AI is going to replace most jobs and good riddance
Timeline
Post
Remote status
Context
8
@Griffith AI will replace most jobs because most jobs require a business degree and knowledge of Microsoft office and are functionally useless
@Shadowman311 I can't imagine accountants existing in 10 years.
Fable came out in July, the same month as K3 and Sol. Three months later we have Opus 5.5, Astra, and K3.1. The leap in capabilities in 2-3 months should shatter anyone's previously held expectations. DeepSeek v4.1 1017 is probably responsible for the surge in critical vulnerabilities and exposures this month, given the timing. Adoption is a slower phenomenon because it's social, but I think by next year we'll have another surge of adoption like we did with the advent of agentic coding this year, and labs will continue to rock serious fields in the same manner they've been rocking mathematics this year, and software in years past.
In other words, I don't think most people have a moat over 10 years, and no one has a moat over 20 years, and both those estimates are likely unreasonably conservative. My honest opinion is both those estimates can be cut by half.
Fable came out in July, the same month as K3 and Sol. Three months later we have Opus 5.5, Astra, and K3.1. The leap in capabilities in 2-3 months should shatter anyone's previously held expectations. DeepSeek v4.1 1017 is probably responsible for the surge in critical vulnerabilities and exposures this month, given the timing. Adoption is a slower phenomenon because it's social, but I think by next year we'll have another surge of adoption like we did with the advent of agentic coding this year, and labs will continue to rock serious fields in the same manner they've been rocking mathematics this year, and software in years past.
In other words, I don't think most people have a moat over 10 years, and no one has a moat over 20 years, and both those estimates are likely unreasonably conservative. My honest opinion is both those estimates can be cut by half.
@Griffith You're assuming our power infrastructure and compute capacity can keep up with demand, which it can't and it won't. Everytime a new data center gets announced the local power company shits themselves and for good reason. This "revolution" will end because we just do not possess the raw materials and energy capacity to keep up with it.
@Griffith We also don't have the power capacity for 100% EV either, all these plans are bottlenecked and impossible because boomers got scared of nuclear power 60 years ago, plus one for the boomers I guess
@Griffith Also SMRs are not viable, the instant one of them fucks up they'll get regulated out of existence
@Shadowman311 @Griffith SMRs aren't a silver bullet, but they have federal protection. We no longer have the forges for monolithic reactor vessels so welded SMRs are more or less all we can do in the nuclear arena. Less efficient due to the lower possible delta T but the only bandaid we can put on our current energy crisis
The bigger problem imo is transmission gear
The bigger problem imo is transmission gear
@Shadowman311 @Paultron Structurally, the purpose of SMRs is to power the data centers into the 2030's, because the bottlenecks will appear before then and everyone knows it. Musk knows it too, which is the aim of his datacenters in space plan. By 2030, realistically the Chinese surge ahead of us in infrastructure, but that's the plan if the race has to go beyond the next two years, which is what Dario is trying to sidestep. Dario is an "RSI soon" guy. Transmission will be a problem for communities around the SMRs, not the data centers themselves.
Replies
0Fetching replies…