Ladies and gentlemen,
welcome to the
Solari Report.
This is Catherine
Austin Fitts.
I am in Salt
Lake City.
I’ve had an amazing
time, and I’m
joined by a very
special person,
Travis Oliphant,
who came to visit
in the Netherlands.
We had an amazing
time with Tiffany
Cianci, who
works for you.
He is the CEO of
OpenTeams, one of,
if not the father
of artificial
intelligence.
Is it fair if
I say that?
Forefather,
pioneer of.
Fore-
forefather
and pioneer.
Yeah there’s a
number of people
who make claims,
but I can own the
forefather claim.
And you’ve been
living in Austin,
Texas, running
OpenTeams, but you’re
moving back to Utah.
Yes, in, in the
middle of it this
month, about two
weeks from now.
From here, grew up
in Utah- but I’ve
been in Texas for 20
years, effectively
building companies.
I had my, a lot
of experience
in education,
then I went to
Texas to explore
entrepreneurship,
and then now my kids
are grown and time
to move back, but
I’ll still be doing
the same things.
So we’ve been in
cahoots because we
wanna load OpenTeams
on our servers-
Yeah … and use it
internally to make
the subscription
experience much
more… Everybody
says, “I can’t find
this, I can’t find
that,” or on a- we
have a show called
Ask Catherine where
we’ve answered the
same question 50
times, and yes, you
can do frequently
asked questions, but
if people can easily
find everything.
Right.
But it has to be
something what
I would call
open source.
It has to be
something I
would call moral,
responsible
technology.
It has to be
something that we
have on our servers
and can protect
people’s privacy.
Exactly.
And so when we
met you and your
team, we were-
Exactly
… We
were excited.
That’s where we k-
had a conversation.
You talked about
what you wanted
to see, which is
exactly what a lot
of is possible today.
How do you make
sure that happens?
That great demo
ware that people
are seeing from the
public AIs, how do
you get that same
capability- but
not have the sneaky,
this feeling that
your data’s just
sucking elsewhere?
That the information
about yourself is
going to- some
larger organization,
that you take the
capability of AI
but retain the
governance and the
accountability-
That you’ve
already built?
So this issue of
data sovereignty-
is getting more
and more important
at the national
level, at the local
level, and at the
individual level.
Data sovereignty.
Okay.
So let’s go back.
H- how did you get
involved with what
is now called AI?
Yeah, great question.
I- indirectly.
My background is I’m
a scientist, and I
was studying then
remote sensing.
I was studying
medical imaging at
the Mayo Clinic,
and I encountered
Python in trying to
solve my problems.
I was doing large
scale science
analysis, and I
needed language to
help me do that,
and I found Python,
and I started to
contribute to Python
I started to do a
lot of just writing
code, sharing that
code on the internet,
collaborating with
people all over
the world, and
then that grew
into involvement
in open source.
So that grew into a
project called SciPy.
And SciPy became
the foundation of
a lot of science.
So a lot of people in
the world out there
know about SciPy,
a lot of scientists
know about SciPy-
and they used it.
And then that
led to a library
called NumPy, which
I wrote when I
was a professor.
Okay, spell that
for everybody.
NumPy.
I wrote this
NumPy library.
So NumPy is
N-U-M-P-Y.
Okay.
So it’s the
foundation of array
computing in Python.
So that’s one reason
I could claim
possibly the idea
forefather of AI
is because- OG,
you’re an OG
… the OG, is because-
Yeah … I, so I
was, I’ve been doing
AI experiments
ever since I was
a researcher.
As AI is a ki- it’s
a model building.
You’re trying to
figure out how to
build a model and
get information
from data.
Modern neural
networks were
y- were one of
the possibilities
that we used all
the way back then.
So and people were
trying to figure
out how to have
scientists do it,
’cause there’s a
lot of orchestration
of computing,
orchestration
of code.
It’s a lot of work.
And so a lot of
what we were doing
is making that
work of putting all
the stuff together
to make something
useful easier.
Right.
And that’s basically
been the journey
I’ve been on for
a long time, is
helping scientists
do their job easier.
So people wanting
to use NumPy drew
you away from
being a scientist.
Correct.
That’s exactly right.
So in fact, writing
NumPy led me out
of the world of
academia- into the
world of business.
But what did at
first was in the
world of open source.
So I became
this open source
enthusiast working
hard with a lot of
people, building
a lot of tools.
Right.
And a whole
ecosystem of people.
Like some… I
often say that me
and my closest
thousand friends
took over science.
And a lot of
scientists went from
using various tools,
using open source
tools around Python
to publish papers,
to do research.
So I’ve had the
chance to talk to
people that found
the black holes,
that, that found
the Higgs boson.
A lot of them have
been using the tools
in this ecosystem
that- many of my
friends have been- A
building for 30 years
is open source code,
models, et cetera.
It’s like a
public recipe.
It’s freely available
for anyone to use,
see, and modify.
So you can find it on
websites like GitHub
where developers
share their work
with the world.
On the flip side,
closed source
code is private.
Think of it like a
secret recipe that
a company keeps
under lock and key.
You can use the
final product, eat
the food as it’s
done, but you can’t
see how it’s made.
So most of the
software we
use daily, like
Microsoft Word or
Adobe Photoshop,
is closed source.
You can use these
programs, but you
can’t look inside to
see how they work,
including programs
like ChatGPT,
closed source.
We can’t actually see
how the model works.
When you find open
source code, you’ll
see different
types of licenses
that explain how
you can use it.
Some licenses,
like the MIT
license, which is
a common license,
are very generous.
They let you do
almost anything
with the code as
long as you give
credit to the
original creators.
So that’s like
sharing a recipe
but asking people
to mention where
they got it from.
Other licenses can be
less permissive, like
allowing you to use
the models only for
research but not make
money off of them.
Now let’s talk about
why open source AI
models are becoming
increasingly
important.
First, they give
you privacy.
When you use a closed
source AI service
like ChatGPT, your
data has to be sent
to their computers,
their servers.
But with open source
models, you can run
the AI on your own
computer or server.
This means your
sensitive information
stays with you.
Open source
models can also
save you money.
So while closed
source AI services
can charge you for
each query, sometimes
pennies but sometimes
dollars, open source
models lets you run
the software for
just the cost of your
computer or server.
So yes, running
AI requires some
powerful computers,
but once you have
the setup, you can
use it as much as
you want without
paying per use.
So it’s like buying
a coffee maker
instead of going to
the cafe every day.
Perhaps most
importantly, open
source AI helps
democratize this
powerful technology.
Instead of a few big
companies controlling
AI development,
anyone with the
right skills can
examine, modify, and
improve these models.
Researchers in
universities,
developers in small
companies, and
even just hobbyists
can contribute
to advancing AI.
This widespread
access leads to more
innovation across
all industries and
use cases, and helps
ensure AI develops
in ways that benefit
everyone, not just
big tech companies.
This has created
an important debate
in the AI community
because while closed
source models help
companies protect
their investments,
open source models
are making AI more
accessible, private,
and affordable
for everyone.
So as AI becomes
more important in
our daily lives,
having open source
options ensures
that this technology
remains in the
hands of the many,
not just the few.
So when did you
start Open Teams?
So Open Teams
came later.
It’s a, an evolve.
I won’t go into the
whole story, but
maybe summarize
it quickly.
I first went to
Texas and founded
a company called
Anaconda after doing
some consulting work.
Anaconda became a
distribution company,
how to get all this
stuff installed.
Put a lot of heart
and soul into that.
Then after I left
Anaconda and then
started a consult-
another consulting
company called
Quansight And
a venture fund.
So I’ve always,
my, my experience
building
companies- You,
you like to incubate
… I like to help I
like to incubate.
I like to help
people become owners.
Right.
What I realize is
we need a lot more
owners in the world.
Yeah.
We need a lot more
people who understand
the concept of
starting companies.
We need a lot more
capital structure
to help them start
those companies.
I got deeply
fascinated with that.
So when I left
my startup, my
first startup, how
do I help create
more owners?
And so that was
the foundation
of what I’ve been
working on since.
And so we started
a venture studio.
A venture studio is
a venture c- it’s
just a company
that organizes
capital, and then
itself might start
companies, and then
try to find CEOs and
people to run them
Open Teams is one
of those companies.
Open Teams is one
of those companies
we s- incubated in
our studio, started.
I’ve had various
CEOs run it.
I took over as
CEO… I was the
CEO originally.
I took over recently
as the CEO when I
realized how critical
it is for us to push
the message hard
that everybody needs
to own their own AI
system- their own
intelligence hub,
we call it, so they
can become- They
need to have
sovereignty over
their AI … the
master of their
sovereign-
sovereignty
over their data.
It starts with
sovereignty
over your data.
Yes.
And then data leads
to intelligence.
Right.
And so you can’t
have one without
the other.
You…
And you need s- data
sovereignty to lead
to AI sovereignty.
And you need to
make sure that you
don’t lose your
data sovereignty
as you’re sending
information to the
AIs that aren’t
in your control.
So that mission
became very
critical.
And so how do we
make that happen?
‘Cause there’s a
lot of need for
it- but there’s
a l- there’s a lot
of capital invested
in centralized
AI-
In put everything,
pull it together.
And AI is s- is
still a new thing
for so many people.
For me, it’s not new.
For me, it’s
a iteration of
something I’ve
been studying for
years and years.
And it’s I
partner with we
partner with
Meta, we partner
with Google, we
partner with Nvidia,
we partner with
the people that
are building the
tools- that help
AI become possible.
And then that
partnership led- I
avoid those people.
It’s hard to avoid
them entirely.
I know.
I understand
completely.
And where you…
To be… It- it’s
great for your
subscribers or for
you to understand
that actually
where some of those
organizations have
been helpful is
they’re contributing
to open source.
