Latest / The Inspired Stories Podcast / Andy Tang's Blueprint for AI Agents, Blockchain, and the Future of Micro-Transactions

Transcript

Andy Tang's Blueprint for AI Agents, Blockchain, and the Future of Micro-Transactions · The Inspired Stories Podcast

0:04
Welcome to another edition of the Inspired Stories podcast,
0:08
where leaders share their experiences so we can learn from
0:11
their successes and be inspired by how they've overcome
0:14
adversity. My name is Anthony Kotaspodi and
0:17
today's guest is Andy Tang. Andy is a partner at Draper
0:21
Associates, a leading early stage venture capital firm based
0:25
in Silicon Valley. Founded in 1985, Draper
0:29
Associates invest in entrepreneurs who are ready to
0:32
reshape entire industries with old technology solutions.
0:37
Andy has extensive experience in dual use tech, AI, biotech, Web
0:42
3 and defense tech. He has helped build over 15
0:46
Unicorn startups including Aqua and Novel.
0:50
Beam and Co founded Draper Dragon, a significant early
0:54
stage investment fund. His strong grasp of emerging
0:58
technologies and deep experience in venture capital have made him
1:02
a trusted advisor to many successful founders.
1:05
Before joining Draper Associates, Andy served on the
1:09
board of multiple tech companies refining his leadership skills
1:13
and industry knowledge and he currently directs VCX and invite
1:18
only Alliance of family Offices which promotes generational
1:22
relationship building, education and Co investing amongst
1:26
members. Now before we get into all that
1:28
good stuff, today's episode is brought to you by my company Add
1:31
Back Benefits Agency, where we offer very specific and unique
1:35
employee benefits that are both great for your team and fiscally
1:39
optimized for your bottom line. One recent client was able to
1:43
add over $900 per employee per year in extra cash flow by
1:48
implementing one of our innovative programs.
1:50
Results vary for each company and some organizations may not
1:53
be eligible. To find out if your company
1:55
qualifies, contact us today at AD back benefits.com.
1:59
All right, back to our guest today, partner at Draper
2:02
Associates, Andy Tang. I appreciate you making the time
2:05
to Share your story today. Thanks for having me so Andy
2:09
you. Received some notable degrees in
2:12
your education? Master's from in Engineering
2:14
from MIT and MBA from Wharton. And as I look at the early part
2:19
of your career, it looks like you were focused more on
2:23
technical roles like at Motorola and Intel.
2:26
I'm curious, at what point did you start to feel the pull to
2:30
get involved in the investment side of the world?
2:34
I think I've always had interest in the business side of the
2:39
world. You know, as an engineer, most
2:43
of us want to create things and many of us are all, I've always
2:49
been curious when you make these things, how do they make an
2:53
impact in the real world? So when I was working at Intel,
2:57
I worked on the Pentium processor first generation and
3:01
that's when I got exposed more to, you know, the marketing
3:03
sales side, competitive landscape.
3:08
And that was really in the, the height of the intranet
3:11
revolution. So I think that's when I got
3:14
really exposed to computing and how that affects Internet and
3:20
the real real world application. And for those who don't
3:25
remember, the Pentium chip was a really big deal at the time.
3:29
Why was it considered such a breakthrough?
3:32
I think it was really the first time you have a computer chip
3:36
that enabled many customers such as Compaq, IBM, Dell, the world
3:45
to make a affordable computer for everyday consumers.
3:49
I remember even my parents and ourselves were able to afford a
3:53
computer when Intel first came out with that Pentium processor,
3:58
right? And at the same time, the power
4:02
you could deliver from a Pentium processor was just more than
4:06
anything you could imagine. You could watch video.
4:08
It was a big deal. You know, you can run
4:10
spreadsheet, you could surf the web.
4:13
So it was really that first price performance the tech
4:21
industry was able to deliver for the everyday person.
4:27
So I want to spend most of our conversation talking about the
4:30
work that you're doing now with Draper, but maybe let's pick one
4:35
career stop along the way in between Intel and Draper.
4:39
That was a really formidable experience for you, picked up
4:44
some really necessary lessons and skills.
4:49
Come again? So what time frame or any time
4:51
frame? Any time frame, but I was we
4:55
talked a little bit about Intel, so something between Intel and
4:58
Draper. What was a career stop along the
5:00
way that you know had a particularly big impact on?
5:03
You. Yeah.
5:04
So between Intel and Draper, I worked on Wall Street.
5:10
I worked for Credit Suisse First Boston and I worked for two of
5:16
the most, I would say interesting bankers of our time,
5:21
you know, during that era, Frank Petrone and George Boutros.
5:26
So they were the tech banker of our of our times.
5:30
That was the very first time I think Wall Street really turned
5:36
to Silicon Valley and said, my gosh, these are great partners
5:39
of ours, could help them raise money, access the capital
5:42
market. This was in the late 90s, early
5:46
2000s and I remember the biggest challenge for us was being a
5:52
West Coast office and and trying to get our clients.
5:57
These are our start up founders, get them access to Wall Street
6:00
capital and then access we did give them.
6:05
And then it was sort of the pendulum swan the other way.
6:08
We got them too much access, raised too much capital, and
6:11
then the rest was historyyourememberthe.com bubble
6:15
melted and that was really tough.
6:19
It was really tough because that whole industry, you know, got
6:22
transformed the the investment banking industry.
6:25
We went from the happy IPO M&A days to just man trying to
6:30
triage and shutting down companies.
6:34
But the reality is that I look back at that time, it really was
6:39
that classic garner, you know, hype curve.
6:43
Any good idea you're going to have investors getting ahead of
6:47
the self, right? And companies didn't have enough
6:49
time to grow into the valuation and over time there's adjustment
6:54
required. But two things always happen.
6:57
Two things. One, investors always
7:01
underestimate how long it takes for adoption to happen for new
7:05
technology, whether it's Internet, whether it's fintech,
7:09
green tech, crypto, dual use, robotics, AI, in all cases, we
7:17
love it and we think it's going to happen right away.
7:20
But no, it doesn't happen right away, right?
7:23
So, and, and I'm not telling anything new, right?
