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
2 291
Obunachilar
-324 soatlar
-167 kunlar
-4530 kunlar
Postlar arxiv
โ€œMachine learning costs, talent and chip shortagesโ€ฆ any AI and machine learning company faces at least one of these challenges, and most face a few at a time,โ€ Pekhimenko told TechCrunch in an email interview. โ€œThe highest-end chips are commonly unavailable due to the large demand from enterprises and startups alike. This leads to companies sacrificing on the size of the model they can deploy or results in higher inference latencies for their deployed models.โ€ With spending on AI-focused chips expected to hit $53 billion this year and more than double in the next four years, according to Gartner, Pekhimenko felt the time was right to launch software that could make models run more efficiently on existing hardware. โ€œTraining AI and machine learning models is increasingly expensive,โ€ Pekhimenko said. โ€œWith CentMLโ€™s optimization technology, weโ€™re able to reduce expenses up to 80% without compromising speed or accuracy.โ€ โ€œFor one of our customers, we optimized their Llama 2 model to work 3x faster by using Nvidia A10 GPU cards,โ€ CentML isnโ€™t the first to take a software-based approach to model optimization. It has competitors in MosaicML, which Databricks acquired in June for $1.3 billion, and OctoML, which landed an $85 million cash infusion in November 2021 for its machine learning acceleration platform. โ€œThe CentML platform can run any model,โ€ Pekhimenko said. โ€œCentML produces optimized code for a variety of GPUs and reduces the memory needed to deploy models, and, as such, allows teams to deploy on smaller and cheaper GPUs.โ€

