uz
Feedback
Continuous Learning_Startup & Investment

Continuous Learning_Startup & Investment

Kanalga Telegram’da oβ€˜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
2 301
Obunachilar
-524 soatlar
-107 kunlar
-3630 kunlar
Postlar arxiv
2024λ…„μ˜ 절반이 κ°”λ‹€. μ–΄λŠμƒˆ λ²€μ²˜νˆ¬μžμžλ‘œμ„œ μΌν•˜λŠ” μ‹œκ°„λ³΄λ‹€ 사업을 ν•˜λŠ” ν¬μ§€μ…˜μ΄ 된 μ‹œκ°„μ΄ 3λ°°κ°€ λ„˜μ–΄κ°„λ‹€. νˆ¬μžμžμ™€ μ‚¬μ—…κ°€μ˜ μ°¨μ΄λŠ” 생각보닀 맀우 큰 것 κ°™λ‹€. 거의 12λ…„μ§Έ 사업을 ν•˜λŠ” μž…μž₯μ—μ„œ λ²€μ²˜μΊν”Όνƒˆλ¦¬μŠ€νŠΈλ‘œμ„œλŠ” κ²°μ½” μ΄ν•΄ν•˜μ§€ λͺ»ν–ˆμ„ 것 같은 점듀이 λ§Žλ‹€. λͺ‡ κ°€μ§€ λ– μ˜€λ₯΄λŠ” 것을 보면, 1) κ³„νšλŒ€λ‘œ λ˜λŠ”κ²Œ 단 ν•œκ°€μ§€λ„ μ—†λ‹€λŠ” 것을 μ•Œκ²Œ λœλ‹€. 10κ°€μ§€ 가섀을 μ‹€ν–‰ν•˜λ‹€λ³΄λ©΄ κ·Έλž˜λ„ λ˜λŠ”κ²Œ 2-3κ°œλŠ” μžˆμ§€ μ•Šμ„κΉŒ μƒκ°ν•˜μ§€λ§Œ, μ‹€μ œλ‘œλŠ” 10개λ₯Ό μ‹œλ„ν•΄λ„ (μ•Όμ†ν•˜λ¦¬λ§ŒνΌ) 단 1κ°œλ„ κ³„νšλŒ€λ‘œ λ˜λŠ”κ²Œ μ—†λ‹€λŠ” 것을 ν™•μΈν•˜κ²Œ λ˜λŠ”κ²Œ λ‹€λ°˜μ‚¬λ‹€. 2) νœ΄κ°€ 등을 λ‚΄μ„œ 자리λ₯Ό λΉ„μš°λ©΄ λ°˜λ“œμ‹œ κ·Έ κΈ°κ°„ λ™μ•ˆμ— λ¬Έμ œκ°€ ν„°μ§„λ‹€. νœ΄μ‹, λ¦¬ν”„λ ˆμ‰¬ 등을 이유둜 짧건 길건 μ•½κ°„μ˜ νœ΄κ°€λ₯Ό μ„ΈνŒ…ν•˜κ³  λ‚˜λ©΄, κ³΅κ΅λ‘­κ²Œλ„ λ°˜λ“œμ‹œ κ·Έ 자리λ₯Ό λΉ„μš°λŠ” κΈ°κ°„ λ™μ•ˆμ— λ¬Έμ œκ°€ ν„°μ§€κ³ , κ·Έλž˜μ„œ νœ΄κ°€λŠ” λ¬Όκ±°ν’ˆμ΄ λœλ‹€. κ·Έλž˜μ„œ μ–Έμ  κ°€λΆ€ν„° 미리 νœ΄κ°€ 기간을 μž‘μ„ 수 μ—†λ‹€. 3) 가끔 μƒκ°ν–ˆλ˜ μ •λ„μ˜ μ„±κ³Όκ°€ λ‚˜μ„œ κ΄€λ ¨λœ ꡬ성원듀과 celebration을 ν•˜κ³  λ‚˜λ©΄, ν•œλ‹¬ 이내에 λ°˜λ“œμ‹œ 큰 λ¬Έμ œλ‚˜ μ˜ˆμƒμΉ˜ λͺ»ν•œ negative eventκ°€ 생긴닀. κ·Έλž˜μ„œ 쒋은 일이 생겨도 μ˜¨μ „νžˆ μ¦κΈ°κ±°λ‚˜ μΆ•ν•˜ν•˜κΈ°κ°€ 쉽지 μ•Šλ‹€. 이런걸 보면 λ‚΄κ°€ ꡳ이 μ™œ 이런 고생길을 ꡳ이 νƒν•΄μ„œ κ°€μ•Όν•˜λ‚˜ 싢닀가도, 가뭄에 μ½©λ‚˜λ“―μ΄ λ‹€λ“€ μ•ˆλ κ±°λΌ μ–˜κΈ°ν–ˆλŠ”λ° 1년에 ν•œλ‘λ²ˆ 정도 λ˜λŠ” 이벀트λ₯Ό κ²ͺ닀보면, 이게 μ°Έ λ§ˆμ•½μ²˜λŸΌ 쀑독성이 μžˆλ‹€λŠ”κ±Έ μ•Œκ²Œ λœλ‹€. 고생길인쀄 μ•Œλ©΄μ„œλ„, λ‹€μ‹œ μΌμ–΄λ‚˜μ„œ ν•  수 밖에 μ—†λŠ” λ³€νƒœ 같은 쀑독성 πŸ™‚ κ·Έλž˜μ„œ μ—¬λŸ¬ 갑둠을박이 μžˆλ”λΌλ„ μ°½μ—…μžλ“€ μ‚¬μ΄μ—λŠ” λ¬΄μ–Έμ˜ κ³΅κ°λŒ€λΌλŠ” 것을 κΉ”κ³  κ°„λ‹€. 각기 μ§€κ³  μžˆλŠ” λ¬΄κ²ŒλŠ” λ‹€λ₯΄λ”라도 큰 짐을 μ§€κ³  κ°€λŠ”, κ·Έ κ³Όμ •μ—μ„œ λ‚˜λ¦„μ˜ 희열 ν¬μΈνŠΈκ°€ μžˆλŠ” λ³€νƒœ 같은 λŒμ—°λ³€μ΄μ’…. λ°•μ§€μ›… λŒ€ν‘œλ‹˜.