7:25
So everybody knows that, Ah, we, you know, we, we underestimate
7:29
how long it takes for, for consumers to adopt.
7:32
The second thing will probably surprise you.
7:34
The second thing is, is that most of us, VCs included, we
7:42
underestimate the impact of these technologies, right?
7:47
Would you ever imagine? That does surprise me, right?
7:51
No, because like, like we're in the middle of the AI hype cycle
7:54
right now. It seems like it's being
7:55
overinflated. So I'm really curious to hear
7:58
your perception on this. So.
7:59
AI, why don't we talk about something that already happened
8:02
you could identify to right. So in 1999 and you and I are
8:07
kind of and you're younger than me, but I, I bet you remember
8:10
what happened in 1999. People came out and said, hey,
8:13
the first time we saw intranet awesome.
8:15
We're gonna we're going to have e-commerce.
8:17
We're going to have video on demand.
8:19
But would you ever imagine in 2025, the level of e-commerce
8:26
penetration, the level of video on demand for Netflix streaming
8:32
video? You could do it on your phone,
8:35
right? You could do it anytime and the
8:38
idea of no buffering. It just happens real time.
8:41
You can watch anything anywhere with one bar even sometimes and
8:47
that I think that is something that we didn't project back in
8:51
the days 1999, but we also didn't think it's going to take
8:55
25 years, right. In fact, some of those fiber
9:00
optics and optical component companies that we took public,
9:06
the amount of infrastructure we build in 2000, it's still
9:10
probably abundant even for application today, right?
9:15
And and that's a fact. All these fiber lines we lit up,
9:18
you know, we have these what we call the DWDM systems, which
9:22
allows you to essentially send different colors or laser across
9:26
fiber lines. So you could, you could have
9:29
multiple channels of communication and increasing the
9:32
bandwidth, right. Those things were, it was
9:35
splitting atoms back in the days.
9:37
We're finally using, I don't know, a couple channels maybe
9:41
right. But I think we were able.
9:42
To so multiple channels on basically the same pipe.
9:45
It wasn't possible before, but now because we're sending
9:49
different wavelengths, different colors.
9:52
And now? You can send them and they're
9:54
not interfering with each. Other correct that was available
9:56
25 years ago, but 25 years ago we had other problems.
10:00
People weren't streaming Netflix on their phones, right?
10:03
But today we are. And finally those technologies
10:06
became useful, right? So it goes to say a lot of these
10:10
technologies were all have always been ahead of themselves,
10:13
right? The applications took a little
10:15
bit longer. In other words, scientists and
10:17
engineers are really good at inventing stuff, but the rest of
10:20
us consumers take take some time.
10:23
And then the idea of streaming video on your phone, you know,
10:27
it, it's it takes some getting used to.
10:29
I mean, these days, Yeah. You know, our, our colleagues
10:33
are Gen. Z, Gen. alpha.
10:36
My kids are Gen. alpha, right? They, they take everything sort
10:39
of for granted that this is just the way it is.
10:42
But no, it has not always been this way, right?
10:45
So, and you talked about AI. So some of the things that I
10:48
think that that could happen, people are not thinking about.
10:51
It's interesting to explore as an early stage venture
10:54
capitalist, right? Because that's essentially my
10:56
job is to, I call it predicting the future, right?
11:00
But I'm cheating a little bit. It's so much easier to predict
11:03
the future. I just can't.
11:04
It's hard to predict the time in which the future is going to
11:07
happen. Interesting.
11:10
OK. So what do you think the future
11:11
of AI holds? What?
11:13
What are you seeing in your crystal ball that maybe the
11:15
average person isn't? I think all of us would work
11:20
differently. It's the, you know, industrial
11:24
revolution, you know, 2 point O. Right.
11:28
I think it's the probably the most transformative technology
11:34
known to mankind since industrial revolution, certainly
11:41
since the intranet. You could argue what's more
11:45
impactful to humanity, intranet or AI, but that's not important
11:49
because they're both pretty impactful.
11:52
But I, I, in my opinion, AI is probably more impactful because
11:56
I think it fundamentally changes how we human in the society view
12:05
work and how we think about production, how we think about a
12:17
fulfilling life, how we think about career and how we think
12:24
about our relationship with each other.
12:28
And to a certain extent you could even think about sort of
12:31
the social contract we have with our government, right, in terms
12:38
of labour and working with each other, because a lot of that is
12:41
going to be done by non human. You know, I think robots and AI
12:49
agents, they're going to take over white color and blue color,
12:54
you know, basic blocking and tackling jobs.
12:58
So you mentioned, you know, you've got kids who are Gen.
13:01
alpha. My kids fit in the same bracket,
13:03
They're 9 and 10 years old respectively.
13:06
As you think about sort of how you want to direct them in their
13:11
educational pursuits and things that they may want to think
13:14
about in terms of career options, how do you wrap your
13:17
head around that? Oh my gosh, this is probably one
13:20
of the most favorite questions people ask when I give
13:24
interviews from our limited partners, our investors to just
13:29
like talking to students and other entrepreneurs.
13:34
That is the Holy Grail. I think they are in a inflection
13:39
point. I think the generic advice that
13:43
were given last 20 years going to computer science, right, that
13:48
has basically took a 180 turn. Google, I think recently just
13:55
had a big layoff, a bunch of, you know, and they're engineers.
14:00
And I think that trend is this is just the beginning.
14:03
I think it's it's happening because we suddenly figured out
14:09
writing code could be done by AI.
14:12
Right. So I think that the direction I
14:15
give to my kids have always been pretty consistent and that is 2
14:19
main things. One is you got to be, you got to
14:22
raise them so that they have the ability to pick up new things.
14:26
Right. The ability to learn new things
14:30
is more important than knowing something and doing something
14:35
sort of a static field of expertise, right?
14:39
That ability to dynamically change course and become the the
14:44
expert in that space. And the second thing is really
14:48
that love for learning really important.
14:51
I think you you got to love to learn new things.
14:55
And that's not necessarily something you would teach.
14:57
I would say that's something you cultivate, right, Really
15:02
allowing kids to explore the things they want to learn.