์•„์ดํฐ ์ถœ์‹œ ํ›„ ์ฒซ 4~5๋…„ ์ •๋„๋Š” ์•ฑ์„ ์ถœ์‹œํ•˜๋Š” ๊ฒƒ๋งŒ์œผ๋กœ๋„ ์œ ์ €์˜ wow๋ฅผ ์ด๋Œ์–ด ๋‚ผ ์ˆ˜ ์žˆ์—ˆ๋˜ ์‹œ์ ˆ์ด ์žˆ์—ˆ๋‹ค. ์ด๋™ํ•˜๋ฉฐ ์ธํ„ฐ๋„ท๋„ ํ•  ์ˆ˜ ์žˆ๊ณ  ํ„ฐ์น˜์Šคํฌ๋ฆฐ์„ ํ™œ์šฉํ•  ์ˆ˜ ์žˆ๋Š” ์Šค๋งˆํŠธํฐ์—์„œ ๊ตฌ๋™์ด ๋œ๋‹ค๋Š” ๊ฒƒ๋งŒ์œผ๋กœ๋„ wow ์˜€๋˜ ๊ฒƒ์ด๋‹ค. ๊ทธ ์ดํ›„ ๊ธฐ์กด ๋ ˆ๊ฐ€์‹œ ์‚ฐ์—…์˜ ํŒจ๋Ÿฌ๋‹ค์ž„์„ ๋ฐ”๊พผ ์„œ๋น„์Šค๋“ค์ด ๋“ฑ์žฅํ•˜์—ฌ wow ๋ฅผ ์„ ์‚ฌํ–ˆ๋‹ค. ์œ ํˆฌ๋ธŒ/๋„ทํ”Œ๋ฆญ์Šค๋Š” TV/๋ฏธ๋””์–ด ์‚ฐ์—…์„ ๋ฐ”๊ฟจ๊ณ , ํŽ˜๋ถ/๊ตฌ๊ธ€์€ ๋ฏธ๋””์–ด/ํฌํ„ธ์‚ฐ์—…์„ ๋ฐ”๊ฟ”๋†“์•˜์œผ๋ฉฐ, ์•„๋งˆ์กด์€ ์‡ผํ•‘/์„œ๋ฒ„์‹œ์Šคํ…œ์„, ํ…Œ์Šฌ๋ผ๋Š” ์ž๋™์ฐจ ์‚ฐ์—…์„, ์šฐ๋ฒ„๋Š” ํƒ์‹œ ์‚ฐ์—…์„, ์—์–ด๋น„์—”๋น„๋ฅผ ํ˜ธํ…” ์‚ฐ์—…์— ํฐ ์˜ํ–ฅ์„ ์คฌ๋‹ค. ์œ„์™€ ๊ฐ™์€ 2000๋…„ ์ดˆ๋ฐ˜ ์ดํ›„ ๊ฒฉ๋™์˜ 20๋…„ ์ •๋„๋ฅผ ๊ฒช์œผ๋ฉฐ, ํ˜„๋Œ€์‚ฌํšŒ์˜ ๋งŽ์€ ์‚ฌ๋žŒ๋“ค์€ ๋ชจ๋ฐ”์ผ ๋ฐ ์›น ์ค‘์‹ฌ์˜ ์„œ๋น„์Šค์— ๊ฝค ๋งŽ์ด ์ต์ˆ™ํ•ด์ง„ ๋“ฏ ํ•˜๋‹ค. ๊ทธ๋ž˜์„œ, ์ƒˆ๋กœ ์ถœ์‹œ๋˜๋Š” ์•ฑ์„ ๋ด๋„ ๊ทธ ๊ฐํฅ์ด ๊ณผ๊ฑฐ ๋Œ€๋น„ ๋ฌด๋ŽŒ์ง€๊ณ  ์žˆ์Œ์„ ๋А๋‚€๋‹ค. ์˜คํžˆ๋ ค, ์š”์ฆ˜์€ ๊ณผ๊ฑฐ์—๋Š” ๋งŒ์กฑํ•˜๋ฉฐ ์‚ฌ์šฉํ•˜๋˜ Big Tech ์‚ฌ์˜ ์„œ๋น„์Šค ์กฐ์ฐจ๋„ ์Šฌ์Šฌ '์ง€๊ฒจ์›Œ์ง„๋‹ค' ๋А๋ผ๋Š” ๋ถ„๋“ค์ด ์ฆ๊ฐ€ํ•˜๋Š” ๋“ฏํ•˜๋‹ค. ๋™์‹œ์— '๊ณผ๊ฑฐ๋ถ€ํ„ฐ ๋งŽ์ด ์ด์šฉํ•˜๋˜ ์„œ๋น„์Šค๋“ค์ด ์ตœ๊ทผ์—๋Š” ์˜ˆ์ „๋งŒ ๋ชปํ•ด์„œ ์‹ค๋ง๊ฐ์ด ํฌ๊ณ , ์–ด์ฉ” ์ˆ˜ ์—†์ด ์ด์šฉํ•˜๊ณ  ์žˆ๋‹ค'๊ณ  ๋งํ•˜๋Š” ๋ถ„๋“ค๋„ ๋งŽ์•„์ง€๋Š” ๋“ฏํ•˜๋‹ค. OpenAI ๊ฐ€ chatGPT ๋ฅผ ์ถœ์‹œํ•˜๋ฉฐ ์ƒˆ๋กœ์šด ํ™œ๋ ฅ์„ ๋ถˆ์–ด๋„ฃ๊ณ  ์žˆ์ง€๋งŒ, ๊ทธ ํŒŒ๊ธ‰๋ ฅ์ด ๊ณผ๊ฑฐ์˜ ์•„์ดํฐ ๋ชจ๋ฉ˜ํŠธ์ฒ˜๋Ÿผ ์—„์ฒญ๋‚˜์ง€ ์•Š์€ ๊ฒƒ๋„ ์‚ฌ์‹ค์ด๋‹ค. (chatGPT๋Š” ์†Œ๋น„์ž ๋ณด๋‹ค๋Š” ์Šคํƒ€ํŠธ์—…์— ํ™œ๋ ฅ์„ ๋ถˆ์–ด๋„ฃ๊ณ  ์žˆ๋Š” ๋“ฏํ•˜๋‹ค. ๊ทธ๋ž˜์„œ ๊ฑฑ์ •์ด๊ธฐ๋„ ํ•˜๋‹ค. ์†Œ๋น„์ž๊ฐ€ ๋ฐ˜์‘ํ•˜์ง€ ์•Š์œผ๋ฉด ๊ฒฐ๊ตญ imapct ๊ฐ€ ์ œํ•œ์ ์ผ ์ˆ˜๋ฐ–์— ์—†๊ธฐ ๋•Œ๋ฌธ์ด๋‹ค) ์š”์ฆ˜์€ ์œ ์ €๋ถ„๋“ค์ด ๋Œ€๋‹ค์ˆ˜ ์„œ๋น„์Šค์— ๊ณผ๊ฑฐ ๋Œ€๋น„ ๋œจ๋œจ๋ฏธ์ง€๊ทผํ•œ ๋ฐ˜์‘์„ ๋ณด์ด๋Š” ์‹œ๊ธฐ์ธ ๊ฒƒ ๊ฐ™๊ธฐ๋„ ํ•˜๋‹ค. ๋‹ค๋งŒ, ์œ ์ €๋“ค์€ ์—ฌ์ „ํžˆ ๋” ๋‚˜์€ ์‚ถ์„ ์‚ด๊ณ  ์‹ถ์–ดํ•˜๋Š” ์š•๊ตฌ๊ฐ€ ๊ฐ•ํ•˜๋‹ค. ๋” ์„ฑ์žฅํ•˜๊ณ  ์‹ถ์–ดํ•˜๊ณ , ๋” ํ–‰๋ณตํ•œ ์‚ถ์„ ์‚ด๊ณ  ์‹ถ์–ดํ•˜๊ณ , ๋” ์ฆ๊ฑฐ์šด ์‹œ๊ฐ„์„ ๋ณด๋‚ด๊ณ  ์‹ถ์–ด ํ•œ๋‹ค. ๋ณธ์ธ ์ธ์ƒ์— ๊ฐ•๋ ฌํ•˜๊ฒŒ ๋‹ค๊ฐ€์˜ฌ ์ˆ˜ ์žˆ๋Š” ์„œ๋น„์Šค๋ฅผ ๋งŒ๋‚˜, ํ•ด๋‹น ์„œ๋น„์Šค์™€ ํ•จ๊ป˜ ๋” ๋งŒ์กฑ์Šค๋Ÿฌ์šด ์‚ถ์„ ์‚ด๊ธฐ๋ฅผ ํฌ๋งํ•˜๋Š” ๋ถ„๋“ค์ด ๊ฝค ์žˆ๋‹ค. ๊ทธ๋ž˜์„œ, ๋ฏธ๋ž˜์—๋Š” ์–ด๋–ค ์„œ๋น„์Šค๊ฐ€ ๋งŽ์€ ์œ ์ €๋กœ๋ถ€ํ„ฐ wow๋ฅผ ์ด๋Œ์–ด ๋‚ผ ์ˆ˜ ์žˆ์„์ง€... ๊ถ๊ธˆํ•˜๋‹ค. ์œ ์ €์˜ ์„ฑ์žฅ๊ณผ ๋งŒ์กฑ์— ๋” ๋งŽ์ด ์ง‘์ฐฉํ•˜๋Š” ์„œ๋น„์Šค๊ฐ€ ๊ทธ wow ๋ฅผ ์ด๋Œ์–ด ๋‚ผ ์ˆ˜ ์žˆ์œผ๋ฆฌ๋ผ ๋ฏฟ์ง€๋งŒ, ๊ณผ๊ฑฐ ๋Œ€๋น„ ์œ ์ €์˜ wow ๋ฅผ ์ด๋Œ์–ด ๋‚ด๊ธฐ ์œ„ํ•œ ๋‚œ์ด๋„๊ฐ€ ๋†’์•„์ ธ์„œ ๊ณผ๊ฑฐ ๋งŒํผ '๊ธ‰์„ฑ์žฅํ•˜๋Š” ์Šคํƒ€ํŠธ์—…'์ด ๋งŽ์ด ๋‚˜์˜ฌ ๊ฒƒ ๊ฐ™์ง€๋Š” ์•Š๋‹ค. ๊ทธ๋ž˜๋„, ์ด๋Ÿฐ ๋‚œ์„ธ๋ฅผ ์ด๊ฒจ๋‚ด๋Š” ๊ฐ€์žฅ ์ข‹์€ ๋ฐฉ๋ฒ•์€ '๋” ํž˜๋“ค์–ด์ง€๋Š” ํ™˜๊ฒฝ'์— ๋ฐ˜์‘ํ•˜๊ธฐ ๋ณด๋‹ค๋Š” '์œ ์ €'์— ์ง‘์ค‘ํ•˜์—ฌ ํ•˜๋ฃจ ํ•˜๋ฃจ ๋” ์ข‹์€ ์„œ๋น„์Šค๋ฅผ ๋งŒ๋“ค์–ด ๋‚ด๋Š” ๊ฒƒ์ด๋ผ ์ƒ๊ฐํ•œ๋‹ค. ๋ฌผ๋ก  ๊ณผ๊ฑฐ์—๋Š” 1๋…„ ๋…ธ๋ ฅํ•˜๋ฉด ๋งŒ์กฑ์„ ์ด๋Œ์–ด ๋‚ผ ์ˆ˜ ์žˆ์—ˆ๋‹ค๋ฉด, ์ตœ๊ทผ์—๋Š” 3~5๋ฐฐ๋Š” ๋” ํ•ด์•ผํ•˜๋Š” ์–ด๋ ค์›€์ด ์กด์žฌํ•˜์ง€๋งŒ ๋ง์ด๋‹ค. ๊ทธ๋ž˜์„œ ์š”์ฆ˜ ๊ฐ™์€ ์‹œ๋Œ€๋Š” Why? ๊ฐ€ ๋” ์ค‘์š”ํ•ด์ง€๋Š” ๊ฒƒ ๊ฐ™๋‹ค. ์™œ ์‹œ์ž‘ํ–ˆ๋Š”์ง€? ๋ฌด์—‡์„ ์œ„ํ•ด ์„œ๋น„์Šค๋ฅผ ๋งŒ๋“ค์–ด ๋‚ด๊ณ  ์žˆ๋Š”์ง€?