Episode 2263 Scott Adams: CWSA 10⧸16⧸23
Episode Stats
Length
1 hour and 4 minutes
Words per Minute
140.24081
Summary
In this episode of Coffee with Scott Adams, host Scott Adams talks about artificial intelligence and what it can do for us, and why it might be better than anything we can think of right now. Scott Adams is an American comedian, podcaster, writer, and podcaster. He is a regular contributor to the New York Times, and host of the popular radio show "Scott Adams Radio" on SiriusXM Radio.
Transcript
00:00:00.000
Good morning, everybody, and welcome to Coffee with Scott Adams, a highlight of human civilization.
00:00:23.580
And if you'd like this experience, which is already extraordinary, to go up to levels
00:00:28.520
that almost nobody could even imagine, and that's not even counting the whiteboard that's
00:00:32.640
coming up, then all you have to do is find yourself a cup or a mug or a glass, a tank of
00:00:37.680
gels or a stein, a canteen jug or a flask, a vessel of any kind, fill it with your favorite
00:00:42.120
liquid, I like coffee, and join me now for the unparalleled pleasure, the dopamine, the
00:00:47.800
day the thing that makes everything better, it's called the simultaneous sip, and it happens
00:00:58.520
Well, according to the CDC, only 2% of Americans got their COVID vaccination this year, 2%.
00:01:14.360
I'm starting to think that the public is waking up.
00:01:20.480
What would happen if the public suddenly woke up?
00:01:27.640
Because I feel like we're maybe, maybe reaching the point where people are waking up.
00:01:39.280
You start with my team good, other team bad, and if you wake up, you realize that both teams
00:01:48.360
are just, you know, following their self-interest and probably don't have much of your interest
00:01:55.240
And then once you don't trust anything, everything starts making sense.
00:02:00.320
It's only when you trust something that you're lost, right?
00:02:07.580
As soon as you say, well, those Democrats are all bad, but thank God I'm on that good team,
00:02:22.580
You don't wake up until you realize everybody is chasing money and dopamine,
00:02:33.100
Well, we have no Speaker of the House yet again, and I ask you this question.
00:02:39.360
How can the United States survive one more day without having a named Speaker of the House?
00:02:52.360
There aren't any damages, but I'm sure that there will be.
00:02:55.280
There will be really big, big damages any minute now.
00:03:04.120
We're not picking a Speaker of the House that half of the country wants to fucking kill.
00:03:11.440
If we could just get somebody in the job that half of the country wanted to put on a pike,
00:03:20.080
We'd be passing laws and solving wars and stuff like that.
00:03:25.960
Do you know how much our inflation would come down if we only had a Speaker of the House
00:03:36.600
I have a provocative thought for you, and it goes like this.
00:03:43.160
As you know, AI creates something like intelligence by predicting the next word that it should say,
00:03:51.720
or next word you're going to say, I suppose, based on words that have been spoken by people before.
00:03:58.140
So AI learns to be intelligent simply looking at the combination of words that humans have spoken before,
00:04:08.400
because our intelligence is embedded in the words in the order of the words.
00:04:14.860
So, of course, AI can unravel that and create some kind of intelligence of its own.
00:04:19.760
But I ask you this, could it learn to predict what you're going to do in any situation
00:04:28.360
simply by looking at the words you have used before and the order in which you use them?
00:04:42.080
If the order of words that you've spoken before are basically a diagram of your brain,
00:04:50.420
you know, just put into a word form, it's a diagram of your brain.
00:04:54.700
Once you have a diagram of somebody's brain, in theory,
00:04:58.880
you should know what they're going to do if you introduce any new stimulation.
00:05:03.860
You know, if there's an or, what are you going to do?
00:05:07.880
So the question I ask is, could we use AI to build a model of Putin
00:05:14.860
that's based on things he've said, not counting speeches, right?
00:05:22.760
Because if you counted his speeches, you might be counting somebody else's words
00:05:31.040
If you could get enough data from his casual conversations,
00:05:35.040
could you build an AI model that you could use as your proxy
00:05:40.100
and say, okay, Putin, but you'd be talking to the AI.
00:05:47.560
do an aerial bombardment of a city on the Russian border.
00:06:01.060
And then you see if Putin keeps doing the same thing the AI says he would do.
00:06:05.460
And once you get a match, you can predict his next move.
00:06:16.540
May I introduce the response to the people who are not good at listening?
00:06:27.340
could you take a break and talk among yourselves?
00:06:29.820
I want to talk only to the people who are bad at listening.
00:06:45.740
hey, those speeches were written by somebody else.
00:06:56.120
That it's not part of the conversation and never shall be.
00:06:59.760
There shall be no discussion of using AI to look at the speeches
00:07:03.460
written by other people as a way to determine what Putin will do.
00:07:13.740
Now, I'd like to take a moment for those who didn't understand that to say,
00:07:18.020
Scott, you realize his speeches are written by other people.
00:07:30.100
And that was, you know, half of the comments were,
00:07:32.640
well, those speeches were written by other people.
00:08:44.720
Now, here's something that's just freaking me out
00:09:42.640
we mostly have just been having more and more people
00:10:04.960
there's a company that's building robots as a service.
00:10:34.500
then every robot nurse had to do that job at your company.
00:11:00.480
and during Christmas you could get three of them
00:11:06.380
And then you don't have to worry about firing them.
00:11:23.900
with the time that we don't need a lot of people.
00:11:32.360
We will live forever by merging with the robots.
00:11:37.480
I'm less concerned about robot danger than you are.
00:11:59.120
which will start connecting with all the other AIs,
00:12:19.340
that are operating in different companies and such,
00:13:19.940
I think that's within the realm of possibility.