At some point around 2008 starting a (tech) company became possible at "ramen" levels - middle class kids could launch on AWS etc. Tech VC struggled to adjust and a new breed of VC (ie ycombinator) saw a gap and exploited it - they needed smart founders who could get to series A with minimal funding. (more fairly they also (slowly?) paid founders much more and encouraged earlier liquidity so founders you know could see earlier payoffs)
anyway, the point was that it became possible to launch and run with small capital investment. But at some point you had to hire more people to do the extra coding bits.
I am wondering if LLMs are about to chnage that. The coding output is so good it could put off hiring a tranche of new devs for months or years. "create a web page to show the cities in yellow where users > 1 poker per month" is something you used to hire someone to do, and correct their work. now you are hiring OpenAI.
How much early phase work can be delegated to OpenAI if you know the right questions? Can the onboarding work through a chat bot? the initial demos? And is the next ycombinator skill set going to be "I know the right questions to ask ChatGPT to allow you to keep lean for another year?"
anyway, the point was that it became possible to launch and run with small capital investment. But at some point you had to hire more people to do the extra coding bits.
I am wondering if LLMs are about to chnage that. The coding output is so good it could put off hiring a tranche of new devs for months or years. "create a web page to show the cities in yellow where users > 1 poker per month" is something you used to hire someone to do, and correct their work. now you are hiring OpenAI.
How much early phase work can be delegated to OpenAI if you know the right questions? Can the onboarding work through a chat bot? the initial demos? And is the next ycombinator skill set going to be "I know the right questions to ask ChatGPT to allow you to keep lean for another year?"