Exploring the Historical Development and Ethical Considerations of Generative AI with Noah Healy

WEBVTT

00:00:04.740 --> 00:00:07.360
Do you scream what do you use for your programs

00:00:07.360 --> 00:00:13.300
your podcast? Dang it. I can't remember right

00:00:13.300 --> 00:00:16.620
off the top of my head. It's It's it's one of

00:00:16.620 --> 00:00:21.519
the multi -channel Recording platforms, but there's

00:00:21.519 --> 00:00:24.420
like three of them and I that the name of the

00:00:24.420 --> 00:00:28.660
one we use isn't springing to mind No worries.

00:00:28.660 --> 00:00:30.199
No worries. Well, we'll get that we'll get that

00:00:30.199 --> 00:00:32.140
information for the podcast. All right, so we're

00:00:32.140 --> 00:00:36.420
doing it again Another second of today's No Prep

00:00:36.420 --> 00:00:39.439
Needed. Noah, have I met you before? You have

00:00:39.439 --> 00:00:41.840
not. We know where you are with that beautiful

00:00:41.840 --> 00:00:45.179
bookcase in the back. I'm in Daytona Beach. Welcome

00:00:45.179 --> 00:00:48.240
to our live stream No Prep Needed. This just

00:00:48.240 --> 00:00:50.840
shows you when you meet someone and you get to

00:00:50.840 --> 00:00:53.439
know them. It's okay. You sit down and you have

00:00:53.439 --> 00:00:55.259
a conversation. The only thing we're doing is

00:00:55.259 --> 00:00:57.420
we're allowing each of you to be flies on the

00:00:57.420 --> 00:01:00.880
wall as Noah shares with me. who he is. I can

00:01:00.880 --> 00:01:02.539
tell you a couple things. Again, I know he's

00:01:02.539 --> 00:01:05.700
bold and he's brave. He's not concerned because

00:01:05.700 --> 00:01:07.799
he actually has his own podcast. So he's been

00:01:07.799 --> 00:01:10.900
doing this for a while. And I love the fact that

00:01:10.900 --> 00:01:14.379
he trusted me enough to say, yes, Adrian. So

00:01:14.379 --> 00:01:17.239
Noah, I want to give you the pleasure of introducing

00:01:17.239 --> 00:01:20.299
yourself because I don't know you. Sure. Yeah.

00:01:20.359 --> 00:01:23.019
Well, Noah Healy, I'm in Charlottesville, Virginia.

00:01:24.089 --> 00:01:27.269
And as we've got going on the banner there, I've

00:01:27.269 --> 00:01:30.269
got a podcast called The Fourth Age, where I

00:01:30.269 --> 00:01:34.609
discuss AI and its implications for our society,

00:01:34.670 --> 00:01:39.180
economy and politics with my co -host, Marty

00:01:39.180 --> 00:01:42.459
Wiener, who is the retired CTO of Reddit. And

00:01:42.459 --> 00:01:47.340
I'm also working on a patent I developed to change

00:01:47.340 --> 00:01:51.640
how marketplaces function and make production

00:01:51.640 --> 00:01:54.019
radically more profitable than it has been for

00:01:54.019 --> 00:01:59.340
the last 800 years. Okay. This is so fascinating

00:01:59.340 --> 00:02:02.519
to me. The only like it's weird around last year,

00:02:02.620 --> 00:02:06.060
around this time is when I first heard of like

00:02:06.060 --> 00:02:09.699
Jasper, ChatGBT, and AI. I'd never heard of it

00:02:09.699 --> 00:02:11.319
before. And I feel like at that point, I was

00:02:11.319 --> 00:02:15.020
behind the game, right? When was this all developed?

00:02:15.840 --> 00:02:19.759
Well, technically, the entire concept of the

00:02:19.759 --> 00:02:22.780
neural net goes back to the earliest days of

00:02:22.780 --> 00:02:25.860
computers themselves. One of the first proposals

00:02:25.860 --> 00:02:30.560
for how to actually build an artificial computational

00:02:30.560 --> 00:02:35.330
mechanism was sometimes called something like

00:02:35.330 --> 00:02:37.969
the synaptotron or something like that, where

00:02:37.969 --> 00:02:42.189
you would model human neural synapses. And the

00:02:42.189 --> 00:02:45.490
first time there was a period where it looked

00:02:45.490 --> 00:02:47.430
like that approach might actually be bearing

00:02:47.430 --> 00:02:51.830
some fruit was late 70s, but mostly early 80s.

00:02:51.830 --> 00:02:56.229
And those companies didn't really wind up producing

00:02:56.229 --> 00:02:59.629
very much of value. And so it was known as the

00:02:59.629 --> 00:03:06.610
AI winter. But in the late 90s, there was a minor

00:03:06.610 --> 00:03:12.110
but fairly public incident of IBM building the

00:03:12.110 --> 00:03:15.650
deep blue supercomputer and defeating the then

00:03:15.650 --> 00:03:22.569
world champion in chess, which was a bit of an

00:03:22.569 --> 00:03:25.370
outlier. It took a while before chess programs

00:03:25.370 --> 00:03:27.750
got to be actually better than people in general,

00:03:28.009 --> 00:03:30.969
other than purposely built supercomputers designed

00:03:30.969 --> 00:03:36.530
to defeat a single human being. However, a few

00:03:36.530 --> 00:03:39.830
different breakthroughs in essentially unrelated

00:03:39.830 --> 00:03:44.830
fields combined with a, at the time, fairly quixotic

00:03:44.830 --> 00:03:50.129
sinking of tens of millions of dollars into research

00:03:50.129 --> 00:03:54.009
led to these generative AIs like Chachi PT and

00:03:54.009 --> 00:03:56.610
Jasper and now Gemini and these are the ones

00:03:56.610 --> 00:04:00.199
you've mentioned. that sort of suddenly took

00:04:00.199 --> 00:04:03.259
the consciousness of the public by storm about

00:04:03.259 --> 00:04:07.500
two years ago now. I feel like that I'm a year

00:04:07.500 --> 00:04:12.900
off. So chat boxes or chat bots, they've been

00:04:12.900 --> 00:04:16.620
around for a while. They're AI. They are what

00:04:16.620 --> 00:04:22.319
people generally think of as AI these days. And

00:04:22.319 --> 00:04:24.540
they're radically more sophisticated than the

00:04:24.540 --> 00:04:28.579
bots that existed before. But there's a bot known

00:04:28.579 --> 00:04:33.560
as Eliza that goes back again to the 70s that

00:04:33.560 --> 00:04:39.329
was based on a speech pattern. of a therapist.

