Our Guest David Weinberger Discusses
AI Has Exposed the Biggest Lie About Humanity
What if AI’s biggest impact is changing how we understand the world and the limits of our own intelligence?
Author, philosopher, and technologist David Weinberger joins Geoff Nielson to explore how artificial intelligence challenges our assumptions about knowledge, human judgment, and business leadership.
Technology is rarely just a tool. It’s a lens through which we interpret the world. David explains how machine learning and large language models find useful patterns in details we often overlook, and what it means to benefit from answers we may never fully understand.
The conversation explores AI ethics, bias, and fairness, asking who decides what these systems should optimize for. From screening job applicants to supporting everyday work, David considers where AI can help and why human values, oversight, and conversation remain essential.
He also challenges rigid five-year strategies and the idea that one leader can navigate complexity alone. Drawing on his experience at Google, David shares how open conversations and diverse perspectives can help organizations uncover ideas and respond to uncertainty.
00;00;01;21 - 00;00;02;12
You guys are ready.
00;00;02;12 - 00;00;04;13
Over here. Yes.
00;00;04;13 - 00;00;04;29
Cool.
00;00;04;29 - 00;00;05;07
by the way,
00;00;05;07 - 00;00;07;11
David, just to give you a sense of kind
of the big thing, like,
00;00;07;11 - 00;00;10;06
I try to keep this fairly unstructured,
but the big things I kind of
00;00;10;06 - 00;00;14;07
want to talk about is I'll start
with just sort of like thinking about AI
00;00;14;07 - 00;00;18;06
and how we should be
conceptualizing it, technology in general,
00;00;18;07 - 00;00;21;21
some of that kind of Heidegger
influenced stuff.
00;00;21;27 - 00;00;22;27
and then we'll talk a little bit
00;00;22;27 - 00;00;25;02
about kind of beautiful particulars,
your upcoming work,
00;00;25;02 - 00;00;27;16
and then maybe a little bit of kind
of everyday chaos and like,
00;00;27;16 - 00;00;29;19
the business implications
of all this stuff.
00;00;29;19 - 00;00;30;24
Oh my God. Okay.
00;00;30;24 - 00;00;33;24
You don't want me mentioning Heidegger,
do you?
00;00;35;00 - 00;00;38;00
AI is everywhere, rewiring everything.
00;00;38;02 - 00;00;43;05
But maybe its most fascinating magic trick
is how it's making us rethink our world
00;00;43;05 - 00;00;44;21
and ourselves.
00;00;44;21 - 00;00;48;19
My guest today is David Weinberger,
veteran technologist, author,
00;00;48;19 - 00;00;50;04
and philosopher.
00;00;50;04 - 00;00;54;13
David sees a world where complexity
is everywhere and are nowhere,
00;00;54;15 - 00;00;59;10
and a complete rethink of our relationship
with technology is absolutely critical.
00;00;59;19 - 00;01;03;08
I'm Jeff Nielsen,
and this is digital disruption.
00;01;06;21 - 00;01;07;28
David, thanks so much for joining us
today.
00;01;07;28 - 00;01;09;26
Really excited to have you on the program.
00;01;09;26 - 00;01;12;08
Maybe just to jump into things.
00;01;12;08 - 00;01;15;24
You know, you've talked a lot
about AI conceptually
00;01;15;25 - 00;01;18;25
and almost, you know, philosophically,
if I can call it that.
00;01;18;27 - 00;01;23;13
So how should we be thinking about
what AI is
00;01;23;19 - 00;01;26;19
and why does that question matter?
00;01;27;25 - 00;01;30;25
Well, it's a good question.
00;01;31;13 - 00;01;34;13
Because technology, historically,
at least in the West,
00;01;34;14 - 00;01;39;16
our dominant major technology has shaped
how we think about ourselves in the world.
00;01;39;22 - 00;01;41;14
You can go back really far.
00;01;41;14 - 00;01;41;25
Certainly.
00;01;41;25 - 00;01;45;10
Go back to the 1700s,
to the enlightenment.
00;01;45;26 - 00;01;50;12
When clocks and watches were the peak
technology, there were unbelievable.
00;01;50;14 - 00;01;51;18
And they really are.
00;01;51;18 - 00;01;54;15
I mean, the watches from back
then are mind blowing.
00;01;54;15 - 00;01;57;04
That human hands could do those things.
00;01;57;04 - 00;01;58;03
And they became the model.
00;01;58;03 - 00;01;59;16
We got the clockwork universe.
00;01;59;16 - 00;02;03;18
The universe literally, literally started
to look like a clock in that it was made
00;02;03;18 - 00;02;07;29
up of little pieces, exquisitely
engineered pieces of pretty simple.
00;02;07;29 - 00;02;10;29
You put them together,
you get a very complex mechanical thing.
00;02;11;16 - 00;02;13;24
And you can go on and on and on from that.
00;02;13;24 - 00;02;15;18
From steam power.
00;02;15;18 - 00;02;18;06
Certainly computers
where in the early 50s,
00;02;18;06 - 00;02;22;12
suddenly everything started
to look like information being processed.
00;02;22;16 - 00;02;24;07
And that is still very much with us.
00;02;24;07 - 00;02;27;17
I mean, I sometimes say, whoa, whoa,
a little slower.
00;02;27;17 - 00;02;31;17
I can't process it as if my brain
is a computer or we're very used to that.
00;02;31;18 - 00;02;34;23
We think about DNA
as information, etc., etc..
00;02;35;19 - 00;02;37;18
We haven't moved entirely away from that.
00;02;37;18 - 00;02;42;03
But that's a really that sort
of thinking changes how we do business.
00;02;42;03 - 00;02;45;03
For example, it changes how we teach.
00;02;45;03 - 00;02;48;15
I think we can talk about that,
but really, I think it's pretty obvious
00;02;48;17 - 00;02;49;22
if you think about it.
00;02;49;22 - 00;02;54;00
And then you get the web, the internet,
when everything starts to look
00;02;54;02 - 00;02;55;00
like a messy web.
00;02;57;09 - 00;02;58;08
Everything
00;02;58;08 - 00;03;01;09
suddenly, oh, actually,
we're dealing with a web here,
00;03;01;11 - 00;03;06;02
and now we have a pretty clearly
our dominant technology.
00;03;06;07 - 00;03;08;02
We're like two minutes into this new era.
00;03;08;02 - 00;03;10;06
But, I mean, it's hard.
00;03;10;06 - 00;03;13;19
I think it's hard to deny
that we're in the a age.
00;03;14;03 - 00;03;17;28
And so what I'm
because my background is in philosophy,
00;03;18;25 - 00;03;21;29
as well as having been in tech
and writing about tech for 40 years.
00;03;21;29 - 00;03;24;29
But, you ever quite shake off.
00;03;26;12 - 00;03;29;12
How you were brought up, so to speak.
00;03;29;15 - 00;03;31;28
My, my interest has been in.
00;03;31;28 - 00;03;33;28
So as a citizen, I'm really interested.
00;03;33;28 - 00;03;37;16
Of course,
in what it's going to do to jobs and,
00;03;38;10 - 00;03;41;09
the questions of bias and inaccuracy,
00;03;41;09 - 00;03;45;02
etc.,
and possibly death of hands of AI laser
00;03;45;02 - 00;03;48;19
bearing robots, which I shouldn't
even make fun of at this point.
00;03;48;23 - 00;03;50;21
Of course I'm interested in that.
00;03;50;21 - 00;03;52;22
But what I'm
00;03;52;22 - 00;03;56;24
o interested in writing about what
I guess intellectually interested in.
00;03;57;25 - 00;03;59;26
Although I think it has a lot of concrete
implications,
00;03;59;26 - 00;04;04;01
is how AI may be shaping
our ideas about ourselves.
00;04;04;06 - 00;04;08;25
And that to it's really early
to be doing this, by the way.
00;04;10;09 - 00;04;13;03
Two days into the, you know, new epoch.
00;04;13;03 - 00;04;16;01
And so I,
00;04;16;01 - 00;04;19;00
I don't know what's going to happen,
but nobody does.
00;04;19;10 - 00;04;23;23
Literally it can't be
we can't know what's going to happen.
00;04;23;23 - 00;04;25;06
It's going to be the result of everything.
00;04;25;06 - 00;04;29;07
It's, you know, the universe
and our lives are chaotic in that way.
00;04;29;19 - 00;04;31;22
Little changes can have gigantic effects.
00;04;31;22 - 00;04;34;01
So it would be foolish to say,
yeah, here's what's going to happen.
00;04;34;01 - 00;04;36;24
And I don't do that and I won't do that.
00;04;36;24 - 00;04;40;18
But I'm interested at poking at it
and thinking about how it might change
00;04;42;12 - 00;04;45;04
now, how we think about things
like or sad things,
00;04;45;04 - 00;04;49;06
So how are you again, given that we're
that we're two minutes into this thing
00;04;49;06 - 00;04;52;20
and it's all evolving
and that, you know, I'm sure
00;04;53;19 - 00;04;56;08
assuming we make it to 30
or 50 years from now,
00;04;56;08 - 00;04;59;23
we'll have a better rearview mirror
than we do windshield right now.
00;05;00;14 - 00;05;05;00
How are you already starting to see it
change our our sense
00;05;05;00 - 00;05;08;08
of like our understanding of ourselves
and our understanding of the universe?
00;05;08;09 - 00;05;12;05
Like, what are some of the patterns
or some of the particulars about AI
00;05;12;05 - 00;05;16;05
and this era
that seem to be fundamentally shifting
00;05;16;05 - 00;05;19;05
from where we were a few years ago?
00;05;19;16 - 00;05;21;03
if you're looking for it,
00;05;21;03 - 00;05;24;07
and especially if you're sort of involved
in this world and,
00;05;24;28 - 00;05;27;18
one way or another, it's important to you.
00;05;27;18 - 00;05;29;21
And, you know, you read it about it,
you think about it.
00;05;29;21 - 00;05;32;16
There's so many interesting things
to think about.
00;05;32;16 - 00;05;34;23
Then, at least for me.
00;05;34;23 - 00;05;38;18
And I think for I imagine for many of us,
we're already starting to pick up
00;05;38;18 - 00;05;43;29
on the vocabulary and language counts
for a lot and many of the concepts.
00;05;44;00 - 00;05;47;28
So, this is not scientific
research of any form, but I've noticed,
00;05;48;11 - 00;05;52;23
you know, just anecdotally,
more people are using the word optimize.
00;05;53;16 - 00;05;55;13
We're not optimized for this.
Going to optimize for that.
00;05;55;13 - 00;05;59;13
Well, that's a big change in how we
of course, we've always had that term.
00;05;59;13 - 00;06;01;28
You've thought about it that way.
But when you're optimizing
00;06;01;28 - 00;06;04;24
you are acknowledging
it's not going to be perfect.
00;06;04;24 - 00;06;06;08
You're going to have to make some choices.
00;06;06;08 - 00;06;09;02
It's built
into the notion of optimizing is
00;06;09;02 - 00;06;12;21
there are multiple ways
to optimize things, different values.
00;06;12;21 - 00;06;17;03
And you you got to pick some
and you recognize
00;06;17;03 - 00;06;18;23
or other values
you're also going to have to meet.