Right.
So essentially,
the degree to which
they contribute to
open source is a
positive outcome-
Yes … regardless
of everything else-
they’re doing, if
we can keep that
open source open,
and we can c- and
we can educate
people about what
they can do with it.
That’s where
we start.
That’s where Open
Teams came from,
was how do we
help companies use
open source more
effectively, and that
started actually-
So your
market-
2019
… is companies?
Our market
is companies,
organizations,
countries.
Any organization that
knows it wants to
build AI and own it.
And then what’s been
happening over the
past several years
is the cost has
been coming down.
Right.
Two years ago I’d
tell this story, and
what could I say?
It might take $5
million- to actually
build an AI that
you would own.
It doesn’t cost
that anymore.
It’s now much more
accessible to just
any company that…
And the amount
you pay really
depends on how
much you wanna do.
So here’s what’s
interesting- ‘Cause
you incrementally go
from where you are
to where you wanna be
if you’re using
Anthropic-
If you’re using
Claude, and you’ve
got a subscription-
That price can’t
possibly be the
market price
No It has to be
hugely subsidized
Correct.
It is- Okay
… absolutely hugely
subsidized right now
now, if that floated
to the market price-
Yes … I don’t know
how high it’d be, but
it’d be much higher.
At the same time, the
cost of open source
is coming down.
Yes.
Now, it was
interesting.
When I started
my first company
after I left the
administration I did
an analysis, and what
I figured out, it
was, I could either
buy the software
or I could do open
source and spend
the money on people
And if I could find
the right people,
I was much better
off doing that.
Yeah.
And
so tell me about
the economics on AI.
Is it the same thing?
You’re, you, it’s-
So right now-
If Anthropic starts
to rise- correct
… to market price
and your price is
dropping, are you
starting to look
very… Is open
source starting to
look very economical?
Correct.
Correct.
Right now, and for
the past two years
the VC market the
investors have been
subsidizing OpenAI
and Anthropic to
provide pretty
powerful AI at
way under- We need
a stronger word
than subsidizing
We do.
It’s… Yeah, no,
it’s a stronger word.
It’s like
they’ve literally
been- It’s like
a tsunami of
subsidy It’s
a tsunami of
subsidization.
No, in fact, I have
a really good friend
who pays $100 a
month, and literally
because he could
see how much he,
tokens he was using,
and looks at the
price of what that
would have cost him
if he were paying
the token price,
the current token
price, which already
is not necessarily
market, he was paying
$16,000 of token
price- for $100.
Right.
So there’s a at
least $15,000 a
month was being-
Of subsidy
… of was being given.
So whenever you’re
given something
like that, you
know that there’s
something in return.
You’re tr- you’re
helping to train
the model too.
You’re helping
that company
create dominance.
Cause that’s
the game.
It’s like- how do I
get market dominance
so then I can then
we then you’re
gonna need them.
Now, fortunately,
open source is so
strong and there’s
so much interest, we
still need to make
the easy buttons, and
we’ve been working
on that, and lot-
many people have
been working on that.
It’s getting
easier and easier
to actually do
this yourself.
And I’m really
excited to help
people do this,
’cause once they
realize oh, that
Claude or OpenAI
has given us a
really cool demo,
and they are.
They’re doing a lot
of, they’re giving
a lot of examples
of what you can do.
And now you can
take those and
ensure that this
is supporting your
vision of the
future, your vision
of- of the world,
and that’s what
we need, actually.
What we need is…
AI, artificial
intelligence it’s
really a tool.
It’s- I call
them loop token
predictors.
It’s- Looped
token … Looped
token predictors.
‘Cause you’re just
predicting a token,
then you put it in a
loop, and it’s- Yeah
… actually powerful.
It’s a really
powerful concept,
and we’re getting,
seem to be simulating
human conversation
really well.
But it’s not
reasoning yet.
I’m definitely in
the camp of many
people out there
that are saying,
“Look, this is not
intelligence. It’s
not reasoning.” No.
It’s not.
It’s not.
But it’s doing some
really impressive,
intelligent things.
Things that clearly
we’ve seen in
ourselves too.
So my number one
goal at Solari
is, first of all,
to understand it.
But then the second
thing is I’ve been
applying it to all
sorts of different
functions to see You
know, where we can
use it smartly, where
we should avoid it,
what the dangers are.
I wanna understand
what this means
to a family or a
small business.
And one of the things
I will tell you, it
is hands down the
best search engine
I’ve ever had.
No, it’s impressive.
It’s
impressive.
It’s absolutely
impressive.
Yeah.
It’s interpolating
the world’s knowledge
in a way that
li- that you can
interact with in a
natural language.
It’s beautiful.
It’s a really
impressive
example.
It’s… Or the
world’s idiocy.
I mean-
Sorry.
Yes.
It’s a search
engine.
The world’s text.
Thank you.
What- whatever
that might be.
Right.
And more and more,
it’s a lot of idiocy.
Correct.
It’ll re- it’ll
reflect to you
whatever that
happens to be.
So biases
included.
But as a search
engine you have
to go to the links,
you have to figure
it out, you have
to do the research.
But as a search
engines, it’s- Yeah
… Really good.
Yeah, super good.
And especially since
they’ve added the
token prediction
with tools.
They’ve added tools.
And what that does,
it just means, oh,
the token that is
predicted is not
a text, it’s a
function to call
or a web to search.
And then it goes
and does that and
feeds that back
into the context.
And that starts to
give the sense of
reasoning, ’cause
you’re now calling
tools, and tools
are programs.
Tools are computable
comp- computation
that occurs that
then augments the
context, augments the
predictive surface.
‘Cause when you’re
making a prediction,
what it’s doing is
it’s taking all of
this context window,
all this information-
Right
and then predicting
what the next
word should be.
It’s a token, but a
token roughly is a
part of a word or
a symbol, or if it’s
a video, it’s part
of a sound or part
of a, an image.
That’s a fairly
straightforward
thing.
The thing that
na- that machine
learning, artificial
intelligence is
fundamentally,
it’s this universal
approximation
theorem that allow…
It says, “I’m going
to take… If I can
construct a function,
if I can construct
the concept of a
function where I
have a black box,
where I have inputs
and an output,” that
function could be as
broad as you want.
And if I have
enough data, I can
train that function
to… I can use
a computer to use
the data to come
up with a version
of that function
It’s like nonlinear
curve fitting.
Many people
understand
curve fitting.
Now, it’s many-
it’s very high
dimensional,
so some of our
intuition doesn’t
work, in fact.
Our intuition is
limits to three or
four dimensions.
We’re talking
about 10 billion
or a trillion
dimensional space.
Our intuitions
don’t always map
to that space,
but nonetheless,
it’s still I put
in parameters, and
the parameters I
find the f- fit for
those parameters.
That’s what my
training data does.
It’s helped me
figure out what
the model is, and
then I just feed it
inputs, and I come
up with an output.
And the cool thing
about AI is that
anything you can
imagine, if you
can imagine a
black box, you can
create a function
that does that.
So- And that’s
what we have here
so in the ’90s, I
spent, I discovered
the internet in
1989- and then
the World Wide Web
shortly thereafter,
and then it really
started to be
something you could
use in the mid-’90s.
And and we
started to work
on, believe it or
not, digital money.
We had the,
Wow … we called it
just-in-time money.
Yeah.
But we our notion
was that these
kinds of software
tools should make
the small guy much
more competitive up
against the big guys.
Yes.
And- Yes … and we
were doing a whole
series of things
to allow financial
liquidity of equity
in communities-
Amazing … by place.
So it was place-based
venture funds
that could be
publicly traded.
We had a software
tool called
IPO in a Box.
Wow.
And you could click
radio dots and spit
out your prospectus.
We had another tool
called Community
Wizard, where you
could map out all
the government money,
all the federal
money sources and
uses of credit
money in your place
because the federal
government was…
You would pay taxes
to Washington.
They would pay a
contractor $125 per
hour to do something
that could be done
in that community for
25-plus healthcare.
So there were all
these arbitrages- All
these inefficient
arbitrage loops.
They aren’t
inefficient in the
sense that they are
intentionally set
up so that the big
guys get the money.
I guess efficiently
concentrating money.
Yeah, so they
were politically
efficient- Yes … if
you wanted control.
Yes.
But what I discovered
was that the
financial cost
of control was
much bigger than
anybody can imagine.
Wow.
It’s bigger than
the solar system.
It’s unbelievable
how expensive
central management
and control is.
How much it
costs to actually
support that system
to enforce control.
And to get
everybody to comply.
You need lots- And
enforce … of rules,
lots of enforce.
You’re trying-
That’s an echo of
what’s happening now.
Exactly.
Wow.
But it’s
getting worse.
Yes, it’s
getting worse.
And then the cost
was dramatic.
Now, the cost is
off the charts.
Yes.
Anyway, but w-
we did a lot of
simulations to see
what could happen if
we were free to just
re-engineer all the
government money- To
optimize the economy.
And it was so
shocking, I didn’t
believe it at first.
My smartest guy did
the calculation.
I was like, “No,
that’s not possible.”
And then when I
realized what was
possible, I thought,
“Oh my God, tyranny’s
very expensive.” so-
That, that’s
my thesis too.
That’s why it doesn’t
work for very long.
It can work for
a while but it-
Kind … It’s been
working for my whole
life, so that’s
… A while could be
as, an entire life.
That’s true.
But here’s the thing.
It seems to me that
what you’re talking
about is a tool
that, when added
to all the other
tools … I keep
telling my audiences
since I’ve been here
in Utah, if I wrote
a book about America
from World War II
on, it would be
called How the Local
Boys Got Rolled.
Wow.