์— ๋Œ€ํ•œ ๋ช…ํ™•ํ•˜๊ณ  ์†”์งํ•œ ๋‹ต์„ ๊ฐ€์ง€๊ณ  ์žˆ๋Š”์ง€๊ฐ€ ๋” ์ค‘์š”ํ•ด์ง€๋Š” ์‹œ๋Œ€๊ฐ€ ์™”๋‹ค. ๊ทธ ์ด์œ ์— ๋Œ€ํ•œ ๋‹ต์„ ์ „์ฒด ํŒ€์ด ๊ณต์œ ํ•˜๊ณ  ์žˆ์–ด์•ผ, wow ๋ฅผ ๋งŒ๋“ค์–ด ๋‚ด๊ธฐ ์œ„ํ•œ ๋…ธ๋ ฅ/์‹œ๊ฐ„์ด ์ฆ๊ฐ€ํ•œ ์‹œ๋Œ€๋ฅผ ๋ฌต๋ฌตํžˆ ์ด๊ฒจ๋‚˜๊ฐˆ ์ˆ˜ ์žˆ๊ธฐ ๋•Œ๋ฌธ์ด๋‹ค. ์„œ๋น„์Šค๋ฅผ ์ด์–ด๋‚˜๊ฐ€๊ณ  ์žˆ๋Š” ๋ณธ์งˆ์  ์ด์œ ๋Š” ๋ฌด์—‡์ธ๊ฐ€? ์šฐ๋ฆฌ ์„œ๋น„์Šค๋Š” ๋ˆ„๊ตฌ๋ฅผ ์œ„ํ•ด ์กด์žฌํ•˜๋ฉฐ ์™œ ์กด์žฌํ•ด์•ผ ํ•˜๋Š”๊ฐ€? ๊นŠ์€ ๋ฐค, ๋ง๊ธ€์„ ๋” ์„ฑ์žฅ์‹œ์ผœ ๋‚˜๊ฐ€๊ธฐ ์œ„ํ•œ ๋ฐฉ๋ฒ•์„ ๊ณ ๋ฏผํ•˜๋‹ค๊ฐ€, ์›๋ก ์ ์ธ ์งˆ๋ฌธ์— ๋Œ€ํ•ด ๋‹ค์‹œ ํ•œ ๋ฒˆ ์ƒ๊ฐํ•ด๋ณธ๋‹ค.

---- Navigating "corporate speak" isn't easy. Here's a helpful guide I put together: "Let me check with my team" = No "Possibly" = No "On my roadmap" = Not happening "This will be done in Q4" = This will be done in Q2 next year "Disagree and commit" = I hate you "Per my last email" = Try reading, for once in your life "Challenging landscape" = We're going out of business, quickly "Digital transformation" = We're going out of business, slowly "Let's circle back" = We'll never speak of this again "Take it offline" = We'll never speak of this again "30,000 foot view" = I don't know what I'm saying "Low hanging fruit" = Easy promotion "Open up the kimono" = HR violation "We use AI" = We don't use AI "We use machine learning" = We don't use machine learning "All hands on deck" = Let's actually try for once, please ๐Ÿ˜‚