00;06;18;23 - 00;06;21;12
And that's why the optimal point
is some type of balancing of those.
00;06;21;12 - 00;06;26;29
And that's very different from a computer
era way of thinking, which is no
00;06;27;00 - 00;06;31;26
put in the information we'll get
get an output input output unambiguous.
00;06;32;28 - 00;06;34;23
We'll
get you know, we'll get an output from it.
00;06;34;23 - 00;06;35;25
No we don't. We get choices.
00;06;35;25 - 00;06;39;13
We get a possibility space,
if you will, possibility spaces,
00;06;39;26 - 00;06;41;28
a term that hasn't arisen as much.
00;06;41;28 - 00;06;44;28
But the idea. I think has
00;06;45;03 - 00;06;48;03
is also part of how
00;06;48;11 - 00;06;50;27
to understand machine learning systems,
00;06;50;27 - 00;06;53;27
the sort of AI that people generally
mean these days.
00;06;55;04 - 00;06;57;09
That's much more real
00;06;57;09 - 00;06;58;20
in my point of view,
from my point of view,
00;06;58;20 - 00;07;01;13
a much more realistic way
to think about things.
00;07;01;13 - 00;07;04;13
Okay, it's we can't have everything.
00;07;05;03 - 00;07;09;01
The world is not set up so that there is
a perfect that satisfies all needs.
00;07;09;01 - 00;07;12;15
It's going to have to be to have to choose
what to optimize
00;07;12;15 - 00;07;14;27
and have a serious discussion
about our values.
00;07;14;27 - 00;07;16;25
What is it that matters?
00;07;16;25 - 00;07;18;14
I think that's a great step forward.
00;07;19;27 - 00;07;22;06
I've noticed it.
00;07;22;06 - 00;07;24;02
We have grandchildren,
00;07;24;02 - 00;07;27;02
one of whom is now four.
00;07;27;09 - 00;07;28;06
And we see them a lot.
00;07;28;06 - 00;07;30;07
We're completely,
00;07;30;07 - 00;07;33;01
blessed to met.
00;07;33;01 - 00;07;34;23
And I found myself looking at how
00;07;34;23 - 00;07;37;23
she was acquiring language from
00;07;39;04 - 00;07;40;18
what I think of as an
00;07;40;18 - 00;07;43;18
AI point of view,
a large language point of view.
00;07;43;19 - 00;07;45;06
So she started.
00;07;45;06 - 00;07;49;18
She's exceptionally bright,
of course, and cute as a button.
00;07;49;27 - 00;07;51;21
She is.
00;07;51;21 - 00;07;55;13
And so I noticed that she was starting
to use words like yet.
00;07;56;08 - 00;08;00;21
And if I were to ask you to define yet,
you certainly could do it.
00;08;00;21 - 00;08;03;02
You probably stumble for a moment
because, okay.
00;08;03;02 - 00;08;04;02
And then you'd get it.
00;08;04;02 - 00;08;06;08
But you did not learn
that word from the definition.
00;08;06;08 - 00;08;09;06
And it is certainly the case that the
three and a half year old didn't learn.
00;08;09;06 - 00;08;12;09
No, they'll be sat down and said,
you know the meaning of the word yet?
00;08;12;11 - 00;08;14;03
Well,
let me tell you yet. It's very handy word.
00;08;14;03 - 00;08;17;07
It means something that hasn't happened
but you think might happen in the future.
00;08;17;10 - 00;08;18;00
Got it.
00;08;18;00 - 00;08;19;01
That did not happen.
00;08;19;01 - 00;08;22;07
That would be a hilarious conversation to
have over the three and a half year old.
00;08;22;13 - 00;08;24;27
Somehow she got it.
00;08;24;27 - 00;08;26;21
She got it right.
00;08;26;21 - 00;08;28;09
And she did that
00;08;28;09 - 00;08;32;00
because she had lots of examples
and she noticed she didn't notice.
00;08;32;00 - 00;08;37;04
But connections were being formed
about where this word connects to reality,
00;08;37;05 - 00;08;39;01
what she's trying to say,
and connects to other words
00;08;39;01 - 00;08;41;12
because she puts it in the right place
as well.
00;08;41;12 - 00;08;44;07
I mean, she's learning like a machine
learning system did.
00;08;44;07 - 00;08;46;16
Now, that's not to say the machine
learning system.
00;08;46;16 - 00;08;51;24
The large language model works
exactly the way that we do that
00;08;51;24 - 00;08;55;26
she does,
or that it's nearly as cute as she is.
00;08;55;26 - 00;09;00;17
And I should love it
like I love my granddaughter, but I.
00;09;01;16 - 00;09;02;25
I notice lots of ways
00;09;02;25 - 00;09;05;25
in which I see things in
00;09;06;18 - 00;09;09;22
large language
model terms or machine learning terms,
00;09;10;25 - 00;09;13;17
and in business this would change.
00;09;13;17 - 00;09;17;05
I think how we think about knowledge
00;09;17;05 - 00;09;20;06
arises, develops and is handled.
00;09;20;14 - 00;09;23;14
It changes ideas about basically rule
based
00;09;24;15 - 00;09;27;20
systems and especially including rule
based management.
00;09;29;09 - 00;09;33;19
I talked to a lot of CIOs, and
everybody knows that running a successful
00;09;33;19 - 00;09;37;16
IT shop is extremely hard,
and AI has only made it worse.
00;09;37;21 - 00;09;38;25
You know the drill.
00;09;38;25 - 00;09;43;29
Immature processes, a fragmented data
ecosystem and an unlimited demand for it.
00;09;44;04 - 00;09;46;12
It all makes for an impossible job.
00;09;46;12 - 00;09;50;09
And yet nobody has invented
a unified operating system for it.
00;09;50;16 - 00;09;53;28
We're literally the last function
without one until now.
00;09;54;10 - 00;09;58;06
That's why Infotech created CIO analytics,
a single platform
00;09;58;06 - 00;09;59;24
to run the IT department.
00;09;59;24 - 00;10;02;07
We call it a pie driven advisory.
00;10;02;07 - 00;10;05;20
It's powered by your data,
our analyst insight,
00;10;05;20 - 00;10;08;21
and it works with your existing tools
right out of the box.
00;10;08;27 - 00;10;12;27
If you're interested, click on the link
in the description to request access
00;10;12;27 - 00;10;15;27
and check out CIO Analytics in Action
for yourself.
00;10;16;01 - 00;10;18;08
An Infotech team lead can get you started.
00;10;18;08 - 00;10;19;10
Same day.
00;10;19;10 - 00;10;23;04
Better context, better insight, better
IT performance
00;10;23;07 - 00;10;25;09
only with API driven advisory.
00;10;25;09 - 00;10;27;02
Only with infotech.
00;10;28;25 - 00;10;31;25
preamble you gave to this is
00;10;31;25 - 00;10;35;24
really interesting to me,
because I found myself in the same place.
00;10;35;24 - 00;10;37;16
By the way,
I've got a five year old and a two year
00;10;37;16 - 00;10;40;18
old I like am comparing their development
to AI all the time.
00;10;40;18 - 00;10;44;12
I've started to think about, you know,
now that we have artificial intelligence,
00;10;44;12 - 00;10;47;12
I'm starting to think
about the human brain more and how
00;10;47;23 - 00;10;50;18
AI is like us, or we're like AI.
00;10;50;18 - 00;10;52;07
But the reason
I'm thinking about the preamble
00;10;52;07 - 00;10;54;29
is because you said,
and it makes complete sense.
00;10;54;29 - 00;10;58;23
This is not the first time
we've had a new technology,
00;10;58;24 - 00;11;02;24
and it's colored the way we see ourselves
and we see the world, right?
00;11;02;26 - 00;11;07;12
And yet we fall into this trap of,
oh, now I understand myself
00;11;07;12 - 00;11;08;13
in a different light.
00;11;08;13 - 00;11;10;16
Now I understand the world
in a different light.
00;11;10;16 - 00;11;15;04
And there's this,
you know, kind of proclivity
00;11;15;04 - 00;11;18;08
to just assume that it's true that we now
have a deeper understanding and we've
00;11;18;26 - 00;11;21;24
figured it out
now we understand intelligence.
00;11;21;24 - 00;11;23;26
But it's funny because, you know,
00;11;23;26 - 00;11;27;26
if you rewind the clock 200 or 300 years,
we would have said, oh, it's mechanical,
00;11;27;27 - 00;11;31;07
the brain is mechanical,
the universe is mechanical, it's all cogs.
00;11;31;16 - 00;11;33;04
And this keeps happening.
00;11;33;04 - 00;11;37;13
And so the reason I bring that up
is, is to what degree is this?
00;11;37;19 - 00;11;41;03
You know, there are actually some
underlying truth here versus it's useful.
00;11;41;10 - 00;11;45;18
And why does that distinction
actually matter
00;11;45;26 - 00;11;49;04
in terms of our interaction
with ourselves in the world?
00;11;51;01 - 00;11;53;13
That is a great question,
00;11;53;13 - 00;11;58;11
which is how podcast guests gain
a little time to think about an answer.
00;11;58;26 - 00;12;03;13
So it is a great question, and I'm going
to make a case that this one is different,
00;12;03;22 - 00;12;07;23
but I've made the case for prior
changes like this.
00;12;07;24 - 00;12;08;27
And this one is different.
00;12;08;27 - 00;12;12;14
including PCs but certainly the web.
00;12;13;06 - 00;12;16;24
So here's
what maybe makes this distinction.
00;12;17;02 - 00;12;20;26
The other ways in which technology
has affected our understanding,
00;12;20;26 - 00;12;23;26
going back at least to,
00;12;23;29 - 00;12;25;26
the clockwork universe.
00;12;25;26 - 00;12;26;26
And you take it back further.
00;12;26;26 - 00;12;31;11
But, is that those ways of thinking
were reductive.
00;12;31;28 - 00;12;34;28
That is, they discovered
00;12;35;10 - 00;12;37;07
and reduction
is a very useful tool. Right.
00;12;37;07 - 00;12;40;23
But they discovered laws, and the laws
00;12;40;23 - 00;12;43;26
are the same for everybody,
so that's great.
00;12;44;06 - 00;12;47;24
It's not Aristotle
by any means who thought things
00;12;47;24 - 00;12;51;06
interacted based on their properties
and they're all different.
00;12;52;04 - 00;12;55;09
But it is
modern science, its beginning of science.
00;12;55;10 - 00;12;58;20
So here, laws that apply to everything,
everything that has mass.
00;12;58;22 - 00;13;02;06
Newton's gravitational attraction
equation works for
00;13;03;05 - 00;13;06;05
and in computing this to jump ahead,
00;13;07;23 - 00;13;10;17
likewise,
you have one program, and the program
00;13;10;17 - 00;13;13;06
can handle all the different sorts of data
that goes in.
00;13;13;06 - 00;13;18;15
And that's a it can only do that
because it is reducing the data to what
00;13;18;15 - 00;13;21;15
it has in common with the program
00;13;21;15 - 00;13;27;10
has been programed to discover and process
the way it is different,
00;13;27;27 - 00;13;31;26
because the web is the idea
that everything is a messy web, which
00;13;31;26 - 00;13;36;11
I think we got from the internet age,
and I think we're carrying forward,
00;13;37;29 - 00;13;39;17
is not reductive.