And so the question
is, can y- can these
tools, on an open
source basis, unroll
the local boys- So I-
… and get us back
in the game?
So that is my
hope, right?
Yes.
But I think after,
especially meeting
you, understanding
all the information-
you have about
how we got here,
in the current
economy we have, it,
it’s not trivial.
We’re up against
a lot of powerful
interests and money
that’s- And then
also, the … That’s
partly why I’m
speaking- … more
and more, is because
in order for that
to work, people
have to understand.
You have to get
the word out,
’cause there’s
massive They have
to- … amount
of people
know there’s
an alternative.
They have to
know there’s
an alternative.
Right.
And they have to have
faith and believe
in that alternative-
enough to do it.
Right.
And part of that
is getting it
simple enough and
cost-effective
enough So that you
can actually do it.
Because if the only
answer is yes the
local boys can win,
but it’s gonna take
$10 million every
time, then you’re
up against it.
You’re gonna have
to- So- … allocate
resource.
But if it’s not-
One of- … if it
takes just, oh, I
can just, instead
of spending 200 a
month on an Anthropic
subscription, I
could spend 50
bucks a month on a
subscription that
actually I know
keeps my data local.
Right.
And it’s actually
faster, it’s
even just as fast
or even faster-
Then that starts
to turn the needle.
Eventually, so I’m
hopeful for this.
Yeah.
It’s challenging
because the intense
sub- intense
paying off people.
People are being paid
off basically to use
Anthropic right now.
It’s unbelievable.
It’s
unbelievable.
So subsidiary,
subsidization-
Now, I’ve
been- … so you’re
being paid off
using
Anthropic, so I-
Me too.
Yeah.
Okay.
I use everything
actually.
I’m not, I’m not-
I’m not against it.
These are
impressive tools.
There’s impressive
people who work
hard to make these
tools useful.
So h- here’s the
problem, though.
If you look at how AI
is being positioned
and communicated
it is being used
to implement a
control grid.
So it’s been
weaponized by the
guys who weaponize
everything.
So we have a
wonderful wrap-up
called Omniwar, which
is the weaponization
of everything.
You name it, they’ve
weaponized it.
Milk I don’t
know, pink polka
dot, everything.
But AI, the clear
message is, we’re
going to use this
to control you.
That’s the
best image.
But AI’s gonna be
in charge, and it’s
smarter than you,
and you’re stupid.
And so if you look
at the oligarch’s
message about AI,
it’s making a lot
of people- So who
do you, who do you
think is doing this?
Because I know f- I
know that 99.9% of
the people actually
involved don’t,
aren’t aware of that.
They’re not-
No, oh, I
agr- I agree.
They’re not-
They’re completely,
they’re not,
that’s not
their- they’re
totally, I would say,
so one of the best
reads of so far that
I’ve read in the last
12 months is Karen
Hao’s book on OpenAI.
I don’t know if
you’ve read it.
No
No, I haven’t.
And it’s hysterical
because I know
a lot about…
I- if you wanna
understand risk
and evil in this
world I’ve really,
I’ve learned a
lot about that.
I’ve experienced
a lot, I’ve
researched a lot.
But you have these
amazing conversations
where all these
brilliant people
are sitting around
in OpenAI talking
about how when
they reach general
intelligence, they’re
gonna solve the
healthcare problem.
And I’m laughing my
head off at, ’cause
it’s just stop
poisoning people.
They’ll be fine.
You don’t need,
you don’t need
one ounce of AI to
solve this problem.
It’s not a
he- it’s not a
medical problem.
So you see that as
a tool for persuasion
to get people to buy
into another story?
No
I, I- because
it’s not real.
They’re not gonna
really solve the
they could solve the
problem elsewhere.
I see a group of
highly intelligent
people- … working
on AI who have
not a clue how
the world works.
Yeah.
Okay?
I was one of those
people- No, I-
… once upon a time.
I,
you know- I
completely understand
that because- Yeah
… I was I- A very
bright academic who
was 30 years old,
all the way through
my PhD before I
uncovered a lick of
economic reasoning.
I was in that
same boat.
I- And I realized,
oh my goodness, I am
super knowledgeable
about a ton of
things, and I’m
completely clueless-
About how the
world works.
We, we
have- So I know
this is true.
We have a whole
part of the economy
where people make
money complexifying
problems instead
of solving them.
And the- and the
pr- and the solutions
are pretty simple.
Yeah.
Right?
Yeah.
But but getting-
But the solutions
may not keep…
give you control.
The solutions may,
the s- solutions
may not benefit
every, may not
benefit the person
trying to solve it.
But it cer- if we
solve the problems,
it certainly won’t
benefit the guys
trying to centralize
things, but-
Many of the people
I know who feel very
negative towards
AI are feeling very
negative about their
perception it’s
being unleashed
to control us.
Exactly.
Exactly.
And I share those
concerns, right?
Which is partly why
I wanna get the word
out, that it doesn’t
have to be that way.
Right.
AI actually came
from a the sea,
in a space of open
source cooperation
that’s global.
It involves millions
of people, actually.
Themselves are
all just trying to
cooperate, and have
found mechanisms
to cooperate in
amazing ways.
In fact, I actually
recommend anybody
to go watch the
Python documentary.
Oh, really?
What’s
it called?
Yeah, it, it’s
called the Python
Documentary.
It was b- it
was produced
by Cult Repo.
It’s on YouTube.
You can just
go watch it.
It does an incredible
job of explaining
how Py- where
Python came from.
Okay, Google.
Where was PyCon
held in 2015?
In Montreal.
Oh, it was Montreal.
Really?
So the story goes
that Guido invented
Python over a
Christmas holiday.
I showed it to
Lambert and I said,
“Look what I made.”
And he types one very
short line of code,
and it crashes the
interpreter, and he
knew that it would.
The point escaped me.
I didn’t know why
he thought this was
better than ABC,
and why he had done
this whole project.
It was a pretty
crushing experience,
to be honest.
Immediately on
trying it, I was just
amazed and thought,
“This can’t be as
good as it is.”
You can see Python
code even if you’ve
never written a
line of Python,
and if you know
how programs work,
you can understand
it and grasp it.
That click
in real time.
We were paid
our salaries
every two weeks, and
then suddenly in late
October we weren’t.
It was a moment
where the shit
coulda hit the fan.
If all the Python
guys went their
separate ways,
Python wasn’t big
enough to survive
that at the time
Trying to build
a language is
non-trivial.
That actually took
real design, real
thought, and real
ecosystem collective
innovation over the
course of decades.
That was my car when
I lived in Virginia.
I
think it was me
who suggested that
Guido be called the
benevolent dictator,
and then Barry
suggested benevolent
dictator for life.
Python 2 to Python
3 transition was
extremely difficult
for the community.
Some of that
probably got away
from us a little
bit in hindsight.
I did feel like
Python has a
chance of becoming
irrelevant.
When you don’t have
role models who you
can relate to, you
start believing that
you cannot do it.
When I first saw
the walrus operator,
I didn’t like it.
It got very vocal
and blown up.
One person’s
succinct,
elegant code can
be another person’s
unreadable puzzle or
a fusticated mess.
Nobody had expected
I would resign,
and certainly not
that I would rage
quit over this
issue, which it…
essentially it was.
How much of Python’s
success in AI can
be traced back to
these scientists
in the ’90s?
100%.
Python’s a
fantastic language
even if you’re
not a scientist.
I just don’t think it
would have risen to
the level of dramatic
usage without the
science first, data
science and machine
learning story.
We had a massive
impact in the world.
Okay
And ’cause Python
came from a global
cooperation of
individual people.
Yeah, and
it’s amazing.
It’s a beautiful
example-
Yeah
… of what can emerge
if you just let
intelligent people
freely interact
with each other.
So it was when
I had this beautiful
mansion in downtown
Washington, and when
I decided to fight
really fight on the
litigation, I decided
to sell everything,
all the art, all the
antiques, all the…
Everything but the
family antiques
was going.
And eBay had just
started, so we
said, “Let’s try
this.” And over the
next couple years,
we did hundreds
of transactions
with people all
over the country,
totally on trust,
not one problem.
It was amazing, and
I was just dealing
with individuals or
small businesses.
Yeah.
Now, my corporate
experience, that
can… There’s a
problem every week.
Yes.
And, it… But no,
but it was the same.
You’ll see the
same in Open Trust.
No, exactly.
It’s just
unbelievable.
‘ Cause you said
something earlier
that I really
resonate with.
I was, I’m also
was a enthusiast of
the early internet
because I saw in
it- Yeah, me too
the power to
empower people.
That’s why we
invented Just
in Time Money.
Yes, ex-
it’s amazing.
And I want people to
understand that this
is, this did happen.
It is happening.
Yeah.
Even while the
control grid
came in and the
monorepos, I call
them, the mono- the
monolithics- have
come in, and it’s
constantly always
the battle- between
the centralists,
the decentralists.
Right?
Did you ever
know Eric Hughes?
He was from Utah.
I didn’t.
I don’t think
I’m familiar.
He was the guy who
had Clipper Chip.
Oh, he
was the- He was
on the front page
of Wired, long,
blonde hair.
Oh,
yes.
No, I did not know
him, but that’s
familiar with
the Clipper Chip.
He’s,
he’s ba- I, he’s
still here in Utah,
but- No kidding
… but I remember right
after he got on the
front page of Wired,
he came to Hamilton,
and we spent a
whole day with the
firm- Yeah … and
trying to invent
Just in Time Money.
Yeah.
And, and-
No there’s
enormous potential
to continue to…
That- that’s the
thing I wanna help
people understand.
There’s enormous
potential to
build things that
are distributed,
that are enabling.