15:06
Yes. You know, a lot of the liberal
15:08
arts education in the last few decades have been shunned
15:15
because of how successful the science and engineering majors
15:20
have been. And I would argue, I think that
15:25
the the table may be turning, but not turning back in the way
15:31
that all just don't go into science and engineering.
15:34
No, no, I'm just saying that think, think there's going to be
15:36
a bit more balanced because you might be able to just pick up
15:40
science and engineering on the fly, or you could assign that
15:47
task to an AI agent or a bot, right?
15:52
But I think the basic education and love love learning is is
15:55
super important. So, you know, you mentioned sort
15:59
of like big picture, like, you know, how does our relationship
16:03
with government, how does our relationship with sort of work
16:06
change, You know, robotics and AI taking over both blue collar
16:10
and white collar jobs at some point down the road?
16:13
Like if you look into your crystal ball, Andy, 20 years
16:16
out, what, what do you think this looks like?
16:18
Yeah, I think 20 years out, if done correctly and you'll have
16:24
different experiments. We're going to have how many
16:27
countries are in the United Nation, call it 200.
16:29
We're going to have 200 social experiments going on, right on
16:33
how governments organize their people to get ready for this
16:38
transformation. AI is just like the industrial
16:43
revolution can put it back in the box.
16:45
Right. The toothpaste is out.
16:48
It doesn't go back. Some people may try to deny it,
16:53
others may embrace it. And the people who embrace it
16:57
may have different ways to go about adopting it.
16:59
How you deal with the people, I think that is it's going to be
17:05
fascinating to see, right? And I, as an American, I really
17:08
hope our leaders do the right thing and kind of manage that
17:12
transition. And I'm not a, I'm not in policy
17:16
making. So I actually don't know how you
17:17
would do it. I just think that 20 years later
17:20
you will see clearly some countries would do a better job
17:23
than others. And I also think one thing is a
17:28
given and that is embracing it is definitely the way to go
17:33
versus trying to fight it. Because Can you imagine if you
17:35
try to fight the internal combustion engine, right?
17:40
Or you try to fight the steam engine and it just doesn't make
17:44
sense, right. If you embrace it, you could
17:48
probably figure out a way to emerge as a winner out of this.
17:55
And I'm happy to also go into the both the blue color and
17:58
white color work kind of what what I see as the the new yes.
18:03
Please, let's talk about that because I, I, I hear more people
18:06
saying, hey, rather than getting into computer programming, get
18:11
your your kids to learn a trade. You know, electrician, plumber,
18:14
like robots aren't going to replace that anytime soon.
18:17
Or do you feel differently about?
18:18
That right, right. I, I probably wouldn't go into
18:20
that level of granularity, but I'll, I'll share with you
18:24
because I actually don't know. I mean, I wish I knew, right?
18:28
I, I slightly worry if people start thinking about in that
18:31
kind of granularity because the, the key is to maintain
18:37
optionality. I mean, you and I spoke a little
18:39
bit about football before the show.
18:42
You know, we're both avid NCAA fans.
18:46
I think the best play is RPO, right?
18:48
Run pass option. You want to maintain your
18:51
maintain your optionality, right?
18:54
Because you don't know, you don't know what the defense is
18:56
going to show you. So you want to be ready, right?
18:59
So I, I won't say, hey, if it's this or that, but I will say
19:05
first let's look at the blue collar work, right?
19:07
The traditional sort of very hands on type type of jobs that
19:12
that appears to be pretty safe. And I agree it's safer than my
19:16
job, right? So the classic thinking is
19:21
robots, right, taking over a lot of the jobs.
19:25
And I I don't disagree, right. And I think we're we're
19:28
basically there, right? And I think it's all depends on
19:31
which industry verticals, at what level they come in, right?
19:35
So that that that sector or that side of the house is probably
19:39
more predictable, which is a lot of the repeatable tasks should
19:46
get automated. And I don't necessarily think
19:50
we're thinking about just bots walking around.
19:52
It could be as simple as a hamburger flipper that doesn't
19:55
look like a bot, but that that takes away that burger flipper
20:00
job you see. If you talk about like
20:05
electrician or plumbing, I think there will be similar type of
20:08
automation that happens that reduces the amount of plumbing
20:14
and electrician types of jobs. But not to worry because I think
20:19
the job will be different and in fact I think the job
20:22
transformation would happen in every industry from plumbers,
20:29
electricians to doctors. So it may be, you know, we still
20:34
need a plumber, but now the plumber has more sophisticated
20:37
tools that they can use to do the job quicker, faster, better.
20:40
Much better put right, so they all have they all have analytics
20:46
that tools that allow them to do things remotely by the time they
20:49
get there, it's a 5 minute job instead of the, you know, 2 hour
20:54
job they need to build right. So his plumber may become a
20:59
quasi blue white collar job. Right.
21:04
And it's, it's going to, and I think it's a, it's actually
21:08
makes a lot of sense. It's, it's, it's similar to
21:11
investment bankers before Excel, right?
21:14
We used to do work using, you know, I don't know, calculators
21:18
and whatever, right? And if you could do your job,
21:20
it's just the moment you have Excel, you could be, you become
21:23
a lot more efficient, right? You're now you have Zoom, you
21:27
have PowerPoint, it's just that side of the house, right?
21:30
The the more physical work haven't been as well invested
21:36
because of the barrier to the AI, barrier to entry for AI, but
21:41
now I think it's getting there. So, and, and the, the
21:46
spreadsheet example is a really good one, right?
21:48
Spreadsheets didn't do away with the need for accountants for
21:52
investment bankers for it. It gave them a tool to be able
21:55
to do their jobs more effectively and more
21:58
efficiently. Do you think the same thing is
22:01
going to happen with AI and robotics tech or is this
22:07
different and there's going to be a deficiency of jobs now
22:12
going forward? So my prediction is this, you
22:15
use a 20 year timeline, which I appreciate because I don't know
22:19
how long it's going to take. Let's just say that the the end
22:23
steady state or that sort of the, there's probably going to
22:27
be some fluctuations. In some countries it will be
22:30
good, some countries not not so good.