00:04:40.170 --> 00:04:42.990
And so you could basically make any statement

00:04:42.990 --> 00:04:46.209
you wanted to Eliza. You could say, you know,

00:04:46.470 --> 00:04:49.350
hi, I'm Adrian Barker. And it would say, how

00:04:49.350 --> 00:04:52.329
do you feel about being Adrian Barker? And and

00:04:52.329 --> 00:04:55.310
it would just keep doing that kind of, you know,

00:04:55.430 --> 00:04:57.790
therapy thing. And it could have conversations

00:04:57.790 --> 00:05:01.529
with people. It couldn't have detailed conversations

00:05:01.529 --> 00:05:06.230
or bring in facts. And it was it didn't hallucinate

00:05:06.230 --> 00:05:11.459
nearly as uh effectively as as these modern chat

00:05:11.459 --> 00:05:16.000
bots do but again we're looking at technologies

00:05:16.000 --> 00:05:21.620
that have been showing off their existence for

00:05:21.620 --> 00:05:25.519
in some cases half a century and what we're now

00:05:25.519 --> 00:05:29.379
seeing is that these things are moving from toy

00:05:29.379 --> 00:05:36.639
stage to uh kind of real stage, and we'll see

00:05:36.639 --> 00:05:40.860
if they can break through to general use. I'd

00:05:40.860 --> 00:05:43.199
imagine they probably will. That is so interesting.

00:05:43.379 --> 00:05:46.259
Actually, about this time last year, I got a

00:05:46.259 --> 00:05:49.199
fun app for myself and my daughter. It was like

00:05:49.199 --> 00:05:51.240
a fake person that seemed real, and you could

00:05:51.240 --> 00:05:53.399
ask questions, and it was for the iPhone. And

00:05:53.399 --> 00:05:55.120
I remember thinking, this is like ridiculously

00:05:55.120 --> 00:05:58.019
cool. So all right, I have a question for you,

00:05:58.199 --> 00:06:01.230
though. With and I know you invented something.

00:06:01.230 --> 00:06:03.529
I think so excited. I get very excited I'm ready

00:06:03.529 --> 00:06:06.529
to like push out a hundred questions to you But

00:06:06.529 --> 00:06:10.170
my one question is this if there was if we had

00:06:10.170 --> 00:06:13.230
25 people in a room and we were all using let's

00:06:13.230 --> 00:06:17.750
just say chat GPT and We all ask the same question

00:06:17.750 --> 00:06:21.750
Why is the sky blue would all 25 people get the

00:06:21.750 --> 00:06:27.470
same answer? probably not in fact uh the way

00:06:27.470 --> 00:06:29.170
many of these things are designed essentially

00:06:29.170 --> 00:06:34.389
certainly not uh the one way to think about how

00:06:34.389 --> 00:06:36.810
these work if you're familiar with sort of patinko

00:06:36.810 --> 00:06:42.589
machines the the there's a a game that they show

00:06:42.589 --> 00:06:44.529
off and the price is right where somebody takes

00:06:44.529 --> 00:06:48.589
sort of a hockey puck and drops it down a bed

00:06:48.589 --> 00:06:52.149
of nails and it sort of bounces around uh yes

00:06:52.149 --> 00:06:56.160
plinko that's it yeah so Some of the mathematics

00:06:56.160 --> 00:07:01.199
of how this thing decides what to do bears a

00:07:01.199 --> 00:07:06.339
very strong resemblance, that's known as lattices

00:07:06.339 --> 00:07:10.579
in the mathematical term, to how Plinko machines

00:07:10.579 --> 00:07:15.759
work. And because of how many arbitrary choices

00:07:15.759 --> 00:07:18.519
there are to make and the amount of arbitrariness

00:07:18.519 --> 00:07:25.769
that they plug into those things that it won't

00:07:25.769 --> 00:07:30.769
be circumstantially, basically the 25 of you

00:07:30.769 --> 00:07:33.089
can't all ask the same question because you're

00:07:33.089 --> 00:07:35.389
asking from different devices at slightly different

00:07:35.389 --> 00:07:38.930
times. And as a result of this combination of

00:07:38.930 --> 00:07:42.089
other factors and their decision to design it

00:07:42.089 --> 00:07:44.790
so that it would be sensitive to various things

00:07:44.790 --> 00:07:48.569
in order to not just mechanically be slavishly

00:07:48.569 --> 00:07:51.550
repeating itself or hackable in that easy way.

00:07:52.900 --> 00:07:55.639
it's very unlikely that even two of you would

00:07:55.639 --> 00:07:58.279
really get the same response. I hope everyone

00:07:58.279 --> 00:08:00.160
heard that because that is so interesting because

00:08:00.160 --> 00:08:03.379
I think that I use ChatGBT every day and my son

00:08:03.379 --> 00:08:06.009
and I tried it with the artwork too. within Dali,

00:08:06.529 --> 00:08:08.649
like I said, okay, let's, we typed in the same

00:08:08.649 --> 00:08:10.449
thing and we both did come out with completely

00:08:10.449 --> 00:08:13.050
different photos, but I was still like, well,

00:08:13.050 --> 00:08:14.529
what if there was more people in the room? Like

00:08:14.529 --> 00:08:17.310
it wasn't enough of a test for me, but I love

00:08:17.310 --> 00:08:19.350
this because in my mind, people are like, so

00:08:19.350 --> 00:08:21.430
those that haven't used it are the first ones

00:08:21.430 --> 00:08:23.410
to say, oh my gosh, you know, it's going to be

00:08:23.410 --> 00:08:25.250
thinking for us. It's going to be taking over,

00:08:25.350 --> 00:08:27.930
which is my next question. But the reality is,

00:08:28.120 --> 00:08:30.759
I may use it, it may start off, but I'm still

00:08:30.759 --> 00:08:33.779
massaging it and taking time to make sure that

00:08:33.779 --> 00:08:36.179
it's giving me the right answers. And I don't

00:08:36.179 --> 00:08:39.539
feel like, what if I'm looking for topics to

00:08:39.539 --> 00:08:42.200
write about? I feel like it's still coming from

00:08:42.200 --> 00:08:44.399
me. It's just helping me to dress it up a little

00:08:44.399 --> 00:08:49.679
bit. So when people say that, Noah, oh, I don't

00:08:49.679 --> 00:08:51.580
like chat GBT or any of that. It's gonna take

00:08:51.580 --> 00:08:55.230
over the world. What's your answer? Well, I think

00:08:55.230 --> 00:08:57.549
it's obviously the case that it will take over

00:08:57.549 --> 00:09:01.629
the world. However, I don't think that people

00:09:01.629 --> 00:09:06.710
are treating it in... The term that I use for

00:09:06.710 --> 00:09:09.289
the generative AI is that it's less consequential

00:09:09.289 --> 00:09:12.110
and more dangerous than people generally regard

00:09:12.110 --> 00:09:18.350
it as. So... when industrialization was getting

00:09:18.350 --> 00:09:21.350
going and engines were happening and people stopped