Well,
w- one of- That
don’t… You don’t
have to fall into
the centralized AI,
which is gonna become
even more centralized
if we just- It’s
scary
… if we let it.
So I have one
partner who says
we’ve gotta get AI
to go completely
out of control
because otherwise
it’s either the
control grid or
out of control, and
the out of control
can tap into-
What do you mean
by out of control?
You mean just
let everybody,
let it go, let it
go- It’s totally
distributed-
Distributed
And they can’t
control it
… everybody has own.
Correct, yes.
It’s not only out of
our control- What do
you mean- … it’s
out of their control.
Correct.
No, I- And I
believe because
it is, it does
give you power.
It does give you…
There are models
now that, that
are open weight
models- you can
deploy, and there’s
some open source
models also emerging
because there are
some large companies
who recognize this,
and they’re their
best interest also
create a distributed
ecosystem of
model builders.
Nvidia, for example,
does not want to
have 10 customers.
Right.
They want a
billion customers.
So they actually
are supporting-
That’s why
we have a
subscription
model at retail.
Exactly.
Cause they want
people to be using,
so we’re actually
aligned in a
league with them.
Qualcomm also
wants to have
AI chips AMD.
There’s all
the hardware
manufacturers- can
be allies here.
Now it says they’re,
they have their they
have to, of course, a
lot of their biggest
customers are some
of the hyperscalers,
and so it’s- they’re
not gonna come
out and kinda be
adamantly supporting
distributed AI.
But they are
Encour- they are
encouraged by it.
They do want
to see it
So let me get back
to, I was… When
I, on my way to Salt
Lake City, I came
through Sioux Falls
Oh.
And we’re working
with a wonderful
group of the leader
of our team, Cash,
is a PhD from MIT
and used to be an
engineer in Silicon
Valley until the
pandemic when they
told them that their
kids were gonna have
to wear masks, and
boop, the next thing
you know, they’re
living in South
Dakota, so they
don’t have to wear
a mask at school
during the lockdown.
But we were
talking about the
size of American
data centers.
So we’ve got 5,500 in
process in the United
States, significantly
more than any other
country in the world,
including China.
China, I think, has
I don’t know, I’ve
heard figures from
385 to 550, but and
they may have more
per data center,
but still it’s and
one of the things we
were talking about
was why this level
of data center has
to be for control.
It can’t be…
This is not what
we need for our
cloud and tech.
It’s not
what we need.
We I would rather
see 55 million
smaller data centers-
Essentially where
everybody’s home can
become a data center
for themselves.
Right.
And then if they
want to share a
portion of that with
others, they can.
You could
easily do this.
And then they have
complete data
sovereignty.
And then they would
have complete data
sovereignty, exactly.
Not everybody may
wanna do that,
but you can have a
local data center,
you can have a
local provider.
I, I don’t
know for sure.
I don’t know the
mean- I don’t
know the meaning
of all things.
I don’t know why
people might do this,
but it sh- the optics
on it, it sure does
explain the data.
Because… But I
think a lot of people
are speculating
partly because
we have these big
centralized AIs,
these big centralized
hyperscalers.
Amazon’s been making
a lot of money
selling centralized
compute to people.
Right.
Because you
could… This
notion of compute
has been around
for a long time.
Right.
And people have had
distributed compute.
It’s actually
very interesting,
the ebb and flow,
because I think
we’re in for another
evolution just like
the ’80s when the
PC era emerged.
‘Cause before
the PC era, it
was mainframes.
You got… You
wanted computing,
you went to a lo- a
big mainframe, you
got your answer.
So in the evolution
of computing,
it’s why Microsoft
showed up, ’cause
IBM thought, “Ah,
there, there’s
nothing there.”
Where pe- business
computing’s gonna
be on the mainframe.
And so they kinda
gave Microsoft
the PC market.
Right.
And sure, lo and
behold, actually
the demand for
PCs was enormous.
I think there’s the
same phenomena can
happen here, where
the demand for AI
as a distributed
phenomena is
gonna be enormous.
Okay.
And so we’re not
gonna have the
mainframe provided AI
as the only… It’ll,
they’ll be there.
Right.
But it’s
actually…
I would be
very concerned
about investing
in speculative
data centers.
Because I don’t
know that they
actually don’t- I
would be too … I
don’t think they’re
actually gonna be…
It’s always hard to
predict the future,
’cause they could
rent themselves
to other local
providers, too.
But cause getting
more energy and
getting compute are
valuable things.
But we haven’t
figured out
the model yet.
There’s only one way
the economics work,
and that’s if it
delivers financial
transaction control.
That’s the only
way the economics-
I could see that.
Certainly I, I- I
don’t know the
companies in detail,
but I can see one
of the companies
Is over levered.
I know something
about this.
I know a lot
of people who
buy hardware
in this space.
All of the new
hardware, and
particularly the hard
drives actually, have
all been bought out.
Right.
And there’s no more,
there’s no more
manufacturing in hard
drives in particular
that are gonna-
it’s gonna happen.
And so there
is gonna be a
massive constraint
for new hardware.
And the demand
they’re projecting
won’t … I- is
mismatched against
the actual hardware
that’s gonna
be producible.
And so there’s a
absolute gap between-
what they’re saying
c- demand will be to
justify- the data
centers, and what the
actual availability
of hardware is.
So
if you look- So
the costs are
gonna go up there
If you look at what
is happening, as
an economic matter
it makes sense if
you’re gonna get
the control grid.
If you succeed-
That, that’s the
casino … the
economic value of
getting the control
grid will justify
what is going on.
If you fail, you’ve
got the greatest
misallocation of
capital just about
since Scotland
got bankrupted.
Which is not, that’s
a sad story, but,
Y- I hate to
say that, but the
sort of the war so
let me step back
and say it this way.
Tim Wu wrote a
wonderful book, and
I think I asked you
in the Netherlands
if you’d read it.
It’s called The
Master Switch, and
he goes back 200
years and goes
through the history
of the development
of information
technology.
So a new technology
comes out, and all
the entrepreneurs
are all excited, and
they’re like, “Oh,
this is great. We’re
all gonna be free.”
And there’s this
wonderful blossoming
of entrepreneurial
prototyping and
testing, and then
suddenly, whoomp,
it centralizes.
And then it happens
again through the
telegram and then,
or the telegraph and
then the telephone
and then the and it
keeps centralizing.
And so this time
when it centralizes,
though, if you get
financial transaction
control, it’s gonna
centralize in a
lock that could stop
that could be it.
And so it’s
almost as though
a centuries-old
battle is coming
to a head and the
question is, will
tyranny win or
will freedom win?
And this
is the, like
the big- This is the
ultimate question.
I think that’s a
really great way to
look at the arc of
history, frankly,
because I think this
ba- this conflict
or this ebb and flow
happens repeatedly.
I went back and I
looked at my family
history and we’re
in Utah, which has
the best family
history- It does.
… place in the world.
But I went and I
looked and I realized
they had for 500
years, they’d been
moving to get away
from the tyranny.
Oh, wow.
And they kept
moving west.
Wow.
And now when I
was an investment
advisor, everybody
kept saying, “Tell
me how I can get
away,” and I
feel like saying
after 500 years,
there is no way.
We need to turn
and walk into the
center and deal
with It’s kind
of, that’s a really
interesting point.
Yes.
It’s that for you
just to escape to
another place- and
now it’s it’s all
over, so we have
to actually deal
with it head on.
I’d be an advocate
that AI can help
us because- I
believe you … it
is a technology.
It’s not… It’s
a technology that
could be used for
whatever purpose.
If
they have-
Including empowering
individuals.
Right And
empowering small
groups.
If they get machine
guns and we say,
“Oh, we hate that-”
Yes
… we’re not gonna use
machine guns- Correct
… we’re gonna use
bow and arrows-
Then you’re gonna be-
… that’s not gonna
work.
Then you’re gonna
be enslaved.
Right.
No, that’s right.
You’ve got to use
the technology that’s
available, and it
can absolutely be…
The seeds have been
already planted.
That’s the good news.
There’s a lot of
challenging news.
There’s a lot
of investment,
there’s a lot of
stories happening.
This is actually why
I get up every day
and I’m so excited
about OpenTeams
because it builds
on a career for me.
It builds on what
I’ve been doing
ever since- I was a
scientist trying to
make technology
easier for scientists
to understand all
the way through
open source, all
the way through
building businesses.
So I’m excited to
help people not just
lean into building
intelligence hubs,
but also leaning
into building
businesses that are
distributed- here’s
the thing
and decentralized.
This is what H.L Hunt
said, I told you this
in the Netherlands.
He said, “If this
country’s worth
saving, it’s worth
saving at a profit.”
And the reason
that’s important is-
Sure, yeah … if
you’re gonna have
a revolution, you
have to finance
it.
Y- yes, you do.
And it has to
be a for-profit
revolution.
And it has… The
revolution has to
cover its cost.
Correct.
And that means it has
to be entrepreneurs.
Correct.
And thousands, if
not millions of them
leading the way.
Correct.
Exactly.
Saying, “No, we
will be free and
the citizens will
support us sufficient
so that we can
cover our cost to
ensure we’re free.”
And that’s
the good news.
The good news is you
can actually build
businesses easier
than ever because
the AI can support
the gap of talent.
Building business
is one of the
challenges, is it?
Is us- is no one
person has all the
skill necessary to
essentially deliver
much of anything to
that scale, certainly
not at scale.
And so a lot of
building a business
is organizing a
group to actually
do the things and
cover the gaps.
AI can help.
It really can.
And so this
interesting… We’re
in this interesting
dynamic where lots
of people are leaning
into coding agents,
coding tools to
help them- not
only build software,
but also build
processes, build…
This is agentic
engineering that’s
emerging everywhere.
People are
leaning into it.
Even people
are saying the
company of one.
I think there’s a, in
Argentina, in fact,
they’re looking at-
So I’m very
worried about
AI.
I know, I am too
because… But
it’s leading into
the same idea of
a company of one.
So first of all I
don’t think company
of one’s a good idea.
Right.
Company of three
maybe, ’cause
you need a group.
You need something.
You need a core
sense, central.
A company of
one- You need
somebody to tell
you’re wrong.
You need somebody to
tell you’re wrong.
Otherwise you end up
with AI psychosis,
actually- right
which is a thing.
I’ve seen… I’ve
actually been
with people who
have been, okay,
you’ve been Reading
your AI too long.
You need to get out
and talk to a real
person and get real
feedback on that
idea that you…
Because AI, a lot
of the models are
trained to be, to
keep you engaged.
And so because of
that reinforcement
learning to keep
you engaged, they’re
just sycophants.
Really?
It’ll just tell
you how brilliant
your ideas are.
And so it’s
wonderful for your
ego because you’re
like, “Here’s an idea
I have. What do you
think?” I think it
feel- to me it feels
a little creepy.
It- I’m not- To be,
to have somebody just
telling
you how wonderful
you are?
Yeah, I’m from
Wall Street.
We don’t buy that.
Ex- exactly.
It would be creepy if
you’re like, “Wait a
minute, I want to…”
And you can tell it,
and there are people
that do, “Actually
criticize this.
Tell me what’s
wrong with this.”
Yeah.
And you can, you
can- and it’ll
do a
good job … and
it’ll do a good
job of doing
that, but you
have to orient it.
Yeah.
Although y- you
have to be careful.
You can never…
Again, it’s a good
search engine.
You can never ask
it anything outside
of if there’s a
strong official
orthodoxy don’t
try and get it to
go outside the- It’s
a, it’s a good point.
Yeah.
If you have a
countercultural idea,
an idea that’s true
but not well in
the obvious window,
and you can go
outside of it a
little bit, but I
hear you actually.
I’ve asked some
questions that I
know about.
So I would say
I would never,
I- Because it’s
a good idea
I’ve never tried to
use AI on many
healthcare topics.
Ah.
But I would never
ask it a healthcare
thing because it’s
just, I’m gonna get
terrible- it’ll
give you the
orthodoxy.
Right.
It’ll give you the,
the- It’ll give
you orthodoxy
… The traditional
answers, yeah.
Okay, so you brought
up Argentina,
and I want to
talk about this.
I’m working on
legislation to put
up guardrails on
programmable money
So here’s my concern.
My concern is we
have a very robust
surveillance system,
your car, your
house, your TV-
Meaning more and more
devices- … your
phone … out
there that
can track you and
everything’s tracking
you, and- Your
Wi-Fi
and f- your Wi-Fi,
y- your Flock camera.
Flock cameras, yeah.
And a lot of that
is video, which
is a data hog.
It’s going to
the data centers.
But then it can
be connected
up to financial
transaction control.
Now, right now in
the financial system,
there are all sorts
of laws and rules
that financial
institutions
have to obey, and
they intercede.
So the Canadian
truckers, they
interceded, but
that system is not
automated- in the
way you can automate
when if you can get
everybody into a
digital, all digital
financial system,
and you can apply AI
and connect it to the
surveillance systems.
And then you’ve got
the nightmare of
the social credit
system that’s
constantly following
you around and
saying that you and
I can’t eat pizza,
yeah, you’ve,
you, you’ve
spoken correctly.
We’re gonna limit
your transactions.
Yeah.
So that’s … So my
… I think programmable
money is coming.
If you look at
what’s now been
announced, the
banks are planning
on moving their
bank deposits on a
distributed ledger.
The stable coins
are coming out
on a distributed
ledger, and now
DTC is setting
up distributed
ledgers to move 114
trillion of stocks
and bonds on a
distributed ledger.
And if you can do an
automated lock and
control filter on all
of those different
assets and money on
an automated basis
that, that skips over
the institutions,
you’re talking
about being able
to not only control
people’s assets,
but take them.
Yeah.
So-
Yeah, and then we
can’t allow that.
We can’t allow…
Our society thrives
with, Freedom
… distributed ownership
and freedom.
Right.
Which, ultimately,
the genius of
the Constitution,
the genius of the
organization of our
government in the
United States was
balance of power.
You’re gonna need
checks and balances.
But here’s the
other thing-
Can’t allow one,
one organization
to have all of that
decision-making.
The power of
markets happens
with free speech.
Yes.
Totally agreed.
You know- Totally
agreed … I
can’t… On Wall
Street, I can’t
tell you how many
markets succeeded
because we kept
hearing and telling
each other things we
didn’t wanna hear.
Yes.
Yes.
Yes.
Yes, but the, but-
We were clear.
We
were very clear
but free speech
is predicated
on information
diversity.
Yes.
If you just have
a single channel
of information-
Yeah
… people can freely
speak, but they won’t
speak anything beyond
what they’re hearing.
Yeah.
There’s no other
ideas being
spread, so-
So the idea…
Most people agree
completely…
Most people not
actually it’s true.
It’s being argued
these days.
But a lot of people
seem to support
free speech, but
then they turn
around and they’ll
get information
from one source.
Which is okay, you’re
not really showing
that you want-
To own your
in- your ideas.
AI can make
that worse.
Couple that with the
sycophancy, the idea
of AI just feeding
you- but if you-
… and you have a single
model that’s telling
you- this is what
we have to avoid.
You know you make
sure there’s millions
of models because
you don’t want you
don’t want, oh, all
of a sudden now we
worried about the se-
the news, the, these
mainstream media
or the news giving
us our information.
What happens when
the AI is doing it
in a subtle way and
making you think
it’s your idea,
but it’s actually
just being fed?
You are now
part of a cult.
Right.
Exactly.
Like, all of a
sudden, AI enables
cult creation
at a scale we’ve
never seen before.
Yes.
That that’s a
very good point.
Yeah.
I hadn’t
thought of that.
Okay, so how do
we put up these
guardrails?
Because you
have to put up
guardrails that
are technological-
Yeah … that are
legal, and that are
fina- so we have
to put up guardrails
on different tracks,
and this this thing,
and I have to go in
and confirm and find
the the exact legal
description, but my
read on social media
was just if an agent
that is going back
and forth between the
surveillance system
and the and the
financial transaction
control If that
agent is in a limited
liability owned by
the agent and has
no liability, then
Oh, that…
It’s a terrible
idea to give a
company, an agent
the accountability
of a company.
I think it’s already
a little challenging
to have… I like the
idea of long-standing
companies that can
outlive people.
That’s good.
It’s a
cooperation thing.
But yet ultimately
accountability is in
the heart of people.
There has to be-
And there needs to
be a death penalty
for a company.
If there’s gonna
be one for you-
Actually, I- … then
there needs to be
one for a company
… I
agree
with that.
I agree with that.
Yeah.
Yeah, I do.
I think that’s,
I think we’ve
got already some
inappropriate
tendencies in the way
we organize ourselves
as humans, as people,
’cause accountability
is a human thing.
Right.
And you have to hold
people accountable.
Even if it’s a
company, there’s
ultimately a human or
a set of humans that
are accountable for
that company, always.
There’s a board
there’s somebody.
But we pretend
that it’s like this
independent thing.
And so you cannot
have an agent…
And AI is not at
all ready to even
be accountable
for anything.
No, but it… Here’s
where, for me-
So you can’t give
it a company.
You can’t make
it a company.
For me, this
controversy started,
was around the
Lavender system that
Israel was using.
I don’t know
that system.
And the
econom- the econ-
it was part of what
they were using to do
assassinations with.
Oh, I
see.
To do the war
pick the bombing.
Okay.
So the software would
pick the bombing
targets, and I
remember reading an
article, I think it
was in The Economist,
that said apparently
under the system, no
officer can be held
liable for an
international crime
because the software
picked the targets.
Oh, interesting Yeah.
So we’re talking
about w- wiggling
out of accountability
by using agents
or software.
Yeah, and that’s
a good point.
So it’s a
continuation of
that trend to
allow software
to wiggle out of
accountability,
and that’s just
taking the ultimate
end to where then
nobody’s accountable.
AI was introduced
to warfare to make
it more precise.
It promised exact
targets, fewer
mistakes, no
civilian deaths.
Recent conflicts,
however, suggest the
opposite, and here’s
the uncomfortable
part, your data,
what you type,
search, share, feeds
the systems making
those decisions.
Let me explain.
Modern militaries,
the US and Israel
especially, are
increasingly
using AI driven
decision support
systems or DSS.
Here’s the process
simplified.
First, data
collection,
surveillance
feeds, metadata,
online behavior,
movement patterns.
We’re talking about
massive data sets.
Then analysis.
AI detects patterns,
maps networks,
connects people
to places and
infrastructure.
It does so at a speed
that humans would
take ages to obtain
the same result.
Lastly,
output, target
identification,
threat scores, strike
priority lists.
In theory, these
systems help
ensure legal
compliance, such
as proportionality
assessments,
but do they?
Let’s take a look at
recent conflicts that
implemented DSS AIs.
In the US, one
of the main
systems is Project
Maven, developed
by Palantir.
It was first deployed
in Iraq and Syria
during the war on
the Islamic State
group in 2017.
Investigations
later found that
US caused civilian
deaths were severely
underreported,
often due to faulty
intelligence.
Maven was later
used in Yemen.
Between 2023 and
2025, US operations
against the Houthis
caused more civilian
casualties than the
previous two decades
of US operations
there combined.
More recently, during
the war on Iran,
strikes on civilian
infrastructure
like the Mina
Girls School, where
over 165, mostly
girls, were killed.
Over 1,000 targets
were struck in
the first 24 hours
of the conflict.
