Comment by bluerooibos
11 hours ago
I completely get the doomer POV, but we've somehow navigated all the previous "dangerous" technologies we've created - electricity, phones, internet - every one of those had similar arguments and concerns of danger.
The optimists' argument:-
Politics:- in general, I think many of the problems in the world today are due to misinformation and lack of education. What happens when we start routing things through an ASI that brings data and logic to the table? What happens when politicians can no longer lie without being caught out live on air? In the UK, local authorities are being flooded with complaints and requests from people; for example, some are doing AI-assisted investigations into accounting "errors".
Science:- I just don't see how the current rate of progress doesn't end up in crazy technologies like perfectly simulated human cells, organs and bodies to the point where we can run experiments virtually and solve all diseases in the next few years. This is happening. Perfect weather predictions far into the future, likewise with earthquakes, etc. Solar panel research explosion resulting in huge efficiency gains, to the point where people no longer need to plug their EV in - car surfaces will be covered in solar panels, as will our windows and roofs. Connecting new homes to the grid will be optional - the same way landline phones are no longer a thing.
I just find it very difficult not to extrapolate all the above.
We got this dump of mathematical breakthroughs from one small team in one company with access to this technology. What happens when this SOTA model is available (and it will continue getting better and cheaper) to everyone working on hard problems - every university on the planet starts cranking out AI-assisted research breakthroughs.
> What happens when politicians can no longer lie without being caught out live on air?
If there is perfect lie detecting technology I could see all kinds of chaos resulting from it. I can't see it only be applied only to politicians, and I think it would be the developers of the technology who decide the use.
I think were we disagree is that you sort of see AI as an extension of technological progress whereas I see it more like an extension of evolution. I view the process of AI training as functioning in a similar way to evolution in that it build circuits into neural networks similar to how evolution built circuits into human brains.
This is great.
>> What happens when politicians can no longer lie without being caught out live on air?
A 5-second delay on a politician's presser. Any lies will be muted in real time and the actual facts presented onscreen. Continue to lie enough, and the politician gets unstreamed.
> but we've somehow navigated all the previous "dangerous" technologies we've created
It's only true if you believe that "putting the burden of living on a dying planet on the future generations" counts as "navigating".