4 ms·
I think automation will succeed where humans hate to do the job because it's too monotonous, risky or unhealthy - fruit picking, Foxconn, mining etc. It also ma
by blinkingled 8y ago
I think automation will succeed where humans hate to do the job because it's too monotonous, risky or unhealthy - fruit picking, Foxconn, mining etc. It also makes the most economic sense to automate those jobs. Knowledge workers' jobs on the other hand - it'll be much easier for employers to put downward pressure on the salaries of those jobs than to replace them with automation due to higher automation costs and risks.
If it works out like that and the displaced workers are supported by their governments with equivalent income that will be a net positive. Those whose skills are still in demand will have jobs, others will have UBI and not many will complain.
However we don't learn anything - UBI will never come in time and there will be a lot of needless suffering and upheaval as a result of hurried, ill-managed drive towards automation.
There is also the whole identity crisis thing - many people equate their worth with their job and not having one will make for a lot of unhappy people.
Interesting times indeed.
- mistrial9 8y agono I think you mishandle the spin.. business will try to create new companies where the AI is doing the work, whatever that year's AI is capable of .. no bargaining, instead straight to "disruption" secondly, the horrible, unanticipated errors will be the catch, while the raw capacities of the machine side will continue to soar.. $0.02