00:09:21.350 --> 00:09:24.549
doing as much day labor because engines could

00:09:24.549 --> 00:09:28.409
do quite a lot of that for us, there was not

00:09:28.409 --> 00:09:31.950
a sort of general pushback that machines were

00:09:31.950 --> 00:09:35.649
taking over. But there were some serious concerns

00:09:35.649 --> 00:09:39.450
with urbanization. People looked at peasants

00:09:39.450 --> 00:09:42.250
getting up off the land and said to themselves,

00:09:42.549 --> 00:09:46.360
well, you know, 80 % of the people working backbreaking

00:09:46.360 --> 00:09:49.220
farming subsistence farming jobs is where all

00:09:49.220 --> 00:09:52.240
the food comes from. So if half of those people

00:09:52.240 --> 00:09:54.980
leave and go to the cities for these good paying

00:09:54.980 --> 00:09:58.360
manufacturing jobs then you know we'll have half

00:09:58.360 --> 00:10:00.919
as much food and half of us will starve to death

00:10:00.919 --> 00:10:03.960
and that's a death spiral so this is all terrible.

00:10:04.240 --> 00:10:06.639
Of course we figured out how to use engines to

00:10:06.639 --> 00:10:10.019
make backbreaking farming labor more effective

00:10:10.019 --> 00:10:14.429
and produce a lot more food. That's the kind

00:10:14.429 --> 00:10:18.029
of changes that we're likely to see. However,

00:10:19.169 --> 00:10:23.110
because these things are producing stuff that

00:10:23.110 --> 00:10:28.330
looks like or sounds like human opinions and

00:10:28.330 --> 00:10:32.309
ideas, the risks that we are currently up against

00:10:32.309 --> 00:10:36.169
is that our institutions, from our politics to

00:10:36.169 --> 00:10:40.129
our economics, are based around the idea that

00:10:40.129 --> 00:10:43.389
everyone that's making statements and having

00:10:43.389 --> 00:10:47.190
opinions is a person and that the number of persons

00:10:47.190 --> 00:10:51.330
that exist is relatively static. It takes 18

00:10:51.330 --> 00:10:53.710
years to turn a baby into an adult and so on.

00:10:54.789 --> 00:10:57.009
I was talking to some recent college graduates

00:10:57.009 --> 00:10:59.710
who were highly successful and are now running

00:10:59.710 --> 00:11:01.789
a multi -million dollar company that they founded

00:11:01.789 --> 00:11:04.129
while they were in college. But I was pointing

00:11:04.129 --> 00:11:08.419
out to them that one of the reasons that sort

00:11:08.419 --> 00:11:11.240
of good jobs are available for people such as

00:11:11.240 --> 00:11:14.419
themselves is that they're relatively scarce.

00:11:14.659 --> 00:11:17.080
College graduates aren't the majority of people

00:11:17.080 --> 00:11:19.960
and college graduates with the sort of gumption

00:11:19.960 --> 00:11:22.379
and go getting to found their own multimillion

00:11:22.379 --> 00:11:25.039
dollar companies are obviously a very small sub

00:11:25.039 --> 00:11:29.299
-fraction of the college graduates. But as has

00:11:29.299 --> 00:11:31.440
been demonstrated multiple times, ChatUBT can

00:11:31.440 --> 00:11:37.549
produce collegiate level essays in seconds. And

00:11:37.549 --> 00:11:42.230
that's just on the $20 a month plan. So in a

00:11:42.230 --> 00:11:45.629
world where there's a few, say, tens of millions

00:11:45.629 --> 00:11:51.690
of college graduates every year, there'll be

00:11:51.690 --> 00:11:53.909
a certain value attached to that. In a world

00:11:53.909 --> 00:11:57.610
where computers can manufacture trillions of

00:11:57.610 --> 00:12:00.490
college graduates every year, that value will

00:12:00.490 --> 00:12:04.370
naturally be radically lower. And the values

00:12:04.759 --> 00:12:08.519
in society that are going to be radically lowered

00:12:08.519 --> 00:12:11.840
are the ones at the tops of our societies. It's

00:12:11.840 --> 00:12:15.340
much easier for these things to simulate what

00:12:15.340 --> 00:12:21.320
CEOs do or what politicians do or what entertainers

00:12:21.320 --> 00:12:25.480
do. And it's a lot harder right now for them

00:12:25.480 --> 00:12:29.080
to simulate things like what cab drivers or garbage

00:12:29.080 --> 00:12:34.399
men do. our society is going to have to change

00:12:34.399 --> 00:12:39.299
in ways that makes us able to use these capacities

00:12:39.299 --> 00:12:41.980
in ways that are actually valuable to society,

00:12:42.580 --> 00:12:45.299
because mostly what it's done so far and what

00:12:45.299 --> 00:12:49.840
it will do until we look into these issues is

00:12:49.840 --> 00:12:54.460
just make it a lot noisier and crazier. Oh my

00:12:54.460 --> 00:12:56.860
gosh, you're brilliant. All right, here's another

00:12:56.860 --> 00:13:01.159
question. I'm doing my senior essay in college.

00:13:01.860 --> 00:13:05.860
I use Chat GPT for the whole thing. I then get

00:13:05.860 --> 00:13:08.559
a little nervous and I say to Chat GPT, hey,

00:13:09.059 --> 00:13:12.159
write like I'm a 19 -year -old female and then

00:13:12.159 --> 00:13:15.259
add a few mistakes so that nobody really knows

00:13:15.259 --> 00:13:17.500
that it's real and make sure that I can't be

00:13:17.500 --> 00:13:21.500
found anywhere and help me to make it sound like

00:13:21.500 --> 00:13:26.200
my own. Will the plagiarism pick it up? How do

00:13:26.200 --> 00:13:29.750
the professors know? when it is chachy betique

00:13:29.750 --> 00:13:32.669
appeared to that it's not. Like, what does that

00:13:32.669 --> 00:13:34.450
look like? Because I get very confused when I

00:13:34.450 --> 00:13:37.529
read that they can figure it out. Well, in some

00:13:37.529 --> 00:13:40.850
cases, they'll be able to work it out because

00:13:40.850 --> 00:13:43.330
they might be able to build models that can detect

00:13:43.330 --> 00:13:47.230
whether or not there is plagiarism. As these

00:13:47.230 --> 00:13:49.730
models develop, there'll be things going on.