00;13;39;17 - 00;13;42;27
It messiness is the opposite of reduction.
00;13;42;27 - 00;13;45;14
Messiness
says anything that's there that connects
00;13;47;06 - 00;13;49;13
because mess has to be connected
one way or another.
00;13;49;13 - 00;13;51;03
Right?
00;13;51;03 - 00;13;52;01
That matters.
00;13;52;01 - 00;13;52;17
The whole
00;13;52;17 - 00;13;56;05
the whole thing depends upon these messy
relationships that cannot be pinned down.
00;13;56;05 - 00;13;59;16
Their changing their manifold,
their differing degrees and topics.
00;13;59;16 - 00;14;01;12
I mean, that's the web.
00;14;01;12 - 00;14;04;29
And we started thinking about the world
that way in many, in many instances.
00;14;05;07 - 00;14;08;04
And I think that I think
because it's not reductive,
00;14;08;04 - 00;14;11;04
it's a more realistic,
00;14;11;23 - 00;14;14;24
you know, our brain
seems to like those generalizations.
00;14;14;24 - 00;14;18;18
They simplify thinking to a point
where it's not just a little chaos
00;14;18;19 - 00;14;21;19
of things happening.
00;14;21;25 - 00;14;23;18
And AI,
00;14;23;18 - 00;14;26;13
machine learning in particular, because
00;14;26;13 - 00;14;27;15
I'm going to stop saying that.
00;14;27;15 - 00;14;29;20
I'm just going to say AI and you'll
00;14;29;20 - 00;14;33;05
because there's lots of types of
AI that were attempts
00;14;33;07 - 00;14;37;05
or at least some some were were attempts
to try to get a general picture
00;14;38;11 - 00;14;39;23
machine learning AI.
00;14;39;23 - 00;14;43;19
The breakthrough was let's
and this is a Lee
00;14;44;14 - 00;14;48;19
researcher in 2007 or so said,
00;14;50;09 - 00;14;51;20
you know AI
00;14;51;20 - 00;14;55;06
usually promising very, very,
very important.
00;14;55;06 - 00;14;58;01
But it's going slower than one would think
it should.
00;14;58;01 - 00;14;59;29
It's another,
00;14;59;29 - 00;15;03;05
type of AI winter,
as these lapses recalled.
00;15;03;13 - 00;15;06;04
And she thought
00;15;06;04 - 00;15;10;03
maybe the issue is
that we got all these genius level,
00;15;10;15 - 00;15;14;03
you know, people who are trying
to write the right algorithms
00;15;14;12 - 00;15;17;12
to get sort of thinking machines
or the next advance.
00;15;17;17 - 00;15;19;15
They're not paying attention to the data.
00;15;19;15 - 00;15;23;06
Maybe we should start with the data,
not with the algorithms.
00;15;23;22 - 00;15;26;18
And that insight
00;15;26;18 - 00;15;28;20
set off, I don't know.
00;15;28;20 - 00;15;31;26
I'm now oversimplifying, of course,
but that was a crucial insight.
00;15;31;26 - 00;15;34;19
And she set up,
00;15;34;19 - 00;15;37;13
the data that that could easily be used,
00;15;38;12 - 00;15;40;14
which was also
00;15;40;14 - 00;15;42;11
a big and helpful step.
00;15;42;11 - 00;15;45;01
So the idea is that
00;15;45;01 - 00;15;50;00
I think a
can I try to explain AI in my terms?
00;15;50;09 - 00;15;52;16
Yeah. Of course.
00;15;52;16 - 00;15;53;03
Be brief.
00;15;53;03 - 00;15;56;03
I'll be very general because,
00;15;57;10 - 00;15;58;12
I'm a writer.
00;15;58;12 - 00;16;01;14
I've worked in a AI companies.
00;16;02;15 - 00;16;03;27
But what do I know?
00;16;03;27 - 00;16;06;27
So here's how I think about it.
00;16;10;03 - 00;16;12;17
So a classic example.
00;16;12;17 - 00;16;15;28
You want a computer to be able
to recognize handwritten numerals.
00;16;16;10 - 00;16;19;09
This was especially important back
when people wrote checks
00;16;19;17 - 00;16;22;21
because, well,
and the checks were famously messy.
00;16;23;14 - 00;16;26;01
So the old way of doing this, of doing
00;16;26;01 - 00;16;29;13
this would be to figure out what the rules
are for each of the numerals.
00;16;29;13 - 00;16;32;13
So one straight line, eight
two circles, etc..
00;16;32;20 - 00;16;34;15
Just what your second grade teacher
or first grade
00;16;34;15 - 00;16;37;11
teacher taught you about the numerals.
But then you look at.
00;16;37;11 - 00;16;40;04
So that's a rules based way.
And if you can get it to work okay.
00;16;41;16 - 00;16;42;06
But if
00;16;42;06 - 00;16;46;03
you look at samples, actual samples
of how people write numerals
00;16;46;15 - 00;16;49;10
and people have collected these,
00;16;49;10 - 00;16;51;04
you look at them, you'll
00;16;51;04 - 00;16;54;12
finding a one that's an upright is rare.
00;16;54;16 - 00;16;58;25
Finding an egg, that's two,
two circles, one smaller, centered.
00;16;59;07 - 00;17;01;28
Almost never.
00;17;01;28 - 00;17;03;07
It just doesn't happen.
00;17;03;07 - 00;17;07;16
So if you instead gather up these examples
and you have a computer,
00;17;08;00 - 00;17;12;04
try to figure out what are the patterns
in the pixels that compose those images
00;17;12;10 - 00;17;15;10
that differentiate
00;17;15;11 - 00;17;18;27
a two from an 8 or 3 from an eight,
because those three and eight
00;17;18;27 - 00;17;22;22
can get pretty hairy
sometimes to distinguish.
00;17;23;06 - 00;17;25;25
And so you have the system
00;17;25;25 - 00;17;28;24
look for pixel patterns,
00;17;28;24 - 00;17;31;14
doesn't care if they make sense or not,
if it helps it
00;17;31;14 - 00;17;34;19
to distinguish those two numerals
or all the numerals, then
00;17;35;21 - 00;17;36;26
you've had a big advance.
00;17;36;26 - 00;17;38;01
And that's what happened.
00;17;38;01 - 00;17;42;27
Except for things
way more complex than than numerals.
00;17;43;19 - 00;17;47;11
One more sentence for me,
the most striking thing about that,
00;17;47;11 - 00;17;51;28
the thing that I'm really is at the bottom
of what I think
00;17;51;29 - 00;17;55;11
is this revolution in our mindset
00;17;55;28 - 00;17;58;28
is that it's not looking for for
00;17;59;02 - 00;18;02;16
it's not looking for what eights have in
common or ones have in common.
00;18;02;16 - 00;18;04;24
That's the old rule based way.
That's what rules want.
00;18;04;24 - 00;18;08;10
They want things that get clustered
because they're saying
00;18;09;14 - 00;18;12;21
machine learning
is looking for what makes differences.
00;18;12;26 - 00;18;15;26
How do you distinguish
an eight from a three?
00;18;16;08 - 00;18;17;15
That's what it's going to do.
00;18;17;15 - 00;18;20;04
It's going to come up with the patterns
that do that.
00;18;20;04 - 00;18;22;18
And so it's really looking at details.
00;18;22;18 - 00;18;26;00
It is not trying to
to find their similarities.
00;18;26;00 - 00;18;28;17
It's trying to find their
differentiations, which is a
00;18;29;29 - 00;18;33;00
a very
different way of thinking about the world,
00;18;33;09 - 00;18;38;08
because we've all been brought up thinking
that truth is in the similarities.
00;18;38;20 - 00;18;42;22
The truth is in the laws that express
a similarity across all things.
00;18;42;25 - 00;18;47;03
Truth is in the categories and words
that we use for things.
00;18;47;19 - 00;18;48;15
What is a vehicle?
00;18;48;15 - 00;18;51;07
Well, we have a definition gets fuzzy,
but we make up our mind.
00;18;51;07 - 00;18;54;06
That's where truth is. That's
where law is.
00;18;54;08 - 00;18;55;29
An AI says, yeah, okay, that's fine.
00;18;55;29 - 00;18;59;16
But if you do what I just described
and you look for the differences,
00;18;59;16 - 00;19;02;18
suddenly knowledge increases.
00;19;02;21 - 00;19;04;21
We know more.
00;19;04;21 - 00;19;08;15
We've been bound
by our limitation of our minds,
00;19;08;25 - 00;19;12;26
which are tiny, you know, 2
pounds or whatever it is of brain
00;19;13;09 - 00;19;16;19
to try to understand a vastly,
vastly complex universe,
00;19;16;22 - 00;19;20;03
which we've done by
and very successfully, obviously,
00;19;20;09 - 00;19;24;03
by reducing things to what they have
in common, looking for laws, looking for,
00;19;24;08 - 00;19;28;06
you know, for categories
and essences of things and the like.
00;19;28;06 - 00;19;29;04
And now we have a tool.
00;19;29;04 - 00;19;33;02
Let's just go ahead and do that of course,
but also has remarkable breakthroughs
00;19;33;04 - 00;19;36;23
like we've never ever seen before
because it's looking at differences.
00;19;38;24 - 00;19;41;21
Thank you and good night.
00;19;41;21 - 00;19;43;22
So what makes it's.
00;19;43;22 - 00;19;44;14
Really interesting.
00;19;44;14 - 00;19;48;07
And I feel like we've only just sort of
scratched the surface of that idea,
00;19;48;09 - 00;19;49;10
the idea of differences.
00;19;49;10 - 00;19;53;02
But maybe this may be too heady
a question, but like, what's the
00;19;53;18 - 00;19;57;08
what's the difference
between similarities and differences?
00;19;57;08 - 00;19;59;15
Because they almost feel like two sides
of the same coin.
00;19;59;15 - 00;20;00;11
If I ask
00;20;00;11 - 00;20;04;20
what makes these eight different versus
what makes them the same as each other?
00;20;05;18 - 00;20;08;03
It feels like I'm asking the same thing
different ways,
00;20;08;03 - 00;20;11;15
but maybe there's something distinct
that I'm not getting at their.
00;20;12;04 - 00;20;15;24
No there is truth in what what you say.
00;20;19;27 - 00;20;24;05
Because one could restate
what I just said about eights
00;20;24;06 - 00;20;28;01
and say no, it's looking for the patterns
that make eights, and those patterns
00;20;28;08 - 00;20;31;26
make all these things the same,
except that they're not all the same.
00;20;33;08 - 00;20;34;17
Look at how we write them.
00;20;34;17 - 00;20;37;19
Or even just look different
fonts, you know, for that matter.
00;20;38;28 - 00;20;40;22
The the important.
00;20;40;22 - 00;20;43;22
So there is
you can flip it and look at it that way.
00;20;44;28 - 00;20;47;19
But the notion that we should the
00;20;47;19 - 00;20;51;26
what makes something what it is, is it's
differences from everything else is Right
00;20;51;28 - 00;20;55;18
the face of 20 least 2500 years of Western
00;20;56;28 - 00;20;58;29
culture, which was all about
00;20;58;29 - 00;21;04;07
very explicitly finding
we don't talk about essences much anymore,
00;21;04;07 - 00;21;09;04
but we did for, you know, more than 2000
years, what exactly?
