4 ms·

Live Stream: How to Start a Hard Tech Startup with Sam Altman, 6:00 EST

7 points by joelg 11y ago

minimaxir 11y ago
Why is this being streamed from an unknown YouTube account vs. an official YC/MIT organization? EDIT: Stream is actually real
- joelg 11y ago
  Not sure. But it's the one that was posted in the official Facebook event.
Jugurtha 11y ago
Preliminary notes jotted down on the fly. If something seems odd or silly, I probably didn't hear well or mixed things up. YC: - 11 years. - 1200 startups. - Recently started to fund "hard tech" (few are funding them). How to start a hard tech company: - What do you need to do differently than a software startup: - The least amount possible. You're not that different. - Same principles apply. - They were first segregated at YC. Turned out to be bad. - Most important change: - YC & the market don't care: Create value or fail. - Don't think you're that different: - "But my idea takes 10 years!": - What can be done in 3 months? - Consistent incremental wins compound. - 10% every week for years. - Find something small and get it done. - Advantage: - 11,000th photosharing app is hard. - Competitive advantage (4 nuclear fusion companies in North America). - Tesla, SpaceX, etc. cf. Michael E. Porter. - Pay attention to what software companies did well/right: - If in doubt about a field, 2 important things (good indicators): - Cost down. - Cycle time is getting shorter. - Difference: - SW companies are usually "clearer" than HW. - It's a great time to start a HW company. Make sure: - Weekly wins. - Keep costs down. - Make peace with the "world": - "Big companies don't play by the rules": - Make peace with that. - No excuses. - Your duty as a founder to get that 10%/small wins consistently. - Keeping cycle times down is important because: - Maintains *momentum*. - Avoids mistakes (typically made by bigger companies). - Establish a culture of winning from the beginning, no matter how small. - Momentum attracts good people: - People want to join a winning team. - How?: - SW startups drive cycle time down by getting closer to the customer. - Detours: - Sometimes you can't reach "final" goal directly. - ex: Tesla - Roadster (revenue) --> Model S --> Model X --> Model 3. - Not from 0 to mass produced car. SpaceX: Not from 0 to colony on Mars. - Even if you're not in YC, you can adopt this model. - More people fail because of "grandeur" than because they did small things. - Raising capital: - Too much money can kill you. (ruins your focus, for example). - Statistically (YC): - Companies that raised $40MM tended to fail. - Business side: - Don't be all hardcore science and bad at everything else. - Someone should take care of the business side. - Not hard: you can do science, you can do business. - It can be boring for some, but it needs to be tended to. - Many will do the thing everyone is doing. Think of a worthy problem. (?) - U.S. Context: - Centuries of abundance and massive growth. - Not so much recently (2%) --> Democracy works badly in this context. - The only hope is innovation (and a more important growth).

7 points by joelg 11y ago

Why is this being streamed from an unknown YouTube account vs. an official YC/MIT organization? EDIT: Stream is actually real

joelg 11y ago
Not sure. But it's the one that was posted in the official Facebook event.

Jugurtha 11y ago

Preliminary notes jotted down on the fly. If something seems odd or silly, I probably didn't hear well or mixed things up. YC: - 11 years. - 1200 startups. - Recently started to fund "hard tech" (few are funding them). How to start a hard tech company: - What do you need to do differently than a software startup: - The least amount possible. You're not that different. - Same principles apply. - They were first segregated at YC. Turned out to be bad. - Most important change: - YC & the market don't care: Create value or fail. - Don't think you're that different: - "But my idea takes 10 years!": - What can be done in 3 months? - Consistent incremental wins compound. - 10% every week for years. - Find something small and get it done. - Advantage: - 11,000th photosharing app is hard. - Competitive advantage (4 nuclear fusion companies in North America). - Tesla, SpaceX, etc. cf. Michael E. Porter. - Pay attention to what software companies did well/right: - If in doubt about a field, 2 important things (good indicators): - Cost down. - Cycle time is getting shorter. - Difference: - SW companies are usually "clearer" than HW. - It's a great time to start a HW company. Make sure: - Weekly wins. - Keep costs down. - Make peace with the "world": - "Big companies don't play by the rules": - Make peace with that. - No excuses. - Your duty as a founder to get that 10%/small wins consistently. - Keeping cycle times down is important because: - Maintains *momentum*. - Avoids mistakes (typically made by bigger companies). - Establish a culture of winning from the beginning, no matter how small. - Momentum attracts good people: - People want to join a winning team. - How?: - SW startups drive cycle time down by getting closer to the customer. - Detours: - Sometimes you can't reach "final" goal directly. - ex: Tesla - Roadster (revenue) --> Model S --> Model X --> Model 3. - Not from 0 to mass produced car. SpaceX: Not from 0 to colony on Mars. - Even if you're not in YC, you can adopt this model. - More people fail because of "grandeur" than because they did small things. - Raising capital: - Too much money can kill you. (ruins your focus, for example). - Statistically (YC): - Companies that raised $40MM tended to fail. - Business side: - Don't be all hardcore science and bad at everything else. - Someone should take care of the business side. - Not hard: you can do science, you can do business. - It can be boring for some, but it needs to be tended to. - Many will do the thing everyone is doing. Think of a worthy problem. (?) - U.S. Context: - Centuries of abundance and massive growth. - Not so much recently (2%) --> Democracy works badly in this context. - The only hope is innovation (and a more important growth).