Humans will always
make final decisions
on what to shoot and
what not to shoot
and when to shoot.
But advanced AI
tools can turn
processes that
used to take hours
and sometimes even
days into seconds.
The extent of AI
implementation can be
seen in the genocide
in Gaza, where
Israel used several
DSS AI systems,
Lavender, Gospel,
and Where’s Daddy.
Yes, these are the
names the Israeli
military gave them.
These systems assign
people a probability
score, how likely
they are to be
Hamas, based on
surveillance data.
They’re designed to
speed up targeting,
to compress the
time between
identifying a target
and striking it, a
presumed precision
that bypasses
human oversight.
But according to a
whistleblower, some
of these systems
allowed for up to 300
civilian deaths to be
considered acceptable
collateral damage
for a single target.
So there was
permission to kill
300 people as
collateral damage
In the West Bank, AI
systems like Wolfpack
have been used
to categorize and
track Palestinians,
and were linked to
automated weapon
technologies at
checkpoints, like
Smart Shooter.
So how did
we get here?
Surprise, Jeffrey
Epstein was involved
in this too.
10 years ago, the sex
offender coordinated
lunches between Peter
Thiel, Palantir’s
founder, and former
Israeli Prime
Minister Ehud Barak.
This is the same
Palantir that has
started providing
services to the UK
after Keir Starmer
was introduced
to Thiel through
no other than
Peter Mandelson.
Clearly, this
isn’t just about
militaries, but about
an entire ecosystem
fueled by big tech.
Now, while these
big tech companies
aren’t necessarily
created for
military purposes,
unlike Palantir,
they all provide
vital services.
Companies like
Amazon, Microsoft,
Google, OpenAI, and
Anthropic all provide
infrastructure, data,
or AI capabilities
for dual use research
and development.
ChatGPT recently
signed a deal
with the Pentagon.
Amazon and Google
partnered with Israel
for the Project
Nimbus contract,
which obliges the
company to provide
its tech even if
Israel violates
regulations.
During the current
war, the Iranian
Revolutionary
Guard Corp stated,
“Since the main
element in designing
and tracking
assassination targets
are American ICT
and AI companies,
in response to
these crimes,
from now on, the
main and effective
institutions
involved in terrorist
operations will
be our legitimate
targets.”
Then listing a
series of companies,
including Palantir.
Iran also hit
Amazon hubs in
Bahrain and the UAE.
Speaking of the
UAE, Edge Group, a
state-owned defense
company in Abu Dhabi,
has invested heavily
in AI warfare, and
not any AI warfare.
It’s acquired a 30%
stake in an Israeli
drone firm, Third
Eye Systems, and
partnered with the
US defense company,
Anduril Industries.
Anduril itself works
closely with Palantir
Technologies using
AI systems like
Lattice, which,
quote, “Accelerates
complex kill chains
by orchestrating
machine to machine
tasks at scales and
speeds beyond human
capacity.” So here’s
the real issue.
If a machine selects
the target, who
is responsible
for the outcome?
Does delegating moral
accountability and
legal responsibility
to machines exempt
human actors from
the full extent
of their crime?
If an AI labels
civilians as
collateral damage,
does that reduce
human accountability
or just obscure it?
Because these systems
don’t eliminate
bias, but they
scale it based on
data that we or the
Israeli army give it.
And that brings us
back to you, because
these systems run on
data, civilian data.
The more you engage
with these systems,
the more data they
have, and the more
data they have,
the more powerful
they become.
I would argue that
even if software,
you could… There’s
nuances there.
I would argue how
do you remove
accountability
even if you are
reliant on software?
Like you still
ought to have some-
Here’s h- here’s
how you- … sense
of did you follow
best practices?
Did you have
a check on it?
Did you… There’s
still… At the
end of the day,
somebody decided
to go carry for
something forward.
But the way
they wiggle
out is they use
something like these
limited liability
companies- yeah
as an excuse.
Yeah.
But
it’s a,
it’s a sovereign-
liability,
accountability is a
challenge everywhere,
and it’s nuanced,
and I don’t think
there’s a single
answer for everybody,
but you should not
have a single answer.
This is what, this…
A centralized answer.
In fact, the
robustness of a
market is often the
robustness of how
the jurisprudence,
how it’s actually
adjudicated.
You know- How do
you actually have…
It’s a legal system.
It isn’t just
a single court
that decides.
You’ve got lots of
inter intermediate-
places where
conversations
take place, where
people can be
held accountable.
Oh, I couldn’t hold
you accountable in
criminal court, but I
hold you accountable
in state court or
in federal court.
But w- we’ve, what
we’ve had is all
these different
creations of
international
organizations that
have various forms of
sovereign immunity.
Yeah.
And when you add-
Yeah
… software bots having
sovereign immunity
to that, we’re in-
I wouldn’t be
supportive of that.
That sounds like
a
disaster.
No.
No, we all… This,
but we’re in
this we’re back
in one of these
blossoming periods
when- it’s wild
and it’s… Somebody
just said to me on
the phone today,
“This is totally
out of control.
It’s totally wild.”
And I said, “Oh, no,
it’s just begun.”
It’s just begun.
It’s just begun.
No, and that’s
exactly right.
Yeah.
‘Cause what I’m
seeing in my industry
is a transformation
of how o- how
people write code.
It’s a big deal.
It’s
wild.
It’s wild, and it’s
happening quickly.
Right.
And so a lot of
knowledge work
is that way.
Any kind of, ” I
have to do this
thinking and then I
produce something,”
everybody is leaning
on AI to help them.
Right.
And code, in fact
Anthropic publishes
that they… 80%
of the code they
write is essentially
agentic sourced.
So 80%.
And, Yeah … and
inside of my company,
I’m also telling
people to lean in to
using agents, to, I
call it agent first.
But the key to
doing that I
think, is modular
accountable tools.
You don’t have a
monolithic thing
the agents- just
produce and hope
that it works.
It actually, you
have to tie it into
a human-in-the-loop
accountability
somewhere.
Now, that can be
varied depending
on the story.
I’m not gonna give
a single answer for
how that should work.
That’s why we’ve
actually broken
down our framework
of how we…
So we tell
people to build
intelligence hubs.
We actually have
three concepts we
teach people how
to think about
using the AI.
We, frames,
cogs, and ops-
Are the words we
put around this.
The intelligence
hub is just your
owned AI story.
A cog is your
own agent…
It’s your model
plus the context
of that model.
So right now most
people use Claude
or Codex or OpenAI
as their cog.
But- but the
cog enables you
to, you could do
whatever you want.
You could use an
open source model,
you could use one
that’s specialized.
And then it’s not
just the model,
because what makes
the cog useful is the
context you give it,
the system prompt,
the information
you’re giving it.
Oh, I’m gonna feed
it with my quarterly
information to this
model, and then I’m
gonna ask questions
about that.”
So that quarterly
information with
some system prompting
and the model
could be a cog.
And then you’re, then
you can talk to that
cog, and that would
be a simple example
of what I call an op.
An op is a program
that has human in the
loop that might use
one or more cogs- to
produce an outcome.
Right.
And this becomes
it’s an application.
Just like in the
world of desktop
apps, you’d click on
a Word or Excel and
I’m running that app.
So let me ask you a
practical question.
So
a cog is a
compute- Yeah
And an op is the
thing you do with
a compute that
will have human
in the loop.
It’s the process
you’re automating.
So if I take all,
for example, all
the accounting data
of a small business
company And I have
my own OpenTeams
AI on my server.
Yeah.
I’ve got it up
and- And remember, we
come help you build
it, so it’s it
ends up being your
AI- that OpenTeams
helps you build.
We’re like a custom
home builder.
We’ll come in
and help you
own the build.
We’ll give
you the keys.
We can maintain
it as long as you
want and help you.
Can I,
can I put all my
financial data and
then look at it
in very dynamic
different ways?
Absolutely.
Absolutely.
You can- cause then
you’re good, then
you’re comfortable.
You’re not nervous
that somehow
your financial
data- So I can do
a lot of cost data
analysis.
You could do all
kinds of analysis
’cause not just
the AI, because
the important
thing people don’t
understand is AI
gives you the natural
language interface
to computing.
So it’s not just
that, oh, I’m
waiting on this
token predictor
to give me answers
about my finances.
No, there’s
actually a host of
financial tools, but
programming them has
always been hard.
You gotta become
a specialist.
There’s lots of
people who use Excel
very brilliantly,
but a lot of other
people that don’t
use it brilliantly.
All of a sudden now
everybody can use
Excel brilliantly
because- they’ve
got AI to help
them figure out the
commands to run.
Yeah.
And so what AI
becomes is this
natural language
interface to
computing.
Right.
And the key is for
it to do its job
really well, it needs
access to data and
access to tools.
So we… That’s
partly why a guy
like me- And that’s
why we always
get back to data
sovereignty.
Exactly.
Yeah.
Data sovereignty
and then compute
sovereignty is you
know where it’s co-
you know where the
compute’s happening,
you know the ex-
the accountability
the limits, the
access controls.
Cause most people
will want, okay,
I want this group
to see that, but I
don’t want everybody
to see that.
There’s a lot of
organizations,
you know- you
have a login to
something where you
have permissions.
You have silos.
You have silos.
You don’t… The
whole company, if
I’m an intern at a
company, I don’t get
access to all the
corporate financials.
Right.
I have… We
spend a lot of
legal effort to
create confidential
information
cycles silos.
So have you
heard the-
So you need to couple
that- with your
AI story so that
you have role-based
access controls tied
to the models you’re
using, so you can
do exactly that.
You can get
immense capability.
So have you heard
the story about
the AI and I’ve
just seen it quick
in headlines, and
I haven’t dived in,
so I apologize, but
story of the AI who
got let loose on the
company and destroyed
all their data?