22:33
But the end steady state, what I expect is a human race.
22:38
All of us work less, right? We will work fewer hours, just
22:43
like my grandparents, you know, my grandparents, they, you know,
22:47
they grew up in China. They were farming, right?
22:49
Or they have farmers working for them.
22:52
I don't think they have this idea of eight to five.
22:54
It's the, you know, this dawn to dusk.
22:57
I don't think they take Sundays off, right?
23:00
It's seven days a week. You just work, work, work, work.
23:02
And you can barely feed yourself, right?
23:06
So my quality of life is better than my parents, and my parents
23:10
better than their parents. I think our kids certainly will
23:14
work fewer hours. Right.
23:17
But I think those hours they work, it's going to be a lot
23:21
more impactful, right? So for instance, we can talk
23:24
about doctors. It's in the, in the US, it's
23:27
notorious. We complain about our doctors
23:30
don't spend enough time with us and doctors complain they work
23:35
too many hours. So what gives, right?
23:38
I think what gives is, you know, I made an investment almost
23:42
seven or eight years ago. Unfortunately, it actually went
23:45
under, right? This.
23:46
So this is a SAS story of investing in the right
23:49
technology too early and they didn't make it.
23:52
So it's AAI doctor. This was in 2016 and they
23:56
essentially build ALON for for doctors, right?
24:02
So it could tell you, you could say, hey, Anthony, you know, AI
24:07
doctor, I'm this, I'm mad I have a headache.
24:11
And the doctor would say, oh, I remember 7 years ago you had a
24:14
headache too. But don't worry, that was
24:16
unrelated. Or if I had, you know, recurring
24:20
headache, you'll remember 7 years ago, three years ago last
24:23
month. You should probably get a get a
24:25
brain scan. Right.
24:27
So you could do all of that and today that's very easily done.
24:32
But what's important is they figured out it's not about.
24:37
So this AI doctor used to score higher than the human doctor
24:41
even seven years ago. They would go take the US, you
24:45
know, medical exam for doctors and it would perform better.
24:47
And of course, because this AI doctor had, you know, wealth of
24:51
information on the Internet and remembers everything that needs
24:53
to be remembered from medical school.
24:56
Human doctors can't do that. So, but what they what this
25:01
company discovered was if you just make this a tool to let the
25:05
doctor use it, the two combined score even better than the
25:10
machine or the human, right. So it really is a win win.
25:15
So I think what's going to happen is now with the advent of
25:20
all the large language model and the small language model, the
25:24
specific models for doctors, doctors will become much more
25:30
productive. Human patients would trust that
25:34
AI doctors have been more for little things.
25:36
I don't need to bother my doctor.
25:38
My doctor could focus on the big things.
25:40
My doctor would become much more accurate in his or her
25:44
diagnosis, right? So really the doctor's job is to
25:49
handle that 5% cases where it may have a difficult delivery to
25:59
the patient and say, hey, you got this illness.
26:03
Yes, it sounds very bad, but we have these options versus AAI
26:09
doctor that may sound like a human, but you know, it's not
26:13
right. So I think at least for me as a
26:14
patient. You still want the bedside
26:18
manner of the human doctor. Absolutely.
26:21
Where, where are we in sort of the timeline?
26:23
Are doctors using AI now in their diagnosis?
26:27
Not or are there not? Yet why?
26:29
What? What's standing in the way of?
26:31
That regulation liability, it's all the non engineering and
26:35
science stuff. The tech is there, it's just the
26:39
human beings aren't comfortable with it from a regulatory or
26:43
from a a like a risk standpoint. Right, right.
26:46
I mean, you can't even do a Zoom call with your doctor unless
26:50
they're licensed in your state right there.
26:53
There's some weird, there's some HIPAA rules about, you know,
26:58
data privacy and compliance. Like I can't.
27:01
So for instance, I have an oncologist I talked to in Texas,
27:04
but I don't think he could just Zoom with me.
27:06
You have to go to Texas, go see him, right?
27:08
So you need that kind of there. There's some patient protections
27:14
in the way of the technology. *** it's not, I'm not saying
27:18
it's all bad. I'm sure the regulations were
27:20
there for a reason. I'm not a policy maker.
27:23
I just think that the tech is ready, the patients are ready.
27:29
So once the regulator open up the floodgate to make sure we're
27:33
fully protected, I think the adoption is, you know, is going
27:36
to happen, you know, and the world will be a better place.
27:41
What's something outside of sort of the AI and robotics realm
27:46
that you're excited about that maybe isn't getting a lot of
27:49
attention? That isn't getting a lot of
27:51
attention. Most yeah.
27:54
So I'll, I'll share with you some of the investment thesis
27:58
I'm, I'm looking at now, since I'm an engineer, I typically
28:03
don't, as opposed to a scientist, I don't take
28:06
scientific risks. I take engineering risks and
28:11
commercialization risks. So what that means is that I
28:15
don't invest in things that require scientific breakthrough.
28:19
So often time my investment thesis involves taking
28:25
interesting things and put them together.
28:28
So in engineering, I call it integration, right?
28:31
So right now I'm looking at integrating AI, robotics and
28:39
blockchain technologies, right? So the idea is this, the idea is
28:45
that, you know, you, you know, we can make these really
28:50
interesting smart AI agents that could probably do a lot of
28:54
things for you already, right? You know, I could ask a lot of
29:00
the questions I will ask are investment analysts.
29:03
I could just ask the AI agent, what's the market size of this?
29:06
What's the market size of that? What's the competitive landscape
29:08
this, What's the competitive landscape of that?
29:11
What's the margin structure? What's the trend?
29:13
So the AI agents can do all of that.
29:17
And on the blockchain side, you know, you have these very
29:24
efficient micro payment technologies that are already in
29:27
place that could help you transact in the, you know, 5
29:31
cents, $0.50 without any trouble, right?
29:35
So the idea is maybe you could combine these AI agents and have
29:44
them transact with each other, right?
29:46
These AI agents would represent you or me, right?
29:51
And they could just go through financial transactions based on
29:55
their need. You and I as humans, we don't
29:57
need to get in, get in the middle of it.