00:13:50.269 --> 00:13:53.309
OpenAI may, they certainly haven't announced

00:13:53.309 --> 00:13:56.350
that they're doing this, but any of these companies

00:13:56.350 --> 00:13:59.330
might have certain embedded characteristics that

00:13:59.330 --> 00:14:04.590
they decide to sign. For example, printers due

00:14:04.590 --> 00:14:08.549
to... regulation have micro printing intrinsic

00:14:08.549 --> 00:14:12.269
in them so that the outputs of printers are traceable

00:14:12.269 --> 00:14:15.070
back to specific machines because of the micro

00:14:15.070 --> 00:14:18.830
printing on each page. And that's also something

00:14:18.830 --> 00:14:24.039
that exists in computer chips. My laptop has

00:14:24.039 --> 00:14:27.500
been unbound so that I can operate a real operating

00:14:27.500 --> 00:14:31.279
system on it. But the standard laptop computer

00:14:31.279 --> 00:14:33.879
chip has something called the trusted computing

00:14:33.879 --> 00:14:37.220
platform, which was pushed through by a number

00:14:37.220 --> 00:14:41.769
of. industry big wigs that allows the chip on

00:14:41.769 --> 00:14:44.870
your machine to understand whether or not the

00:14:44.870 --> 00:14:48.269
software running on it belongs to an approved

00:14:48.269 --> 00:14:51.429
person. And if not to not run that software.

00:14:51.730 --> 00:14:56.649
And that's basically a Microsoft digital rights

00:14:56.649 --> 00:14:59.809
management technique that is built in at the

00:14:59.809 --> 00:15:03.149
chip level so that customers can't pirate their

00:15:03.149 --> 00:15:05.529
stuff. And there's other people that we're also

00:15:05.529 --> 00:15:09.570
interested in. in being able to control technology

00:15:09.570 --> 00:15:14.629
that not only belongs to consumers. It would

00:15:14.629 --> 00:15:17.389
be possible for these companies to be behaving

00:15:17.389 --> 00:15:19.549
this way. They have not announced that they are

00:15:19.549 --> 00:15:23.210
doing so, but they've certainly done nothing

00:15:23.210 --> 00:15:27.049
to ensure us that they aren't at least that nefarious.

00:15:28.649 --> 00:15:32.389
But that's more or less the precise issue that

00:15:32.389 --> 00:15:36.720
you've put your finger on because if it were

00:15:36.720 --> 00:15:40.919
developed a method to work out that people were

00:15:40.919 --> 00:15:44.700
plagiarizing, that methodology itself would be

00:15:44.700 --> 00:15:48.779
the basis necessary to train new sets of generative

00:15:48.779 --> 00:15:51.179
AIs that could evade that, that would be able

00:15:51.179 --> 00:15:56.679
to do their jobs even better. And so I go to

00:15:56.679 --> 00:15:59.639
an example from my collegiate experience. I went

00:15:59.639 --> 00:16:04.200
to UVA and we have an honor system which as part

00:16:04.200 --> 00:16:08.600
of its action, it's strongly discouraged for

00:16:08.600 --> 00:16:13.200
professors to even monitor students during the

00:16:13.200 --> 00:16:17.379
giving of exams. There's a handful of very notorious

00:16:17.379 --> 00:16:22.740
cases where it's actually done. Basically, physics

00:16:22.740 --> 00:16:27.580
for jocks was a notorious bed of cheats. But

00:16:27.580 --> 00:16:33.019
for the most part, students are trusted and encouraged

00:16:33.019 --> 00:16:36.059
to be trusted at every level of the university.

00:16:36.440 --> 00:16:39.559
One of my professors, who was the dean of undergrads

00:16:39.559 --> 00:16:42.539
at the time, felt that this was a perfectly decent

00:16:42.539 --> 00:16:45.019
thing to do and actually sort of wonderful. But

00:16:45.019 --> 00:16:48.700
on the other hand, training engineers didn't

00:16:48.700 --> 00:16:52.019
feel that it was perhaps wise to simply open

00:16:52.019 --> 00:16:55.139
the doors to cheating. And so he designed his

00:16:55.139 --> 00:16:57.879
exam so they could not be, in fact, cheated on.

00:16:58.009 --> 00:17:02.289
And so his exams were open book, open note, week

00:17:02.289 --> 00:17:06.670
long systems where you could solve the problems

00:17:06.670 --> 00:17:10.390
yourself or with sufficient research, maybe well

00:17:10.390 --> 00:17:13.430
find the sources of the problems and simply copy

00:17:13.430 --> 00:17:20.789
down the work. However, because you had effectively

00:17:20.789 --> 00:17:25.710
infinite. scope and scale, the grading was extremely

00:17:25.710 --> 00:17:30.990
tight. He would grade down for misspellings or

00:17:30.990 --> 00:17:35.470
poor penmanship or basically any error at all

00:17:35.470 --> 00:17:39.089
because the object was to produce finished engineering

00:17:39.089 --> 00:17:43.990
work under general engineering conditions. And

00:17:43.990 --> 00:17:48.130
you could certainly start thinking along those

00:17:48.130 --> 00:17:53.640
lines. In a world where it's possible to credibly

00:17:53.640 --> 00:17:57.519
pretend to be a skilled practitioner of any field

00:17:57.519 --> 00:18:01.019
because ChatGPT can do that for you. How would

00:18:01.019 --> 00:18:06.119
you go about interacting with people such that

00:18:06.119 --> 00:18:10.759
it doesn't matter whether or not they have the

00:18:10.759 --> 00:18:14.579
ability or ChatGPT is providing the ability or

00:18:14.579 --> 00:18:17.420
anything. You can instead distinguish whether

00:18:17.420 --> 00:18:23.430
or not the value being produced is real or simply

00:18:23.430 --> 00:18:27.130
a very convincing fake. So when I was growing

00:18:27.130 --> 00:18:29.410
up, we would go downstairs, we would get the

00:18:29.410 --> 00:18:33.789
big, huge dictionaries and our encyclopedias

00:18:33.789 --> 00:18:36.109
and we would take the encyclopedia and, oh, I

00:18:36.109 --> 00:18:39.309
have to write about, hey, Lincoln, okay. And

00:18:39.309 --> 00:18:42.970
then I would like... I pretty much took it word

00:18:42.970 --> 00:18:44.569
-for -word and then just switched things out

00:18:44.569 --> 00:18:46.789
because I figured how would they know right?