μ™Έμ‹μ—…μ—μ„œ μ‚°μ „μˆ˜μ „ κ²ͺμ–΄λ³Έ μ‚¬λžŒλ“€μ΄ ν›Œλ₯­ν•œ 쉐프가 사μž₯으둜 μ„±μž₯ν•˜κΈ° μœ„ν•΄μ„œ ν•΄μ£ΌλŠ” μ‘°μ–Έλ“€. https://youtu.be/_Oo3Ji4MVZU?si=gm_czz69_vzuKJaw

μ •μΉ˜κ°€ (μ‡Όν† μΏ  νƒœμž)μ—μ„œ 사상가 (ν›„μΏ μžμ™€ μœ μΉ˜ν‚€)λ₯Ό 거쳐 경제인 (μ‹œλΆ€μ‚¬μ™€ μ—μ΄μ΄μΉ˜) κΉŒμ§€. μ‹œλΆ€μ‚¬μ™€ μ—μ΄μ΄μΉ˜ κ΄€λ ¨ μ—¬λŸ¬ μ˜μƒμ΄ μžˆμ§€λ§Œ 이창민 κ΅μˆ˜λ‹˜ μ˜μƒμ„ λ³΄μ‹œλ©΄ κ·Έκ°€ μ–Όλ§ˆλ‚˜ μœ„λŒ€ν•œ μΈλ¬Όμ΄μ—ˆλŠ”μ§€ μ•Œ 수 μžˆμŠ΅λ‹ˆλ‹€. 이창민 κ΅μˆ˜λ‹˜μ˜ 일본정독 μ˜μƒλ“€μ€ 사싀 4ν™” 뿐만 μ•„λ‹ˆλΌ μ „λΆ€ λ‹€ κ·€ν•˜κ³  쒋은 λ‚΄μš©μž…λ‹ˆλ‹€. https://n.news.naver.com/mnews/article/023/0003844258 https://youtu.be/QnRT2tfRMAA?si=t3fHb0hROGJZDuiq&t=2220 (이창민의 일본정독 4ν™”: 37λΆ„λΆ€ν„°)

일본은 μ™œ 미ꡭ의 침범을 λ°›κ³  λ‚œ 뒀에 ν™• λ³€ν™”ν•  수 μžˆμ—ˆλŠ”κ°€? μ§€λ°©μ˜ 힘. μ§€λ°©μ˜ λ²ˆμ£Όλ“€μ΄ μ „μŸν•˜κ³  νŒ¨λ°°ν•˜λ©΄μ„œ λ°°μ›Œμ•Όν•œλ‹€λŠ” 것을 κΉ¨λ‹«κ³ , 그듀끼리 κ²½μŸν•˜λ©΄μ„œ λ°œμ „ν•  수 μžˆμ—ˆμŒ. 각 μ§€λ°©λΌλ¦¬μ˜ 열망이 λΉ λ₯Έ λ³€ν™”λ₯Ό λ§Œλ“€μ—ˆμŒ. 집단 지성이 μ‹œλ„ˆμ§€κ°€ λ‚˜λ €λ©΄ 쀑앙 μ§‘κΆŒμ΄ μ•„λ‹ˆλΌ μ„œλ‘œ 달라야 μ‹œλ„ˆμ§€κ°€ λ‚˜μ˜¬ 수 있음. μ„œμ–‘κ³Ό ꡭ지전을 치λ₯΄λ©΄μ„œ 제ꡭ이 될 것이냐 식민지가 될 κ²ƒμ΄λƒλΌλŠ” 선택을 μ§μ‹œν–ˆμŒ. 세상을 μ–΄λ–»κ²Œ 바라볼 것이냐

https://maily.so/capitaledge/posts/99ed5dd7 맀트리슀λ₯Ό νŒλ§€ν•˜λ˜ μΊμŠ€νΌλŠ” 헐값에 λ§€κ°λ˜μ—ˆκ³  μ‹€λ¦¬μ½˜λ°Έλ¦¬μ˜ μ‹ λ°œλ‘œ μ£Όλͺ©λ°›μ•˜λ˜ μ˜¬λ²„μ¦ˆλŠ” 상μž₯ ν›„ μ£Όκ°€κ°€ 99% ν•˜λ½ν•œ μƒν™©μž…λ‹ˆλ‹€. κ·ΈλŸΌμ—λ„ λΆˆκ΅¬ν•˜κ³  μ—¬μ „νžˆ μ†ŒλΉ„μž μ‹œμž₯에 μ§‘μ€‘ν•˜λŠ” νˆ¬μžμžλ“€μ΄ μžˆμŠ΅λ‹ˆλ‹€. DTC λͺ¨λΈμ΄ 어렀움을 κ²ͺ은 것이지, μ†ŒλΉ„μž μ„Ήν„°μ—μ„œλŠ” μ—¬μ „νžˆ μƒˆλ‘œμš΄ 기업듀이 νƒ„μƒν•˜κ³  λΉ λ₯΄κ²Œ μ„±μž₯ν•˜λ©° μ†ŒλΉ„μžλ“€μ˜ 지갑을 μ—΄κ²Œ ν•˜λŠ” 사둀듀이 κ³„μ†ν•΄μ„œ λ‚˜νƒ€λ‚˜κ³  있기 λ•Œλ¬Έμž…λ‹ˆλ‹€.

수읡λ₯  μ–˜κΈ°λŠ” μ•„λ‹ˆκ³  μƒλ°˜κΈ°μ— 느꼈던 점을 λ˜μ§šμ–΄λ΄€μŠ΅λ‹ˆλ‹€. 글이 μ–΄λ ΅κ²Œ μ“°μ˜€λ˜ 만큼, 아직 덜 깨달은 것을 μ˜λ¬΄κ°μ— μž‘μ„±ν•œ 것 κ°™μŠ΅λ‹ˆλ‹€. 무언가 λͺ…λ£Œν•˜κ²Œ 말할 수 μžˆλ‹€λŠ” 건 속이 μ‹œμ›ν•œ μΌμ΄μ§€λ§Œ, 그러기 μœ„ν•΄μ„œλŠ” 정말 λ§Žμ€ κ³ λ―Όκ³Ό 곡뢀가 μˆ˜λ°˜λ˜λ„€μš”. μ–Έμ  κ°€λŠ” 쉽고 μ‹¬ν”Œν•˜κ²Œ 해보고 μ‹ΆμŠ΅λ‹ˆλ‹€. (딸깍) https://freebutdeep.substack.com/p/34-2024 #투자

Shopify의 λͺ¨λ“  λ―Έμ…˜μ€ 6μ£Όλ§ˆλ‹€ κ²€ν† λ©λ‹ˆλ‹€. κ·Έλž˜μ„œ ν† λΉ„, κΈ€λ Œ, μ €λŠ” ν•œ 방에 λͺ¨μ—¬ νšŒμ‚¬μ˜ λͺ¨λ“  λ―Έμ…˜μ„ κ²€ν† ν•©λ‹ˆλ‹€. λ”°λΌμ„œ νšŒμ‚¬μ˜ μ–΄λ–€ λ―Έμ…˜λ„ μ΅œλŒ€ 6μ£Ό λ™μ•ˆ κΆ€λ„μ—μ„œ λ²—μ–΄λ‚  수 μ—†μŠ΅λ‹ˆλ‹€. 변경이 ν•„μš”ν•˜λ‹€λ©΄ λ°”λ‘œ λ³€κ²½ν•©λ‹ˆλ‹€. λŒ€λΆ€λΆ„μ˜ νšŒμ‚¬μ—μ„œ λͺ‡ 달, λͺ‡ λΆ„κΈ°, λͺ‡ 년이 κ±Έλ¦¬λŠ” 일을 Shopifyμ—μ„œλŠ” μ΅œλŒ€ 6μ£Ό λ§Œμ— 끝낼 수 μžˆμŠ΅λ‹ˆλ‹€." (Source: 2023 Shopify Investor Day)

https://youtu.be/lPrRVnxHqvE?si=TSZZMu0led8YWZ81 ν•˜λ£¨ν•˜λ£¨ μ€‘μš”ν•œ 것에 μΆ©μ‹€ν•œ 것. 그런 ν•˜λ£¨λ“€μ΄ λˆ„μ λ˜λŠ” 것.