00;21;09;05 - 00;21;11;00
Because everything has to be for some.
00;21;11;00 - 00;21;11;29
We have it in our head.
00;21;11;29 - 00;21;16;22
We've had it in our heads
since ancient Greeks that things half
00;21;16;23 - 00;21;20;06
the realest of things
is what they essentially are.
00;21;20;13 - 00;21;22;17
And that essence
is the same for all things.
00;21;22;17 - 00;21;25;05
So you end up reducing
things that are different.
00;21;26;03 - 00;21;28;19
You forget, you ignore their differences,
and we only look
00;21;28;19 - 00;21;32;28
at what they have in common and say that's
what in essence really is.
00;21;33;00 - 00;21;36;10
And we have expended a huge amount of time
and energy over
00;21;36;10 - 00;21;39;29
course of thousands of years
trying to figure out nailing them down.
00;21;41;01 - 00;21;42;20
There's value to that.
00;21;42;20 - 00;21;44;17
But you were also throwing out
00;21;44;17 - 00;21;47;17
all the information that makes one thing
different from another.
00;21;48;05 - 00;21;50;29
Laws do this all gravitational law.
00;21;50;29 - 00;21;53;11
I'm not going to argue with Newton.
00;21;53;11 - 00;21;56;05
There is a limit.
00;21;56;05 - 00;21;57;18
One of the two greatest geniuses
00;21;57;18 - 00;22;00;20
in Western history.
00;22;02;02 - 00;22;04;01
Aristotle one.
00;22;04;01 - 00;22;04;22
Yeah.
00;22;04;22 - 00;22;07;22
Cheap opinion.
00;22;08;05 - 00;22;10;03
His law of gravitation.
00;22;10;03 - 00;22;12;02
Absolutely. Of course he nails it.
00;22;12;02 - 00;22;14;18
He got it right.
It's the mass of the thing.
00;22;14;18 - 00;22;19;13
When you're thinking about things as mass,
you are reducing them to the single
00;22;19;29 - 00;22;22;29
most elemental thing that
00;22;23;03 - 00;22;26;26
everything physical thing has in common,
hearing off everything else.
00;22;26;26 - 00;22;29;00
And that's
the power of the gravitational law.
00;22;29;00 - 00;22;34;05
It also works
in very specific circumstances
00;22;34;21 - 00;22;39;03
when there's, for example, in space,
when the gravitational effects
00;22;39;03 - 00;22;43;04
of the rest of the universe aren't don't
matter to what you're trying to measure,
00;22;43;06 - 00;22;44;22
although they're still there.
00;22;44;22 - 00;22;47;10
And so there's a certain abstraction
of view even.
00;22;47;10 - 00;22;50;21
And just saying,
well, here's here's how gravity works.
00;22;50;22 - 00;22;53;15
Yes, but that law won't work.
00;22;53;15 - 00;22;56;16
We hope that law won't give us a precise
and accurate answer
00;22;56;16 - 00;23;01;01
anywhere in the universe, because universe
consists of mass all over the place,
00;23;01;02 - 00;23;04;02
and all of
it has some degree of attraction.
00;23;04;18 - 00;23;07;18
So it's vastly simplifying,
extremely useful.
00;23;08;08 - 00;23;11;08
Now we don't have to
because we have machines,
00;23;11;09 - 00;23;14;19
can look at the particulars
and details of individual things
00;23;15;21 - 00;23;16;08
and use
00;23;16;08 - 00;23;19;08
that to advance knowledge
at the expense of human
00;23;19;08 - 00;23;22;23
understanding, because it doesn't care
whether we understand the patterns.
00;23;22;23 - 00;23;27;06
It only cares if the patterns help it do
the job we've given it, which is to,
00;23;27;11 - 00;23;30;06
you know, sort the good eggs
00;23;30;06 - 00;23;33;06
from bad eggs
or whatever it is that we've given that,
00;23;34;20 - 00;23;36;24
Digital Disruption has partnered
with McLean
00;23;36;24 - 00;23;39;24
and company
to bring you the human centric AI program.
00;23;39;24 - 00;23;41;00
If you're a knowledge worker,
00;23;41;00 - 00;23;44;01
you've probably noticed
that your colleagues are all using AI,
00;23;44;10 - 00;23;47;24
and it's not always pretty,
but it's a solvable problem.
00;23;47;26 - 00;23;50;11
It's not a tools problem,
it's a people problem.
00;23;50;11 - 00;23;52;23
And yet we keep training on tools.
00;23;52;23 - 00;23;56;07
McLennan Companies
Human centric AI program is a judgment
00;23;56;07 - 00;24;00;05
first learning experience built to drive
AI impact at scale.
00;24;00;16 - 00;24;04;00
Three self-paced learning paths,
plus live, expert
00;24;04;00 - 00;24;07;04
led sessions and content
that evolves as AI does.
00;24;07;11 - 00;24;10;24
Better judgment,
better outputs, better outcomes.
00;24;10;29 - 00;24;13;04
Check it out
at the link below to learn more.
00;24;14;25 - 00;24;18;12
that framing
of you don't need to understand it.
00;24;18;19 - 00;24;19;19
These are the patterns.
00;24;19;19 - 00;24;24;21
It's true that that feels like it's
one of the, I guess, conceptual
00;24;24;21 - 00;24;29;14
sea changes that we're dealing with
or even coping with right now.
00;24;29;14 - 00;24;32;02
what to
you are kind of the implications of that.
00;24;32;02 - 00;24;36;28
I'm deliberately framing it as fairly
seismic in terms of our understanding.
00;24;36;29 - 00;24;38;04
Would you agree with that?
00;24;38;04 - 00;24;41;10
And like,
what does that mean for us mere mortals?
00;24;44;04 - 00;24;46;09
I agree with.
00;24;46;09 - 00;24;47;01
I agree with you.
00;24;47;01 - 00;24;49;02
I think it is an ethical change.
00;24;49;02 - 00;24;53;29
That's why I think this is a bigger change
than most of the other revolutions
00;24;53;29 - 00;24;57;07
that we've been through,
including clockwork universe,
00;24;58;02 - 00;25;01;12
which reinforced the idea of simple rules
governing everything,
00;25;02;13 - 00;25;05;22
through computing,
through the webbing of the web.
00;25;06;11 - 00;25;10;03
Because this one decisively,
I think, I think empirically,
00;25;10;03 - 00;25;14;04
we have evidence that machines
00;25;14;13 - 00;25;18;19
that don't know anything but work
because they're not conscious
00;25;18;27 - 00;25;22;21
but are able to handle gigantic,
00;25;23;27 - 00;25;25;03
inconceivable amounts
00;25;25;03 - 00;25;28;03
of particulars of data,
00;25;28;05 - 00;25;30;16
routinely come up with
00;25;30;16 - 00;25;33;06
better knowledge than we do,
00;25;33;06 - 00;25;36;09
can can figure out things
we can't figure out
00;25;37;19 - 00;25;39;27
whether it's protein folding
00;25;39;27 - 00;25;42;26
or speed up those processes,
00;25;43;00 - 00;25;45;08
weather patterns and the rest.
00;25;45;08 - 00;25;47;18
Basically, everything on the surface
of the Earth affects the weather,
00;25;47;18 - 00;25;51;14
and AI isn't gathering all of that,
but it's gathering more than we did.
00;25;51;17 - 00;25;54;25
And it advances because it is not limited
by human understanding,
00;25;54;26 - 00;25;58;18
because the patterns express
the individuality of events.
00;25;58;27 - 00;26;01;29
if we if this is what we pick up from AI,
00;26;02;27 - 00;26;05;24
that I think
00;26;05;24 - 00;26;09;09
in some ways I think
it may be a great lifting of a burden.
00;26;09;10 - 00;26;10;10
I mean, science will continue.
00;26;10;10 - 00;26;13;13
We'll keep on trying to understand things
and good and support it more
00;26;13;13 - 00;26;18;14
and put more money into it and stop being
such jerks about it, about its outcomes.
00;26;18;22 - 00;26;21;10
I absolutely this is not an either or.
00;26;22;26 - 00;26;23;18
Science is.
00;26;23;18 - 00;26;26;18
Yeah. But if we
00;26;26;21 - 00;26;31;08
recognize that
and that we benefit from acknowledging
00;26;32;01 - 00;26;34;10
the finitude of our of our minds,
00;26;34;10 - 00;26;39;25
that changes the picture of the universe
for us and relieves us
00;26;39;25 - 00;26;42;25
of a tremendous burden
00;26;43;03 - 00;26;46;05
of thinking that we are
the rational animals, our essence.
00;26;46;17 - 00;26;47;28
I mean, literally,
this is what we thought,
00;26;47;28 - 00;26;49;07
you know, for thousands of years
00;26;49;07 - 00;26;51;28
we thought our essence was
we are the rational animals.
00;26;51;28 - 00;26;52;13
Okay, well,
00;26;52;13 - 00;26;55;14
we are rational and we are animals,
but there are a lot of other things.
00;26;55;17 - 00;26;58;17
Our destiny is to understand the universe.
00;26;59;29 - 00;27;02;08
Wow. Is that a big destiny?
00;27;02;08 - 00;27;04;27
And I think there may be some relief
from saying,
00;27;04;27 - 00;27;08;01
actually,
the universe is so much more complicated.
00;27;08;28 - 00;27;11;09
We can't do that. We can do other things.
00;27;11;09 - 00;27;15;17
that's not for us
and shouldn't feel bad about it.
00;27;15;17 - 00;27;17;11
That we don't understand the universe yet.
00;27;17;11 - 00;27;19;07
But someday we will, because we won't.
00;27;19;07 - 00;27;21;07
Not in the sense
in which we want to, I don't think.
00;27;23;21 - 00;27;24;29
There's
00;27;24;29 - 00;27;28;23
I don't know, there's like a
there's a sadness to that though as well.
00;27;28;23 - 00;27;30;10
And, and the
00;27;30;10 - 00;27;31;27
I guess the imagery
that was coming to mind
00;27;31;27 - 00;27;34;09
while you were talking about that,
and maybe this is too esoteric,
00;27;34;09 - 00;27;37;08
but it reminded me of A Hitchhiker's
Guide to the Galaxy and like,
00;27;37;10 - 00;27;39;05
you know, the answer machine
where it's like, you know,
00;27;39;05 - 00;27;42;06
what's the you know, you know,
what's the meaning of life?
00;27;42;07 - 00;27;43;06
Yeah. 42 right.
00;27;43;06 - 00;27;47;05
But is that is
is that the direction we're heading?
00;27;47;05 - 00;27;49;10
And is that a good or a bad thing?
00;27;49;10 - 00;27;54;02
Because the implication I'm hearing
is we have limits as humans.
00;27;54;02 - 00;27;58;05
So what we have to do now is feed
the machine a pile of data
00;27;58;05 - 00;27;59;19
and ask for the answer
00;27;59;19 - 00;28;02;23
and trust in the machine
that it's giving us the right answer.