29:59
For example, I really like your podcast, right?
30:05
But for you to set up a subscription, that's just too
30:08
much, you know, overhead, right? Why not set up some sort of
30:15
interesting structure where you can just take 5 cents $0.10 a
30:20
dollar every time somebody listen to a podcast?
30:24
Right. And so it would somehow be
30:26
hooked into Spotify or Apple Podcasts and so on my website.
30:30
And every time, yeah, you listen to an hour of it, I get paid a
30:33
nickel. Or something and then and then
30:34
the pricing could be different could be just based on like I
30:37
might have an AI agent I have I give the agent a budget, right?
30:41
The budget may be 50 bucks a year.
30:44
The agent is going to know over time.
30:46
Andy really likes business podcast, right?
30:49
Loves macroeconomics, political podcast, but it's not as much
30:54
interested in entertainment as a kind of boring guy.
30:56
It was just like sports and sports and business.
30:59
That's my thing, right? So the AI agent figures out over
31:03
time, right? Andy's would be happy to pay
31:06
$0.05 per hour, right? And then maybe over time it
31:12
creeps up to six cents per hour of of the podcast I listen to.
31:16
And that just happens in the background without me knowing,
31:19
right? And if I keep listening to more,
31:21
the agent's going to say, oh, I bet he really support this
31:25
initiative, right? He loves this.
31:28
You know, I happened to also be a animal welfare advocate.
31:32
So this animal podcast, he really is probably underfunded
31:37
and he will probably play a dollar an hour because there are
31:41
no other supporters. He really wants to step up.
31:43
So those are the things I think is suitable for AI sort of
31:49
integration, AI plus blockchain. Right.
31:53
And then I'll give you another one.
31:55
I used to work for Abd, one of the biggest power generators in
32:00
the world. The challenge has always been,
32:04
how do we generate power where it's consumed?
32:09
And you say, well, that's easy, we got solar panels.
32:11
But the problem is when you generate solar power from a
32:14
solar panel, that power logically goes back to the grid,
32:19
right? You sell it back to the grid and
32:21
the grid sell it to your neighbor, right?
32:25
Because the grid is this thing that has a, a steam power
32:29
generation. It, it literally boils water,
32:32
you know, a couple 100 miles away, right?
32:35
You, you, you, you fire some gas turbine, boils water and that
32:40
turns into a, a, a steam turbine and generates mechanical power,
32:45
converts that to electricity right there.
32:48
And then you lose about 30 to 40% of your, you know, energy,
32:53
right? And then you, and then, I mean,
32:55
it's not, that's just the beginning.
32:58
By the time it gets to a transmission distribution to
33:00
your house, you lose another 30%, right?
33:04
So essentially you're down to, I don't know, 30% of what you
33:07
started with. So really the goal is you and
33:11
your neighbor start trading energy.
33:14
Right. You won't.
33:16
Be able to send it from one of us directly to the other, Yeah.
33:19
And then and then and then so, so the, the challenge is not in
33:25
the physical connection, but it's in the, it's in the
33:28
finance, right? How do I trade with my neighbor?
33:33
That's where AI agent steps in again, right?
33:35
So you have essentially the AI agent talk to each other and
33:39
say, hey, you know, I want to sell my electricity.
33:42
My boss is not home. I got all this thing generated
33:46
from my solar panel, but you guys seem to be home kids
33:49
playing computer games and, you know, dad cooking, right?
33:55
So why don't I send over some electricity to you?
33:57
How much do you want to buy? The agent can look at the spot
34:00
price and figure it out. So conceptually, this is an idea
34:04
that you're really excited about.
34:06
Are you invested in tech that is working on this?
34:09
Well, so I, I gave a couple of examples of what could we have
34:14
investments and I would probably have one investment in the, the
34:20
decentralized solar trading. But I think it will happen more.
34:25
And I tell you a lot of these things would happen in this
34:27
cycle with the current administration.
34:29
They really lifted the blockchain technology trade in
34:33
the US, so you're probably going to see more of that.
34:37
And you think unlocking blockchain access is important
34:42
to these microtransactions that you're talking about?
34:44
You don't think it can be done sort of on the current
34:47
infrastructure of the the existing payment system?
34:49
It's too expensive, right? And then it really, I think, the
34:55
beauty. Sorry, sorry, Andy, I, I want to
34:56
come back to that point for a second.
34:58
It's too expensive because of the current players who control
35:03
how money can move around Visa, MasterCard, you know, the Swift
35:08
system, ACH, etcetera. And so with blockchain, now you
35:12
wipe out or greatly reduce those costs.
35:15
Now you've got a cheaper infrastructure to do it.
35:17
Yeah. You enable peer-to-peer, right?
35:20
Really two things. One is just the overhead, the
35:23
legacy technology is just more expensive.
35:25
They, they, it's not like they're trying to, I don't think
35:28
they're gouging us, right. I just think that the technology
35:31
is more expensive. They need to charge you 3% for
35:34
them to make a profitable living as a public company.
35:38
And, and I respect that. So it's really just the next
35:42
generation of technologies doesn't take 3% and that is
35:46
really helpful for microtransactions and high
35:49
frequency transfers. That's one.
35:53
And the second thing is about the open network.
35:57
The idea of open network is really important because if you
36:01
have an open network that allows people to develop these agents,
36:04
that could trade, because I bet you're going to have 1000 agents
36:08
launched by people and 10 of them really work well.
36:12
And those are the 10 that survived.
36:14
And I don't know if solar trading is one of them.
36:16
That might be another sort of pinhead idea that doesn't work,
36:20
right. But the, the, the important
36:22
thing is you got to try these seemingly crazy ideas because if
36:27
if one or one or two of them worked, they could transform
36:30
humanity, right? So the idea is blockchain and AI
36:33
agent, those are the transformative technologies.
36:36
They happen to become available around this time.
36:39
So when you start combining them, you see some interesting
36:42
things and then you throw robotics into the mix, right?
