๊ฒฝ๋˜๊ธฐ๋„ ํ•จ. ๋ถˆํ™•์‹ค์„ฑ์ด ๋†’์„์ˆ˜๋ก ์•ผ์ƒํ•™์Šต์ด ์ค‘์š”ํ•จ. ์ฆ‰ ํ•™์Šตํ•˜๋Š” ๋ฐฉ๋ฒ•์„ ํ•™์Šตํ•ด์•ผ ํ•œ๋‹ค ๊ฐœ๋ฐœ์ž์˜ ์‹ค๋ ฅ๊ณผ ๋…„์ฐจ ์ผ๋ฐ˜์ ์œผ๋กœ ๊ฐœ๋ฐœ์ž์˜ ๊ฐ€์น˜๋Š” ๊ทธ ์‚ฌ๋žŒ์˜ ๊ฒฝ๋ ฅ(๋…„์ฐจ)๋กœ ํŒ๋‹จ. ๊ฒฝ๋ ฅ์ด ๋ช‡๋…„ ๋˜์ง€ ์•Š์•˜์„ ๋•Œ๋Š” ์“ธ๋งŒํ•œ ๊ธฐ์ค€์ด ๋  ์ˆ˜ ์žˆ์Œ. ํ•˜์ง€๋งŒ ์ผ์ • ์ •๋„ ๋…„์ฐจ๊ฐ€ ์ง€๋‚˜๋ฉด ๊ฒฝ๋ ฅ์œผ๋กœ ์ฑ„์šฉํ•  ๋•Œ์™€ ๊ด€์‹ฌ์‚ฌ๋กœ ์ฑ„์šฉํ•  ๋•Œ์™€ ์ƒ๊ด€์„ฑ์€ ๊ฑฐ์˜ ๋™์ผํ•จ(5๋…„์ฐจ์™€ 10๋…„์ฐจ๋Š” ์„ฑ๊ณผ ์ธก๋ฉด์—์„œ ํฐ ์ฐจ์ด๊ฐ€ ์—†์Œ). ๊ฒฐ๊ณผ์ ์œผ๋กœ ๊ฒฝ๋ ฅ์ด๋ผ๋Š” ์š”์†Œ๋Š” ์กฐ์ง์— ์†ํ•ด๋ฅผ ์ค„ ์ˆ˜ ์žˆ๋Š” ๋ฐฉ์‹. ๊ฒฝ๋ ฅ๊ณผ ์‹ค๋ ฅ์€ ๋น„๋ก€ํ•˜์ง€ ์•Š๋Š”๋‹ค. ๋‹จ ๊ฒฝํ—˜์˜ ํญ์ด ๋„“๊ณ  ๋‹ค์–‘ํ•˜๋‹ค๋ฉด ๋ณ€๋ณ„๋ ฅ์ด ์žˆ๋‹ค. ๊ฐœ๋ฐœ์ž ์ฑ„์šฉ ๋ฐฉ๋ฒ• ๊ตฌ์กฐํ™”๋œ ์ธํ„ฐ๋ทฐ: โ€œ์ง€๋‚œ ํ”„๋กœ์ ํŠธ์—์„œ ๋™๋ฃŒ๊ฐ€ ์–ด๋ ค์›€์„ ๊ฒช์„ ๋•Œ ์–ด์ฉ ํ–‰๋™์„ ํ•˜์…จ๋Š”์ง€ ๊ตฌ์ฒด์ ์œผ๋กœ ์˜ˆ๋ฅผ ๋“ค์–ด์ฃผ์„ธ์š”โ€ ์ž‘์—… ์ƒ˜ํ”Œ ํ…Œ์ŠคํŠธ ์‹ค์ œ๋กœ ๊ฐ™์ด ์ผํ•ด๋ณด๊ธฐ(Trial Sourcing) ๊ฐ™์ด ์ผํ•  ์‚ฌ๋žŒ๋“ค์ด ๊ฐ™์ด ์ธํ„ฐ๋ทฐ์— ์ฐธ์—ฌํ•˜๊ธฐ ์ด๋ฏธ ์ฑ„์šฉํ•œ ์ธ๋ ฅ์„ ์–ด๋–ป๊ฒŒ ํ•  ๊ฒƒ์ธ๊ฐ€ ๊ตฌ์„ฑ์›์ด ์š”์ฆ˜์— ์–ผ๋งˆ๋‚˜ ๊ณต๋ถ€ํ•˜๊ณ  ์ˆ˜๋ จํ•˜๋А๋ƒ๋กœ ๊ทธ์ง์›์˜ ์„ฑ๊ณผ๊ฐ€ ๊ฒฐ์ •. ์กฐ์ง์€ ๊ฐœ์ธ์ด ์ „๋ฌธ์„ฑ์„ ์ข€ ๋” ๋ฐœ์ „์‹œ๊ณ  ๊ด€๋ฆฌํ•  ์ˆ˜ ์žˆ๊ฒŒ ์ตœ๋Œ€ํ•œ ์ง€์›ํ•˜๋Š”๊ฒƒ์ด ํšจ๊ณผ์ . ํ•˜์ง€๋งŒ ์‹œ์Šคํ…œ๊ณผ ๋ฌธํ™”์— ๋ฌธ์ œ๊ฐ€ ์žˆ๋‹ค๋ฉด ์•„๋ฌด๋ฆฌ ํ›Œ๋ฅญํ•œ ๊ตฌ์„ฑ์›๋„ ๋ฌปํ˜€๋ฒ„๋ฆฌ๊ธฐ ์‰ฝ๋‹ค. ๋ฐ˜๋Œ€๋กœ ํ›Œ๋ฅญํ•œ ์‹œ์Šคํ…Œ๊ณผ ๋ฌธํ™”์†์—์„œ ํ‰๋ฒ”ํ•œ ๊ตฌ์„ฑ์›์ด ์„ฑ๊ณผ๋ฅผ ๋‚ผ ์ˆ˜ ๋„ ์žˆ์Œ ๊ฐœ๋ฐœ์ž ๊ฐœ๊ฐœ์ธ์€ ๋…„์ˆ˜๋งŒ ๋Š˜๋ฆด๊ฒƒ์ด ์•„๋‹ˆ๋ผ ๋ณธ์ธ์˜ ๊ธฐ๋Ÿ‰์„ ํ–ฅ์ƒ์‹œํ‚ฌ ์ˆ˜ ์žˆ๋Š” ์˜๋„์  ์ˆ˜๋ จ์ด ํ•„์š”. ์ด ์ˆ˜๋ จ์—๋Š” ์งง์€ ํ”ผ๋“œ๋ฐฑ์ด ํ•„์ˆ˜. ์ž๊ธฐ ๊ณ„๋ฐœ ์ž๊ธฐ ๊ณ„๋ฐœ์€ ๋ณต๋ฆฌ๋กœ ๋Œ์•„์˜จ๋‹ค(์ž๊ธฐ ๊ณ„๋ฐœ์„ ํ•˜๋ฉด ํ•  ์ˆ˜๋ก ๊ฐ€์†๋„๊ฐ€ ๋ถ™๋Š”๋‹ค) ์ž์‹ ์ด ์ด๋ฏธ ๊ฐ€์ง€๊ณ  ์žˆ๋Š” ๊ฒƒ๋“ค์„ ์ตœ๋Œ€ํ•œ ํ™œ์šฉ ์™ธ๋ถ€์˜ ์ž๊ทน์„ ์ตœ๋Œ€ํ•œ ๋นจ๋ฆฌ ์ฒดํ™” ์ž์‹ ์„ ๊ฐœ์„ ํ•˜๋Š” ํ”„๋กœ์„ธ์Šค ๋งŒ๋“ค๊ธฐ ๋” ์ผ์ฐ, ๋” ์ž์ฃผ ์‹คํŒจํ•˜๊ณ  ์‹คํŒจ์—์„œ ๋ฐฐ์šฐ์ž ์ž์‹ ์˜ ๋Šฅ๋ ฅ์„ ํ‚ค์›Œ์ฃผ๋Š” ๋„๊ตฌ๋ฅผ ์ ์ง„์ ์œผ๋กœ ๊ฐœ๋ฐœ ์กฐ์ง์˜ 3๊ฐ€์ง€ ์ฐจ์›์˜ ์ž‘์—… A : ์›๋ž˜ ํ•˜๊ธฐ๋กœ ํ•œ ์ผ B : A๋ฅผ ์ž˜ ํ•˜๊ธฐ ์œ„ํ•œ ์ž‘์—… C : B๋ฅผ ์ž˜ ํ•˜๊ธฐ ์œ„ํ•œ ์ž‘์—…, ๋ถ€ํŠธ ์ŠคํŠธ๋ž˜ํ•‘, ์ง€์ˆ˜์  ํ–ฅ์ƒ ์กฐ์ง์˜ ํ˜•ํƒœ ์›Œํฌ ๊ทธ๋ฃน : ์ผ์„ ๋‚˜๋ˆ ์„œ ํ•˜๋Š” ๋ถ„์—… ์กฐ์ง ํŒ€ : ์ง‘๋‹จ์˜ ์ง€๋Šฅ์„ ๋†’ํ˜€ ์ƒ์‚ฐ์„ฑ์„ ํ–ฅ์ƒ์‹œํ‚ค๋Š” ์กฐ์ง ํ•™์Šต ํ”„๋ ˆ์ž„ VS ์‹คํ–‰ ํ”„๋ ˆ์ž„ ์‹คํ–‰ ํ”„๋ ˆ์ž„ : ์„ฑ๊ณผ(โ€˜์ž˜ํ•˜๊ธฐโ€˜)๊ฐ€ ์ค‘์š”ํ•จ ํ•™์Šต ํ”„๋ ˆ์ž„ : ๋ฐฐ์šฐ๊ธฐ(โ€˜์ž๋ผ๊ธฐโ€˜)๊ฐ€ ์ค‘์š”ํ•จ ํ•™์Šต์— ์œ ๋ฆฌํ•œ ์กฐ๊ฑด ๋ชฉํ‘œ๊ฐ€ ๋ถ„๋ช…ํ•˜๊ณ  ๊ฐ๊ด€์ ์œผ๋กœ ์ •ํ•ด์ ธ ์žˆ์œผ๋ฉฐ ๋ณ€๊ฒฝ๋˜์ง€ ์•Š์Œ ์„ ํƒํ•  ์ˆ˜ ์žˆ๋Š” ํ–‰๋™๊ณผ ์ข…๋ฅ˜๊ฐ€ ์œ ํ•œํ•จ ๋งค ์ˆœ๊ฐ„ ์ž์‹ ์ด ๋ชฉํ‘œ์— ์–ผ๋งˆ๋‚˜ ๊ทผ์ ‘ํ–ˆ๋Š”์ง€ ์•Œ ์ˆ˜ ์žˆ์Œ ์ฃผ๋กœ ๋‹ซํžŒ ์‹œ์Šคํ…œ ์†์—์„œ ํ™œ๋™ ๊ณผ๊ฑฐ์˜ ์„ ํƒ๊ณผ ๊ฒฐ๊ณผ์— ๋Œ€ํ•œ ๊ตฌ์กฐํ™”๋œ ๊ธฐ๋ก์ด ๋งŽ์Œ ํ•™์Šต์— ๋ถˆ๋ฆฌํ•œ ์กฐ๊ฑด(์•”๋ฌต์ง€, ์ง๊ด€์ด ์ž‘๋™ํ•˜๋Š” ํšŒ์ƒ‰ ์˜์—ญ) ์•„๋ž˜์—์„œ ํ•„์š”ํ•œ ๋Šฅ๋ ฅ ๋…์ฐฝ์„ฑ(Originality) ์‚ฌํšŒ์  ๋ฏผ๊ฐ์„ฑ(Social Perceptiveness) ํ˜‘์ƒ(Negotiation) ์„ค๋“(Persuation) ํƒ€์ธ์„ ๋•๊ณ  ๋Œ๋ณด๊ธฐ(Assisting and caring others) ์ฝ”๋”์™€ ๊ฐœ๋ฐœ์ž์˜ ์ฐจ์ด : AI๊ฐ€ ๋Œ€์ฒดํ•  ์ˆ˜ ์žˆ๋Š”๊ฐ€? ๊ฐ€์žฅ ํ•™์Šตํ•˜๊ธฐ ํž˜๋“  ์ง์—…์ด ์‚ด์•„๋‚จ๋Š”๋‹ค ๋‹ฌ์ธ์ด ๋  ์ˆ˜ ์žˆ๋Š” ์กฐ๊ฑด ์‹ค๋ ฅ์„ ๊ฐœ์„ ํ•˜๋ ค๋Š” ๋™๊ธฐ ๊ตฌ์ฒด์ ์ธ ํ”ผ๋“œ๋ฐฑ์„ ์ ์ ˆํ•œ ์‹œ๊ธฐ์— ๋ฐ›์„ ์ˆ˜ ์žˆ์–ด์•ผ ํ•จ ์ž‘์—…์˜ ๋‚œ์ด๋„์™€ ์‹ค๋ ฅํ–ฅ์ƒ ๊ฐ€์žฅ ์ž˜ ๋ชฐ์ž…ํ•˜๊ธฐ ์œ„ํ•ด ์ ์ ˆํ•œ ๋‚œ์ด๋„๊ฐ€ ํ•„์š”. ์–ด๋–ค ๋‚œ์ด๋„๊ฐ€ ์ ์ ˆํ•œ์ง€ ํŒ๋‹จํ•˜๊ธฐ ์œ„ํ•ด ๋ฉ”ํƒ€์ธ์ง€ ๋Šฅ๋ ฅ์ด ํ•„์š”. ๊ตฌ์„ฑ์›์˜ ์‹ค๋ ฅํ–ฅ์ƒ์„ ์œ„ํ•œ ๋ฆฌ๋”์˜ ์—ญํ•  ๊ตฌ์„ฑ์›์ด ์Šค์Šค๋กœ ๋ชฐ์ž… ์ƒํƒœ๋ฅผ ์กฐ์ ˆํ•  ์ˆ˜ ์žˆ๋Š” ๋Šฅ๋ ฅ์„ ํ‚ค์›Œ์ค˜์„œ ์ž๊ธฐํšจ๋Šฅ๊ฐ์„ ๋†’ํ˜€์ค€๋‹ค. ์˜๋„์  ์ˆ˜๋ จ์˜ ์ผ์ƒ์  ์˜ˆ์‹œ (๊ตฌ์ฒด์  ์˜ˆ๋“ค์ด ๋งŽ์ด ๋‚˜์˜ด...) ์‹ค์ˆ˜๋Š” ์˜ˆ๋ฐฉํ•˜๋Š” ๊ฒƒ์ด ์•„๋‹ˆ๋ผ ๊ด€๋ฆฌํ•˜๋Š” ๊ฒƒ์ด๋‹ค. ์‹ค์ˆ˜๋Š” ๋ฐœ์ƒํ•œ๋‹ค. ์‹ค์ˆ˜๊ฐ€ ๋‚˜์œ ๊ฒฐ๊ณผ๋ฅผ ๋‚ด๊ธฐ ์ „์— ์กฐ๊ธฐ์— ๋ฐœ๊ฒฌํ•˜๊ณ  ๋น ๋ฅธ ์กฐ์ทจ๋ฅผ ์ทจํ•˜์ž. ์ดํ›„์˜ˆ๋Š” ์‹ค์ˆ˜๋ฅผ ๊ณต๊ฐœํ•˜๊ณ  ์‹ค์ˆ˜์— ๋Œ€ํ•ด ์–˜๊ธฐํ•˜๊ณ  ์‹ค์ˆ˜์—์„œ ๋ฐฐ์šฐ์ž. ์‹ค์ˆ˜ ์˜ˆ๋ฐฉ๋ณด๋‹ค ๊ด€๋ฆฌ์— ๋น„์ค‘์„ ๋” ๋‘˜ ์ˆ˜๋ก ๊ธฐ์—…์˜ ํ˜์‹  ์ •๋„๊ฐ€ ๋” ๋†’๋‹ค. ์™œ๋ƒํ•˜๋ฉด ์‹ค์ˆ˜๊ฐ€ ์—†์œผ๋ฉด ํ•™์Šตํ•˜์ง€ ๋ชปํ•˜๊ธฐ ๋•Œ๋ฌธ. ์‹ค์ˆ˜ ๊ด€๋ฆฌ๋ฅผ ํ•˜๋Š” ๋ฌธํ™”์ผ ์ˆ˜๋ก ํ•™์Šต์„ ๋” ์ž˜ํ•œ๋‹ค. โ€œ๋” ๋›ฐ์–ด๋‚œ ์Šค์ผ€์ดํ„ฐ๊ฐ€ ๋” ์ž์ฃผ ์—‰๋ฉ๋ฐฉ์•„๋ฅผ ์ฐง๋Š”๋‹คโ€ ๋‚˜ํ™€๋กœ ์ „๋ฌธ๊ฐ€์— ๋Œ€ํ•œ ๋ฏธ์‹  ์–ด๋–ค ๊ธฐ์ˆ ์  ์‹ค์ฒœ๋ฒ•์ด๋ผ๋„ ๊ทธ๊ฑธ ํ˜„์‹ค์—์„œ ์ ์šฉํ•˜๊ธฐ ์œ„ํ•ด์„œ๋Š” ์‚ฌํšŒ์  ์ž๋ณธ(์‹ ๋ขฐ)๊ณผ ์‚ฌํšŒ์  ๊ธฐ์ˆ (์†Œํ”„ํŠธ ์Šคํ‚ฌ)์ด ํ•„์š”. ์‚ฌํšŒ์  ๊ธฐ์ˆ ์„ ๋†’ํžˆ๊ธฐ ์œ„ํ•ด ๋งˆ์ดํฌ๋กœ ์ธํ„ฐ๋ ‰์…˜์— ์ฃผ์˜ํ•œ๋‹ค(๊ธฐ๋ก, ๋ณต๊ธฐ, โ€˜๊ทธ๊ฒƒ์ด ์ตœ์„ ์ด์—ˆ์„๊นŒโ€™ ํšŒ๊ณ ํ•ด๋ณด๊ธฐ) ํ•จ๊ป˜ ์†Œํ”„ํŠธ์›จ์–ด ๊ฐœ๋ฐœ์„ ์ž˜ ๊ด€๋ฆฌํ•˜๊ธฐ ์œ„ํ•œ 3๊ฐ€์ง€ ๋Šฅ๋ ฅ ๋ณต์žกํ•œ ์ƒํ™ฉ์„ ์ดํ•ดํ•˜๋Š” ๋Šฅ๋ ฅ : ์‹œ์Šคํ…œ์  ์‚ฌ๊ณ (Systems Thinking) ๊ด€์ฐฐํ•˜๋Š” ๋Šฅ๋ ฅ : ์ผ์ฐจ์  ์ธก์ •(First-Order Measurement) ํ–‰๋™ํ•˜๋Š” ๋Šฅ๋ ฅ : ์ผ์น˜์  ํ–‰๋™(Congruent Action) ๋ณ€ํ™”๋ฅผ ๊ธฐ๋Œ€ํ•˜๊ธฐ(Anticipating Change) ํ”„๋กœ์ ํŠธ ์„ฑ๊ณต์š”์†Œ ๊ด€๋ฆฌ : ๊ด€๋ฆฌ๋ฐฉ์‹, ๊ด€๋ฆฌ์ž ์‹œ์Šคํ…œ : ์กฐ์ง์ฒด๊ณ„ ์‚ฌ๋žŒ ๋„๊ตฌ ํ˜‘๋ ฅ์„ ํ†ตํ•œ ์ถ”์ƒํ™” ๋‹ค๋ฅธ ์‹œ๊ฐ์„ ๊ฐ€์ง„ ๋‘ ์‚ฌ๋žŒ์ด ํ˜‘๋ ฅํ•˜๊ธฐ(ํŽ˜์–ด ํ”„๋กœ๊ทธ๋ž˜๋ฐ) ์‹ ๋ขฐ์ž์‚ฐ์ด ๋†’์€ ์กฐ์ง์€ ์ปค๋ฎค๋‹ˆ์ผ€์ด์…˜ ํšจ์œจ๊ณผ ์ƒ์‚ฐ์„ฑ์ด ๋†’์Œ ์‹ ๋ขฐ(communication trust)์ž์‚ฐ์„ ๋†’ํžˆ๋Š” ๋ฐฉ๋ฒ• ํˆฌ๋ช…์„ฑ ๊ณต์œ  ์ธํ„ฐ๋ ‰์…˜ ์ดํ•ด(์ƒ๋Œ€์„ฑ์„ ์ธ์ •ํ•˜์ž) : ์‚ฌ๊ณ  ๊ณผ์ •์— ๋Œ€ํ•œ ์ดํ•ด, ์–ด๋–ค ๊ฐ€์น˜๊ด€์„ ๊ฐ€์ง€๊ณ  ์žˆ๋Š”์ง€, ์–ด๋–ค ๋ฐฉ์‹์„ ์„ ํ˜ธํ•˜๋Š”์ง€ ํ•˜ํ–ฅ์‹ VS ์ƒํ–ฅ์‹ ๋ฌธ์ œ๋ฅผ ๋ถ„๋ฅ˜ํ•˜์ž ์ž˜ ์ •์˜๋œ ๋ฌธ์ œ(well-defined) : ๋””์ž์ธ์ด ํ•„์š” ์ •์˜๋˜์ง€ ์•Š์€ ๋ฌธ์ œ(ill-defined) : ๋””์ž์ธ ๋ถˆํ•„์š” ์‹ค์ œ๋กœ๋Š” ํ•˜ํ–ฅ์‹๊ณผ ์ƒํ–ฅ์‹ ๋ฐฉ๋ฒ•์„ ๋ฒˆ๊ฐˆ์•„ ๊ฐ€๋ฉด์„œ ์‚ฌ์šฉํ•œ๋‹ค. ์กฐ์ง์ด ๋ฌธ์ œ๋ฅผ ํ•ด๊ฒฐํ•˜๋Š” ๋ฐฉ๋ฒ•(ํ˜‘๋ ฅ ๋ชจ๋ธ) ์˜์‚ฌ์†Œํ†ต์˜ ๋ฌธ์ œ : ๋ฐ”ํ†ตํ„ฐ์น˜์— ๋„ˆ๋ฌด ๋งŽ์€ ๋น„์šฉ์ด ๋ฐœ์ƒ ๋ชจ๋“  ์‚ฌ๋žŒ๋“ค์ด ๋ชจ๋“  ๋‹จ๊ณ„์˜ ์ „๋ฌธ๊ฐ€๊ฐ€ ๋˜๋„๋ก