00;28;02;28 - 00;28;07;00
And maybe that's right,
but it certainly feels,
00;28;07;05 - 00;28;10;24
I don't know, somewhere
between scary and dangerous. So.
00;28;10;24 - 00;28;13;18
So let me frame it up to you this way.
00;28;13;18 - 00;28;15;19
Is that where we're going?
00;28;15;19 - 00;28;20;20
And is there some nuance around that
that we need to be conscious of
00;28;21;02 - 00;28;24;15
if we're going to actually integrate
00;28;24;15 - 00;28;27;22
this as a tool for human
understanding and decision making.
00;28;30;02 - 00;28;33;02
Let me give you an example.
00;28;34;02 - 00;28;39;14
A. Because I just want to say again
quickly,
00;28;39;18 - 00;28;42;29
we will I'm pretty convinced
we will continue trying
00;28;42;29 - 00;28;46;20
to understand the universe in our terms,
which requires reducing things,
00;28;47;16 - 00;28;49;27
and helpful and true ways.
00;28;49;27 - 00;28;54;15
I hope we will, however, now also
recognize that while we're doing that,
00;28;54;20 - 00;28;55;19
we can't do it all.
00;28;55;19 - 00;29;00;11
We're the world is essentially chaotic
and unpredictable by human minds.
00;29;00;11 - 00;29;03;14
We do better predicting it
by looking at patterns and
00;29;04;14 - 00;29;07;14
and even in a flicker patterns
in the data.
00;29;07;19 - 00;29;10;19
So I'm hoping
00;29;10;25 - 00;29;12;07
no idea if I'm right.
00;29;12;07 - 00;29;16;28
It's a hope that if this becomes
our way of thinking, the way that,
00;29;16;29 - 00;29;20;00
you know, processing information,
we saw that everywhere.
00;29;22;22 - 00;29;25;22
That it will.
00;29;26;07 - 00;29;26;28
Change your mind
00;29;26;28 - 00;29;29;28
about things like fairness
and I hope in better ways.
00;29;30;16 - 00;29;33;16
So I picked fairness because it is
00;29;33;29 - 00;29;38;17
the simplest of it's
such a simple moral idea
00;29;39;07 - 00;29;42;22
that it has not been taken
seriously very often
00;29;42;22 - 00;29;44;18
in the entire history of Western
philosophy.
00;29;44;18 - 00;29;48;04
Sometimes, yes, but morality and justice
00;29;48;27 - 00;29;51;17
have been worked over a lot.
00;29;51;17 - 00;29;52;21
Fairness seems too simple.
00;29;52;21 - 00;29;55;01
You can explain it to a four year old.
00;29;55;01 - 00;29;56;04
So when,
00;29;57;23 - 00;29;58;10
our four year
00;29;58;10 - 00;30;02;10
old granddaughter is at the eight year
olds birthday party, and the eight year
00;30;02;10 - 00;30;05;10
old gets a bigger piece
and she gets a little rupees, she will.
00;30;05;11 - 00;30;09;01
This is not based on a life experience,
but it probably could have been.
00;30;09;18 - 00;30;14;00
She probably would say,
why did I get such a small piece?
00;30;14;00 - 00;30;17;24
My sister got a big one,
and you likely will say to her,
00;30;18;25 - 00;30;21;01
well,
00;30;21;01 - 00;30;23;12
sweetie,
00;30;23;12 - 00;30;27;21
your sister is bigger than you, and,
you know, bigger people get more food.
00;30;27;21 - 00;30;30;28
But, it's her birthday,
so she's special today.
00;30;31;10 - 00;30;34;04
And you have just explained
fairness to her.
00;30;34;04 - 00;30;38;05
You haven't done it by saying abstractly.
00;30;39;04 - 00;30;42;00
Well, fairness means treating people
exactly the same.
00;30;42;00 - 00;30;44;09
Unless there's a relevant difference.
00;30;44;09 - 00;30;46;03
It's basically what fairness is.
00;30;46;03 - 00;30;47;06
But that's what you did.
00;30;47;06 - 00;30;50;29
You gave her two differences
and she had some type of intuition
00;30;50;29 - 00;30;53;18
or she didn't and still,
you know, was angry.
00;30;53;18 - 00;30;55;11
And you probably give her
a little more cake.
00;30;55;11 - 00;30;58;19
Anyway, fairness is incredibly simple.
00;31;00;16 - 00;31;04;03
AI. I think if we if we take up,
00;31;04;18 - 00;31;07;27
if we spend more time paying attention
to the particulars and seeing them
00;31;07;27 - 00;31;11;24
as actually incredibly important
as opposed to the general rule,
00;31;12;00 - 00;31;13;28
then we look at fairness
a little bit differently.
00;31;13;28 - 00;31;15;28
The case of the cake is an easy one.
00;31;15;28 - 00;31;18;28
In real life, you know, a pick one.
00;31;25;06 - 00;31;27;01
I can go back to the 60s.
00;31;27;01 - 00;31;31;22
We're still arguing over 1960s laws
governing fairness about Di.
00;31;32;22 - 00;31;33;10
Yeah, sure.
00;31;33;10 - 00;31;34;28
Treat everybody exactly the same.
00;31;34;28 - 00;31;35;29
Oh, okay.
00;31;35;29 - 00;31;36;26
Unless it's.
00;31;36;26 - 00;31;39;04
But what is the relevant difference here?
00;31;39;04 - 00;31;43;15
You will start arguing
about the particulars of the case.
00;31;43;28 - 00;31;46;28
Say it's a higher in your office.
00;31;47;18 - 00;31;50;00
Then the discussion is going to be about
00;31;50;00 - 00;31;53;00
for a particular person.
00;31;53;05 - 00;31;54;06
Is going to be about
00;31;54;06 - 00;31;58;07
the importance in your organization
of having more diversity.
00;31;59;14 - 00;32;02;12
It's going to be about the person as well.
00;32;02;12 - 00;32;04;13
Of course, it's
going to be very particular.
00;32;04;13 - 00;32;06;08
That's
what you're going to be arguing about.
00;32;06;08 - 00;32;09;24
That's good,
because fairness is really complicated.
00;32;10;05 - 00;32;13;10
Unless you're talking about a four year
old birthday cake slice.
00;32;13;21 - 00;32;17;04
And if we grow to accept that, then we see
00;32;17;11 - 00;32;21;07
fairness as a conversation
about what you want to optimize for.
00;32;21;19 - 00;32;24;08
If I can go back to that,
you can think about it
00;32;24;08 - 00;32;27;08
very clearly in those terms, I think.
00;32;27;08 - 00;32;32;19
And that's good because it sort of gives
validity to the particulars
00;32;32;20 - 00;32;36;14
and doesn't try to sweep them all
under a rug of generalization.
00;32;39;07 - 00;32;39;28
there's there's.
00;32;39;28 - 00;32;40;27
A challenge here.
00;32;40;27 - 00;32;43;15
And, you know,
you mentioned that you mentioned the word
00;32;43;15 - 00;32;45;07
bias earlier as one of the,
00;32;45;07 - 00;32;48;26
you know, kind of armchair
things you're thinking about with AI is
00;32;49;14 - 00;32;52;23
when we use the phrase
relevant particulars
00;32;52;28 - 00;32;56;21
or relevant details,
relevance is now doing an awful lot of
00;32;56;23 - 00;33;00;25
heavy lifting, because who gets to decide
what's relevant and what's not relevant,
00;33;00;25 - 00;33;03;10
and who gets to decide
what we're optimizing for.
00;33;03;10 - 00;33;06;27
And suddenly this optimization
and this answer machine
00;33;06;27 - 00;33;11;19
is only as good at answering
what we think it should be answering.
00;33;11;19 - 00;33;17;22
And it seems to make the question
extremely, extremely important.
00;33;17;22 - 00;33;22;21
And having some sort of,
you know, common understanding and values
00;33;22;21 - 00;33;25;22
for the question, common values for,
00;33;26;17 - 00;33;29;19
you know, the actual model itself
and what we choose to,
00;33;30;02 - 00;33;33;12
you know, emphasize in the training data
and the inference and all the kind of
00;33;33;13 - 00;33;35;00
the nitty gritty stuff. So
00;33;36;15 - 00;33;37;29
what do we what do we.
00;33;37;29 - 00;33;38;25
Do with all that?
00;33;38;25 - 00;33;42;15
And how like, is that
a surmountable problem?
00;33;42;15 - 00;33;45;22
Or is that just a limitation of trying
00;33;45;22 - 00;33;49;13
to create some sort of universal machine
00;33;49;27 - 00;33;52;27
in a very messy world?
00;33;55;13 - 00;33;56;17
So there's a very bad
00;33;56;17 - 00;34;00;16
outcome of this,
which would be for us to say,
00;34;00;26 - 00;34;03;26
yeah, fairness,
issues of morality, fairness,
00;34;04;00 - 00;34;05;21
what our company ought to be,
00;34;05;21 - 00;34;08;12
what the vision for our company
should be, or what our strategy should be
00;34;08;12 - 00;34;10;12
just to feed into the machine,
will stick things out.
00;34;10;12 - 00;34;13;12
That's the computer error,
00;34;14;01 - 00;34;16;21
error, error.
00;34;16;21 - 00;34;19;10
And it was computers in the 1950s
when they were introduced,
00;34;19;10 - 00;34;22;17
were ridiculed, feared and ridiculed
for just that reason.
00;34;23;04 - 00;34;23;28
There's a fate.
00;34;23;28 - 00;34;26;04
It should be a famous
movie called Desk Set.
00;34;26;04 - 00;34;28;12
Katharine Hepburn. And what's his name?
00;34;28;12 - 00;34;30;05
Sorry. Old age.
00;34;30;05 - 00;34;32;22
Got memory holes all over the place
for names in particular,
00;34;34;01 - 00;34;36;22
in which the guy is a computer expert.
00;34;36;22 - 00;34;39;12
He goes into the into the company there.
00;34;39;12 - 00;34;43;08
A company has a great set of librarians,
all of whom are women,
00;34;43;27 - 00;34;46;27
not which was of the time.
00;34;47;07 - 00;34;50;07
And so there is a very clear,
00;34;50;28 - 00;34;53;25
discussion
about sex going on gender and sex, etc.
00;34;53;25 - 00;34;55;20
roles in this movie.
00;34;55;20 - 00;35;00;21
And the argument against the guy and
his computer is it's he's got no humanity.
00;35;00;21 - 00;35;03;16
It's just going to run some rules.
And then of course, at the end.
00;35;03;16 - 00;35;04;05
Sorry. Get away.
00;35;04;05 - 00;35;07;05
The ending though, a computer just starts
spewing out paper.
00;35;07;06 - 00;35;08;01
Chaotically.
00;35;08;01 - 00;35;10;19
And librarians win.
00;35;10;19 - 00;35;11;05
That's the.
00;35;11;05 - 00;35;12;02
That's all.
00;35;12;02 - 00;35;14;05
We may end up with that. With AI.
00;35;14;05 - 00;35;17;18
That would be a mistake
if we if if we trust it, if we think it's
00;35;17;18 - 00;35;21;22
a computer, all the fashion computer,
then yeah, that's where we end up.
