Shipping useful software that solves people’s problems requires finding the place on the tradeoff spectrum of short term vs long term value

If you sharpen your axe forever you never chop down the tree

If you start hacking without sharpening your axe you will move too slow to chop down the tree

AI is creating a barbell of extremes on this spectrum

A lot of work we previously invested in to sharpen our axe we can reliably guess will be solved by ai and thus we now should be hacking

Other work to sharpen that previously may have seemed like a waste of time will have compounding value as AI progresses

It’s a fun time where we all must be thinking from first principles to figure out what to prioritize. And we cannot judge what strategy is right based on recent results. The correct strategies and incorrect strategies can easily look flipped based on recent trends

Experienced this barbell first hand on Saturday and Sunday. Great essay on the tension to build pre and post AI.