ํ•˜๊ฑฐ๋‚˜ VS ๋ฐ”ํ†ต ํ„ฐ์น˜์˜ ๋น„์šฉ์„ ์ค„์ธ๋‹ค ์‚ผํˆฌ์••์‹ ์˜์‚ฌ์†Œํ†ต, ์€์—ฐ์ค‘์— ์ •๋ณด๊ฐ€ ์กฐ์ง์— ์Šค๋ฉฐ๋“ค๋„๋ก ํ•œ๋‹ค ํ•œ๋ฒˆ์— ์ฒ˜๋ฆฌ๋˜๋Š” ์ผ์˜ ์–‘์„ ์ค„์ธ๋‹ค ์ „๋ฌธ๊ฐ€ํŒ€์ด ์‹คํŒจํ•˜๋Š” ์ด์œ  : ์ „๋ฌธ๊ฐ€๋“ค์˜ ego ๋•Œ๋ฌธ์— ํ˜‘๋ ฅ์„ ์•ˆํ•˜๋ ค ํ•œ๋‹ค ํ˜‘๋ ฅ์„ ์œ„ํ•ด์„œ๋Š” ์‹ ๋ขฐ๊ฐ€ ๋ฐ”ํƒ•์ด ๋˜์–ด์•ผ ํ•จ ์พŒ์† ํ•™์ŠตํŒ€ ํ•™์Šต ํ™˜๊ฒฝ์„ ๋งŒ๋“ค ์ˆ˜ ์žˆ๋Š” ๋ฆฌ๋”๊ฐ€ ํ•„์š”(ํ•˜์ง€๋งŒ ํ•™์Šต ์˜์ง€๊ฐ€ ์žˆ๋Š” ๊ตฌ์„ฑ์›์ด ์—†๋‹ค๋ฉด -.-) ํŒ€์›๋“ค์ด ์‹ฌ๋ฆฌ์ ์œผ๋กœ ๋ณดํ˜ธ๋ฐ›๊ณ  ์žˆ๋Š”๊ฐ€ ์ƒˆ๋กœ์šด ๊ฒƒ์„ ์ œ์•ˆํ•˜๊ณ  ์‹œ๋„ํ•˜๋Š”๋ฐ ์—ด๋ ค์žˆ๋Š”๊ฐ€ ์‹คํŒจ์— ๊ด€๋Œ€ํ•œ๊ฐ€ ์ž ์žฌ์  ๋ฌธ์ œ๋ฅผ ์ง€์ ํ•˜๊ณ  ์‹ค์ˆ˜๋ฅผ ์ธ์ •ํ•˜๋Š”๋ฐ ๋ถ€๋‹ด์„ ๋А๋ผ๋Š”๊ฐ€ ์• ์ž์ผ ํ™•๋ฅ ๋ก  ๊ด€์‹ฌ์‚ฌ์˜ ์„ž์ž„(mingling of concern)์„ ํ†ตํ•ด ์„œ๋กœ์„œ๋กœ ๋งŽ์€ ๊ฒƒ์„ ๋ฐฐ์šธ ์ˆ˜ ์žˆ๋‹ค. ํŒ€์›๋“ค์„ ์ตœ๋Œ€ํ•œ ์„ž์ด๋„๋ก ํ•œ๋‹ค. ํ•œ๋ช…์ด๋ผ๋„ ์ค‘์š”ํ•œ ๋ฐœ๊ฒฌ์„ ํ•˜๋ฉด ๋‚˜๋จธ์ง€ ๋ชจ๋“  ๊ตฌ์„ฑ์›์ด ๊ณต์œ (copy)ํ•  ์ˆ˜ ์žˆ๋„๋ก ํ•œ๋‹ค. ์ง๋ ฌ์  ์กฐ์ง vs ๋ณ‘๋ ฌ์  ์กฐ์ง : ์ง๋ ฌ์  ์กฐ์ง์—์„œ๋Š” ๊ฐ€์žฅ ์•ฝํ•œ ๊ฐœ์ฒด์˜ ๋Šฅ๋ ฅ์ด ๊ทธ ์กฐ์ง์˜ ๋Šฅ๋ ฅ(AND ์กฐ๊ฑด), ๋ณ‘๋ ฌ์  ์กฐ์ง์€ ๊ฐ ๊ฐœ์ฒด ๋Šฅ๋ ฅ์˜ ํ‰๊ท ์น˜๊ฐ€ ๊ทธ ์กฐ์ง์˜ ๋Šฅ๋ ฅ์น˜(OR์กฐ๊ฑด) ์• ์ž์ผ์€ ์ข‹์€ ์ผ์— ๋Œ€ํ•ด์„œ๋Š” โ€˜๊ทธ๋ฆฌ๊ณ โ€™ ํ™•๋ฅ ์„ โ€˜๋˜๋Š”โ€™ ํ™•๋ฅ ๋กœ ๋ฐ”๊พธ๊ณ  ๋‚˜์œ ์ผ์— ๋Œ€ํ•ด์„œ๋Š” โ€˜๋˜๋Š”โ€™ ํ™•๋ฅ ์„ โ€˜๊ทธ๋ฆฌ๊ณ โ€™ ํ™•๋ฅ ๋กœ ๋ฐ”๊ฟ‰๋‹ˆ๋‹ค. ์• ์ž์ผ : ํ•จ๊ป˜ ์ž๋ผ๊ธฐ ๋ถˆํ™•์‹ค์„ฑ์ด ๋” ๋†’์€ ํ”„๋กœ์ ํŠธ์— ์ ํ•ฉ ๋ถˆํ™•์‹ค์„ฑ์ด ๋‚ฎ์€ ํ”„๋กœ์ ํŠธ๋Š” ๋น„์ฆˆ๋‹ˆ์Šค์  ๊ฐ€์น˜๊ฐ€ ๋–จ์–ด์ง ๋ฏธ๋ž˜์˜ ๋ถˆํ™•์‹ค์„ฑ์€ ์ž์—ฐ์Šค๋Ÿฌ์šด ๊ฒƒ โ€˜ํ•™์Šตโ€˜๊ณผ โ€˜ํ˜‘๋ ฅโ€™์ด ๋ถˆํ™•์‹ค์„ฑ์— ๋Œ€ํ•œ ๋Œ€์•ˆ์ด๋‹ค โ€œ๊ณ ๊ฐ(๋ชจ๋“  ์ดํ•ด ๋‹น์‚ฌ์ž)์—๊ฒŒ ๋งค์ผ(์ž์ฃผ) ๊ฐ€์น˜๋ฅผ ์ „ํ•˜๋ผโ€ ์• ์ž์ผ ์„ฑ๊ณต์š”์ธ ๊ณ ๊ฐ์ฐธ์—ฌ ๋ฆฌํŒฉํ† ๋ง ์ž๋™ํ™” ํ…Œ์ŠคํŠธ ์ฝ”๋“œ ๊ณต์œ  ๋›ฐ์–ด๋‚œ ์• ์ž์ผ ์ฝ”์น˜ ๋‹น์‹ ์˜ ์กฐ์ง์— ์ƒˆ ๋ฐฉ๋ฒ•๋ก (์• ์ž์ผ)์ด ๋จนํžˆ์ง€ ์•Š๋Š” ์ด์œ  ์กฐ์ง์˜ ๋‹น์‹ ์— ๋Œ€ํ•œ ์‹ ๋ขฐ๊ฐ€ ๋ถ€์กฑํ•˜๊ธฐ ๋•Œ๋ฌธ ๋ฌธํ™”์  ํ’ํ† , ์ƒ์„ฑ์  ๊ณผ์ •

The Buffett Index: Why Warren Buffett might be selling more Apple soon @chamath 's insight: One way to tell if @WarrenBuffett has gotten disengaged from a company is the number of mentions in his annual letter to shareholders the facts: -- Buffett first bought $AAPL in 2016 -- his position was worth ~$175B at the end of 2023 -- 2020 letter: 11 $AAPL mentions -- 2021: 10 mentions -- 2022: 2 mentions -- 2023: 1 mention -- in Q4 23, Berkshire sold close to ~$2B worth of $AAPL (~1% of its ~$175B position) So here's the precarious setup for $AAPL: 1) Apple has a lot of headwinds (flat iPhone revenue, regulatory scrutiny, etc.) 2) Annual letter mentions have gone down 3) Buffett just started selling BONUS: Buffett followed a similar pattern with his Wells Fargo position! -- Buffett first bought $WFC in 1989/1990 -- $WFC was at one point one of Berkshire's largest holdings -- annual shareholder letter mentions went to 0 starting in 2016 -- Buffett started selling around late 2017/early 2018, and closed his position in 2022 *note: mentions = written mentions, not including appearances in charts or tables **note: not investment advice