00;35;22;06 - 00;35;28;23
If, however,
we recognize that for it to help us,
00;35;30;09 - 00;35;31;18
the things
that's not going to help us with
00;35;31;18 - 00;35;35;05
because they're value driven
as almost everything is,
00;35;37;09 - 00;35;38;16
but as we are,
00;35;38;16 - 00;35;41;16
let's say it's a, it's a, you're using it
00;35;41;21 - 00;35;44;29
AI for going through job applications.
00;35;46;00 - 00;35;49;17
If you notice that G seems to be all white
guys, it's
00;35;50;11 - 00;35;52;24
I don't know why I came up
with that example, but it's all white
00;35;52;24 - 00;35;56;09
guys are getting passed
through here to the interview stage.
00;35;56;18 - 00;35;57;15
Then you're going to look at it
00;35;57;15 - 00;36;00;27
and you can start thinking about what
the other factors are that it's missing,
00;36;00;27 - 00;36;04;13
and you're going to intervene
so that it will pay attention.
00;36;04;13 - 00;36;08;04
So we'll stop paying attention to
some things which are proxies for gender.
00;36;08;12 - 00;36;13;13
And you're going to say, well, actually,
we don't care that much about grad school.
00;36;13;13 - 00;36;15;26
We're really looking
for innovative thinking,
00;36;15;26 - 00;36;17;23
and we don't care where it comes from.
00;36;17;23 - 00;36;20;24
And if some grad school seem
00;36;20;25 - 00;36;24;00
to statistically increase,
you're having the right conversation.
00;36;24;00 - 00;36;27;27
If it's about what, how,
what to optimize your new car for
00;36;27;28 - 00;36;31;08
and you're trying to it's got
you got a list of 15 things in there.
00;36;31;08 - 00;36;32;07
Contradictory.
00;36;32;07 - 00;36;35;17
You can't have it
both be set for speed and for minimal
00;36;35;17 - 00;36;38;20
ecological impact, their safety, etc..
00;36;38;22 - 00;36;40;18
So you're going to have a conversation.
00;36;40;18 - 00;36;42;17
That's what I hope happens out of this,
00;36;42;17 - 00;36;45;07
that it gets turned back
to human conversation
00;36;45;07 - 00;36;48;11
that now recognize and maybe in the guise
of what we have to tell the machine.
00;36;48;11 - 00;36;51;12
But these are human values,
and that conversation
00;36;51;14 - 00;36;56;21
is a deeply human one,
and it needs to happen among people.
00;36;57;06 - 00;36;59;08
It can't happen among machines.
00;36;59;08 - 00;37;02;08
So it could go either way, but just say,
oh, let the computer do it,
00;37;02;08 - 00;37;05;12
in which case we'll get dreadful
answers sometimes.
00;37;05;24 - 00;37;09;11
Or we can recognize that
these are mostly dimensional.
00;37;10;17 - 00;37;12;00
Excuse me,
00;37;12;00 - 00;37;15;00
multidimensional, value laden decisions
00;37;15;04 - 00;37;18;04
that you need
00;37;18;11 - 00;37;20;16
need to argue and discuss.
00;37;20;16 - 00;37;22;24
I mean, ultimately, I think fairness.
00;37;22;24 - 00;37;26;13
To go back to that example,
I think I think fairness ultimately is an
00;37;27;08 - 00;37;29;26
a good faith conversational negotiation.
00;37;29;26 - 00;37;31;06
That's what it should be.
00;37;31;06 - 00;37;33;05
And when you fall back on no, it's unfair.
00;37;33;05 - 00;37;34;21
And that's all you say.
00;37;34;21 - 00;37;37;21
That's a single dimensional view of it.
00;37;37;21 - 00;37;40;14
Then everybody loses.
00;37;40;14 - 00;37;43;11
If we learn from these machines
00;37;43;11 - 00;37;45;22
that,
00;37;45;22 - 00;37;48;14
you need to pay attention
to the particulars, then
00;37;48;14 - 00;37;51;03
we're in much better shape.
00;37;51;03 - 00;37;53;22
So I want to
I want to put a fine point on this.
00;37;53;22 - 00;37;58;00
So if you're if we're using this
for decision making, let's use
00;37;58;00 - 00;38;03;11
the examples of business leaders
who are already using AI and mass for
00;38;03;21 - 00;38;06;24
God knows what, what what are like
00;38;06;26 - 00;38;09;26
some of the right ways and wrong ways
00;38;10;08 - 00;38;14;16
to use this as a new technology
based on everything you just said.
00;38;15;18 - 00;38;18;19
It's easiest when the outcome is clear.
00;38;19;07 - 00;38;22;07
So supply chain management,
00;38;24;04 - 00;38;26;10
there's basically one outcome.
00;38;26;10 - 00;38;27;02
You know what?
00;38;27;02 - 00;38;30;02
You know what the value is.
00;38;30;05 - 00;38;33;05
Although there are other values
hidden in there like,
00;38;33;18 - 00;38;36;19
non exploitative value chains.
00;38;37;03 - 00;38;41;17
You know, if you're getting
your material raw materials from slaves
00;38;41;29 - 00;38;44;24
that maybe should be factor in as, as,
00;38;44;24 - 00;38;47;22
you know, parameter that matters.
00;38;47;22 - 00;38;49;26
That aside, assuming you know,
00;38;49;26 - 00;38;54;08
that's not a computer problem, let's say
you have nice ethical supply chain then.
00;38;54;11 - 00;38;57;10
Yeah, it'll do great at that.
00;38;57;27 - 00;39;00;12
I've somebody's going to think of a way
in which that could go wrong.
00;39;00;12 - 00;39;03;02
And I'll say, yeah, no,
you're right about that.
00;39;03;02 - 00;39;06;01
if after careful testing, you're,
00;39;07;10 - 00;39;11;22
sifter for job applications
turns out to be putting forth
00;39;11;22 - 00;39;14;10
not only qualified candidates
with candidates that are not,
00;39;16;00 - 00;39;18;20
benefiting from systemic bias.
00;39;18;20 - 00;39;19;23
And this is this is a tough one.
00;39;19;23 - 00;39;23;02
I don't know if this can be done
then pure first pass,
00;39;24;24 - 00;39;26;21
of sorting
00;39;26;21 - 00;39;27;12
makes some sense.
00;39;27;12 - 00;39;30;22
I'd still be a little nervous about it
because there are so many human factors
00;39;30;22 - 00;39;34;08
involved and sociological factors.
00;39;36;04 - 00;39;42;02
The more closely tied,
where the outcome is clearer,
00;39;42;05 - 00;39;45;05
the value is clear
is that you're aiming for,
00;39;46;21 - 00;39;50;00
and the human values
00;39;50;00 - 00;39;53;06
are either relatively simple or settled.
00;39;53;16 - 00;39;55;23
You don't have to argue
about them anymore.
00;39;55;23 - 00;39;58;08
Then AI can be helpful.
00;39;58;08 - 00;40;00;00
Very helpful in medicine.
00;40;00;00 - 00;40;02;18
I mean, just talk about business.
00;40;02;18 - 00;40;06;02
But clearly many other uses
I think are much less controversial.
00;40;07;26 - 00;40;10;26
Well, it's interesting,
and I'm going to be
00;40;10;27 - 00;40;12;26
deliberately
a little bit facetious about this,
00;40;12;26 - 00;40;16;02
but what I heard
you say earlier and I'll rephrase, is
00;40;16;11 - 00;40;21;04
AI is really great for automating things
that don't require human values.
00;40;21;05 - 00;40;23;25
And by the way,
everything requires human values.
00;40;23;25 - 00;40;24;08
Yeah.
00;40;26;01 - 00;40;29;01
Well, so.
00;40;30;01 - 00;40;33;01
So let me clarify that Sure.
00;40;34;03 - 00;40;38;09
As I said, I'm like,
you know, being facetious for a fact.
00;40;38;14 - 00;40;40;10
I know,
00;40;40;10 - 00;40;42;16
you know, you're being
00;40;42;16 - 00;40;45;16
a really good question.
00;40;45;28 - 00;40;50;15
So protein
folding is the value of that is,
00;40;50;19 - 00;40;54;17
I think, pretty clear to Vance's
medical issues
00;40;54;17 - 00;40;57;17
and health issues.
00;40;57;25 - 00;41;02;02
And I that that's,
00;41;02;03 - 00;41;06;08
I think, the only human value
that's involved in it.
00;41;06;08 - 00;41;09;08
We want it to be accurate, of course, but,
00;41;11;11 - 00;41;14;17
so I don't have much trouble with that.
00;41;15;28 - 00;41;20;06
If it turns out that protein folding
somehow enables industries to do terrible,
00;41;20;07 - 00;41;25;24
terrible things, then I not going to blame
the protein folder folding.
00;41;26;03 - 00;41;29;08
I may regret that we ever did it,
but it's not the protein.
00;41;29;09 - 00;41;31;19
It's a human problem.
00;41;31;19 - 00;41;34;14
Just like a car company
trying to figure out
00;41;34;14 - 00;41;37;24
what value is going to its car
that's a human problem.
00;41;38;03 - 00;41;41;23
Or a company trying to decide
whether particular potential hire,
00;41;43;08 - 00;41;47;16
how die fits into that or shouldn't
fit into it.
00;41;47;24 - 00;41;49;15
That's not a computer problem.
00;41;49;15 - 00;41;52;15
That's a human problem.
00;41;55;16 - 00;41;58;06
You know, if
00;41;58;06 - 00;42;00;21
I think I can leave it there,
you can think of lots of
00;42;00;21 - 00;42;03;21
AI and machine learning Sure. where
00;42;04;02 - 00;42;06;02
or even
00;42;06;02 - 00;42;09;14
people are going to like this, even
having it right things for us.
00;42;09;21 - 00;42;13;22
So they're terrible examples of this
where definitely don't want it.
00;42;14;10 - 00;42;20;06
But I know certainly no managers
who have to write too many.
00;42;20;09 - 00;42;21;18
They're being exploited.
00;42;21;18 - 00;42;22;07
They have to. Right?
00;42;22;07 - 00;42;23;21
They have too many people
reporting to them,
00;42;23;21 - 00;42;27;12
and they have to write up reports
on the people of all of them.
00;42;28;02 - 00;42;30;07
And it's very hard.
00;42;30;07 - 00;42;31;07
It's very time consuming.
00;42;31;07 - 00;42;36;11
And so they speak with their,
you know, speak with their AI.
00;42;36;21 - 00;42;39;18
They tell it what they
think of the person.
00;42;39;18 - 00;42;44;08
And the AI pops out
a rough draft and the person edits it.
00;42;46;27 - 00;42;49;05
First and has control over it.
00;42;49;05 - 00;42;51;07
I know it's
hard for me to be against that.
00;42;51;07 - 00;42;52;26
You can see how it could go bad.
00;42;52;26 - 00;42;55;28
Somebody just takes it, takes it and turns
it and never reads them.
00;42;56;22 - 00;42;59;03
Likewise for homework, for that matter.
00;42;59;03 - 00;43;02;07
But it's a real aid there,
and the human ultimately
00;43;02;07 - 00;43;05;07
has complete control over what it says.
00;43;07;28 - 00;43;08;13
Right.