36:46
So you can suddenly have you, you don't know what you're going
36:49
to get when you have these enabling technologies becoming
36:52
available for the first time. Let's take a step back for a
36:57
second, Andy. And I mean, I want to talk a
36:59
little bit more about sort of the big picture of Draper.
37:02
There's Draper University, there's Draper Dragon, Draper
37:04
Associates. What's what's kind of the bigger
37:07
vision that you and Draper are driving towards?
37:11
So just as an introduction, so, so Tim Draper is the founder of
37:17
these entities. So he has been in business.
37:20
He's a third generation venture capitalist who's came from, you
37:24
know, his grandfather, right? So so he started the Draper
37:28
platform 40 years ago. And over a time he had created
37:32
many pieces of these puzzles to serve entrepreneurs.
37:37
So for instance, so Draper Associates was the the oldest
37:42
entity he started, which is just a pure investment platform,
37:46
early stage venture capital. Over time, he added on pieces.
37:50
So he added on Draper University, which is a school
37:54
for entrepreneurs globally, right?
37:57
And Draper Dragon was one of the funds he and I created to do
38:02
international investments. And then there's also ATV show
38:05
called Meet the Drapers, where we have a Shark Tank style, the
38:11
TV show for entrepreneurs who come and pitch to the investors
38:15
on Meet the Drapers. And the difference from Shark
38:20
Tank is that these are actually venture backable ideas versus
38:24
product, right? So these are the key components.
38:29
We have Venture Fund and Draper Associates, that's our flagship
38:31
fund, early stage seed Series, A venture fund.
38:37
And then we have Draper University which provides
38:39
entrepreneur education for young people, call it 20 to 30
38:44
something year old, first time or early entrepreneurs who just
38:48
want to learn about that Silicon Valley magic.
38:51
And we have been in business since 2012 and have over 5000
38:58
people who have gone through the program and drawn from 102
39:02
countries. We recently finished our 30th
39:07
cohort at Draper University and we have produced successfully
39:11
produced over 7 Unicorn companies from scratch.
39:16
Unicorn being defined as over a billion dollars in value.
39:19
Yeah, in technology, technology companies over a billion dollar
39:24
valuation and meet the Drapers. We're on our Season 8 and we
39:30
currently have 20 million viewers around the world.
39:33
Where can we find this show? It's on drapertv.com.
39:39
OK, onlinedrapertv.com I'd like to hear a little bit about VCX
39:46
this invite only alliance of family offices.
39:50
What was the inspiration for starting that and and what what
39:53
a members? What value do they get?
39:55
Yeah. So so this was started, I want
39:59
to say 2013 or 14. And the idea was, you know, we
40:05
have limited partners in our venture fund.
40:07
These are basically our investors, right?
40:10
And we noticed over time we have, we our investors include
40:14
pension funds, foundation, sovereign wealth funds.
40:19
And then we noticed this one category family office has been
40:22
growing very fast. Then I realized what's happening
40:26
was we noticed a lot of the generational transition from
40:33
sort of family score business, like grandfather, father made
40:37
money in paper, pulp, industrial, real estate,
40:40
whatever. And then second generation took
40:43
over, ran the core business. The third generation took over
40:45
and they want to do something different.
40:47
And often time, I would say probably 100% of the time, they
40:51
want to get into tech investing. So what we did was we first
40:56
started this really, you know what I call it technology
41:01
retreat. We can program just to show our
41:05
family office partners how we think about investing in
41:10
technology. And it's a program it started
41:14
over, you know, sort of three day duration.
41:18
We have now successfully condensed down to one day, right
41:21
And that one day we talked about you know macroeconomics, we
41:26
talked about different asset classes, portfolio construction
41:30
for family office and then specifically we hone in on
41:34
venture as an asset class. And then we talked about our
41:37
best practice in making investments.
41:40
And then we talked about the how to select entrepreneurs and
41:46
CEOs, and then we actually put them through an actual series of
41:51
pitches. When we interview entrepreneurs,
41:53
they get to see how we kind of make our decisions.
41:57
So it's a technology retreat is a way to just bring more
42:01
awareness into family offices, how to make technology
42:05
investments with the emphasis on venture capital.
42:09
So what are you looking for in an entrepreneur and a founder?
42:13
What What are some of the tell tale signs that this is a guy or
42:17
a gal we want to be involved with?
42:20
Yeah. So we really try to narrow down
42:24
to 2 main things, right? And these are two main things I
42:30
think you'll find in every company just to keep it very
42:35
high level. That is, we look for
42:39
opportunities in the very large market, very large market,
42:44
right? Venture by nature is, is highly
42:49
risky. So the success rate sometimes is
42:53
not so high, right. So the idea you want to is if
42:57
you are able to make a successful venture, you want to
43:00
make sure it really counts, right.
43:02
So often time the size of the exit is directly driven by the
43:07
size of the opportunity. So that's one thing essentially
43:12
looking for a big market or big potential market and we can kind
43:16
of dig in some other characteristics.
43:18
But I want to kind of leave, leave that and then go to the
43:21
other side of the equation. The other side of the equation
43:23
is I could change business plans, but I can't change
43:28
founders or people. So the team is super important.
43:32
We want to find, ideally a group of people.
43:36
Right. Who are passionate about the
43:39
mission, but are also experts in their space and these people
43:46
could essentially take you into that market they're going after.
43:50
So that's sort of the very simple formula for early stage
43:55
investments. Is that primarily what you do
43:58
are early stage investments or are you also getting involved
44:01
with companies that have kind of proven themselves a bit more and
44:04
they're just looking for some growth capital?
44:06
Good question. So we do growth this as well,
44:08
especially growth capital for our own companies.
44:12
So, so I'll just describe this in the slightly nerdy but
44:18
technically accurate language for both early stage and late
44:22
stage criteria. So in statistics, there's
44:25
something called the false positive and false negative
44:27
error, right? So early stage, I want to make
44:31
sure I minimize false negative error translation.
44:37
I want to make sure if it's a good company, I don't
44:40
inadvertently make the mistake of turning it down.
44:43
That's called a false negative, right?
44:45
You could probably sort of realize why that is because if
44:50
it's the next Robin Hood or Coinbase, you would hate to turn
44:54
it down at the seat stage, right?