โ€œ์ฐฝ์—…๊ฐ€์˜ ๊ธธ์— ๋“ค์–ด์„œ๋ฉด ํšŒ์‚ฌ๊ฐ€ ์—ฌ๋Ÿฌ๋ถ„์˜ ๊ฐ€์กฑ๋ณด๋‹ค ์ค‘์š”ํ•ด์ ธ์š”. ๊ฐ€์กฑ๋“ค์˜ ๋Œ€์†Œ์‚ฌ๋ฅผ ์ฑ™๊ธธ ์ˆ˜ ์—†๊ณ , ์ž๋…€๋“ค์—๊ฒŒ ๊ตฟ๋‚˜์ž‡ ํ‚ค์Šค๋ฅผ ํ•  ์ˆ˜ ์—†๊ฒŒ ๋  ๊ฑฐ์˜ˆ์š”. ๋˜ ํ˜„๊ธˆ ์ž์‚ฐ์ด ํ•„์š”ํ•ด ์ข‹์€ ์ง‘๊ณผ ์ข‹์€ ์ฐจ๋Š” ํฌ๊ธฐํ•ด์•ผํ•  ๊ฑฐ๊ณ ์š”. ์นœ๊ตฌ๋‚˜ ์ง€์ธ๋“ค์€ ์—ฌ๋Ÿฌ๋ถ„์„ ์ดํ•ดํ•˜์ง€ ๋ชปํ•˜๊ณ , ์ธ์ƒ์—์„œ ๊ธธ์„ ์žƒ๊ณ  ๋ฐฉํ™ฉํ•œ๋‹ค๊ณ  ์ƒ๊ฐํ•  ๊ฑฐ์˜ˆ์š”. ๋ถ€๋”” ๋ฐฉํ™ฉ์ด ์งง์•„์•ผํ•  ํ…๋ฐ ํ•˜๊ฒ ์ฃ .โ€ โ€œํŒ€์›๋“ค์˜ ์›”๊ธ‰์„ ์ค„ ์ˆ˜ ์—†๋Š” ์‹œ๊ธฐ๊ฐ€ ์˜ฌ ๊ฑฐ์˜ˆ์š”. ์ด ์—ฌ์ •์—์„œ ๋ˆ„๊ตฐ๊ฐ€๋Š” ๋‹น์‹ (์ฐฝ์—…๊ฐ€)์„ ๊ณ ์†Œํ•  ๊ฑฐ์˜ˆ์š”. ์—ฌ๋Ÿฌ๋ถ„์„ ๊ณ ์†Œํ•˜๋Š” ์‚ฌ๋žŒ์ด ๊ณต๋™ ์ฐฝ์—…์ž์ผ ์ˆ˜๋„ ์žˆ๊ณ  ํˆฌ์ž์ž์ผ ์ˆ˜๋„ ์žˆ์–ด์š”. ๋˜ ์ž์ฃผ ํ‹€๋ฆด ๊ฒƒ์ด๊ณ , ์ด ๋ฌธ์ œ๋“ค์„ ํŒ€์›๋“ค์ด ๋‹ค ๋ณด๊ฒŒ ๋  ๊ฑฐ์˜ˆ์š”. ์ชฝํŒ”๋ฆฐ ๋‹จ๊ณ„๋ฅผ ๋„˜์–ด์„œ๋ฉด ๋‚ด๊ฐ€ ํ•˜๋Š” ๊ฒŒ ๋งž๋‚˜, (๋Œ€ํ‘œ์ง์—์„œ) ๋‚ด๋ ค์™€์•ผ๊ฒ ๊ตฌ๋‚˜๋ผ๋Š” ์ƒ๊ฐ๋„ ๋“ค๊ฒŒ ๋˜์ฃ . ํŒ€์›๋“ค๋„ ๊ณ„์† ์—ฌ๋Ÿฌ๋ถ„์„ ์‹ค๋ง์‹œํ‚ฌ ํ…Œ์ง€๋งŒ, ๊ทธ๋Ÿผ์—๋„ ๊ณ„์† ์˜ˆ์ˆ˜์ฒ˜๋Ÿผ ์‚ฌ๋ž‘์„ ํผ์ค˜์•ผ ํ•ด์š”. ํŒ€์›๋“ค์„ ์œ„ํ•ด ๊ฒฉ๋ฆฌ ์‹œ์ผœ์•ผ ๋งˆ๋•…ํ•œ ์ง์›์—๊ฒŒ๋„ ์›ƒ์–ด์•ผ ํ•˜๋Š” ์‹œ๊ธฐ๊ฐ€ ์žˆ์Šต๋‹ˆ๋‹ค.โ€ (์ด์Šน๊ฑด ๋น„๋ฐ”๋ฆฌํผ๋ธ”๋ฆฌ์นด ๋Œ€ํ‘œ, ์ •์ฃผ์˜ ์ฐฝ์—…๊ฒฝ์ง„๋Œ€ํšŒ ๋ฐ๋ชจ๋ฐ์ด 2023) ์ •๋ง ๋ฆฌ์–ผํ•œ ์ฐฝ์—…์ž์˜ ์‚ถ. ์ง€๊ธˆ ์ƒํ™ฉ์ด ๋„ˆ๋ฌด ํž˜๋“ ๋ฐ ๋ฌด์Šจ ๋ถ€๊ท€์˜ํ™”๋ฅผ ๋ˆ„๋ฆฌ๊ฒ ๋‹ค๊ณ  ์ด๊ฑธ ํ•˜๋‚˜ ์‹ถ์ง€๋งŒ ๋˜ ๋‚˜๋ฆ„์˜ ์ด์œ ๊ฐ€ ์ƒ๊ฒจ์„œ ์ง€์†ํ•˜๋Š” ๊ฒƒ์ด์ฃ . ๊ฐœ์ธ์ ์œผ๋กœ๋Š” ์ด ์‚ถ์ด ๋‹น์—ฐํ•˜๊ธฐ๋ณด๋‹ค๋Š” ์ ์ฐจ ๊ฐœ์„ ๋์œผ๋ฉด ์ข‹๊ฒ ์Šต๋‹ˆ๋‹ค. ์ฐฝ์—…์ž์— ๋Œ€ํ•œ ์กด์ค‘๊ณผ ๋ณดํ˜ธ๋„ ์ปค์ ธ์•ผ ํ•˜๊ณ ์š”.