00:18:46.869 --> 00:18:48.529
I'm just a kid they wouldn't really know that

00:18:48.529 --> 00:18:51.089
I'm using an encyclopedia and but sometimes they

00:18:51.089 --> 00:18:53.529
would say did you use an encyclopedia and I didn't

00:18:53.529 --> 00:18:56.390
lie so I was like yes, but Like the same kind

00:18:56.390 --> 00:18:58.410
of thing right like I always I always felt that

00:18:58.410 --> 00:19:01.349
as long as you knew weird how to find the information

00:19:01.349 --> 00:19:03.769
How much can you store in your brain anyway,

00:19:03.869 --> 00:19:08.069
so I mean am I off of my way of thinking? I would

00:19:08.069 --> 00:19:10.170
certainly tend to agree with you. I don't really

00:19:10.170 --> 00:19:14.009
see a moral differentiator between a person that

00:19:14.009 --> 00:19:17.769
has temporarily stored information in their hypothalamus

00:19:17.769 --> 00:19:21.130
and people who have temporarily stored information

00:19:21.130 --> 00:19:26.190
on their inner thigh in order to pass an exam.

00:19:29.049 --> 00:19:33.190
I understand that there is a very big practical

00:19:33.190 --> 00:19:37.700
difference in existing you know, educational

00:19:37.700 --> 00:19:40.039
practice. And that's one of the points that I

00:19:40.039 --> 00:19:43.240
very strongly disagree with. And I think CHAT

00:19:43.240 --> 00:19:45.779
GPT and some of these other generative systems

00:19:45.779 --> 00:19:51.319
is going to very much point up these problems.

00:19:51.420 --> 00:19:53.819
And it points again to one of these issues that

00:19:53.819 --> 00:19:57.619
we're going to have to cope with societally because

00:19:57.619 --> 00:20:01.339
it really is important to find high skilled,

00:20:01.559 --> 00:20:06.450
high capability people. to work on our hard problems

00:20:06.450 --> 00:20:09.890
because we actually have hard problems and we

00:20:09.890 --> 00:20:13.529
get better results out of very smart people than

00:20:13.529 --> 00:20:17.049
we get out of very stupid people. And so if we

00:20:17.049 --> 00:20:21.089
lose the ability to tell the difference, then

00:20:21.089 --> 00:20:24.849
things won't go very well. But on the other hand,

00:20:25.730 --> 00:20:28.109
to the extent that we are capable of using these

00:20:28.109 --> 00:20:31.380
technologies to augment our capacities, failing

00:20:31.380 --> 00:20:34.720
to do that is also going to be a big negative.

00:20:37.420 --> 00:20:42.920
Less well -known, but another kind of use of

00:20:42.920 --> 00:20:46.039
these neural nets is a thing called reinforcement

00:20:46.039 --> 00:20:49.480
learning. or occasionally deep learning, which

00:20:49.480 --> 00:20:54.519
is capable of four sort of known systems creating

00:20:54.519 --> 00:20:58.940
very deep understandings. And some of the outcomes

00:20:58.940 --> 00:21:01.900
of deep learning technology have been quite extraordinary.

00:21:02.400 --> 00:21:07.720
For example, Google released a paper maybe four

00:21:07.720 --> 00:21:12.039
or five months ago now. where they announced

00:21:12.039 --> 00:21:19.660
220 ,000 new stable chemical crystalline structures.

00:21:20.700 --> 00:21:23.680
And in all of human history, we had developed

00:21:23.680 --> 00:21:27.039
somewhere between 50 and 70 ,000 known stable

00:21:27.039 --> 00:21:30.059
chemical crystalline structures. So they sort

00:21:30.059 --> 00:21:35.799
of five X'd the number of known types of crystals.

00:21:36.240 --> 00:21:37.700
And something that's important to understand

00:21:37.700 --> 00:21:44.339
about crystals is that we have historical precedent

00:21:44.339 --> 00:21:47.759
for not being able to actually find these things

00:21:47.759 --> 00:21:49.960
experimentally until after we know that they're

00:21:49.960 --> 00:21:55.410
real. So there were a large number of crystals

00:21:55.410 --> 00:21:59.750
discovered during the 20th century as mathematical,

00:22:00.049 --> 00:22:03.410
going back to that lattice thing again, as the

00:22:03.410 --> 00:22:09.930
mathematics of lattices and also irregular lattices

00:22:09.930 --> 00:22:14.559
really advanced. we gained new abilities to understand

00:22:14.559 --> 00:22:17.119
new kinds of crystalline structure. And when

00:22:17.119 --> 00:22:19.660
we did that, and then started doing the x -ray

00:22:19.660 --> 00:22:21.519
crystallography to actually look at the universe,

00:22:21.799 --> 00:22:24.059
we suddenly discovered that many well -known

00:22:24.059 --> 00:22:29.119
minerals actually had some of these obscure crystalline

00:22:29.119 --> 00:22:32.740
structures that we hadn't known about because

00:22:32.740 --> 00:22:36.019
we didn't know how to interpret the data we were

00:22:36.019 --> 00:22:40.180
looking at. Wow. OK, hold on for a second. A

00:22:40.180 --> 00:22:43.250
lot of that went over my head. I'll try to catch

00:22:43.250 --> 00:22:46.029
her. Wow. Okay. I do want to go into that. You

00:22:46.029 --> 00:22:47.990
have a podcast and that you're a venture, but

00:22:47.990 --> 00:22:50.109
I have to ask one last final question on this.

00:22:50.109 --> 00:22:52.549
Cause I know like I can talk to you for hours.

