I have never considered Pinterest vs other tech company algorithms/business models. One would think Amazon would know a thing or two about the "sell people stuff" model and possibly create a platform for social selling. Maybe they missed their chance with Vine back in the day. 😏
I reckon the wins in enterprise will be knowledge arbitrage where young(er) people will more quickly figure out how to quietly automate the essence of their work, and capture the free time for themselves rather than sharing it for others (their boss especially) to use.
I see it as capitalism vs central planning. A thousand thousand thousand enterprising young employees will much sooner find a way to maximise their own outcomes, and more likely find a way to take personal credit for efficiency gains, while CEOs will be entirely misguided in their efforts at coordinated AI. (no formal political education was involved in forming this take)
I haven't seen anyone making that argument that explicitly before, but I think that seems pretty right? That mandates to use AI in various ways will probably have pretty strange results, but a lot of the bigger changes will just come from people figuring out how to do things way faster or better than they did before.
Yeah exactly, something like that. I like the idea of rampant experimentation leading to strange (maximal) results.
Maybe the best mandate is a general remit to try AI things? That seems like it is mostly in action (which maybe undermines the initial point, somehow with AI it is all a lot more inorganic and forced - AI foie gras)
I thought initially that there would be some reluctance to let people do that (eg, a stigma around writing stuff with chatgpt), but it seems like that's actually gone away in a hurry. Which, I actually don't love for something like that, but do think is probably useful to experiment a little more widely, with more interesting ideas.
One thing he mentions that I've always kind of wondered about, is if half of TikTok's success comes from removing the timestamps of the videos, so you can never tell how old something is, and they can recycle the entire library all the time, rather than just needing to show you fresh stuff.
Related to humans doing the drudge work, in a sense that's what AI data centers are.
Just like our brain does a lot of unconscious autonomic work, LLMs are blissfully unaware of the data centers staffed with people doing all the heavy lifting of electricity, climate control, failover, disaster recovery etc.
I don’t think of it as bad, just an interesting parallel.
I was talking with some friends last night about the idea that “learn a trade” is a rising meme. For the foreseeable future, getting your water heater fixed will still require that a human come to your house and not a robot. But it’s plausible that tradespeople will make use of quasi-robots and gizmos to do the dirtiest work in the same way surgeons use advanced technology to do arthroscopic surgery.
Yeah, there's some chart that seems to periodically circulate about which percent of which jobs could get replaced, and it's a strange list to see things like cooks and mechanics on "most likely to not be disrupted" end.
"The truth may somehow be worse: Rather than turning us into batteries, the machines will make us Jira admins."
I have never considered Pinterest vs other tech company algorithms/business models. One would think Amazon would know a thing or two about the "sell people stuff" model and possibly create a platform for social selling. Maybe they missed their chance with Vine back in the day. 😏
God, can you imagine? An Amazon-built twitter, where all of your friends or whatever are people who are buying the same stuff as you?
I hate it, and want to write fan fic about it.
I reckon the wins in enterprise will be knowledge arbitrage where young(er) people will more quickly figure out how to quietly automate the essence of their work, and capture the free time for themselves rather than sharing it for others (their boss especially) to use.
I see it as capitalism vs central planning. A thousand thousand thousand enterprising young employees will much sooner find a way to maximise their own outcomes, and more likely find a way to take personal credit for efficiency gains, while CEOs will be entirely misguided in their efforts at coordinated AI. (no formal political education was involved in forming this take)
I haven't seen anyone making that argument that explicitly before, but I think that seems pretty right? That mandates to use AI in various ways will probably have pretty strange results, but a lot of the bigger changes will just come from people figuring out how to do things way faster or better than they did before.
Yeah exactly, something like that. I like the idea of rampant experimentation leading to strange (maximal) results.
Maybe the best mandate is a general remit to try AI things? That seems like it is mostly in action (which maybe undermines the initial point, somehow with AI it is all a lot more inorganic and forced - AI foie gras)
I thought initially that there would be some reluctance to let people do that (eg, a stigma around writing stuff with chatgpt), but it seems like that's actually gone away in a hurry. Which, I actually don't love for something like that, but do think is probably useful to experiment a little more widely, with more interesting ideas.
My favorite description of TikTok was written by the person whose column he is guest writing https://kylechayka.substack.com/p/essay-how-do-you-describe-tiktok
One thing he mentions that I've always kind of wondered about, is if half of TikTok's success comes from removing the timestamps of the videos, so you can never tell how old something is, and they can recycle the entire library all the time, rather than just needing to show you fresh stuff.
Related to humans doing the drudge work, in a sense that's what AI data centers are.
Just like our brain does a lot of unconscious autonomic work, LLMs are blissfully unaware of the data centers staffed with people doing all the heavy lifting of electricity, climate control, failover, disaster recovery etc.
True, but, is that bad? Or I guess, should we care? (Is it weird that the answer isn't definitely no?)
I don’t think of it as bad, just an interesting parallel.
I was talking with some friends last night about the idea that “learn a trade” is a rising meme. For the foreseeable future, getting your water heater fixed will still require that a human come to your house and not a robot. But it’s plausible that tradespeople will make use of quasi-robots and gizmos to do the dirtiest work in the same way surgeons use advanced technology to do arthroscopic surgery.
Yeah, there's some chart that seems to periodically circulate about which percent of which jobs could get replaced, and it's a strange list to see things like cooks and mechanics on "most likely to not be disrupted" end.