51% of Zendesk's revenue is international 47% of Cloudflareโ€™s revenue is international 46% of HubSpotโ€™s revenue is international 45% of MongoDB's revenue is international 38% of Asanaโ€™s revenue is international 34% of Twilioโ€™s revenue is international 22% of Oktaโ€™s revenue is international Go Global as soon as you have customers there Pull ahead

์ด ์‚ฌ๋žŒ์€ ์ž๊ธฐ๊ฐ€ ํ”Œ๋ ˆ์ดํ•˜๊ณ  ์žˆ๋Š” ๊ฒŒ์ž„์˜ ํŒ์ด ์–ด๋–ป๊ฒŒ ์„ค๊ณ„๋˜์–ด์žˆ๋Š”์ง€ ๋ถ„๋ช…ํ•˜๊ฒŒ ์•Œ๊ณ  ์žˆ๋„ค https://www.youtube.com/watch?v=Czmkes4yAOA

Long ๋ฒ ์ด๋น„๋ถ€๋จธ vs Short MZ BofA ๋Š” ์ตœ๊ทผ ๋ณด๊ณ ์„œ์—์„œ Long ๋ฒ ์ด๋น„๋ถ€๋จธ, Short ๋ฐ€๋ ˆ๋‹ˆ์–ผ (MZ) ํˆฌ์ž ์•„์ด๋””์–ด๋ฅผ ์ œ์‹œํ•ฉ๋‹ˆ๋‹ค. ์ง€๊ธˆ์˜ ๊ณ ๊ธˆ๋ฆฌ, ๊ณ ๋ฌผ๊ฐ€ ํ™˜๊ฒฝ์—์„œ ๋ฏธ๊ตญ๋‚ด ์†Œ๋น„ ์ง€์ถœ์˜ ์ฃผ๋„๊ถŒ์€ 65์„ธ ์ด์ƒ์˜ ๊ณ ๋ น์ธต์ด ์™„๋ฒฝํžˆ ์ฅ๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค. ํ•œ๊ตญ์˜ ๋ฒ ์ด๋น„๋ถ€๋จธ ์„ธ๋Œ€์ธ ์ €ํฌ ๋ถ€๋ชจ๋‹˜๋“ค์„ ๋ณด๋”๋ผ๋„ ๋Œ€๋ถ€๋ถ„ ์ž๊ฐ€๋กœ 1 ์ฃผํƒ ์ด์ƒ์„ ๋ณด์œ  ์ค‘์ž…๋‹ˆ๋‹ค. ๋Œ€์ถœ ๊ฑฐ์˜ ์—†์Šต๋‹ˆ๋‹ค. ์ž์‚ฐ ๊ฐ€๊ฒฉ์˜ ์ธํ”Œ๋ ˆ์† Wealth effect ๋ฌด์‹œ ๋ชปํ•ฉ๋‹ˆ๋‹ค. ๋ฐ˜๋ฉด, MZ ์„ธ๋Œ€๋Š” ์ฃผ๋กœ ์“ฐ๊ณ  ์ฃฝ์ž๋Š” ์„ธ๊ณ„๊ด€, ์ตœ๊ทผ ๋ช‡๋…„๊ฐ„ ๋ถ€๋™์‚ฐ ์˜๋Œ์กฑ, ์ฝ”์ธ/์ฃผ์‹ ๋นšํˆฌ์กฑ์˜ ํ•ฉ๋ฅ˜๋กœ ๋ถ€์ฑ„ ๋ถ€๋‹ด์ด ์ƒ๋‹นํ•ฉ๋‹ˆ๋‹ค. ์ˆจ๋งŒ ์‰ฌ์–ด๋„ ์ด์ž์˜ Snowball effect ๋ฅผ ๊ฒฝํ—˜ํ•˜๊ณ  ์žˆ์ฃ . ๋”์šฑ์ด ๋ฒ ์ด๋น„๋ถ€๋จธ๋Š” ๋ˆ์ด ๋งŽ๋“  ์ ๋“  ์—ฌ์ „ํžˆ ์ €์ถ•์— ๋Œ€ํ•œ ๋ฏฟ์Œ์ด ์žˆ์Šต๋‹ˆ๋‹ค. ๋ฏธ๊ตญ๋„ Cash return ์ด 5% ๋ฅผ ์ƒํšŒํ•ฉ๋‹ˆ๋‹ค. Cash is king ์ธ ์ž‘๊ธˆ์˜ ์ƒํ™ฉ์€ ์„ธ๋Œ€๊ฐ„ ๊ฒฉ์ฐจ๋ฅผ ๋˜๋Œ๋ฆฌ๊ธฐ ํž˜๋“ค ์ •๋„๋กœ ๋ฒŒ๋ฆฌ๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค. ์Šน์ž์™€ ํŒจ์ž๊ฐ€ ๋ช…ํ™•ํžˆ ๊ฐˆ๋ฆฌ๋Š” ๊ฒŒ์ž„์ด๋ผ๋ฉด, ์„นํ„ฐ์™€ ์‚ฐ์—…๋‚ด ํˆฌ์ž ์•„์ด๋””์–ด๋กœ ํ™œ์šฉํ•ด ๋ณผ๋งŒ ํ•ฉ๋‹ˆ๋‹ค.