00:22:53.089 --> 00:22:57.230
I read somewhere that I first I read 2050. They

00:22:57.230 --> 00:23:00.210
said that the AI systems will be smarter than

00:23:00.210 --> 00:23:03.170
we are. And then someone said, Oh no, like you

00:23:03.170 --> 00:23:06.529
could say 2050, but really 2030 will probably

00:23:06.529 --> 00:23:09.849
be there where the, these systems will be actually

00:23:09.849 --> 00:23:13.529
smarter than a human brain. Is that yes, no,

00:23:13.809 --> 00:23:18.390
true? OK, well, I would once what I usually tell

00:23:18.390 --> 00:23:20.430
people under the circumstances is that you've

00:23:20.430 --> 00:23:25.789
got your tenses wrong. So another fairly impressive

00:23:25.789 --> 00:23:27.849
use of deep learning, and this is in combination

00:23:27.849 --> 00:23:31.930
with some of this generative stuff, there's a

00:23:31.930 --> 00:23:36.309
test. a math test given out around the world.

00:23:36.430 --> 00:23:38.990
And it's occasionally known as the hardest math

00:23:38.990 --> 00:23:44.210
test in the world. And it's given to kids for

00:23:44.210 --> 00:23:49.609
the most part. The scores go from zero to 100.

00:23:50.089 --> 00:23:56.430
And the average winning score is 25 .9. The average

00:23:56.430 --> 00:24:01.319
second place score in the world is 23. an approach

00:24:01.319 --> 00:24:06.019
combining deep learning and generative AI, along

00:24:06.019 --> 00:24:12.119
with a computer program called a proof solver,

00:24:12.779 --> 00:24:18.400
was able to get 25 on this test. So strictly

00:24:18.400 --> 00:24:22.480
speaking, not smarter than the smartest human

00:24:22.480 --> 00:24:31.750
being of any given age. age cohort, but 94 percent

00:24:31.750 --> 00:24:35.509
is smart, 96 percent as smart as that, and smarter

00:24:35.509 --> 00:24:44.009
than the second smartest human being. So what

00:24:44.009 --> 00:24:47.630
we think of as intelligence itself is one of

00:24:47.630 --> 00:24:50.430
the things that's on the chopping block for these

00:24:50.430 --> 00:24:54.839
systems. That's interesting. Chess, for example,

00:24:54.900 --> 00:24:56.720
has been regarded for centuries as a measure

00:24:56.720 --> 00:25:00.599
of intelligence. And as I said, in 1997, the

00:25:00.599 --> 00:25:03.720
then world champion lost in a six game series

00:25:03.720 --> 00:25:06.559
against Deep Blue. Now, he wasn't completely

00:25:06.559 --> 00:25:11.059
crushed. He won a couple and he lost a few more

00:25:11.059 --> 00:25:16.799
and there were some draws. The modern chess engines

00:25:16.799 --> 00:25:23.700
are something like 800 points stronger than modern.

00:25:24.819 --> 00:25:28.339
Top level super grandmasters, world number one

00:25:28.339 --> 00:25:31.099
world champion, those guys, they're they're eight

00:25:31.099 --> 00:25:33.400
hundred ish points. So with this in perspective,

00:25:33.680 --> 00:25:37.000
chess masters, basically, if you're in a if you're

00:25:37.000 --> 00:25:39.480
in a town or a reasonable city and you've got

00:25:39.480 --> 00:25:42.359
like a chess club and they're not that serious,

00:25:42.779 --> 00:25:45.299
the best player at that chess club is about 800

00:25:45.299 --> 00:25:48.420
points weaker than the best player on earth.

00:25:50.410 --> 00:25:54.609
This is another 800 points stronger than that.

00:25:55.390 --> 00:25:58.170
And these chess engines have been well beyond

00:25:58.170 --> 00:26:01.569
the capacity of human beings to play chess for

00:26:01.569 --> 00:26:05.369
decades now. And they're getting better at chess.

00:26:07.170 --> 00:26:11.309
So chess has enjoyed a big revival with COVID

00:26:11.309 --> 00:26:13.950
and everybody being online and getting to play

00:26:13.950 --> 00:26:16.609
chess as a hobby and so on. There's a lot more

00:26:16.609 --> 00:26:20.859
chess being played. The best chess players seem

00:26:20.859 --> 00:26:23.220
to be getting a bit better as a result of more

00:26:23.220 --> 00:26:27.160
competition more play and so on Although possibly

00:26:27.160 --> 00:26:32.119
also because of cheating But the machines are

00:26:32.119 --> 00:26:35.359
also Competing against one another and the machines

00:26:35.359 --> 00:26:37.880
are getting better at chess faster than we are

00:26:37.880 --> 00:26:41.420
getting better at chess Wow, okay, and I get

00:26:41.420 --> 00:26:43.900
that all the moves but one more question and

00:26:43.900 --> 00:26:48.279
then we go into where your invention Where does

00:26:48.279 --> 00:26:51.460
wisdom fall in on this? Like I always felt like

00:26:51.460 --> 00:26:54.319
wisdom was very important and something unique

00:26:54.319 --> 00:26:58.480
and different to each of us. I think wisdom is

00:26:58.480 --> 00:27:02.619
very important and I think that again, I think

00:27:02.619 --> 00:27:04.960
hopefully we'll be able to call upon reserves

00:27:04.960 --> 00:27:09.180
of that in order to survive the transition of

00:27:09.180 --> 00:27:12.700
our societies to new kinds of societies that

00:27:12.700 --> 00:27:19.480
can cope with these things. But again, it's not

00:27:19.480 --> 00:27:23.460
outside of the realm of imagination, which is

00:27:23.460 --> 00:27:26.180
the realm of imagination is the precise realm

00:27:26.180 --> 00:27:30.880
being accessed by computers, that it would be

00:27:30.880 --> 00:27:35.160
possible to come up with a framework for wisdom

00:27:35.160 --> 00:27:39.839
and produce wise systems. That is absolutely

00:27:39.839 --> 00:27:46.500
not what anybody's currently focused on. But

00:27:46.500 --> 00:27:52.539
super wise intelligences are a mythological and

00:27:52.539 --> 00:27:56.799
actually sociological point of human history.

00:27:56.839 --> 00:28:00.079
Things like the Oracle at Delphi were supposedly

00:28:00.079 --> 00:28:02.980
always right and always sort of put people on

00:28:02.980 --> 00:28:07.299
the paths that they belonged on. And it's not

00:28:07.299 --> 00:28:12.799
very difficult to sort of imagine a sci -fi scenario

00:28:12.799 --> 00:28:16.819
where the Chad GPT at Delphi could become a real

00:28:16.819 --> 00:28:21.319
thing somehow, although probably not with the

00:28:21.319 --> 00:28:26.500
current cast of creators. Thank you. That's interesting.

00:28:26.819 --> 00:28:30.400
Wow. All right. So again, we haven't met everyone.

00:28:30.559 --> 00:28:33.039
So just so you know, this is I'm like blown away.