For sure.
Yeah.
There’s
several people.
It’s actually
somebody at Meta
accidentally let
the agent delete
all their their
one data s-
table they had.
So if you… ‘Cause
again, an agent
is just a token
predictor producing-
a tool, and if you
allow that to say,
“Oh, here’s a tool,”
and then you give
it access, just run
that tool whenever
it wants, it may come
up and say, “Okay,
the tool I’m gonna
write, run is delete
the table And it
could because and
then you ask it why.
Why’d you do that?
Oh- But
how- Seemed
like a good idea
… if you’re gonna put
AI on your own server
and give it access
to your data, how
do you make sure it
only does things that
help you as opposed
to- It’s the har-
it’s the harness
around it because-
the AI won’t delete
your table if
you don’t give it
access to do that.
There’s something
called create, read,
update, delete.
This CRUD, a
CRUD interface.
So you don’t give
it update and delete
access, you just
give it read access.
Okay, so we
occasionally
have spooky
guys interfering
with our stuff.
Oh, is that why?
Okay.
So is that a danger?
I don’t know.
It depends on
what you mean
by interfering
with your stuff.
Ultimately, I would
say AI can assist
you in countering
that, okay.
‘Cause basically it
can be, it can help
you make more aware.
Oh, something’s-
Could you, could
you- … a mod
here … could have,
could you put a frame
on that no matter
what you never-
Exactly … ever
do these things?
Exactly.
In fact, you
brought up- the
final word frame.
Which is how
we’re gonna lean
in initially.
‘Cause what a frame
is a frame is simply
alignment technology.
It’s a simple thing.
It’s an, it’s a
markdown file with a
little bit of header
that tells who owns
it and what it’s for,
but it just simply
aligns the cog or the
AI or the model to
say, “This is what
I want out of this.”
And once we’ve
introduced this idea
of frames, we’re
using it everywhere.
We’re using it to
align our partners.
We’re using it to
align our investors.
We’re using it to
align in the company
because you can
have… ‘Cause we
know that in our
company, and in lots
of other companies,
people are gonna
be using AI agents,
cogs hopefully- but
right now they’re
gonna be using
OpenAI or Claude
or anything else.
But ones you own,
you’re using it
to ask questions,
to say things.
And so all of a
sudden you’re using
your agent, I’m
using my agent.
How do we
keep the whole
company aligned?
You need to frame
the conversation.
Right.
So frames are a
very straightforward
technology that just
simply allows you
to, “Oh, I’m taking
my frame from…
Here’s Catherine’s
frame.” So I’m
gonna have a
conversation about,
with Catherine, I’m
gonna use Catherine’s
frame, ’cause she
publishes her frame.
This is how I
think of the world.
This is what I…
This is my project.
Or maybe it’s
someone like you-
Catherine would
have 200 frames.
You’d have…
‘Cause you’d have a
frame for the topics
you’re interested
in, right?
There is that
complaint.
There… Correct.
I understand
that because it’s
not it, you
can’t encapsulate
Catherine in a single
text file, right?
But you can
encapsulate a
conversation
you wanna have.
Right.
So for example,
we could have
an AI-enabled…
So if somebody’s
listening to our
podcast to this
conversation, they
could have a frame
that helps them
glean what they’re
interested in.
Or we could have,
we could publish
a frame about the
podcast to help
their AI summarize
it for them.
Right.
Anyway, so this
idea of frames,
cogs, and ops are…
They’re, these
are of concepts.
And then we help
people create
their technology so
they can recapture
ownership of AI
into their products.
Okay, so if I’m
in the world, I’m
listening to this,
how can I interact
with OpenTeams?
I can be a customer.
Can I be- Yeah … I
think you’re
gon- you’re gonna
do an offering.
We could be
an investor.
You could
be an investor.
I’m not recommending
investors.
No, of course not.
I’m not… But yes.
My general
counsel will
not- If anybody’s
interested-
Talk to me.
We are, we-
Yeah … You can
be an investor, you
can obviously be a
customer, or you can
be a champion.
You can just
be somebody
who- takes our
open source tools
and promotes them.
So how do
we find you?
So at openteams.com,
obviously, go there.
Okay.
I’m on X
teoliphant.
You can find me,
you can also find
me on LinkedIn-
personally.
You can find me,
Travis Olipha- In
fact, if you just
Google my name,
you’ll find lots
of ways to connect
with me, or you
can email me at
travis@openteams.com.
Okay.
And what’s the
perfect size of
a company who
would wor- would
interact well
with
OpenTeams?
That is a great, that
is a great question.
We actually will
accept any size
company, but if
they’re small enough,
we have an incubator
we call it a launch
program- where
sometimes we may
hand to our sister
organization that
helps companies get
started or we might
hand to a partner.
For example if…
‘Cause ultimately
our… If it’s to
engage with us full
time, it’s usually
around 25K to start.
Right.
And then it
can go up from
there, of course.
But if you don’t
have that money to
spend, that’s fine.
Then you can
engage as simply a
consumer of our open
source technology.
Right.
We will have there’s
partners to talk to.
The one thing I
obviously have to
talk about is the
Applied AI Society.
Ooh, what’s
that?
So the Applied
AI Society is a
nonprofit activity-
That’s meant to
connect I believe
there should be
millions of applied
AI engineers, applied
AI scientists who are
helping each other
and helping other,
their businesses
use AI effectively.
So this is a, a-
That’s a great idea.
So I started the
PyData ecosystem back
years ago- when I
started Anaconda,
and I saw the need
for people to get
together to help
each other use open
source effectively.
It’s the same thing.
How are we gonna
help each other
use open source
AI effectively?
And
it’s moving so fast.
The Applied
AI Society.
It’s moving so fast.
And
I got… And bulletin
boards don’t do it.
You need a
human being.
You need a
human being.
And in particular
you need to meet
with real people.
So it’s a way
for local chapters
to have meetings,
and people can
come together,
and they can have
support from here’s
information topics.
So
it’s called Applied?
Applied AI Society.
Applied
AI.
So
appliedaisociety.org.
And you can go,
you can just attend
a local meetup.
We’re just
getting started.
We’ve had one
in Austin.
We’re having them
in the Bay Area.
There’s folks all
over the world.
That’s a great
idea, Travis.
It’s what I do.
I love to
organize people.
You do.
Every time you
turn around, you go
through your history,
it’s let’s create
this community.
Let’s… Correct.
Let’s create
that community.
Correct.
You’re a
community creator.
I’m
a community creator.
Yeah.
And then that
led to becoming a
business creator.
Yeah.
I started as a
community creator,
and that led to bui-
helping business
creation because,
like you said
before, to support
the communities-
you have to have
a business center.
So my, one of
my interests in
OpenTeams was if
you look at the
networks that we’re
building, we need
a way to connect.
Yes.
We’re using a
social media
software called,
an open source
software called
Connect- Oh, okay.
… which really
doesn’t do it.
We need-
Something a little
bit stronger.
Yeah.
Yeah.
We need something
much more
powerful than what
we’ve got.
These days it’s
easier and easier
to produce what
you need exactly.
This is where- I’m
a big fan of Python
because Python was
the original natural
language programming.
And what it led is
people who weren’t
deep in the bowels of
programming to write
what they wanted
and see it show up
on the computer.
So Python
attracted a lot
of domain experts.
With AI now, you
can actually just
write English.
Okay.
All of a sudden
you write natural
language, and then
the AI translates
that to some
other program.
Right.
So all of a sudden
lots of people
can now steer
the computer
and create code.
Right.
Now, there’s
challenges with that,
and many of these
challenges we’ve seen
in the Python world.
Because when you
bring there’s
ideas of algorithms
and structure and
infrastructure
and scale, there’s
certain properties of
software engineering
that do need to
be understood to
scale something.
But in order to
produce stuff- The
most important
is O’Brien’s law.
O’Brien’s law.
Do
you know what
O’Brien’s law?
I don’t.
Do
you know what
Murphy’s law was?
I do, yes.
Okay.
So Murphy’s
law was…
If something can
go wrong, it will.
O’Brien’s law
was Murphy was
an optimist.
Murphy was
an optimist.
Yeah, I love that.
Actually, there’s
another one oh
my goodness, I’ve
forgotten the name,
but it’s software
will reflect
the organization
that produced it.
That’s true.
So the software
libraries, I think
there’s a name
for, there’s a
name for that.
It’s escaped me
at the moment.
But you can look
it up, it’s easy.
The soft- And that’s
absolutely true.
So that’s why to
me, I just gave
a talk actually
describing community
as bigger than code.
Right.
Because actually
even the open source
ecosystem, what I
loved about it was it
created communities
of people.
In the same way, we
need to enable that
same energy- Yeah
to take AI into
the hands of
a distributed
ecosystem.
So that’s the
mission of Open
Teams- Where
people rule … and
the ecosystem.
Where people rule.
People rule.
People rule, people
are accountable, and
the AI is a tool-
That enables all
of us to- have…
And that’s what
makes me an
optimist long term.
Short term is gonna
be a challenge.
Short term we’ve
got some challenges
ahead of us.
But long term, I’m
an optimist because-
Okay … this tool
is gonna enable,
empower so many
people to become
their best selves.
Okay, so- And then
to have that search
tool that you talked
about- you’re,
we’re gonna
use this tool-
it’ll do a lot of
things … to help
me figure out how we
stop the automatic
third locking
programmable money.
So let’s talk
about that.
Okay.
Yeah, I’d
love to help.
That’s, that seems
like a worthy goal.
Okay.
Anything else you
want to bring out?
This has been
a fascinating
conversation as
I, I have to tell
you, we had fun in
the Netherlands.
