ThoughtStorms Wiki

Interesting. I'd forgotten about this.

I tend to feel I was completely blind-sided by the arrival of Transformers and LanguageModels. Which is true. BUT ... this answer from 2017 still holds up. At least I didn't say anything which has been shown to be completely wrong, in it.

Quora Answer : How powerful will AI become over the next 20 years?

Dec 29, 2017

It will be more powerful than you can possibly imagine today. But less autonomous than you fear.

So, AI will be doing extraordinary and terrifying things. It will be accurately recognising everything you do, it will be accurately predicting your patterns of behaviour. Quite plausibly it will be able to "read your mind" (ie. make good predictions about what you're thinking based on scans of brain activity). It will engage in more or less unfoolable scrutiny based on collating multiple pieces of information about you.

It will plausibly converse like a human. It will drive cars, and trucks, and aircraft and delivery drones, and pipe-inspecting robots (including inside your body to inspect you medically) and viable droids. AI will be in the laboratories, formulating new hypotheses about new medicines and materials. I confidently expect we'll be seeing the first new hypotheses in physics and cosmology, perhaps in economics and social sciences too, which are basically conjectured by machines rather than people.

Corporations and other large institutions will continue to evolve towards human / AI hybrids. The team of creative humans driving the organization will continue to shrink, while more and more of the information processing is delegated to the machines. AIs will run the large bureaucracies. And run them better than humans could. AIs will do most other jobs better than humans too. We will face a great challenge in finding enough jobs for people in large, largely automated, corporations.

OTOH, AI will still do exactly what its human owners / programmers want it to. We won't see what people call "General" intelligence, or AIs forming their own agenda or rebelling against humans. They'll remain securely "in the box". When AIs do something bad (and they absolutely will) it will be because of badly intentioned humans behind the scenes telling them to do exactly that.

No Backlinks