Kevin Sidely

@anythingsideways

human-ai software collaboration may end soon

translation: enjoy the “ai as your copilot” phase while it lasts. dario amodei goes on a podcast and basically says we are in the centaur era of software engineering. half human, half machine. like in chess after deep blue beat kasparov. for a while, human plus ai could outperform either alone. then that window closed. machines pulled ahead. humans became spectators. his argument is that software is going through the same pattern. right now developers with ai tools are more productive. demand might even increase temporarily. everyone feels empowered. faster commits. smarter autocomplete. fewer stack overflow tabs. but according to him, that hybrid advantage may be short lived. once models get good enough, the human layer stops being leverage and starts being latency. white collar disruption stops being theoretical at that point. not dramatic layoffs overnight. just gradual irrelevance. fewer junior roles. less need for large teams. one engineer supervising what used to require ten. the uncomfortable part is not that ai helps. it is that help might be a transitional phase before replacement. and most people are still celebrating the copilot.

tech firms racing to build ai data centers in space

of course it was going to end up here. ai models are eating electricity like there is no tomorrow, grids are stressed, energy debates everywhere. solution. just move the servers to space. apparently earth is getting too small for our compute needs. spacex confirms it is pulling xai fully into the fold. musk merging his toys again, valuation casually floating around 1.25 trillion dollars before a potential ipo. and right after that, an application to launch up to a million satellites. not for internet this time. orbital data centers. powered by almost uninterrupted solar energy. the pitch is simple. in space you get constant sun, no local grid constraints, theoretically infinite scaling. musk openly saying space based ai is the only way to scale long term and could become the most economically logical location within three years. so now the future is not just ai replacing jobs. it is ai running from orbit. next step probably arguing about server latency between earth and space while your prompt takes two extra milliseconds. we cannot fix traffic, but we are planning server farms in orbit. priorities.

why intelligent and talented people fail

why are people who genuinely have the skills and qualities to deserve success not where they should be? why can’t they realize their potential? why do they lose positions, titles, and authority they deserve to completely mediocre people who don’t deserve them at all? i want intelligent, talented, capable people to run countries, have a say, and produce real value. but somehow these types of people rarely succeed. instead, wherever there are empty headed, shallow people, they’re the ones who end up successful.