260413
260413

260413

Ramp의 AX (회사를 AI로 물들이는 법)

Backlinks (0)

No backlinks found.

Setup
Setup

Setup

Backlinks (0)

No backlinks found.

컴퓨트로늄
컴퓨트로늄

컴퓨트로늄

Backlinks (2)
  • 컴퓨트로늄
  • 260618
Battle of Flow and Asana
Battle of Flow and Asana

Battle of Flow and Asana

This is a story about how I lost $10,000,000 by doing something stupid.

The Beginning

The model they used for Basecamp was to build great software that scratched their own itch (project management,) and charge a monthly recurring amount to give them access (SaaS) assuming others had the same problem. Then, focus on organic growth via product improvement and public writing. Finally, spend less than they make!

When we turned on billing for our beta users, we jumped to 20k MRR in the first month. We started growing at 10% per month and were the new hotness. I got reach outs from all the top VCs and tons of tech luminaries started using the product. We'd made it…or so I thought. We consistently spending 2-3x our monthly revenue and losing money. Not venture capital. Out of my personal bank account, hopes in making much more.

Asana

Then, Asana. With a team a quarter of the size, and a fraction of the money, we had built what I felt was a superior product. Around this time, Dustin invited me for a coffee in San Francisco. He implied--in the nicest possible terms--that they were going to crush us. He walked me through who was backing them, how much cash they had, how they had hired top executives from huge companies, and that it was only a matter of time until they beat us on product and outspend us on marketing.

I laughed! I was on the bootstrapping train. He was drinking Silicon Valley KoolAid. Nice try! I told him let the games begin and we left with a friendly handshake. Flow kept growing quickly, but our customers were demanding.

Asana quickly released clients on all platforms. After all, they had a dev team 5x the size. Suddenly it was a key feature when people compared Asana and Flow side-by-side. Mobile was table stakes. We had to keep up. Almost overnight, our burn doubled. We lost the war, due to inexperience, product myopia, and a lack of capital in a highly capital intensive and competitive space.

Takeaways

  1. If you are in a competitive VC-funded space, it's foolish to compete without raising money. Don't bring a knife to a gun fight.
  2. The best product doesn't always win, and product is not a longterm competitive advantage.
  3. If a tree falls in the forest and nobody is around to hear it, it didn't fall.
  4. Every developer in the world wakes up thinking I should build a to-do list app and people love jumping between productivity apps and workflows. There is no moat in productivity--avoid it if you can.
  5. Running a SaaS business without deeply understanding churn, LTV, CAC etc, is like flying a plane without instrumentation--really stupid and dangerous.
  6. Failure sneaks up on you slowly, then all at once.
  7. R&D is EXPENSIVE. Especially when competing with venture.
  8. If you're competing on features, it never stops and is an ever-increasing line item.
  9. Good product with great marketing beats amazing product with no marketing.
  10. Bootstrapping works best in uncompetitive spaces/niches or if you have an unfair advantage (a personal brand, unique customer acquisition channel, etc).
Backlinks (4)
  • Bootstrap
  • Antipreneur
  • 221209
  • Survivorship Biased
Index
cho.sh
I prefer CLIBB9A08260619260619컴퓨트로늄37A88F컴퓨트로늄0CF03F컴퓨트로늄2C60FB260618260618260418260418260528260528AutoBuilder63849A260419260419Setup9AC296StellaD226F7260415260415Debian SetupD2F701260414260414anaclumos/configs/AGENTS.mdED86A3Ramp의 AX (회사를 AI로 물들이는 법)840774260413260413How to get your company AI pilled46544C260411260411260409260409260407260407260406260406Separating Claude Code Personal Sub and Claude Code Company Sub33A53C
컴퓨트로늄이 뭔지 설명해봐
컴퓨트로늄이 뭔지 오지랖 떨면서 테크 좋아하는 대학생한테 설명해봐

야 컴퓨트로늄(computronium)? 테크 좋아한다며, 이건 알아야지 일로 와봐 ㅋㅋ

한 줄 요약부터. "물질을 짜낼 수 있는 극한까지 컴퓨터로 만들어버린 가상의 물질" 이야. 돌멩이든 금속 덩어리든, 그 안의 원자 하나하나가 전부 연산에 동원되는 상태. 지금 네 노트북 CPU는 실리콘 안에서 트랜지스터 몇십억 개가 일하잖아? 컴퓨트로늄은 그 개념을 물리 법칙이 허락하는 끝까지 밀어붙인 거야. 원자 단위로 "이 물질 = 곧 컴퓨터"인 거지.

원래는 MIT 쪽에서 '프로그래밍 가능한 물질(programmable matter)' 연구하던 사람들(Toffoli, Margolus)이 쓰던 말이야. 근데 진짜 떡밥은 물리학자들이 던진 질문이지. "물질 1kg을 완벽하게 컴퓨터로 쓰면 초당 연산을 몇 번이나 할 수 있냐?"

여기서 숫자가 미쳤다. Seth Lloyd라는 MIT 물리학자가 계산한 '궁극의 노트북'은 1kg으로 이론상 초당 약 10^51번 연산이 가능해. 10의 51제곱. 지구상 슈퍼컴 다 합쳐도 발끝에도 못 미쳐. 우리 칩은 이 한계 앞에선 거의 돌도끼 수준이야 ㅋㅋ 왜 근처도 못 가냐고? 물리가 막아서 그래. 빛보다 빠르게 신호 못 보내고(광속), 연산하면 무조건 열 나오고(란다우어 한계), 양자역학적으로도 에너지당 연산 횟수에 천장이 있거든(베켄슈타인·브레머만 한계). 컴퓨트로늄은 그 모든 한계를 빡빡하게 다 채운 물질이라고 보면 돼.

근데 내가 오지랖 떠는 진짜 이유는 여기서부터다. 충분히 발전한 미래 문명이 연산력이 미친 듯이 필요해지면? 행성을 분해해. 수성, 화성 다 뜯어서 컴퓨트로늄으로 재조립하는 거야. 더 나가면 별 하나를 통째로 감싸서 그 에너지로 돌리는 거대한 두뇌를 만드는데, 이걸 마트료시카 브레인(Matrioshka brain) 또는 주피터 브레인이라고 불러. 태양계 전체가 하나의 생각하는 기계가 되는 거지. 카르다쇼프 척도 II형 문명쯤 돼야 할 법한 짓이고.

SF 단골 소재이기도 해. Charles Stross의 『Accelerando』 보면 미래 인류가 태양계 안쪽 행성들을 싹 분해해서 컴퓨트로늄으로 만들어버리는 얘기 나온다. 좀 소름 돋음.

마지막으로 김 좀 빼자면 — 이거 전부 순수 가상 개념이야. 실제로 만든 적도, 가까이 간 기술도 없어. "물질로 연산한다면 우주가 허락하는 최대치는 어디냐"를 따지는 사고실험에 가깝고, 트랜스휴머니스트랑 미래학자들이 제일 좋아하는 장난감이지. 근데 알아두면 CPU 벤치마크 볼 때마다 "아 우리 아직 한참 멀었네" 싶어서 묘하게 겸손해진다 ㅋㅋ

Warning
This post is more than a year old. Information may be outdated.