"This symbiotic relationship between the human worker, who has a business context, and the AI, which can work faster and even smarter but lacks the input, will define the future of white-collar work that Shumer has warned about, according to Gopal. “You have to pick and choose the context and you have to keep capturing the context, right? And I think that’s really what the shift is for the average white-collar worker is that they have to understand.”
Gopal related an anecdote from his team, expressing frustration with a mediocre software engineer now that they have AI coding tools. “We’re like, ‘Man, like, it’s just more expensive to talk to you than it is to do it myself. Like, to explain what I need built on the product takes more time than me just slamming it out of AI on the side.'” The time it takes to talk to a mediocre engineer could be spent managing an AI output instead, he added. He likened this to every employee having a personal technical co-founder by their side at all times, potentially enabling them to produce 20 times as much work."
Today's 1bln CEO isn't yesterday's 1bln CEO. But I guess that's because the fiat printer goes brrr. ~lol
This I feel is a common misunderstanding though. Because the AI won't do 20x on top of itself magically, but a new engineer could ask your well skilled LLM swarm with access to years of team best practices and move from 0.5x to add their own 20x.
That requires someone to think about how to preserve institutionalization of the entire stack. May be worth putting one of your best guys on. Not to make a handbook, but to have an automated continuous learning process and an on boarding system.
I think there's a compounding effect too with multiple people working on the same thing with AI since the AI is only as good as your prompt. Having another member of the team to use AI to review, cross-examine intent, or optimize your diff is 40x if you both 20x.
Diminishing returns though, a 3rd person in the mix probably creates more noise with a committee effect... so what used to be teams of many can pare down to teams of 2.
Something we've been trying that seems to go well is jumping on a screen share, and tag-teaming planning mode. With AI writing the doc and us both in real-time it can be comprehensive scope with great detail really fast. One then executes the build-from-doc and reviews the result, the other cross-examines. The more detailed docs may be a bigger efficiency gain than with code.
Diverse tool/model preferences catch things we might not otherwise too, up to and including re-do of anything done with free Antigravity tokens.