๐Ÿ“ข Introducing Voyage AI Founded by a talented team of leading AI researchers and me ๐Ÿš€๐Ÿš€. We build state-of-the-art embedding models (e.g., better than OpenAI ๐Ÿ˜œ). We also offer custom models that deliver ๐ŸŽฏ+10-20% accuracy gain in your LLM products. ๐Ÿงต We think embeddings are an under-loved, but incredibly important, part of everyoneโ€™s retrieval stack. So we set out to build just better embeddings. Voyage embeddings are SOTA on MTEB and 9 held-out benchmarks covering industry domains. Learn more here: https://lnkd.in/gk59ZE-T We partnered with @LangChainAI to help enhance their official chatbot, live at chat.langchain.com. Both our base and fine-tuned embeddings improve both the retrieval and response quality, and the latter is deployed! Check out https://lnkd.in/gZ2aH4KU for more details. Try Voyage embeddings for free โ€“ your first 5000 documents are on us! Create an account at voyageai.com and check out our API docs at https://docs.voyageai.com . To fine-tune your own custom model, email โœ‰๏ธ contact@voyageai.com for early access!

๋‚ด๊ฐ€ ์ผ๋ณธ์— ๊ทผ๋ฌดํ•  ๋‹น์‹œ ๋ฐ˜๋„์ฒด ํˆฌ์ž๊ฐ€ ์ด๋ค„์กŒ๊ธฐ ๋•Œ๋ฌธ์— ๋„์ฟ„ ์ง€์‚ฌ์—๋Š” ์ด๋ณ‘์ฒ  ํšŒ์žฅ์˜ ์ž๋ฌธ์—ญ์„ ํ•ด์ฃผ๋Š” ๋ฐ˜๋„์ฒด ๊ด€๋ จ ์ „๋ฌธ๊ฐ€๋“ค์ด ์ž์ฃผ ๋ฐฉ๋ฌธํ–ˆ๋‹ค. ์ž๋ฌธํ•œ ์ผ๋ณธ์ธ ๋Œ€ํ•™๊ต์ˆ˜์—๊ฒŒ โ€œ์šฐ๋ฆฌ ํšŒ์žฅ๋‹˜์ด ๋ฐ˜๋„์ฒด๋ฅผ ์–ผ๋งˆ๋‚˜ ์•Œ๊ณ  ๊ณ„์‹ญ๋‹ˆ๊นŒ?โ€ ๋ผ๊ณ  ๋ฌผ์–ด๋ดค๋‹ค. ๊ทธ๋ถ„์€ ์›ƒ์œผ๋ฉด์„œ โ€œ์ดˆ๋“ฑํ•™์ƒ ์ˆ˜์ค€์ž…๋‹ˆ๋‹คโ€๋ผ๊ณ  ๋‹ตํ–ˆ๋‹ค. โ€œ๊ทธ๋Ÿฐ๋ฐ ์™œ ์ฃผ์œ„ ์‚ฌ๋žŒ๋“ค์˜ ๋ฐ˜๋Œ€๋ฅผ ๋ฌด๋ฆ…์“ฐ๊ณ  ๋ฐ˜๋„์ฒด ํˆฌ์ž๋ฅผ ๋ฐ€์–ด๋ถ™์ด์‹ค๊นŒ์š”?