이건 λΆ„λͺ… μ€€λΉ„λœ 자만이 μ•Œ 수 μžˆλŠ” κ°μ΄μ—ˆμ„ν…Œλ‹€. "λΆ€μ²œμ„ λ– λ‚˜ ν°λ¬Όμ—μ„œ 놀 λ•Œκ°€ λλ‹€λŠ” 감이 μ™”λ‹€." https://www.joongang.co.kr/article/25260802

κ³΅λΆ€λΏλ§Œμ΄ μ•„λ‹ˆλΌ, μ–΄λ €μ„œλΆ€ν„° μ›¬λ§Œν•œ 건 λ‹€ 직접 κ²°μ •ν–ˆλ‹€. λ¬Όλ‘  무슨 λ¬Έμ œκ°€ 생기면 λΆ€λͺ¨λ‹˜μ΄ λ‚˜μ™€ 같이 ν•΄κ²° 방법을 μ°Ύμ•˜κ³ , ν‰μ†Œ λŒ€ν™”λ₯Ό 많이 ν–ˆλ‹€. κ·Έλž˜λ„ μ—¬μ „νžˆ 결정은 λ‚΄ λͺ«μ΄μ—ˆλ‹€. μ§„λ‘œλ‚˜ 인생 λ°©ν–₯도 λ‚΄κ°€ μ„ νƒν–ˆκ³ , λΆ€λͺ¨λ‹˜μ€ μ‘΄μ€‘ν–ˆλ‹€. κ·ΈλŸ¬λ‹ˆ 되렀 μ•žμœΌλ‘œ μ–΄λ–»κ²Œ μ‚΄μ•„μ•Ό ν• μ§€ 생각이 λ§Žμ•˜λ‹€. λ‚΄κ°€ 뭘 μ’‹μ•„ν•˜κ³  뭘 μ›ν•˜λŠ”μ§€ 늘 골똘히 μƒκ°ν•˜λ©° 닡을 μ°Ύμ•˜λ‹€. 그게 μ‚¬μ—…μ΄μ—ˆλ‹€. μ–΄μ©Œλ©΄ 혼자만의 μ—΄λ“±κ°μ΄μ—ˆμ§€ λͺ¨λ₯Έλ‹€. μ–΄μ¨Œλ“  μ–Ήμž–μ„ λ•Œκ°€ 적지 μ•Šμ•˜μ§€λ§Œ, κ·Έλ•Œλ‚˜ μ§€κΈˆμ΄λ‚˜ λΆˆνŽΈν•œ 상황을 맞λ‹₯뜨리면 였히렀 긍정적 μžκ·Ήμ„ λ°›λŠ”λ‹€. κ°€λ Ή 2021λ…„ 'λ§ˆλ—‘ν‚΄'이 처음 νˆ¬μžλ°›μ„ 즈음 "'룩 뢁(λΈŒλžœλ“œ μŠ€νƒ€μΌμ„ λ³΄μ—¬μ£ΌλŠ” 사진집)' μ—†λŠ” λΈŒλžœλ“œκ°€ 무슨 λΈŒλžœλ“œλƒ"λŠ” μ†Œλ¦¬λ₯Ό λ“€μ—ˆλ‹€. 아무리 업계 κ΄€ν–‰μ΄λΌμ§€λ§Œ, 룩 뢁 ν•„μš”μ„±μ„ λͺ» λŠλΌλŠ”λ° 남이 ν•˜λ‹ˆκΉŒ 무쑰건 따라 ν•˜κ³  μ‹Άμ§„ μ•Šμ•˜λ‹€. κ·Έλ“€ μž…λ§›μ„ λ§žμΆ”λŠ” λŒ€μ‹  "λ‚΄κ°€ λ§žλ‹€λŠ” κ±Έ 보여주겠닀"κ³  λ‹€μ§ν–ˆλ‹€. κΈˆμˆ˜μ €λŠ”μ»€λ…• 두 λ”Έ λŒ€ν•™ 학비도 λΉ λ“―ν•œ ν‰λ²”ν•œ 맞벌이 κ°€μ •μ˜ κ³ μ‘Έ ν•™λ ₯인 λ‚΄κ°€ μ„±κ³΅ν•˜λŠ” 사둀λ₯Ό 보여주고 μ‹Άμ—ˆλ‹€. 내겐 그런 μ‘΄μž¬κ°€ μ—†μ—ˆκΈ°μ—, λΉ„μŠ·ν•œ κΏˆμ„ κΎΈλŠ” μ΄λ“€μ—κ²Œ κΈΈλΌμž‘μ΄κ°€ 되고 μ‹Άλ‹€. λ™λŒ€λ¬Έμ€ 같은 λ””μžμΈμ„ μ΅œμ†Œ 두 μž₯ 사야 ν–ˆκΈ°μ—, μ›ν–ˆλ˜ 15만 μ›μ§œλ¦¬ λ©‹μ§„ λ””μžμΈ λŒ€μ‹  κ°œλ‹Ή 2만~3만 μ›λŒ€ κ°’μ‹Ό 점퍼 두 μž₯을 골라 λΈ”λ‘œκ·Έμ— μ˜¬λ Έλ‹€. μ•ˆ νŒ”λ Έλ‹€. μΆ• μ²˜μ§„ λ‚˜λ₯Ό μ§€μΌœλ³΄λ˜ μ–΄λ¨Έλ‹ˆκ°€ "μ–Όλ§ˆλ©΄ λ˜κ² λŠλƒ"κ³  λ¬Όμ—ˆλ‹€. "30λ§Œμ›. " κ·Έ 돈으둜 λ™λŒ€λ¬Έμ— 달렀가 거친 상인듀 μƒλŒ€ν•˜λ©° 점퍼 두 μž₯을 샀닀. λ§‰μ°¨λŠ” μ§„μž‘μ— 끊기고 μˆ˜μ€‘μ—” 4000~5000μ›λΏμ΄μ—ˆλ‹€. μΉ΄νŽ˜μ—μ„œ 유자차λ₯Ό μ•žμ— 두고 λΆ€μ²œκ°€λŠ” 첫차λ₯Ό 기닀리며 ν¬μŠ€νŠΈμž‡μ— μ΄λ ‡κ²Œ 썼닀. "성곡할 κ±°λ‹€. " 이 λ©”λͺ¨λŠ” μžκΈ°μ‹€ν˜„μ  μ˜ˆμ–Έμ΄ 됐닀. 6κ°œμ›”μ―€ μ§€λ‚˜λ‹ˆ 옷 μ‚΄ 돈 말고도 100λ§Œμ›μ―€ 남아 μƒˆ μ‹œλ„λ₯Ό ν–ˆλ‹€. 