44:57
So you so you want to minimize false negative at the late
45:01
stage? I want to minimize false
45:03
positive. And the reason is I'm typically
45:06
writing a larger check. So for me, I actually want to
45:08
make sure if I'm investing in, let's say the growth round of
45:13
Robin Hood Coinbase, it is priced properly.
45:17
It is indeed going to become the next deck of corn, right?
45:21
Because at that point, I'm kind of paying up, I'm paying a
45:23
higher evaluation. So I'd rather avoid making
45:26
mistakes. So the mistake there is to make
45:31
the wrong investment and the early stage, the mistake is to
45:35
not make the investment when you're supposed to be making
45:38
investments. Does that make sense?
45:39
So we looked at. The makes a lot of sense.
45:41
Sure you you don't want to miss out on the what could be the
45:44
next big thing, but they're harder to spot at that stage
45:46
because they're just kind of a kernel of an idea.
45:49
So maybe you're making more more bets at that stage, saying we
45:54
know that a smaller percentage of these are going to make it
45:56
through, but fear of missing out.
45:58
I I've been make sure like I'm, I'm involved in this.
46:01
If it does hit Yeah, right. And then those later stage
46:04
companies, you're writing a bigger check.
46:05
The valuation is not so much in your favor at that point.
46:08
So you want to be a little bit more cautious there.
46:10
Right. Absolutely perfect.
46:12
Perfectly put. Yeah.
46:14
No. You know somebody who's done and
46:18
seen as much as you have, Andy, obviously you've seen a lot of
46:20
successes. You've been involved in a lot of
46:22
unicorns. I'm sure that you've had your
46:24
fair share of challenges, whether they're personal or
46:27
professional. I'd be curious to hear about a
46:29
particular instance, a big challenge that you've had to
46:32
work through, and what you learn kind of going through that
46:35
process. Yeah, I I'll share personal
46:41
obstacle I went through and that it kind of changed the way I I
46:45
look at the world. So I was diagnosed with cancer
46:48
in 2011 and and unfortunately, you know, there there is no
46:56
evidence of disease now, but I think what that changed me the
47:00
the way. So I think how I've changed from
47:03
there was AI became a lot more interested in the, the, the
47:10
confluence of computer science and biology, right?
47:15
I got to really learn about America's healthcare system in
47:21
our Pharmaceutical industry and drug development and realize
47:26
they're, they're about to go through a transformative change
47:32
in those industries. And that's how I actually got
47:34
into the AI doctors. I looked at, you know, AI based
47:39
drug discovery companies, right? Because I, I feel like coming
47:44
from the electronics industry and a electrical engineer, you
47:49
know, one of my love for science is quantum mechanics, right?
47:52
I love the idea of being able to study atoms and electrons and
48:00
particles down to the molecular level, right?
48:04
And the ability to predict statistically where these atoms
48:10
are going to be, that's physics, that's engineering.
48:15
But when it comes to biology and medicine, people say it's an
48:20
art. I think it's more of a science
48:22
now, but I would like to see the art and the science become
48:27
engineering and then become manufacturing and becoming more
48:33
of a retail product, right? Have you ever had an iPhone
48:37
issue? You go back to the Apple Store
48:38
and you say, hey, how does this work?
48:40
How's and almost 100% of the time, right, they'll come up
48:47
with a reason and say, all right, all right, this is
48:48
because you need the new software upgrade.
48:50
Oh, because this phone is actually broken.
48:53
There's a component. And then I bet if they take that
48:56
phone back to their manufacturer, they'll say, Oh,
48:58
yeah, this component went through the diffusion chamber in
49:01
the fab. And then there was 1, you know,
49:04
gas valve didn't turn on, so it didn't provide the right doping
49:08
for the chip. So the electrons weren't moving
49:10
as fast. So you have a short circuit
49:12
here. That's why the application
49:13
doesn't work for the client. There is a deterministic
49:17
solution for a consumer, a retail problem, right?
49:22
On the healthcare side, if you go to your doctor and say,
49:24
doctor, my head hurts, right? First of all, I don't think the
49:27
doctor, doctor probably tell you, right?
49:29
Great, you know, stress, dehydration, rest, take some
49:34
Tylenol and let's observe, right?
49:37
You certainly can go back to the manufacturer, your mom and dad
49:39
and say hey mom, you know I want a refund your mom and that get
49:42
the hell out of here. So I think my goal is to move
49:50
our healthcare system closer to manufacturing and retail
49:56
experience, right? And away from this more of a art
50:02
and science, you know, and I think we're now probably closer
50:07
to engineering, but I wanted to be at the manufacturing and
50:11
retail experience. It's it's inexpense.
50:14
It's affordable, right? It's deterministic and it's not
50:17
scary, right? So it's not a.
50:20
How do you do that? How do you get there?
50:22
I think it's entrepreneurship and venture capital.
50:26
I, I, I really think it's, it's really.
50:28
So those are two ingredients, but what does some of the in
50:30
between stepping stones look like do you think?
50:32
Yeah, I, I think a lot of the technologies are, I wouldn't say
50:38
they're available. I think they're getting there,
50:40
right. So for instance, like the
50:41
diagnosis, right. I think for what I went through
50:45
as a cancer patient, that the amount of time it takes for
50:49
uncertainty, the tests, the invasive tests, a lot of that
50:54
should be could be shortened. You know, the, the amount of
50:57
time if you ever gone to the doctor, get an X-ray, it takes,
51:00
you know, at least 24 hours. And then you look at the
51:04
radiology report, you're like, what the heck is that?
51:07
By the way, every radiology report at the end of it, you'll
51:10
have the sentence may need attention, right?
51:14
That's both scary and unhelpful. You know, of course it needs
51:18
attention, right? So my radiology report was
51:22
ignored for three, two to three years because the doctor wasn't
51:27
paying attention. There was a tumor there.
51:28
The doctor basically just overlooked it.
51:32
And you know who discovered it? It's somebody who took an
51:35
economic interest in me. It's an insurance underwriter.