00;43;08;13 - 00;43;12;01
And the the operative piece being that
it's a first draft and that it's edited
00;43;12;01 - 00;43;15;01
and that there's some sort
of human judgment being applied on it.
00;43;15;04 - 00;43;16;13
in this case, yes.
00;43;16;13 - 00;43;19;13
Yes. of of sorting a
00;43;20;18 - 00;43;23;18
male chickens from female chickens.
00;43;23;21 - 00;43;26;00
Chicks.
00;43;26;00 - 00;43;30;10
We there are chicken sectors
who know how to do that.
00;43;30;11 - 00;43;33;11
And it's very important
because they grind up the male ones
00;43;33;17 - 00;43;36;17
because they not going to produce eggs
or whatever.
00;43;36;19 - 00;43;37;18
Sorry, I'm vegetarian.
00;43;37;18 - 00;43;39;23
This whole thing makes me upset.
00;43;39;23 - 00;43;42;15
So there is a long tradition,
hundreds of years of people
00;43;42;15 - 00;43;45;22
trained as six chicken sectors who can
00;43;46;22 - 00;43;49;22
look at the chicken and tell what a sex
is, and they can't tell you how.
00;43;49;22 - 00;43;52;22
And the only way you get trained
as a chicken fixer is
00;43;52;27 - 00;43;55;04
you stand next to one
and you make a judgment.
00;43;55;04 - 00;43;59;12
You say, first time you say male
and second sex or says, no, we're down.
00;43;59;12 - 00;44;01;12
Next one you say female no.
00;44;01;12 - 00;44;03;07
And you do this over and over
and over again.
00;44;03;07 - 00;44;06;08
And in a machine learning
sort of way that you cannot express,
00;44;06;11 - 00;44;10;00
you get like
the good ones are like 98% accurate.
00;44;10;05 - 00;44;12;19
If AI gets
00;44;12;19 - 00;44;15;19
gets 100%, 99% accurate.
00;44;15;23 - 00;44;18;20
I feel bad for
I do feel bad for the chickens exercise
00;44;18;20 - 00;44;21;20
because it's a long, noble profession.
00;44;21;23 - 00;44;24;00
And not a transferable skill
00;44;24;00 - 00;44;26;16
as far as I know.
00;44;26;16 - 00;44;28;24
I do, I feel bad for them, but that's,
00;44;30;15 - 00;44;33;13
AI will do that, and we won't complain.
00;44;33;13 - 00;44;35;17
I think
00;44;35;17 - 00;44;38;01
I'm chuckling
over your use of the word noble
00;44;38;01 - 00;44;40;08
to describe the chicken sector profession.
00;44;40;08 - 00;44;43;08
But we we Are you kidding me?
00;44;43;13 - 00;44;44;28
mystical. Yeah. is.
00;44;44;28 - 00;44;48;19
They do something with great accuracy,
and they cannot tell you why.
00;44;49;15 - 00;44;50;20
Right, that right.
00;44;50;20 - 00;44;55;02
I mean, it's That's it's it is interesting
and it is very kind of machine
00;44;55;02 - 00;44;56;08
learning coded.
00;44;56;08 - 00;44;58;22
There's there's something else
I wanted to press on though.
00;44;58;22 - 00;45;00;22
That seems like
it needs to be brought into the fold.
00;45;00;22 - 00;45;04;08
And you've you've mentioned it already,
which is just the,
00;45;04;26 - 00;45;07;19
the messiness and the chaos
00;45;07;19 - 00;45;10;22
that maybe was inherent all along.
00;45;10;22 - 00;45;14;13
But certainly our perception of the world
is that it's getting messier
00;45;14;13 - 00;45;15;12
and more chaotic.
00;45;15;12 - 00;45;18;09
And so I want to press on that.
00;45;18;09 - 00;45;21;28
Like, to what degree is the world
actually getting more chaotic versus
00;45;21;28 - 00;45;24;28
our understanding of it
is getting more chaotic? How
00;45;26;15 - 00;45;29;00
how is
all of that changing for us right now?
00;45;29;00 - 00;45;32;23
Where is it going
and how do you see, you know,
00;45;32;24 - 00;45;35;25
the AI era as impacting that
00;45;36;27 - 00;45;39;00
Just a simple little question. Jeff.
00;45;39;00 - 00;45;42;00
just just straightforward. Yes.
00;45;43;08 - 00;45;46;00
Oh, it's not a yes or no.
00;45;46;00 - 00;45;47;11
just groping for time.
00;45;47;11 - 00;45;49;25
So the world is certainly
getting more complex.
00;45;49;25 - 00;45;51;04
I think that's just objectively true.
00;45;51;04 - 00;45;53;28
And I don't know who would disagree with
that.
00;45;53;28 - 00;45;54;28
Would you.
00;45;54;28 - 00;45;58;08
So will you accept that I you know
00;45;59;17 - 00;46;00;12
I think so.
00;46;00;12 - 00;46;03;19
But but you know, to my point,
it seems like two things are happening
00;46;03;19 - 00;46;05;17
at once. It's getting more complex.
00;46;05;17 - 00;46;09;21
And our tools for understanding
the world are getting more complex.
00;46;09;21 - 00;46;13;21
And they're enabling us
to understand complexity now that maybe
00;46;13;22 - 00;46;15;12
we would have overlooked in the past.
00;46;15;12 - 00;46;19;18
So it's tough for me to figure out what's,
you know, in the eye of the subject
00;46;19;18 - 00;46;22;18
versus
what's truly going on in the universe.
00;46;22;27 - 00;46;23;07
Yeah.
00;46;23;07 - 00;46;26;07
And so the other part of that answer,
I think, is
00;46;26;08 - 00;46;28;14
the world is getting more
complex. And that and thus
00;46;29;21 - 00;46;31;02
I'm going to say, yeah,
00;46;31;02 - 00;46;34;02
more chaotic.
00;46;34;24 - 00;46;38;00
There's more people, more people,
more stuff, more interactions, more moving
00;46;38;00 - 00;46;42;21
parts and more interdependencies,
which is, you know, at the heart of,
00;46;44;00 - 00;46;44;19
chaos.
00;46;44;19 - 00;46;47;19
I hope and think this is, again, so
00;46;48;01 - 00;46;50;18
me and the book I have coming out,
00;46;50;18 - 00;46;53;18
I it's,
00;46;53;25 - 00;46;56;02
explicitly takes an optimistic view of how
00;46;56;02 - 00;46;59;22
AI may be changing our idea
of how the world works, in part
00;46;59;25 - 00;47;03;09
because I didn't want to be depressed
for the five years that I was writing it.
00;47;03;22 - 00;47;04;28
Since I don't know what's going to happen,
00;47;04;28 - 00;47;06;18
I could choose to say
it's all going to fall apart
00;47;06;18 - 00;47;09;26
or try to present a more positive,
hopeful thing.
00;47;10;14 - 00;47;13;15
But I also, if he made me guess,
I would guess
00;47;13;21 - 00;47;17;20
true without giving any evidence,
I think this might actually happen.
00;47;17;26 - 00;47;19;21
So it gives us a different picture.
00;47;19;21 - 00;47;22;17
Lets us acknowledge the chaos
00;47;22;17 - 00;47;27;24
that has always been with us, the ways
in which our best attempts to manage
00;47;27;24 - 00;47;31;19
the world have involved
reducing it, throwing out,
00;47;32;19 - 00;47;33;29
differences
00;47;33;29 - 00;47;38;10
among things, trying to find similarities,
facing those because we just can't
00;47;38;16 - 00;47;44;03
handle the multidimensional
differential truth and we just can't.
00;47;44;12 - 00;47;48;15
And so we have
this great shortcut of rules
00;47;48;15 - 00;47;52;13
and classes of things,
and we can it's fantastic.
00;47;52;13 - 00;47;55;17
And it's obviously insanely useful
because, you know, I
00;47;55;17 - 00;47;59;01
look at your background and we seem to be
building skyscrapers these days.
00;47;59;09 - 00;48;02;26
That's
that's rule based to a very large degree.
00;48;04;05 - 00;48;08;07
But we also know that virtually
everything in our life is chaotic.
00;48;08;07 - 00;48;11;06
Most of it doesn't matter very much
like who's going to be in the sidewalk,
00;48;11;06 - 00;48;13;21
in the crosswalk with you
when you cross the street.
00;48;13;21 - 00;48;16;25
Some point it may turn out
to be incredibly important,
00;48;17;15 - 00;48;20;15
for whatever reason, grisly or not.
00;48;20;17 - 00;48;22;01
But no, that doesn't matter.
00;48;22;01 - 00;48;24;28
But it is chaotic. If we recognize
00;48;26;10 - 00;48;27;09
during the computer age
00;48;27;09 - 00;48;28;27
everything started
to look like information,
00;48;28;27 - 00;48;31;18
which is helpful,
but misleading because it's not.
00;48;31;18 - 00;48;34;18
It's stuff.
00;48;34;19 - 00;48;38;17
If we end up looking at the world
as containing so many different
00;48;38;17 - 00;48;44;11
particulars, noting the very positive role
that particulars play in our lives.
00;48;44;22 - 00;48;48;06
So our four year old
00;48;49;06 - 00;48;52;06
granddaughter,
00;48;52;14 - 00;48;55;16
I get a little dopamine hit
just thinking about her.
00;48;55;16 - 00;48;58;12
I'm getting one right now.
00;48;58;12 - 00;49;00;16
Because of her particularity
00;49;00;16 - 00;49;03;16
is that she's not a generic granddaughter.
00;49;04;21 - 00;49;06;17
I think about her.
00;49;06;17 - 00;49;09;01
I swear, I'm getting flooded
with dopamine.
00;49;09;01 - 00;49;10;07
I'm thinking about her.
00;49;10;07 - 00;49;12;13
She has a very quiet smile.
00;49;12;13 - 00;49;15;23
She's incredibly shark for,
you know, a four year old
00;49;15;26 - 00;49;19;17
and funny and good natured
and all those things are why I love her.
00;49;19;20 - 00;49;23;09
Not. It's not a grant,
not a generic granddaughter.
00;49;23;12 - 00;49;28;17
That's an easy example, but it extends
across our entire universe of experience.
00;49;29;04 - 00;49;34;24
What matters to us, what we care about,
which is at the essence of consciousness.
00;49;34;24 - 00;49;37;27
I would say you can't be conscious
that nothing matters to you.
00;49;37;27 - 00;49;41;19
You wouldn't be able to organize anything,
so to speak, would just be
00;49;41;21 - 00;49;42;28
flickers and blobs.
00;49;43;27 - 00;49;44;19
What matters to
00;49;44;19 - 00;49;47;19
us are almost always the particulars,
00;49;48;23 - 00;49;51;23
almost always.
00;49;53;02 - 00;49;54;21
The nail sticking in your tires.
00;49;54;21 - 00;49;55;20
A particular.
00;49;55;20 - 00;49;56;19
It matters to you
00;49;56;19 - 00;49;59;06
that you've got a flat flat,
and now you have to do something about it.
00;49;59;06 - 00;50;02;12
That's pretty general
and that's important to budget that
00;50;02;12 - 00;50;05;12
nail that did it to you.