단일 ν’ˆλͺ©μ΄ μ•„λ‹ˆλΌ μ—¬κΈ°μ €κΈ°μ„œ μ½”νŠΈΒ·μŠ€μΉ΄ν”„ λ“±μœΌλ‘œ 직접 μ½”λ””ν•΄ ν™”λ³΄μ²˜λŸΌ 찍어 μ˜¬λ ΈλŠ”λ°, μ›” λͺ‡μ²œλ§Œμ› 맀좜이 λ‚  만큼 λŒ€λ°•μ΄ 났닀. 1λ…„μ―€ μ§€λ‚˜μž 고객듀이 μ΄λŸ°μ €λŸ° μš”κ΅¬λ₯Ό λ‹΄μ•„ "직접 λ””μžμΈν•΄λ‹¬λΌ"κ³  ν•΄μ„œ μ œμž‘κΉŒμ§€ λ›°μ–΄λ“€μ—ˆλ‹€. νŒ¨μ…˜ μ „λ¬Έ νˆ¬μžμ‚¬ λ“± μ—¬λŸ¬ 곳이 투자λ₯Ό μ œμ•ˆν•΄μ™”λ‹€. κ²½μ˜ λŠ₯λ ₯의 ν•œκ³„λ₯Ό λΌˆμ €λ¦¬κ²Œ λŠκΌˆκΈ°μ— κ°€μž₯ λ§Žμ€ λˆμ„ νˆ¬μžν•˜κ² λ‹€λŠ” 곳을 κ³¨λžλ‹€. λ§€μΆœμ€ 150μ–΅(2021), 500μ–΅(2022), 1000얡원(2023)으둜 κ°€νŒŒλ₯΄κ²Œ μ˜¬λžλ‹€. ν•˜μ§€λ§Œ 60% μ§€λΆ„νˆ¬μžλ₯Ό ν•œ νˆ¬μžμ‚¬μ™€μ˜ κ°ˆλ“±μœΌλ‘œ 끝내 λ‚΄κ°€ ν‚€μš΄ 'λ§ˆλ—‘ν‚΄'κ³Ό μ§€λ‚œν•΄ κ²°λ³„ν–ˆλ‹€. μ†μƒν–ˆλ‹€. ν•˜μ§€λ§Œ 과거에 μ‚¬λ‘œμž‘νžˆλ©΄ μ˜€λŠ˜μ„ μ‚΄ 수 μ—†λ‹€. 미래둜 λ‚˜μ•„κ°ˆ 수 μ—†λ‹€. λ‚΄ μ„ νƒμ˜ κ²°κ³Όλ‹ˆ λˆ„κ΅¬ 탓도, μ–΄λ–€ λ³€λͺ…도 ν•˜κ³  μ‹Άμ§€ μ•Šμ•˜λ‹€. κ²Œλ‹€κ°€ μž„μ‹ κ³Ό 겹쳐 μ‹ κ²½ μ“Έ μ—¬λ ₯도 μ—†μ—ˆλ‹€. κΉ”λ”ν•˜κ²Œ 내렀놨닀. μ§€λ‚œ 3μ›” 첫 아이 μΆœμ‚° 30λΆ„ μ „κΉŒμ§€λ„ νœ΄λŒ€ν°μœΌλ‘œ 업무 μ²˜λ¦¬ν•˜κ³ , 산후쑰리원에선 직원듀이 보내쀀 λ””μžμΈμ„ μ»¨νŽŒν–ˆλ‹€. μ‰¬λŠ” 게 편치 μ•Šμ•˜λ‹€. λŒ€ν‘œκ°€ μ§„μ‹¬μœΌλ‘œ λˆ„κ΅¬λ³΄λ‹€ μ—΄μ‹¬νžˆ 일해야 직원듀이 λ”°λΌμ˜¨λ‹€κ³  μƒκ°ν•΄μ„œλ‹€. κ·Έλ ‡κ²Œ μ§€λ‚œ 5μ›” 50λ…„ 역사 νŒ¨μ…˜ κΈ°μ—…(μ„Έμ •)의 투자λ₯Ό λ°›μ•„ λ‚΄ 이름을 λ”΄ μƒˆ λΈŒλžœλ“œλ₯Ό λŸ°μΉ­ν–ˆλ‹€. 눈 λœ¨λŠ” μˆœκ°„λΆ€ν„° μž λ“€κΈ° μ§μ „κΉŒμ§€, 심지어 μΆœν‡΄κ·Ό μ „κ³Ό 후도 μ—…λ¬΄μ˜ μ—°μž₯이닀. μ €λ…μ—” ν•œλ‘ μ‹œκ°„ SNS 라이브둜 고객과 μ†Œν†΅ν•˜κ³ , λ‹€μŒλ‚  이쀑 λ°˜μ‘μ΄ 쒋은 κ±Έ 골라 μž…κ³ λŠ” μž μ›λ™ μ§‘μ—μ„œ λ‚¨νŽΈκ³Ό ν•¨κ»˜ 좜근길 μΉ΄νŽ˜μ— λ“€λŸ¬ 데일리룩을 μΈμŠ€νƒ€κ·Έλž¨μ— μ˜¬λ¦°λ‹€. 이 κ³Όμ •μ—μ„œ λ””μžμΈΒ·μƒ‰μƒ λ“± 고객 μš”κ΅¬λ₯Ό μ΅œλŒ€ν•œ 빨리 λ°˜μ˜ν•΄ μƒν’ˆμœΌλ‘œ λ‚΄λ†“λŠ”λ‹€. μ§μ›λ“€μ€ νž˜λ“€κ² μ§€λ§Œ μž˜ν•˜λŠ” λΈŒλžœλ“œκ°€ λ„ˆλ¬΄ λ§Žμ€ μš”μ¦˜ μ΄λ ‡κ²Œ ν•˜μ§€ μ•ŠμœΌλ©΄ 살아남을 수 μ—†λ‹€. μ΄λ ‡κ²Œ 일에 λͺ°λ‘ν•˜λ©΄μ„œ μ™œ 아이λ₯Ό λ‚³μ•˜λŠλƒκ³  λ¬»λŠ” μ‚¬λžŒλ“€μ΄ μžˆλ‹€. λ‚œ 컀리어 μ„±μž₯ λͺ»μ§€μ•Šκ²Œ 행볡이 μ€‘μš”ν•˜λ‹€. 퇴근 ν›„ λ‚¨νŽΈκ³Ό κΉ€μΉ˜μ°Œκ°œμ— μ†Œμ£Ό ν•œμž”ν•˜λ©΄μ„œ λŒ€ν™”ν•˜λŠ” 일과 μΌμƒμ˜ κ· ν˜•μ΄ 늘 기쁘고 ν–‰λ³΅ν–ˆλ‹€. λ‹Ήμž₯의 성곡을 μœ„ν•΄ 아이λ₯Ό μ„ νƒν•˜μ§€ μ•ŠλŠ”λ‹€λ©΄ λ‚˜μ€‘μ— κ³΅ν—ˆν•˜κ³  ν›„νšŒν•  κ±° κ°™μ•˜λ‹€. 사업을 ν‚€μ›Œμ˜€λ©΄μ„œ 늘 κ²½κ³„ν•˜λŠ” 지점이기도 ν–ˆλ‹€. μ„±κ³΅ν•΄μ„œ λΆˆν–‰ν•΄μ§€μ§€ 말자. λˆμ΄λ“ , 컀리어든, λͺ…μ˜ˆλ“ , κ·Έ 무엇에도 λ‚΄ 삢이 μž‘μ•„λ¨Ήνžˆμ§€ μ•Šμ•˜μœΌλ©΄ μ’‹κ² λ‹€. κ·Έλ ‡κ²Œ λ‹€μΈμž‡μ„ ν•œ 번 더 μ„±κ³΅μ‹œμΌœ, 첫 번째 성곡이 운이 μ•„λ‹ˆμ—ˆλ‹€λŠ” κ±Έ 증λͺ…ν•˜κ³  μ‹Άλ‹€. https://www.joongang.co.kr/article/25260802