51:40
The insurance underwriter looked at it because I was applying for
51:44
life insurance. Most of us have no business in
51:48
go and read their medical I I assume you haven't pulled your
51:51
medical records and try to read it right?
51:54
I wouldn't even know how to do it, no.
51:55
Exactly right. So when I applied for life
51:58
insurance, the insurance agent said, hey, you know, fill out
52:00
this form and I did whatever, I didn't know what he was doing.
52:03
And then a few days later I came back with a 70 page report.
52:08
Now it's probably 700 pages. 70 pages were when I was quote UN
52:11
quote, healthy, a healthy 39 year old.
52:16
And so he said, well, what's this?
52:19
And I said, what do you mean? And I started reading as a
52:21
layman, I started reading. I was like, this needs
52:23
attention. This needs attention, right?
52:27
So, so I think there's, there are a lot of low hanging fruit.
52:30
There's a lot of things AI could just do and just alert you.
52:33
Our doctors are busy. And I and I went to Stanford
52:38
Medical Center. This is not like I Andy went
52:40
through some, you know, county clinic, right?
52:43
Andy went to Stanford medical for my primary care.
52:48
So it these things happen, you know, it's, it's not due to lack
52:51
of talent, right? It's due to the lack of.
52:54
We're human beings. We get distracted, we miss
52:56
things. Absolutely, absolutely right.
52:58
There's no finger pointing. It's just, you know, we all need
53:01
a little help. I want to create an AI agent
53:03
that could help me better become a better investor.
53:05
You know I sort. Of Andy, you, you talked about
53:09
how your cancer diagnosis generated this, this interest in
53:15
computer science and biology sort of coming together.
53:18
But I'm curious what impact it had on you sort of on a personal
53:22
side, like mentally, emotionally, like how did you
53:26
sort of deal with this heavy news?
53:28
Yeah. So I think the biggest challenge
53:30
is actually the, the, the trustworthy medical system.
53:40
That's probably the biggest challenge, right?
53:43
I suddenly realize our doctors are spread thin and they cannot
53:53
be fully, I, I cannot fully rely on them, right?
53:58
And again, it's not finger pointing, it's just a reality.
54:01
You know, they got, they got lives, they got kids, they got
54:04
busy stuff going on. I can't rely on them if if you
54:09
want. And which is a kind of
54:10
interesting conclusion to reach because I know if I tell you,
54:16
hey, you cannot rely on your, you know, professor for
54:21
learning. You cannot rely on your whatever
54:27
pick, you know, your plumber for this, for that you say, yeah, of
54:31
course I cannot. But if I say you cannot rely on
54:34
your doctor for your health, you'd be like, what?
54:36
What am I going to do? But that's the, that's the state
54:39
of American healthcare. You have to take ownership,
54:43
right? And most of us just don't have
54:46
the time, the resources and, and even know how to do it.
54:51
And that's why I think using AI is a way to go.
54:54
And so I, I, I personally dealt with it just over time, right,
54:59
that kind of. Any support systems?
55:02
Yeah, I don't know outlets that you found that were helpful.
55:05
I, I said concierge medicine is a stop gap, right?
55:09
Essentially you got to pay it. I mean, I hate to say it, but
55:13
you know, it's just, that's our system.
55:16
If you, if you don't pay, you get the minimum sort of level of
55:21
service. And that to me that is not good
55:23
enough for anybody. I don't care how rich or poor
55:26
you are. I just think that it's not
55:28
adequate because for what happened to me, right, I, I
55:31
don't think it's good enough. I think you got it.
55:34
You got to either pay or ramp up AI.
55:37
So everybody's got someone sort of doing the risk management for
55:42
their healthcare. Andy, how would you characterize
55:47
your superpower? I think my superpower is super
55:52
transparent, direct and analytical.
55:57
That sums it up pretty neatly. I just got one more question
56:01
then for you, Andy. But before I ask it, excuse me,
56:04
I want to do 2 things. I want to invite anybody
56:06
listening today to go ahead and hit the follow button on your
56:08
favorite podcast app. We've had a really great
56:10
interview here today with Andy Tang from Draper Associates.
56:14
And I want you to continue to get more wonderful content like
56:17
this, Andy. I also want people know the best
56:20
way I had to get get in touch with you directly or to continue
56:23
to follow your story. What would that be?
56:25
LinkedIn. I'm very active on LinkedIn, so
56:28
a lot of thoughts that I share with you here either have been
56:32
published on LinkedIn or will continue to.
56:34
I'll continue to share. I found that to be a great
56:37
community of people and it's a it's a decent publishing
56:41
platform for thoughts. And we'll include a link to that
56:46
in the show notes here on our site, folks.
56:48
But if you're just listening, you can look up Andy Tang and
56:52
it's a Draper Associates on LinkedIn.
56:54
You, you'll be able to find him pretty easily.
56:56
So last question for you, Andy, you and I have had a nice
56:58
conversation here today. I hope that we keep in touch.
57:01
Let's say in a year from now we reconnect and you're excited and
57:05
you're celebrating something. What's that thing?
57:10
You said in a year, right? I am hoping I could play better
57:18
tennis with my son. I have an 11 year old and he
57:21
keeps beating me. So I need to figure out a way to
57:24
get better at tennis. And it's not an easy he's
57:28
growing, you know, he's spending more time on tennis.
57:30
So I got to figure out a way to be more efficient at acquiring
57:34
skills. I love this answer.
57:39
You're a man who's battled cancer.
57:41
You're a man who's helped build multiple unicorns.
57:45
You're investing in some of the leading tech in the world.
57:49
You are right on the bleeding edge of everything.
57:51
And the thing that you want to be celebrating a year from now
57:54
is beating your 11, you know, soon to be 12 year old son in
57:57
tennis. That's awesome that that shows a
58:00
a great focus on life priorities.
58:04
Thank you and it was great speaking with you.
58:06
Thank you for the. Yeah, thank you, Andy.
58:08
I want to thank you for setting aside both the time and the
58:11
energy to Share your story today.
58:12
I'm grateful for it. And folks, that's a wrap on
58:15
another episode of the Inspired Stories podcast.
58:18
Thanks for learning with us today.