Cathie woods latest update claims ai productivity boom is starting now not just hype
Key data: 43% of ceos save 8+ hours per week using ai. only 5% of employees report same
her argument is gap will close as tools spread. then well see real gdp growth acceleration (7 8% by end of decade)
From ml perspective the question is whether current models can actually drive that productivity gain or if were still in impressive demos phase
the 90s productivity paradox is relevant. computers everywhere but productivity flat for years. took a decade for businesses to figure out how to use them
Are we in same phase with ai? tools exist but workflows not figured out yet
her inflation claim is interesting. if ai drives 7 8% real gdp without inflation that would be historically unusual
From research side what needs to be true: models better at reasoning (seeing progress), tools easier to use (still technical), businesses restructure workflows (barely started), regulatory concerns addressed
some coding agents hitting 76% on swe bench now. saw cursor, verdent, antigravity and a few others at that level recently. impressive progress but theres still huge gap between solving isolated github issues and actually augmenting workers at scale in production environments
Also productivity gains often mean job displacement. 43% of ceos saving 8 hours probably means roles getting cut
[link] [comments]