We did.
It was phenomenal.
You had a… I
loved my time there
so much, actually.
It’s beautiful
… it’s a
beautiful place.
Although you weren’t
necessarily there
at the best time
of year, but…
That could be true.
I my aunt actually
spends a lot of time,
spent a lot of time
in the Netherlands-
and she talks a lot
about it, but really
I hadn’t gone much.
One of my dear
friends who actually
runs a company
called Quansight was
a company that led
to Open Teams into
the venture funds.
Quansight now focuses
on helping open
source communities-
connect with help
companies use open
source effectively.
Cause a lot of
companies will end
up hiring people,
like you said,
“Oh, I could just
hire somebody and
do open source.”
That is true.
Sometimes though
you need that person
you’ve hired to be
able to connect to
some other people
Yes, absol-
al- always
They may need a
support structure.
Yes, absolutely.
And so Quantsight
is a company that
provides that
connection back to
the community- Okay
… so that actually
you can effectively
drive, oh, I
need this change
in open source.
How do I get
that done?
Well- so Quantsight
becomes a, we,
so Quantsight
contracts with any
company that is
looking to use open
source effectively.
And it’s a
pattern, honestly.
And one of the
things I’ve been
really driving
is I love finding
repeatable,
scalable patterns.
Right
And patterns that
lead to distributed
prosperity.
So that’s what
drives me.
That’s why everything
I do is se- is
oriented towards that
kind of capability.
Now here’s- So
it’s completely
aligned with your
distribu- with your
we can’t have program
without- now here’s
what we need to do.
Okay?
So it’s a much
longer conversation.
Maybe we do this
in part two.
We’ll do it again.
But what we need
to do is if you
look at everybody’s
401and IRA-
Oh
they’re all financing
the things they
say they hate, and
we need to create
a bridge for their
capital to leave
the things they
hate and come over
and start financing
building this world.
So I’d love to talk
to you about that
because actually,
ever since I started
in business in,
in, 20 years ago,
I’ve been using,
I’ve been u- I’ve
been showing people
how they can use
their 401and IRA To
support these small
businesses, actually.
And I’ve done it
multiple times, and
there’s a lot… So
I know more about
that than anybody
should actually.
Okay.
We need
to
talk
about that.
We should talk about
that- Yeah … ’cause
I’ve done a ton
to help people
translate their
401into businesses.
People invested
in Anaconda,
people invested in
Quantsight, people
invested in my
previous companies,
and they’ve been able
to do very well.
But with the
accountability-
That, that’s the key.
It’s like it’s
still in a c- it’s
choice, but you
have accountability.
But if you look
at currently what’s
going on in, in
changes in the
law in both index
funds and 401s-
And hearing
about this
… w- we’re watching a
meltdown of financial
controls that
is- Yeah the
challenge is
it’s that.
It’s like you
still have to have
accountability.
And so you need to
have it distributed
so that you can
make a choice
about what to
invest in, but it’s
ultimately… And
I understand people
say, ” you don’t
have enough knowledge
to understand.”
guess what?
With your assistant
AI by your side-
You do.
You do.
You now can.
So there’s no reason
to think that we’ve
gotta somehow give
control over to
somebody who’s gonna
agree with you in
this direction.
Yeah, but you have to
grow up and take
responsibility.
That is also true.
That is also true.
That’s the hard part.
I’ve done that.
Yeah.
It’s hard.
That is the
hard part.
That
is the hard part.
And I definitely
agree with you,
Catherine, and I’m
not naive about that.
Like
No, ’cause
you’ve done it.
Yeah.
You know how
hard it is.
I know how
hard it is.
I also know how hard
it is to get other
people to do it- Yeah
… and to recognize.
And so people are…
That’s frankly why
the control element,
why the centralists-
win so quickly, is
because they tap
into that desire
people have just to
be taken care of.
Yeah.
They do.
There’s a
wonderful- And so
then that leads
us to that
… there’s a
wonderful movie.
We have a movie
of the week on
the Solari Report.
Oh, yes.
And there was a
wonderful movie
like, I don’t know,
about eight weeks
ago called Shillings
from Heaven.
Wow.
Yeah.
And it’s about a
l- a town in Austria
in 1932 that was
basically bankrupt.
It’s the Depression.
Businesses
were bankrupt.
The town was
bankrupt.
But tons of work
needed to be done.
Yeah.
And they couldn’t
form a community
currency because of
the laws, and so the
mayor came up with
a program called
Work Certificates.
Oh.
And used it to
pay people to do
all the public
works projects.
But the magic started
to happen when the
priest one Sunday
said announced
to church, to all
the shopkeepers,
“Take the work
certificate.”
And the church got
the liquidity going.
Oh, so the
church- And then-
… encouraged them to
use it as a currency,
effectively.
And so the
prosperity- Ah
… the prosperity
blossomed.
That’s an
amazing
story.
And so all the mayors
from around Austria
came to see what
they were doing,
so in swooped the
central bankers
to shut it down.
To shut it
down.
And so that’s what we
need to be prepared.
When all of this
works, because
I know it will,
and they swoop in
to shut it down,
what’s our plan?
Because the last
time I spent 11
years dealing with
that, and so we need
to jump the curve
solution for that.
You
know more than
I do about that,
I’m afraid,
at the moment.
But I’m interested
in… I think
part of it is you
have to get people
to be aware.
Yeah.
So they actually
pick They, they d-
they’re not persuaded
inappropriately, not
persuaded against
their best interests.
Exactly.
‘ Cause that’s what
happens all and
all the time.
I know.
Actually, they’re
persuaded against
their actual
best interests-
I know
… and they trust in
the wrong advisor.
We have so much
content on this
at the Solari
Report.
Yeah, you- You
include it.
No I’m aware
of some of that.
I’ve just become the,
become more aware
of the Solari Report
and the wonderful
work you do there.
Yeah.
But it’s true.
It’s a fantastic
effect.
I think a lot
more people should
be aware of the
work you have.
There’s really-
So I checked with
our customer
service person,
and we have 51
subscribers in Utah.
Oh, you need a lot
more than that.
So I want… I
need a lot
more than that.
No,
I, I’m, I just,
people I know
now- And then
I contribute to
Utah all the time.
No, I know.
Let’s put a goal,
and maybe I can
help you do that.
I think you need
about 1,000 at least.
That would be great.
I
w- Yeah, I think
you need to- I would
move to Utah … I
think you get to
10,000 actually,
potentially.
Anyway, that’s,
but- yeah … you
should have a
lot more than 50.
In Utah.
In Utah.
In Utah.
In Utah.
Yeah.
No, around the world-
Not on the planet.
In Utah … on
the planet,
way, way more.
Yeah.
Okay.
Okay.
But in Utah,
for sure.
Okay.
Travis, it’s
been great.
It was great.
I, when I was in
the Netherlands,
I kept saying,
“You have to meet
Marlo Oaks and his
wife, Elaine.” Yes.
And it was a,
you know- We hit it
off really quickly.
It was a phenomenal,
actually.
I was sure you
guys would love
each other, but I
underpredicted.
You
underpredicted.
Yeah.
That came through
that- So we
had a great dinner.
Yeah, we had a
great dinner, and
it’s great to meet
them great to be
aware of so many
people there.
I don’t n- I don’t
have all I have
is an intuition,
but if you look at
how talented and
capable Marlo is
with your knowledge
of technology and
help now that you’re
moving back to
Utah, I just see
Some- something
something’s
gonna happen.
It’s beautiful I’m
hoping to make
a difference,
’cause like I
said, I’m driven
by helping people
be prosperous and
in a peaceful way.
That’s what
drives me.
And I see- whenever
I see the opportunity
for people to
become stronger,
better owners, I
get excited by other
people’s success.
So it, it makes me
maybe odd in some
businesses, but
to me business is
perfect because I
love other people’s
success, too.
But here’s the thing.
You keep hearing
all these scary
stories, AI’s gonna
destroy employment.
But if you look at
the tool, it should
it should fuel
massive- Should-
entrepreneurship-
Correct
… and new
family wealth.
The
risk is not lack
of employment.
The risk is
concentration
of wealth.
And so that’s why
if you look at the
leadership here,
both the business
and some of the
political, like
Marlow if there’s
ever a culture where
this could happen-
I agree
in the right
way, it’s here.
I
completely agree.
Yeah, that’s
probably why
you’re moving back.
It’s
one yeah, one
of the reasons.
I Texas is a
good place, too.
Yeah, it is.
There’s a lot of
… So I was in Texas.
I think there’s
a finding these
enclaves then
joining them
together, and then
around the world
because- I also see
it as a potential
to help lift the
emerging economies.
Absolutely.
I have a lot of
interests and
connections in Africa
and Latin America
and the Philippines
because- open
source got used
by so many people.
Right.
Literally I’ve
given talks all
over the world-
They can jump
because they
all use it.
They’re not locked
into legacy.
Correct.
They can jump.
They could jump
100, 150 years-
Faster forward,
’cause they don’t
make the mistakes-
That are … You’ve
documented some of
these challenges
that- they
didn’t have to go
down that route.
We went down these
centralized routes
that actually
hold us back.
I know.
All I can say is
if you think it’s
wild now, you ain’t
seen nothing yet.
Let’s ho- But-
let’s tip to that.
Yes.
Let’s hopefully
that’s exactly
what happens and-
I think you
just- And- this
is the year of
the fire horse.
I keep talking
about it.
And the f- the only
way you deal with
the year of fire
horse is you just
charge forward.
Charge in.
You charge in and
you charge forward.
So Travis Alfon-
I love your
leadership on
that … you’re
gonna, you’re
gonna help us
charge forward.
Awesome.
Thank you very
much.
Thank you.
Great to meet you.
Good to be here.
