Comment by sriram_malhar
5 days ago
There may be many disruptive technologies, but none come remotely close to AI when it comes to rate of change. Crypto has been around for a while, and hasn't really made a dent to the world
We had friends over for dinner a couple days back; between us we had two computer scientists, one psychologist, one radiologist, one doctor. Each of us were in turn astonished and somewhat afraid of the rapid pace of change. In a university setting, students are routinely using Claude and ChatGPT for everything from informal counseling to doing homework to generating presentations to doing 'creative' work (smh).
At the end of they day, we all agreed that we were grateful that we are at the tail end of our working life, and that we didn't have to deal with this level of uncertainty
AI feels particularly disruptive now because it’s new and we don’t know how it will affect society yet.
But people surely felt the same way about gunpowder, the steam engine, electricity, cars, phones, planes, nukes, etc.
Or look at specific professions that software has negatively affected in recent decades. Not a lot of people use travel agents anymore, for example.
I’m not saying that the negative effects are good. But that’s just the nature of technological advancement. It’s up to society to adapt and help out those who have been most negatively affected.
They're not really comparable.
AI is not only a general-purpose technology (such as, e.g., electricity or computers), but also the only one that can self-improve. Also, its potential for diffusion is much bigger, because unlike for electricity and computers, barriers are much more easily overcome [1].
But I'd suggest looking at this not as a separate "technology" like airplanes or smartphones, but rather as a software breakthrough. Everything that worried people about software [2] in the 20th and the early 21st century - or anything that came to pass on a smaller scale - is now much more worrisome.
[1] Pretty much any person or machine today that has access to electricity and computers can use AI, thanks to subscription models, cloud computing, VC subsidies, web access, etc.
[2] For example, large-scale automation and job loss, mass surveillance, robot swarms, etc.