โ€ํ•˜๊ณ  ๋‹ค์‹œ ๋ฌผ์œผ๋‹ˆ ๊ทธ๋ถ„์€ โ€œํšŒ์žฅ๋‹˜์˜ ์‚ฌ์—…์— ๋Œ€ํ•œ ๊ฐ๊ฐ์€ ๋™๋ฌผ์  ๋ณธ๋Šฅ์ž…๋‹ˆ๋‹ค. ์•„๋ฌด๋„ ๋ง๋ฆด ์ˆ˜ ์—†์Šต๋‹ˆ๋‹คโ€๋ผ๊ณ  ๋‹ตํ•ด ํ•จ๊ป˜ ์›ƒ์—ˆ๋‹ค. ๊ฒฐ๊ณผ์ ์œผ๋กœ ํ•œ๊ตญ์˜ ์ „์ž์‚ฐ์—…์ด ์ผ๋ณธ์„ ์ถ”์›”ํ•  ์ˆ˜ ์žˆ๋˜ ๊ฑด ์˜ค๋„ˆ ์œ„์ฃผ์˜ ์ง€๋ฐฐ๊ตฌ์กฐ์—์„œ ์‹ ์†ํ•˜๊ณ  ์žฅ๊ธฐ์ ์ธ ์˜์‚ฌ๊ฒฐ์ •์„ ํ–ˆ๊ธฐ ๋•Œ๋ฌธ์ด๋‹ค. ๊ณผ๊ฑฐ ์ด๊ฑดํฌ ํšŒ์žฅ์€ ์ฃผ์žฌ์›์ด ๊ท€๊ตญํ•  ๋•Œ ๋ฐ˜๋“œ์‹œ ์ผ๋ณธ์˜ ์ „์ž์ œํ’ˆ์„ ํ•˜๋‚˜์”ฉ ์‚ฌ์™€์„œ ๋ถ„์„ํ•˜๋ผ๊ณ  ์ง€์‹œํ–ˆ๋‹ค. ๋‚˜์ฒ˜๋Ÿผ ๊ธฐ์ˆ  ๋ฌธ์™ธํ•œ์ธ ์‚ฌ๋žŒ์—๊ฒŒ๋Š” ๊ณคํ˜น์Šค๋Ÿฌ์šด ์ผ์ด์—ˆ๋‹ค. ์‚ผ์„ฑ์ „์ž๊ฐ€ VTR ๊ฐœ๋ฐœ์— ๋‚˜์„  ์ดˆ๊ธฐ์—๋Š” ๋ถˆ๋Ÿ‰๋ฅ ์ด ๋„ˆ๋ฌด ๋†’์•„ ์šด๋™์žฅ์— ์ œํ’ˆ์„ ์Œ“์•„์„œ ๋ถˆ์„ ์ง€๋ฅด๊ธฐ๋„ ํ–ˆ๋‹ค. ๋„์‹œ๋ฐ”์˜ ํ˜‘์กฐ๋ฅผ ์–ป์–ด ๋ผ์ธ์˜ ์กฐ์žฅ๋“ค์„ ๋„์‹œ๋ฐ” ๊ณต์žฅ์— ๋ณด๋‚ด ์‹ค์Šต์„ ์‹œํ‚ค๊ธฐ๋กœ ํ•œ ์ ์ด ์žˆ๋‹ค. ํ•œ๋ฐ ์‹ค์Šต์ด ์ง„ํ–‰๋˜๋˜ ์ค‘ ๋„์‹œ๋ฐ”์—์„œ ๋”๋Š” ํ•œ๊ตญ์˜ ์—ฐ์ˆ˜์ƒ์„ ๋ฐ›์ง€ ์•Š๊ฒ ๋‹ค๊ณ  ํ†ต๋ณดํ–ˆ๋‹ค. ๋„์‹œ๋ฐ” ๊ฐ„๋ถ€๋กœ๋ถ€ํ„ฐ ์ „ํ•ด ๋“ค์€ ์‚ฌ์œ ๋Š” ์–ด์ด๊ฐ€ ์—†์—ˆ๋‹ค. 3์ฐจ ์‹ค์Šต์ƒ๋“ค์ด๋‚˜ 1ยท2์ฐจ ์‹ค์Šต์ƒ๋“ค์ด ํ•˜๋Š” ์งˆ๋ฌธ์ด ๋™์ผํ•˜๋‹ค๊ณ  ํ–ˆ๋‹ค. ๊ฐ™์€ ์งˆ๋ฌธ์ด ๋‚˜์˜ค๋Š”๋ฐ ๋น„์‹ผ ๊ฒฝ๋น„๋ฅผ ๋“ค์—ฌ ํ•ด์™ธ ์—ฐ์ˆ˜๋ฅผ ํ•  ํ•„์š”๊ฐ€ ์žˆ๋А๋ƒ๋Š” ์–˜๊ธฐ์˜€๋‹ค. ์กฐ๊ธˆ ๋ชจ์š•์ ์œผ๋กœ๋„ ๋“ค๋ ธ์ง€๋งŒ ํ‹€๋ฆฐ ์ง€์ ์ด ์•„๋‹ˆ์—ˆ๋‹ค. ๊ฐœ์ธ์˜ ์ด์ต๋ณด๋‹ค ์ง‘๋‹จ์˜ ์ด์ต์„ ์šฐ์„ ์‹œํ•˜๊ณ , ์ƒˆ๋กœ์šด ์ง€์‹์€ ๊ณต์œ ํ•˜๋Š” ์ผ๋ณธ์ธ์˜ ์‚ฌ๊ณ ๋ฐฉ์‹์œผ๋กœ๋Š” ์šฐ๋ฆฌ ์—ฐ์ˆ˜์ƒ๋“ค์˜ ๊ฐœ์ธ์ฃผ์˜์  ์‚ฌ๊ณ ๋ฐฉ์‹์ด ์ดํ•ดํ•˜๊ธฐ ์–ด๋ ค์› ์„ ๊ฒƒ์ด๋‹ค. ์š”์ฆ˜ ์•„ํ‚คํ•˜๋ฐ”๋ผ ๋งค์žฅ์˜ ์•ž์ž๋ฆฌ๋ฅผ ์ฐจ์ง€ํ•˜๊ณ  ์žˆ๋Š” ์‚ผ์„ฑ, LG TV๋ฅผ ๋ณด๋ฉด 40๋…„ ์ „์— ์ฐฝํ”ผํ–ˆ๋˜ ๊ทธ๋•Œ์˜ ์ผ์ด ๋– ์˜ค๋ฅด๊ณค ํ•œ๋‹ค.

์‹ ๋™์•„์—๋Š” ๊ฐ€๋” ์ข‹์€ ๊ธ€๋“ค์ด ์žˆ๋‹ค. https://shindonga.donga.com/economy/3/03/13/4503748/1