In case you missed it: Chrome is adding a Gemini Nano AI model right inside your browser via a new window.ai API. Spent some time this weekend putting together a simple open source browser extension using Chrome’s experimental local LLM API and Cartesia's super-fast TTS model ⚑️ It can: β†’ Summarize highlighted text (Cmd+Shift+S) β†’ Explain content as if you're 5 years old (Cmd+Shift+E) ... all on-device! Code: https://lnkd.in/gyWPHyK7 https://www.linkedin.com/posts/varunshenoy_ai-ugcPost-7213965233600495616-K4CE?utm_source=share&utm_medium=member_ios

λͺ¨λ“  λ§Œλ‚¨/λ―ΈνŒ…μ— μ΅œμ„ μ„ λ‹€ν•΄μ•Όν•œλ‹€. κ·Έλƒ₯ μ§€λ‚˜μΉ  μˆ˜λ„ μžˆμ—ˆλ˜ λ―ΈνŒ…μ΄(큰 μ•„μ  λ‹€ 없이 μ§„ν–‰λ˜μ—ˆλ˜ λ―ΈνŒ…μ΄), λ‚˜μ€‘μ— 쒋은 의미의 큰 impact κ°€ λ˜μ–΄ λŒμ•„μ˜€λŠ” 경우λ₯Ό μ’…μ’… κ²½ν—˜ν•œλ‹€. μΌλ‘€λ‘œ, μœ μ €λΆ„κ³Ό μ§„ν–‰ν–ˆλ˜ λ―ΈνŒ…μ΄ λ‚˜μ€‘μ— λ‚˜λΉ„νš¨κ³Όκ°€ λ˜μ–΄ 투자 유치둜 이어지기도 ν•˜κ³ , 큰 B2B Deal 둜 이어지기도 ν•œλ‹€. 그런 κ²½ν—˜μ„ λͺ‡ 번 ν•œ μ΄ν›„λ‘œλŠ”, '이것 저것 μž¬μ§€ 말고, λͺ¨λ“  λ―ΈνŒ…/λ§Œλ‚¨μ— μ΅œμ„ μ„ λ‹€ν•˜μž' μƒκ°ν•˜λ©° λ―ΈνŒ…μ„ μ§„ν–‰ν•˜κ³  μžˆλ‹€. λ―ΈνŒ…μ„ μ§„ν–‰ν•˜κ³  μ‚¬λžŒμ„ λ§Œλ‚  λ•Œ, μŠ΅κ΄€μ μœΌλ‘œ λ‹€μ§ν•˜λŠ” 뢀뢄은 μ•„λž˜μ™€ κ°™λ‹€. 1. μƒλŒ€λ°©μ„ μ„£λΆˆλ¦¬ νŒλ‹¨ν•˜μ§€ μ•ŠλŠ”λ‹€. (μƒλŒ€λ°©μ„ Judge ν•˜λ©΄ μ•ˆλœλ‹€. λ‚˜μ˜ νŽΈν˜‘ν•¨κ³Ό μ–΄λ¦¬μ„μŒμ„ λ“œλŸ¬λ‚  뿐이며, 큰 기회λ₯Ό λ†“μΉ˜λŠ” 지름길이닀) 2, λ‚˜λ₯Ό ν•„μš”λ‘œ ν•˜λŠ” μ‚¬λžŒ/μžλ¦¬κ°€ 있으면 μ΅œλŒ€ν•œ κ°„λ‹€. κ°€μ„œ μ΅œμ„ μ„ λ‹€ν•œλ‹€. (λ°”μœμ²™ν•˜μ§€ μ•Šκ³  νŠ•κΈ°μ§€ μ•ŠλŠ”λ‹€. μ‹œκ°„μ΄ μ•ˆλ λ•ŒλŠ” μ΅œλŒ€ν•œ μ–‘ν•΄λ₯Ό κ΅¬ν•œλ‹€) 3. μ˜ˆλ―Όν•˜κ±°λ‚˜ κΉŒλ‹€λ‘­κ²Œ κ΅΄μ§€ μ•ŠλŠ”λ‹€. (μƒλŒ€λ°©μ„ ν”Όκ³€ν•˜κ²Œλ§Œ ν•  뿐이닀) 4. λ„ˆλ¬΄ λ‚΄ μ•„μ  λ‹€λ§Œ λ‚΄μ„Έμš°μ§€ μ•ŠλŠ”λ‹€. μƒλŒ€λ°©μ—κ²Œ λΆ€νƒν•˜κ³  싢은 λ‚΄μš©μ΄ μžˆμ„ λ•Œμ—λŠ”, 1) μ–‘ν•΄λ₯Ό κ΅¬ν•˜κ³ , 2) μ†”μ§ν•˜κ²Œ 도움을 κ΅¬ν•œλ‹€. 5, λ―ΈνŒ… 쀑 μ†Œκ°œλ°›λŠ” μ‚¬λžŒμ΄ μžˆλ‹€λ©΄ (예: 이 λΆ„μ—κ²Œ 연락해 λ³΄μ„Έμš”) , μ—°λ½μ²˜ 받은 ν›„ μ΅œλŒ€ν•œ λΉ λ₯΄κ²Œ μ—°λ½ν•˜κ³ , μ΅œμ„ μ„ λ‹€ν•΄ λ§Œλ‚œλ‹€ (μ†Œκ°œν•΄μ€€ 뢄이 ν”Όν•΄λ°›μ§€ μ•Šκ²Œ ν•œλ‹€)Β  6. νšŒμ‹ μ€ μ΅œλŒ€ν•œ 빨리 ν•œλ‹€. 7. λŒ€ν™” 쀑 남 이야기 ν•˜μ§€ μ•ŠλŠ”λ‹€ (남 μ΄μ•ΌκΈ°λŠ” 쒋은 이야기라 해도 μ™ λ§Œν•˜λ©΄ μ•ˆν•˜λŠ” 것이 더 μ’‹λ‹€) 8. 아이컨택 μž˜ν•˜κ³ , μ˜¨μ „νžˆ μ§‘μ€‘ν•œλ‹€. (이야기에 μ§‘μ€‘ν•˜μ§€ μ•ŠμœΌλ©΄ λ°”λ‘œ ν‹°κ°€ λ‚œλ‹€) 9. μƒλŒ€λ°©μ΄ λΆˆνŽΈν•  수 μžˆλŠ” ν‘œν˜„/행동은 μ΅œλŒ€ν•œ ν”Όν•˜κ³ , μ •μΉ˜/인쒅/쒅ꡐ λ“± λ―Όκ°ν•œ 뢀뢄은 화두에 μ˜¬λ¦¬μ§€ μ•ŠλŠ”λ‹€. 10. μ„ μ˜μ˜ 도움을 μ œμ•ˆν•  λ•Œμ—λ„, 지킬 수 있고 & μ‹€ν–‰ν•  수 μžˆλŠ” κ²ƒλ§Œ μ œμ•ˆν•œλ‹€. (μˆœκ°„μ˜ μ •μ˜κ°(?)에 μ‚¬λ‘œμž‘ν˜€ κ³΅μˆ˜ν‘œ 날리면 μ•ˆλœλ‹€) 11. 인사 μž˜ν•˜κ³ , 청결함을 μœ μ§€ν•œλ‹€. 특히 μ°½μ—…μžλŠ” λͺ¨λ“  λ§Œλ‚¨μ— μ΅œμ„ μ„ λ‹€ν•  ν•„μš”κ°€ μžˆλ‹€. νšŒμ‚¬μ˜ μ„±μž₯에 λͺ¨λ“  κ°€λŠ₯성이 μ—΄λ €μžˆμ„ 수 μžˆμŒμ€, μ°½μ—…μžμ˜ νƒœλ„μ— λ‹¬λ €μžˆλ‹€.