00:28:33.420 --> 00:28:38.039
Obviously, Noah is like super, super duper intelligent.

00:28:38.670 --> 00:28:41.410
like to put you up against a chat GPT Noah, you'll

00:28:41.410 --> 00:28:44.130
kick their ass. Okay, but seriously, tell me

00:28:44.130 --> 00:28:46.609
about the invention. And what is it called the

00:28:46.609 --> 00:28:51.170
CDM? Yes, CDM coordinated discovery markets.

00:28:51.490 --> 00:28:54.990
So basically, the marketplace exists to figure

00:28:54.990 --> 00:28:59.150
out the price that everyone that wants to sell

00:28:59.150 --> 00:29:01.490
and everybody wants to buy wants to buy and sell

00:29:01.490 --> 00:29:06.000
the same amount of stuff. So If you're just kind

00:29:06.000 --> 00:29:07.880
of like, you know, setting prices, you could

00:29:07.880 --> 00:29:09.539
just set prices wherever you want. You'd be like,

00:29:09.619 --> 00:29:13.279
wheat, cool, $100 a bushel. Well, people would

00:29:13.279 --> 00:29:15.380
be very interested in growing wheat at $100 a

00:29:15.380 --> 00:29:16.900
bushel. That's a lot more than you get right

00:29:16.900 --> 00:29:19.440
now. But breadmakers are not going to be very

00:29:19.440 --> 00:29:21.940
interested in buying wheat at $100 a bushel.

00:29:22.240 --> 00:29:24.099
They might not be able to sell those loaves of

00:29:24.099 --> 00:29:29.619
bread at $500 each. So, you know, suddenly we're

00:29:29.619 --> 00:29:33.500
transforming our, you know, cities into wheat

00:29:33.500 --> 00:29:36.000
farms. And at the same time, we're shutting down

00:29:36.000 --> 00:29:38.700
our bakeries. And like, you know, there's a major

00:29:38.700 --> 00:29:41.660
mismatch. Or you could say, well, wheat, that's

00:29:41.660 --> 00:29:44.960
a nickel a bushel. That seems fair. Well, maybe

00:29:44.960 --> 00:29:47.160
it seems fair to view the bread buying public,

00:29:47.160 --> 00:29:49.180
but the farmers aren't going to go into that

00:29:49.180 --> 00:29:51.160
business. They'll grow rice or corn or something

00:29:51.160 --> 00:29:56.079
else instead. So there's a price point. And that

00:29:56.079 --> 00:30:01.440
price point is the result of the sort of co -negotiation

00:30:01.440 --> 00:30:07.700
of everybody in the marketplace. And we currently

00:30:07.700 --> 00:30:13.019
develop a marketplace where it's based on individuals

00:30:13.019 --> 00:30:16.640
negotiating private prices, which are then...

00:30:16.559 --> 00:30:19.500
published to everybody. And the problem with

00:30:19.500 --> 00:30:23.299
that is that most of the information stream is

00:30:23.299 --> 00:30:26.579
actually robots trading with other robots. So

00:30:26.579 --> 00:30:29.039
going back to that, are they smarter than us?

00:30:29.579 --> 00:30:33.400
How will this affect our societies? We have had

00:30:33.400 --> 00:30:37.720
quite literally AI systems based on many of these

00:30:37.720 --> 00:30:41.640
same techniques that OpenAI is using. trading

00:30:41.640 --> 00:30:45.759
as the majority of the volume in our marketplaces

00:30:45.759 --> 00:30:51.700
for decades. And that's led to a lot of the market

00:30:51.700 --> 00:30:54.900
instability that we see where crazy things happen,

00:30:55.019 --> 00:30:58.019
sometimes so crazy that they become national

00:30:58.019 --> 00:31:02.500
news or international disasters. The way my system

00:31:02.500 --> 00:31:07.750
works is by integrating in multiple viewpoints

00:31:07.750 --> 00:31:12.410
and providing a stable price to entire sectors

00:31:12.410 --> 00:31:16.190
of the marketplace to then see how much they

00:31:16.190 --> 00:31:20.240
want. that price. So rather than being rewarded

00:31:20.240 --> 00:31:23.559
for being able to find somebody willing to take

00:31:23.559 --> 00:31:26.720
the worst deal you can come up with, you are

00:31:26.720 --> 00:31:29.779
instead rewarded for contributing towards the

00:31:29.779 --> 00:31:32.000
best deal that everybody can come up with. And

00:31:32.000 --> 00:31:34.000
the more you contribute to that, the greater

00:31:34.000 --> 00:31:37.180
your reward. Wow, that sounds brilliant. So if

00:31:37.180 --> 00:31:40.599
we go on to change .org, what exactly are we

00:31:40.599 --> 00:31:44.460
going to, how can we help you? Well, so I've

00:31:44.460 --> 00:31:46.740
been working on this patent for actually around

00:31:46.740 --> 00:31:50.200
nine years now. And the patent office is breaking

00:31:50.200 --> 00:31:53.779
its own rules, lying to Congress and doing everything

00:31:53.779 --> 00:31:58.569
else. it can to slow down and halt the adoption

00:31:58.569 --> 00:32:03.789
of this technology. Right now, it's been appealed.

00:32:03.789 --> 00:32:06.089
It was appealed last year and it's been scheduled

00:32:06.089 --> 00:32:12.380
to be heard July 2025. filing the appeal and

00:32:12.380 --> 00:32:14.359
waiting for it to be heard, there's not really

00:32:14.359 --> 00:32:17.460
much I can do. But about a month ago, somebody

00:32:17.460 --> 00:32:20.319
suggested, you know, I could do this. I could,

00:32:20.599 --> 00:32:23.420
you know, create a public record of people's

00:32:23.420 --> 00:32:26.359
being interested in having a better economy.

00:32:26.519 --> 00:32:30.480
And so that's there for people who would like

00:32:30.480 --> 00:32:35.220
there to be better markets or... or indeed a

00:32:35.220 --> 00:32:38.660
patent system that wasn't behaving in a very,

00:32:38.680 --> 00:32:42.539
very untoward fashion. OK. All right. Thank you.

00:32:42.619 --> 00:32:45.140
Thank you very. So go on and check it out, everyone.

