That's not at all true, or at the very least, it requires a raft of other assumptions which may not hold, to all stay true, in order for the consequence to stay true. Like for example the cost of rare earth metals, good trading relationships between nations, and so on.
> we've gotten much better at getting solutions to scale
I'd agree, but only for a very narrow set of "solutions". Scaling something like twitter is just in no way whatsoever equivalent to scaling something like self-driving cars.
> There's a very real risk it'll happen faster than people can adapt. The people who's jobs will be first to be automated out are more likely living paycheck-to-paycheck. Those other, incremental, technological advances he compares against in the blog post took years for the labor force to adapt. How are we going to handle this?
There is a risk, but there are risks of many things happening. My guess is that we're not going to handle this. Just as in software development we have the YAGNI principle, I don't think it's a good idea to make policy based on future predictions that are very cloudy at best. Sure, there's a risk it will happen, but there are actual problems here today demanding resources and time...
Sure, those lines aren't monotonically decreasing/increasing but the trend isn't likely to reverse direction any time soon.
>I'd agree, but only for a very narrow set of "solutions". Scaling something like twitter is just in no way whatsoever equivalent to scaling something like self-driving cars.
There'll definitely be a costly hardware buy-in for a lot cases, but some applications of AI might only need a camera - say, a replacement for a cashier at a grocery store.
>There is a risk, but there are risks of many things happening. My guess is that we're not going to handle this. Just as in software development we have the YAGNI principle, I don't think it's a good idea to make policy based on future predictions that are very cloudy at best. Sure, there's a risk it will happen, but there are actual problems here today demanding resources and time...
I think it's worth discussing because waiting for the bad outcome to happen before trying to handle it could be too little too late. Even if the percentage risk of an economic shock is low, the magnitude of the potential damage is high.
The average pre-tax income of the bottom 50% has stagnated since 1980 at about $16,000 per adult [1]
These people are more likely to live paycheck-to-paycheck and more likely to work jobs which might be targeted by automation. I think it's a recipe for disaster.
Exactly, look at mobile phones, that took years to scale up. /s
As much as many might believe that moral outrage drives policy, the reality is that in the US, industry drives policy.
The transportation industry is powerful, and the transportation labor force is incredibly weak. I think we'll see the national government push for accelerating autonomous vehicles long before we see them push for banning them.
I think autonomous vehicles mean there won't be a transportation industry, per se. There would have to be some barrier to entry that's not readily apparent now. After all, "Uber for self-driving semis" under an Amazon style aggregator ...
Why are you then confident that safety numbers will carry the day? When someone is killed by a driverless car, that's going to be an emotional issue, and the appetite for cold hard data may be limited.
If safety numbers were really what people were after, it's hard to see how the US would have the gun culture that it does.
Engineers put far too much faith in data, and it's why they're often so bad at politics. Look at the incoming president; what people want is frequently more important than what is.
The upshot of what I read was - these folks are, in effect, test pilots whether they know it or not, and even a good driver could have died.
Some trends will always hold true; cost of hardware goes down, capabilities of software goes up. I don't think anyone is saying it'll happen in 7.8 years, but as the HN group is aware, we've gotten much better at getting solutions to scale. There's a very real risk it'll happen faster than people can adapt. The people who's jobs will be first to be automated out are more likely living paycheck-to-paycheck. Those other, incremental, technological advances he compares against in the blog post took years for the labor force to adapt. How are we going to handle this?