쒋은 삢을 μœ„ν•œ μ•ˆλ‚΄μ„œ https://m.yes24.com/Goods/Detail/108551375 1. μŠ€ν† μ•„μ£Όμ˜μžλ“€μ˜ λͺ©ν‘œλŠ” λͺ¨λ“  감정을 λͺ°μ•„λ‚΄λŠ” 것이 μ•„λ‹ˆλΌ 뢀정적인 κ°μ •λ§Œμ„ λͺ°μ•„λ‚΄λŠ” 것이닀. μ‹ κ²½μ•ˆμ •μ œλ‘œ μ–»λŠ” 무감각 μƒνƒœκ°€ μ•„λ‹ˆλΌ, λΆ„λ…Έ, μŠ¬ν””, λΆˆμ•ˆ, 두렀움 λ“±μ˜ 뢀정적 감정이 μ΅œμ†Œν™”λœ λ™μ‹œμ— 기쁨 λ“±μ˜ 긍정적 감정이 가득찬 마음 μƒνƒœ. 1. μžκΈ°κ°€ κ°€μ§„ κ²ƒμ—μ„œ 기쁨을 찾으며 λ‚΄λ©΄μ˜ 기쁨 외에 λ‹€λ₯Έ 기쁨을 바라지 μ•Šμ„ 것이닀. 2. μš•λ§μ— κ΄€ν•΄μ„œ) λ§Œμ‘±ν•  쀄 λͺ¨λ₯΄λŠ” λ§ˆμŒμ„ κ·Ήλ³΅ν•˜μ§€ μ•Šκ³ λŠ” 쒋은 μ‚Ά, 의미 μžˆλŠ” 삢을 μ‚΄ 수 μ—†λ‹€λŠ” κ²ƒμ΄μ—ˆλ‹€. 끝없이 더 κ°€μ§€λ €λŠ” μ„±ν–₯을 λ‹€μŠ€λ¦¬λŠ” 쒋은 방법이 μ§€κΈˆ κ°€μ§„ 것에 λ§Œμ‘±ν•˜λŠ” 것. 3. μš°μ •κ³Ό λΆ€ λ“± 삢이 μ„ μ‚¬ν•˜λŠ” 쒋은 것은 즐기되 거기에 μ§‘μ°©ν•΄μ„œλŠ” μ•ˆλœλ‹€κ³  λ³΄μ•˜λ‹€. μŠ€ν† μ•„ μ² ν•™μžλ“€μ€ 삢이 μ£ΌλŠ” 선물을 μ¦κΈ°λŠ” λ°μ„œ μ’…μ’… λ²—μ–΄λ‚˜ μ§€κΈˆ 즐기고 μžˆλŠ” 것을 μ–Έμ œλ“  μžƒμ„ 수 μžˆλ‹€λŠ” 사싀을 자주 μˆ™κ³ ν•΄μ•Όν•œλ‹€κ³  λ³΄μ•˜λ‹€. 2. μŠ€ν† μ•„ 철학을 μˆ˜λ ¨ν•˜λŠ”λ°μ—λŠ” λΉ„μš©μ΄ λ“ λ‹€. μ‚Άμ˜ 철학을 κ°–μ§€ μ•Šμ•˜μ„ λ•Œ 치λ₯΄λŠ” λΉ„μš©μ€ 이보닀 더 크닀. λ¬΄μΉ˜ν•œ 것을 μ’‡μœΌλ©° ν•˜λ£¨ν•˜λ£¨λ₯Ό 보내닀 κ²°κ΅­ μ†Œμ€‘ν•œ 삢을 λ‚­λΉ„ν•˜λŠ” μœ„ν—˜μ΄ 더 큰 λΉ„μš©μ΄λ‹€. 3. μš°λ¦¬λŠ” 운λͺ…이 μ“΄ 연극에 λ“±μž₯ν•˜λŠ” λ°°μš°μ— μ§€λ‚˜μ§€ μ•ŠλŠ”λ‹€. μš°λ¦¬λŠ” μ—°κ·Ήμ˜ 배역을 선택할 수 μ—†λ‹€. μ–΄λ–€ 배역이 μ£Όμ–΄μ§€λ“  μ΅œμ„ μ„ λ‹€ν•΄ μ—°κΈ°ν•  뿐이닀. 4. μ„±κ°€μ‹  μ‚¬λžŒμ„ λŒ€ν•  λ•ŒλŠ” λ‚˜λ₯Ό μ„±κ°€μ‹œμ— μ—¬κΈ°λŠ” μ‚¬λžŒλ„ μ‘΄μž¬ν•œλ‹€λŠ” 사싀을 κΈ°μ–΅ν•˜λΌ. μƒλŒ€μ˜ 잘λͺ»μ— 짜증이 λ‚  λ•ŒλŠ” μž μ‹œ 멈좰 μžμ‹ μ˜ 결점에 λŒ€ν•΄μ„œ μƒκ°ν•˜λΌ. 그러면 μƒλŒ€μ˜ 결점에 κ³΅κ°ν•˜λ©° κ΄€λŒ€ν•˜κ²Œ λŒ€ν•  수 μžˆλ‹€. 5. λ‹€λ₯Έ μ‚¬λžŒμ— λŒ€ν•œ λ‚˜μ˜ 생각을 ν†΅μ œν•˜λ©΄, 그듀이 우리 삢에 λ―ΈμΉ˜λŠ” 뢀정적인 영ν–₯을 μ€„μΌμˆ˜ μžˆλ‹€. μ£Όλ³€ μ‚¬λžŒμ˜ 행동과 말, 생각과 κ³„νšμ— λŒ€ν•΄ μΆ”μΈ‘ν•˜λŠ”λ°μ— μ‹œκ°„μ„ λ‚­λΉ„ν•˜μ§€ λ§μ•„μ•Όν•œλ‹€. 그듀에 κ΄€ν•΄ μ œλ©‹λŒ€λ‘œ μƒμƒν•˜κ³  μ§ˆνˆ¬ν•˜κ³  μ‹œκΈ°ν•˜κ³  μ˜μ‹¬ν•΄μ„œλŠ” μ•ˆλœλ‹€. 마λ₯΄μΏ μŠ€μ— λ”°λ₯΄λ©΄ ν›Œλ₯­ν•œ μŠ€ν† μ•„μ£Όμ˜μžλŠ” 곡곡의 이읡을 μœ„ν•΄μ„œκ°€ μ•„λ‹ˆλ©΄ λ‹€λ₯Έ μ‚¬λžŒμ˜ 생각에 λŒ€ν•΄ μƒκ°ν•˜μ§€ μ•ŠλŠ”λ‹€. 6. 우리 내면에 뢄노와 증였, λ³΅μˆ˜μ‹¬μ΄ 일어날 λ•Œ ν•  수 μžˆλŠ” μ΅œμƒμ˜ λ³΅μˆ˜λŠ” κ·Έ μ‚¬λžŒκ³Ό λ˜‘κ°™μ΄ 되기λ₯Ό κ±°λΆ€ν•˜λŠ” 것이닀. 7. μ‚¬λžŒλ“€μ΄ λΆˆν–‰ν•œ μ΄μœ λŠ” κ°€μΉ˜μžˆλŠ” 것이 무엇인지 λͺ¨λ₯΄κΈ° λ•Œλ¬Έμ΄λ‹€. μ œλŒ€λ‘œ λͺ¨λ₯΄κΈ° λ•Œλ¬Έμ— μžμ‹ μ„ ν–‰λ³΅ν•˜κ²Œ ν•΄μ£ΌλŠ” 것이 μ•„λ‹ˆλΌ λΆˆμ•ˆν•˜κ³  λΉ„μ°Έν•˜κ²Œ λ§Œλ“œλŠ” 것을 μ’‡μœΌλ©° ν•˜λ£¨ ν•˜λ£¨λ₯Ό 보낸닀. 8. 자유λ₯Ό μ§€ν‚€λ €λ©΄ λ‚˜μ— κ΄€ν•œ μ‚¬λžŒλ“€μ˜ ν‰νŒμ— λ¬΄μ‹¬ν•΄μ•Όν•œλ‹€. μ‚¬λžŒλ“€μ˜ λΆˆμΈμ •μ΄ μ‹«λ‹€λ©΄ 그만큼 μ‚¬λžŒλ“€μ˜ 인정에도 μ΄ˆμ—°ν•΄μ•Όν•œλ‹€. 9. μŠ€ν† μ•„ μ² ν•™μžλ“€μ€ 자유λ₯Ό μ€‘μ‹œν–ˆλ‹€. μš°λ¦¬κ°€ κ°€μ§„ νž˜μ„ νƒ€μΈμ—κ²Œ λ„˜κ²¨μ£ΌλŠ” ν–‰μœ„λŠ” 무엇이든 κΊΌλ Έλ‹€. μ‚¬νšŒμ  μ§€μœ„λ₯Ό μ’‡λŠ”λ‹€λŠ” 것은 우리λ₯Ό μ§€λ°°ν•˜λŠ” νž˜μ„ λ‹€λ₯Έ μ‚¬λžŒλ“€μ—κ²Œ λ„˜κ²¨μ£ΌλŠ” 것과 κ°™λ‹€. 10. 죽은 λ’€ λͺ…성은 κ³΅ν—ˆν•œ λͺ…성이닀. μ£½μ€μžλŠ” λͺ…성을 λˆ„λ¦΄ 수 μ—†λ‹€. 미래 μ„ΈλŒ€κ°€ ν•œλ²ˆλ„ λ³Έ 적 μ—†λŠ” μ‚¬λžŒμ„ μΉ­μ°¬ν•  것이라 κΈ°λŒ€ν•˜λŠ” 것도 λΆ€μ§ˆμ—†λ‹€. 11. λΆ€λ₯Ό κ°€μ§„λ‹€κ³  μŠ¬ν”” μ—†λŠ” 삢을 μ‚΄ 수 μ—†μœΌλ©° λŠ™μœΌλ©΄ λΆ€κ°€ 우리λ₯Ό μœ„λ‘œν•  μˆ˜λ„ μ—†λ‹€. λΆ€λŠ” 물질적 μ‚¬μΉ˜λ₯Ό μ–»κ³  감각적 μΎŒλ½μ„ λˆ„λ¦¬κ²Œ ν•˜μ§€λ§Œ μ§„μ •ν•œ λ§Œμ‘±μ„ 가져닀주지도 μŠ¬ν””μ„ λ¬Όλ¦¬μΉ˜μ§€λ„ λͺ»ν•œλ‹€. 12. ν‰λ²”ν•œ μ‚Ά, 기본만 κ°–μΆ˜ μƒν™œμ—μ„œ 기쁨을 μ°ΎλŠ” λŠ₯λ ₯이 μ€‘μš”ν•˜λ‹€. 13. 두렀움을 ν”Όν•˜λŠ” 3κ°€μ§€ 방법 1. μžμ‹ μ˜ ν’ˆμ„±μ—μ„œ μ΅œμ„ μ„ λ‹€ν•΄ 기쁨을 μ°ΎλŠ”λ‹€. 2. 그것을 μžƒμ„ λ•Œλ₯Ό μ€€λΉ„ν•œλ‹€. μ§€κΈˆ λˆ„λ¦¬λŠ” 것듀이 ν–‰μš΄μ˜ μ‚¬κ±΄μž„μ„ κΈ°μ–΅ν•œλ‹€. 3. 눈이 λ†’μ•„μ§€μ§€ μ•Šλ„λ‘ μœ μ˜ν•œλ‹€. μ΅œκ³ κ°€ μ•„λ‹ˆλ©΄ 기쁨을 λͺ» λŠλΌλŠ” μ‚¬λžŒμ΄ λ˜μ§€ μ•Šλ„λ‘ μ‘°μ‹¬ν•œλ‹€. 14. μ‰½κ²Œ 얻을 수 있고 λˆ„κ΅¬λ„ μ•—μ•„κ°ˆ 수 μ—†λŠ” 것을 즐겼기 λ•Œλ¬Έμ— μ‚Άμ—μ„œ 즐길 것이 μ•„μ£Ό λ§Žμ•˜λ‹€. 그듀은 μžˆλŠ” κ·ΈλŒ€λ‘œμ˜ μžμ‹ μ„ 즐겼으며 μ§€κΈˆ μ΄λŒ€λ‘œμ˜ μ‚Άκ³Ό 세상을 즐겼닀. 이것은 κ²°μ½” μž‘μ€ μ„±μ·¨κ°€ μ•„λ‹ˆλ‹€. 15. μš΄μ„ μ •λ³΅ν•˜κΈ°λ³΄λ‹€ 자기 μžμ‹ μ„ μ •λ³΅ν•˜λΌ. 기쑴의 μ§ˆμ„œλ₯Ό 바꾸기보닀 μžμ‹ μ˜ μš•λ§μ„ λ°”κΎΈλ €κ³  λ…Έλ ₯ν•˜λΌ. μžμ‹ μ˜ 생각 외에 μ–΄λ–€ 것도 μ™„λ²½νžˆ ν†΅μ œν•  수 μ—†μŒμ„ 믿으라. 문제 해결에 μ΅œμ„ μ„ λ‹€ν–ˆλ‹€λ©΄ 더 이상 ν•  수 μžˆλŠ” 일이 μ—†λ‹€κ³  믿으라.