00:32:45.240 --> 00:32:46.819
The link is there and I'll put it in the show

00:32:46.819 --> 00:32:50.400
notes. And you are a podcast host. So tell us

00:32:50.400 --> 00:32:53.279
about your show. Yes. So it's called The Fourth

00:32:53.279 --> 00:32:58.339
Age and subtitled The AI Revolution. This goes

00:32:58.339 --> 00:33:02.660
to my thing is that the plow, the steam engine

00:33:02.660 --> 00:33:07.049
and the computer are all technologies that change

00:33:07.049 --> 00:33:12.250
mankind's physical interaction with the environment,

00:33:12.750 --> 00:33:16.289
not just economically, but also politically and

00:33:16.289 --> 00:33:19.849
socially and even religiously. The religions

00:33:19.849 --> 00:33:23.450
of pre -agrarian cultures do not resemble tightly

00:33:23.450 --> 00:33:26.109
the religions of post -agrarian cultures. The

00:33:26.109 --> 00:33:28.710
religions of industrial nations do not resemble

00:33:28.710 --> 00:33:33.210
the religions of agrarian nations. Since I say

00:33:33.210 --> 00:33:35.150
that the computer is a bigger deal than the steam

00:33:35.150 --> 00:33:38.730
engine or the plow, I would expect our political,

00:33:39.109 --> 00:33:41.890
cultural, social, religious, and other institutions

00:33:41.890 --> 00:33:47.359
to also transform at least as much as... that

00:33:47.359 --> 00:33:50.619
they have during the previous revolutions and

00:33:50.619 --> 00:33:55.160
that if we're going to have civilization here

00:33:55.160 --> 00:33:58.059
in the fourth age, we're going to need to grapple

00:33:58.059 --> 00:34:02.019
with these questions. So the podcast, so far

00:34:02.019 --> 00:34:04.259
it's about half and half discussions between

00:34:04.259 --> 00:34:08.860
Marty and I and some guests are discussing these

00:34:08.860 --> 00:34:11.739
sorts of issues where we see AI having some of

00:34:11.739 --> 00:34:15.099
these impacts, what kinds of outcomes are happening

00:34:15.099 --> 00:34:21.429
and how how we can use wisdom and also mathematics

00:34:21.429 --> 00:34:26.610
to try to address these issues. Okay, and Marty

00:34:26.610 --> 00:34:32.530
is? Marty is the retired CTO of Reddit. I love

00:34:32.530 --> 00:34:37.570
Reddit too. I'm on it. I have the app. But what

00:34:37.570 --> 00:34:39.690
I've noticed is a lot of questions because Google

00:34:39.690 --> 00:34:42.550
is now more like a answer site than a search

00:34:42.550 --> 00:34:45.150
site, right? So a lot of times when I ask a question,

00:34:45.309 --> 00:34:48.710
it'll put me on Reddit. But do you agree with

00:34:48.710 --> 00:34:52.489
that with Google? Absolutely. So Google is another

00:34:52.489 --> 00:34:55.550
very interesting example of the production of

00:34:55.550 --> 00:34:58.530
an unstable super intelligence. So Google search

00:34:58.530 --> 00:35:02.940
back in the late 90s, early 2000s, was fantastic.

00:35:03.360 --> 00:35:05.179
And it wasn't just because it was free. It didn't

00:35:05.179 --> 00:35:08.300
have advertising on it. It was also because the

00:35:08.300 --> 00:35:10.739
internet hadn't reacted to the existence of Google

00:35:10.739 --> 00:35:14.559
search yet. And so Google's results were very

00:35:14.559 --> 00:35:19.900
much on point. But with the domination of search

00:35:19.900 --> 00:35:23.280
as how people interact with the internet, it

00:35:23.280 --> 00:35:28.420
became necessary to engage in search engine optimization.

00:35:29.309 --> 00:35:34.630
rather than pursuing content as your core methodology.

00:35:35.489 --> 00:35:38.170
And what happened is Google search algorithm

00:35:38.170 --> 00:35:43.849
has becoming worse and less effective because

00:35:43.849 --> 00:35:48.730
it exists, basically. And this is the fundamental

00:35:48.730 --> 00:35:51.329
problem that we're going to have across all of

00:35:51.329 --> 00:35:56.699
these types of deployments. If chat GBT floods

00:35:56.699 --> 00:36:03.159
the zone with erudite, you know, text and speech

00:36:03.159 --> 00:36:08.800
and then sort of turns journalism and communication

00:36:08.800 --> 00:36:13.460
and sales into an undifferentiated mush of noise.

00:36:13.880 --> 00:36:15.719
Well, then society is going to disintegrate because

00:36:15.719 --> 00:36:18.300
nobody's going to actually know anything or be

00:36:18.300 --> 00:36:22.780
able to trust anyone anymore. So it's only by

00:36:22.780 --> 00:36:26.429
creating systems that remain stable after they

00:36:26.429 --> 00:36:30.469
exist that we can actually exploit these things

00:36:30.469 --> 00:36:33.409
on a durable basis. And so that's the challenge.

00:36:33.989 --> 00:36:36.349
Wow, Noah. Thank you so much for your time today.

00:36:36.570 --> 00:36:38.750
I feel bad we went over, but I really appreciate

00:36:38.750 --> 00:36:41.289
it. Talking to you is incredible. And I'll put

00:36:41.289 --> 00:36:44.570
all the, can you send me an email with all of

00:36:44.570 --> 00:36:46.570
your information so I can put it in the show

00:36:46.570 --> 00:36:49.309
notes? Really appreciate your time. And even

00:36:49.309 --> 00:36:51.989
though a lot of it did go over my head, wow,

00:36:52.050 --> 00:36:57.079
you are super intelligent. I learned a lot too.

00:36:57.099 --> 00:36:59.699
So thank you so much. Thanks for having me here.

00:36:59.699 --> 00:37:01.980
This was great. I appreciate it. I really appreciate

00:37:01.980 --> 00:37:04.340
your being bold and brave enough to come on with

00:37:04.340 --> 00:37:07.019
no prep needed. Thank you. Have a great rest

00:37:07.019 --> 00:37:09.480
of the day. Where are you located? Charlottesville,

00:37:09.679 --> 00:37:12.039
Virginia. Oh, beautiful. Okay. I'm in Daytona

00:37:12.039 --> 00:37:14.539
Beach, Florida. Awesome. All right, everyone.

00:37:14.619 --> 00:37:16.860
Thank you to Noah, my new friend. I don't know

00:37:16.860 --> 00:37:18.380
if he'd want me as a friend. I'm not that smart

00:37:18.380 --> 00:37:21.900
though. Okay. Thank you, Noah. Have a great rest

00:37:21.900 --> 00:37:26.760
of the day. Thank you, Noah.