https://youtu.be/E4TldCRLyoo?si=NAU-4ckVk4DsAcGG AI Startup๋“ค์ด ๊ณ ๋ฏผํ•ด์•ผํ•  ์ฃผ์ œ๋“ค 1. AI๋กœ ์ธํ•ด ๋ณ€ํ™”๋  ๊ธฐํšŒ - ๋ฏธ๊ตญ ์†Œํ”„ํŠธ์›จ์–ด ํšŒ์‚ฌ๋“ค์˜ ๋งค์ถœ $0.5t - ์ธ๊ฑด๋น„ ๋ฒ ์ด์Šค๋กœ ์ง€์ถœํ•˜๋Š” ๋น„์šฉ: $3.5t - ์ด ์ค‘ 10%๊ฐ€ AI๋กœ ๋Œ€์ฒด๋œ๋‹ค๋ฉด ์•ฝ $0.35t๋กœ ๊ธฐ์กด์˜ ์†Œํ”„ํŠธ์›จ์–ด ๋งค์ถœ๊ณผ ๋น„์Šทํ•จ. ์—ฌ๊ธฐ์— AI๋กœ ์ƒˆ๋กญ๊ฒŒ ๋งŒ๋“ค์–ด์งˆ ๊ธฐํšŒ๊นŒ์ง€ ๊ณ ๋ คํ•˜๋ฉด ์‹œ์žฅ์€ ๋” ์ปค์ง. - ์ตœ๊ทผ Klarna๊ฐ€ AI๋ฅผ ์ ๊ทน ํ™œ์šฉํ•ด์„œ ๊ณ ๊ฐ์‘๋Œ€์˜ ๋น„์šฉ์€ ์ค„์ด๊ณ  ๋งŒ์กฑ๋„๋Š” ๋†’์ž„. https://www.klarna.com/international/press/klarna-ai-assistant-handles-two-thirds-of-customer-service-chats-in-its-first-month/ 2. ๋‹น์‹ ์ด ํ‘ธ๋Š” ๋ฌธ์ œ๊ฐ€ ํŠน๋ณ„ํ•œ ๋„๋ฉ”์ธ์ธ๊ฐ€? ์•„๋‹ˆ๋ฉด ํŠน๋ณ„ํ•œ Workflow์ธ๊ฐ€? - ๋‹น์‹ ์ด ๋ณด์™„/๋Œ€์ฒดํ•˜๋ ค๋Š” ์—…๋ฌด๋Š” ์–ผ๋งˆ์งœ๋ฆฌ ์—…๋ฌด์ด๊ณ  ์–ผ๋งˆ๋‚˜ ๋งŽ๊ณ  ํฐ ๊ธฐํšŒ์ธ๊ฐ€? - AI๋กœ ๋Œ€์ฒดํ•  ์ˆ˜ ์žˆ๋Š”๊ฐ€ ์•„๋‹ˆ๋ฉด AI๋กœ ํ›จ์”ฌ ๋” ๋‚ซ๊ฒŒ ๋งŒ๋“ค ์ˆ˜ ์žˆ๋Š”๊ฐ€? 3. ๋น ๋ฅธ Feedback loop๋ฅผ ๋งŒ๋“ค ์ˆ˜ ์žˆ๋Š”๊ฐ€? - ์Šคํƒ€ํŠธ์—…์ด ๊ฐ€์ง„ Unfair Advantage๋Š” ์†๋„์ธ๋ฐ, AI ์Šคํƒ€ํŠธ์—…์—๊ฒŒ ์†๋„๋ž€ ์ œํ’ˆ-๋ฐ์ดํ„ฐ-๋ชจ๋ธ-์ œํ’ˆ๊ฐœ์„  ์‚ฌ์ดํด์˜ ์†๋„. - ์ œํ’ˆ์„ ๋งŒ๋“ค์–ด์„œ ๊ณ ๊ฐ์˜ Uniqueํ•œ ๋ฐ์ดํ„ฐ, ํ”ผ๋“œ๋ฐฑ์„ ๋ฐ›์•„์„œ ๋‚ด๊ฐ€ ๊ฐ€์ง„ ๋ชจ๋ธ์ด ์ง€์†์ ์œผ๋กœ ๋‚˜์•„์ง€๋ฉด์„œ ๋ชจ๋ธ์˜ ๊ฐœ์„ ์ด ๊ณ ๊ฐ์˜ ๊ฒฝํ—˜ ๋งŒ์กฑ๊ณผ ์ง€๋ถˆ๋กœ ์ด์–ด์ง€๋Š”๊ฐ€? ๊ทธ๋ฆฌ๊ณ  ๊ทธ ์†๋„๊ฐ€ ์•„์ฃผ ๋น ๋ฅธ๊ฐ€?

1. Problem definition & Market sizing - How much is it the task? -> How many/big? - Could it be automated by AI or Could it be much better with AI? 2. Build your product - Gain customers data and improve your AI - Is your data unique your competitors cannot access? 3. Feedback loop - Based on customers data and feedback, is your AI much better?

https://www.sequoiacap.com/article/follow-the-gpus-perspective/ The financial dynamics surrounding the use of GPUs (Graphics Processing Units) in the AI industry. It highlights a significant cost consideration: for every $1 spent on a GPU, an equivalent amount, approximately $1, is required for energy costs to operate the GPU within a data center. With Nvidia projected to achieve $50 billion in GPU revenue by year-endโ€”a figure deemed conservative based on analyst forecastsโ€”this scenario suggests that data center expenditures could reach around $100 billion. Furthermore, the article underscores the necessity for end users of GPUs, such as Starbucks, X (formerly known as SpaceX), Tesla, Github Copilot, or emerging startups, to achieve a 50% margin to sustain their operations and investments in GPU technology[1]. Sources [1] AIโ€™s $200B Question https://www.sequoiacap.com/article/follow-the-gpus-perspective/