Data, Size of Models, and Efficiency No Priors: Apple recently announced models running on devices at around three billion parameters. How does this relate to your focus on efficiency? Karan Goel: The initial wave of companies focused on scaling data and compute, leading to great models. The second wave focuses on efficiency, enabling powerful models to run cheaply and repeatedly at scale. Our technology aims to make high-capability models run efficiently on smaller devices, moving from data centers to the edge. Recent Launch of Text to Speech Product No Priors: Cartesia recently launched its text-to-speech product, which is impressive in terms of performance. Can you tell us more about that launch? Karan Goel: We wanted to put our technology to work, demonstrating efficient audio generation. Text-to-speech is a longstanding field, yet there’s still room for improvement. Our goal is to create engaging, high-quality speech that people would want to interact with for more than 30 seconds. Multi-modality & Building Blocks No Priors: How has your pioneering work on SSMs impacted your approach to multi-modality or speech? Albert Gu: Multi-modality hasn't been the primary motivation, but different modalities present different challenges, influencing the design of our models. SSMs are versatile building blocks that can be applied to various modalities, including audio and text. What’s Next at Cartesia? No Priors: What’s next for Cartesia? Will you focus on Sonic and audio, or explore other modalities? Karan Goel: We’re excited about Sonic, showing the potential for real-time, low-latency audio generation. We aim to bring Sonic on-device, enhancing its capabilities and efficiency. Long-term, we’re developing multimodal models that can converse intelligently, integrating audio and text reasoning. Our goal is to build versatile models that run efficiently on various devices. Latency in Text to Speech No Priors: Latency is a significant issue in text-to-speech applications. How does multimodality address this? Karan Goel: Reducing latency is crucial. Orchestrating multiple models adds overhead, making it inefficient. We aim to create seamless, integrated models that handle multimodal tasks efficiently, reducing latency and improving performance. Choosing Research Problems Based on Aesthetic No Priors: How do you choose research problems? Albert Gu: I choose problems based on aesthetic appeal. Some solutions feel elegant and correct, driving my interest. This approach guided the development of SSMs and continues to influence my work. Product Demo No Priors: Can you show us a demo of your product? Karan Goel: Sure. Here’s our Sonic text-to-speech model running on a MacBook in real-time. [Demo showing the model generating speech quickly and efficiently.] This demonstrates the speed and quality of our model. Cartesia Team & Hiring No Priors: How large is your team, and are you hiring? Karan Goel: We have 15 people and 8 interns. We’re hiring across the engineering stack, especially for model roles. We’re excited to have talented people join us in building the future. No Priors: Thank you for joining us. Karan Goel & Albert Gu: Thank you.

No Priors: Welcome back to No Priors. We're excited to talk to Karan Goel and Albert Gu, the co-founders of Cartesia and authors behind revolutionary models such as S4 and Mamba. They're leading a rebellion against the dominant architecture of Transformers, and we’re excited to talk to them about their company today. Welcome, Karan and Albert. Karan Goel & Albert Gu: Thank you, nice to be here. Use Cases for Cartesia and Sonic No Priors: Tell us a little more about Cartesia, the product, and what people can do with it today. Karan Goel: Definitely. We launched Sonic, a really fast text-to-speech engine. Some exciting use cases include interactive, low-latency voice generation. We've seen a lot of interest in gaming, where players can interact with characters and NPCs in real-time. Another area is voice agents, where low latency is crucial. We’ve managed to reduce latency by 150 milliseconds and aim to shave off another 600 milliseconds over the next year. Karan Goel & Albert Gu’s Professional Backgrounds No Priors: Can you talk about your backgrounds and how you ended up starting Cartesia? Albert Gu: Karan and I both came from the same PhD group at Stanford. I worked on sequence modeling during my PhD, starting with problems at DeepMind. I became interested in recurrent models around the same time Transformers gained popularity. Karan and I collaborated on the S4 model, which showed the effectiveness of state space models. Recently, I proposed Mamba, which has shown promising results in language modeling. I’ve also started as a professor at CMU, continuing research while also working on Cartesia. Karan Goel: I grew up in India, from an engineering family. I initially aimed to be a doctor but switched to engineering due to low aptitude in biology. After IIT and grad school, I ended up at Stanford, working on reinforcement learning. Chris Ray, our PhD adviser, was skeptical about reinforcement learning, leading me to explore other areas. Albert and I worked together on various projects, including S4, which drew me into the field. State Space Models (SSMs) versus Transformer-Based Architectures No Priors: Could you tell us more about SSMs and how they differ from Transformer-based architectures? Albert Gu: SSMs originated from recurrent neural networks and focus on processing sequences one at a time, updating beliefs or states with new information. This approach can be loosely inspired by brain functions. SSMs are particularly good at modeling perceptual signals, like raw waveforms and pixels, which are less compressible than text data. Recent models like Mamba have improved at handling text data. One key advantage of SSMs is their linear scaling, providing constant time processing for new tokens, unlike Transformers' quadratic scaling. However, both models have their strengths and weaknesses, and hybrid models combining SSMs and Transformers have shown promising results. Domain Applications for Hybrid Approaches No Priors: Are there specific domains where you see applications for these hybrid approaches? Albert Gu: Currently, most applications are focused on text, as that's where interest lies. However, SSMs have been applied to DNA modeling, pre-training on long DNA sequences for tasks like detecting point mutations. Text to Speech and Voice No Priors: How did your research lead to focusing on text-to-speech and voice generation? Albert Gu: We wanted to demonstrate the versatility of our models in practical applications. Audio seemed like a natural fit due to its real-time processing requirements. We thought it would be a cool first application. Karan Goel: Audio and multimodal data are interesting because they are information-sparse, allowing for fast processing and large context handling. Audio, in particular, has many emerging commercial applications, like voice agents and gaming. Efficient models can also be deployed on smaller hardware, pushing inference closer to the edge rather than relying on data centers.

https://youtu.be/Ti6Gft5xO1s?si=BMf4_KB8Ln2J2iPD Experts tell you what will not happen large change is driven by entrepreneur
https://youtu.be/Ti6Gft5xO1s?si=BMf4_KB8Ln2J2iPD Experts tell you what will not happen large change is driven by entrepreneur led companies.