00;50;05;22 - 00;50;09;14
Now, I've lost my little rush of dopamine
because I'm thinking about flat tires.
00;50;09;22 - 00;50;12;20
But, Well, and let me,
let me maybe take it
00;50;12;20 - 00;50;15;24
in the opposite direction of dopamine
because the dopamine is great.
00;50;15;24 - 00;50;17;13
I'm thinking about
your granddaughter is great.
00;50;17;13 - 00;50;22;19
But the chaos, the messiness,
one of the implications of it to me
00;50;22;19 - 00;50;26;29
is that it seems like there's this,
like, information overload.
00;50;27;01 - 00;50;29;05
We're just exposed
00;50;29;05 - 00;50;32;27
to more and more information, more
and more choices, more and more decisions,
00;50;33;10 - 00;50;38;09
and expected to make those decisions
more and more quickly.
00;50;38;12 - 00;50;42;05
And if I can paint a little bit
of a picture here and I'm dramatically
00;50;42;05 - 00;50;45;18
oversimplifying, but,
you know, I can picture an era where,
00;50;46;12 - 00;50;49;23
you know, you used to be a farmer
and you would farm the same crop
00;50;49;23 - 00;50;51;12
for your entire adult life.
00;50;51;12 - 00;50;53;02
And, you know,
the weather would change every season
00;50;53;02 - 00;50;55;05
or what have you,
but that was your vocation.
00;50;55;05 - 00;50;58;05
Or you're a shoe cobbler and,
you know, you kind of do this for,
00;50;58;17 - 00;51;00;12
you know, 50 years or so.
00;51;00;12 - 00;51;03;27
And then, you know, we go into an era
where you build a model T
00;51;03;27 - 00;51;07;22
or a refrigerator and you refresh
the model every ten years or so.
00;51;08;05 - 00;51;12;05
And it just seemed
and now, you know, with AI, with software,
00;51;12;13 - 00;51;15;25
everything is just getting faster
and faster.
00;51;15;25 - 00;51;18;19
And there's more to consider. And
00;51;20;02 - 00;51;20;24
the reason I say
00;51;20;24 - 00;51;23;24
the opposite of dopamine
is because it feels overwhelming, right?
00;51;23;24 - 00;51;26;24
Like like the natural
human response is stress.
00;51;27;02 - 00;51;29;15
How do I do this? It's challenging.
00;51;29;15 - 00;51;31;08
And, you know,
I work with a lot of business leaders
00;51;31;08 - 00;51;34;11
and a lot of technology leaders,
and the stress is palpable.
00;51;34;13 - 00;51;38;10
Like if you're thinking
about managing technology
00;51;38;11 - 00;51;43;14
for a company or you're thinking
about making decisions, it's just it's
00;51;43;14 - 00;51;49;09
faster and faster and it's overwhelming
and it's feeling really crazy.
00;51;49;17 - 00;51;52;24
And the word that came to my mind just,
just now is, is, I guess, unsustainable.
00;51;52;25 - 00;51;55;25
Like,
can we really keep up with this pace?
00;51;56;07 - 00;52;01;07
So I don't know, I'm curious if that's
one of the things you're looking at.
00;52;01;07 - 00;52;05;03
And I guess the other piece
there is that the tools we're using
00;52;05;29 - 00;52;09;15
don't really seem like they're necessarily
keeping off side of AI, right?
00;52;09;16 - 00;52;13;11
Like the the era of having
like a five year strategy or, you know,
00;52;13;12 - 00;52;17;01
you build this product for 20 years
is out the window.
00;52;17;01 - 00;52;21;25
So so how if you're not feeling dopamine,
if you're feeling stressed, like,
00;52;21;26 - 00;52;23;01
what do you do about it.
00;52;23;01 - 00;52;25;01
Do you you talked about optimism before.
00;52;25;01 - 00;52;26;13
Do you have a message of hope
00;52;26;13 - 00;52;30;11
or guidance for these leaders
who are dealing with these challenges?
00;52;36;22 - 00;52;38;10
yeah.
00;52;38;10 - 00;52;41;10
So everything you say is true.
00;52;41;26 - 00;52;46;13
I don't know how how we come out of it,
what we evolve into.
00;52;46;16 - 00;52;49;17
It may be that my generation.
00;52;49;28 - 00;52;53;03
Certainly. Well, certainly my generation.
00;52;53;06 - 00;52;54;03
Well may not.
00;52;54;03 - 00;52;57;22
We haven't quite absorbed
the internet yet.
00;53;01;03 - 00;53;03;12
It may be that your generation
00;53;03;12 - 00;53;06;12
will be learning from the
00;53;07;01 - 00;53;10;01
the next one who are brought up on this,
00;53;10;23 - 00;53;13;13
this I.
00;53;13;13 - 00;53;16;14
So I think you are pointing
00;53;16;15 - 00;53;19;15
to a fact of reality.
00;53;19;28 - 00;53;22;28
Reality has always been like this.
00;53;24;13 - 00;53;27;13
We just have coping methods
00;53;27;16 - 00;53;30;14
for blocking out particulars.
00;53;30;14 - 00;53;33;14
The overwhelming, the chaos by,
00;53;34;26 - 00;53;37;27
learning not to pay attention to it
when it doesn't matter to us.
00;53;39;07 - 00;53;43;01
We haven't yet done that yet
with this tech because it's firing stuff.
00;53;43;10 - 00;53;45;09
The same thing was said about computers,
by the way.
00;53;45;09 - 00;53;47;05
Right? The same fear.
00;53;47;05 - 00;53;48;13
And maybe they were right.
00;53;49;11 - 00;53;52;11
We we, you and I both grew up in the.
00;53;52;15 - 00;53;53;14
Well, maybe not you.
00;53;53;14 - 00;53;55;28
Actually, I grew up in the computer error.
00;53;55;28 - 00;53;57;00
Kids today.
00;53;57;00 - 00;54;00;02
You don't know what it was like back then.
00;54;00;07 - 00;54;03;07
And absolutely,
there was a pushback against that.
00;54;05;06 - 00;54;08;03
But we got used to
00;54;08;03 - 00;54;09;18
thinking that way and dealing with it.
00;54;09;18 - 00;54;13;08
So you use five year
00;54;13;08 - 00;54;16;28
plans as an example and strategies,
and I completely agree with you.
00;54;16;28 - 00;54;19;28
But what you're pointing at is,
00;54;20;22 - 00;54;21;24
the reality of the world.
00;54;21;24 - 00;54;25;02
It's always been like it's always been way
more surprising
00;54;25;06 - 00;54;28;25
than we have been willing to acknowledge,
because it's been it's scary.
00;54;29;22 - 00;54;31;18
It's scary.
00;54;31;18 - 00;54;32;21
Still is.
00;54;32;21 - 00;54;35;00
You know, the,
00;54;35;00 - 00;54;38;00
black swan idea.
00;54;38;07 - 00;54;39;05
Sure. Everybody here knows it.
00;54;39;05 - 00;54;43;14
You know, the business
sometimes the gigantic out of the blue,
00;54;44;09 - 00;54;48;13
a big piece of damage is done
to a crucial piece of the business.
00;54;49;24 - 00;54;50;19
And that
00;54;50;19 - 00;54;53;14
that was a bestseller
should be a bestseller.
00;54;53;14 - 00;54;55;22
The question is,
why did we need to hear that?
00;54;55;22 - 00;54;58;04
Doesn't
everybody know that? That happens now?
00;54;58;04 - 00;55;01;06
The book goes on beyond pointing that out.
00;55;02;07 - 00;55;04;29
But yeah, of course,
we've always been subject to black swans
00;55;04;29 - 00;55;06;17
because life is like that.
00;55;06;17 - 00;55;10;08
But black swans are the big events
that drop in, you know, fireball hits
00;55;10;09 - 00;55;13;09
a production plant.
00;55;14;06 - 00;55;17;06
But every day is series of little tiny
00;55;17;27 - 00;55;19;03
black swans.
00;55;19;03 - 00;55;21;04
Don't know what's going to happen
when we deal with them.
00;55;21;04 - 00;55;25;29
So a five year problem with a five year
plan is not, I think, not simply
00;55;25;29 - 00;55;28;29
that the pace is picked up
and there's more chaos.
00;55;29;15 - 00;55;32;15
It's always a bad idea.
00;55;32;22 - 00;55;36;09
You know, strategy for business
was not a thing for war.
00;55;36;25 - 00;55;40;23
The idea of strategy for war
was not a thing till the 18th century,
00;55;41;12 - 00;55;43;28
when suddenly, mainly Germans wrote books
00;55;43;28 - 00;55;47;08
about here's big patterns basically.
00;55;47;08 - 00;55;49;12
And yeah, yeah,
we know that they get disrupted.
00;55;49;12 - 00;55;50;26
But this is helpful.
00;55;50;26 - 00;55;54;08
It is only surprisingly recently
00;55;54;08 - 00;55;57;20
that businesses started
thinking, oh, he's calling it strategy.
00;55;59;09 - 00;56;00;18
Peter Drucker had to change
00;56;00;18 - 00;56;04;00
the title of a book in 1960,
very early 60s,
00;56;04;22 - 00;56;07;04
because he used the word business strategy
in it.
00;56;07;04 - 00;56;09;14
And the publisher
said, no, nobody's going to go for that.
00;56;09;14 - 00;56;10;24
I mean, that's more stuff.
00;56;10;24 - 00;56;12;23
And he literally had to change
the title of his book.
00;56;12;23 - 00;56;15;13
This is Peter Crocker, of all people.
00;56;15;13 - 00;56;18;13
You think he'd be able
to push his publishers around?
00;56;19;25 - 00;56;23;15
Strategy is a utopian idea.
00;56;23;16 - 00;56;27;13
I mean, it's nice thing to think about,
but I think you do way better
00;56;27;13 - 00;56;29;16
if you stay open
to the developing particulars.
00;56;29;16 - 00;56;33;22
And that means, for example, changing,
if not the actual leadership structure,
00;56;34;06 - 00;56;35;22
how leadership functions,
00;56;35;22 - 00;56;38;22
what it means to make a decision,
and where the ideas come from
00;56;39;10 - 00;56;43;29
which often should come or would be coming
from all the ranks of the business
00;56;43;29 - 00;56;44;27
and from customers.
00;56;44;27 - 00;56;48;28
If they had a way of talking with one
another unfettered.
00;56;49;00 - 00;56;50;28
Maybe not about business all the time,
00;56;50;28 - 00;56;53;24
because that's not the only place
for business ideas come from.
00;56;53;24 - 00;56;57;06
And if there is a mechanism
for filtering those discussions
00;56;57;06 - 00;57;01;06
with permission,
of course, up into the management layer,
00;57;02;08 - 00;57;05;08
AI obviously can help the filtering part.
00;57;06;05 - 00;57;09;09
But most many businesses, most businesses
00;57;09;09 - 00;57;12;09
I think, view having a corporate wide,
00;57;14;10 - 00;57;16;14
internet where people can talk about
00;57;16;14 - 00;57;19;14
whatever they want,
even if it's not business.
00;57;21;13 - 00;57;21;29
That is.
00;57;21;29 - 00;57;23;11
That's where new ideas come from.
00;57;23;11 - 00;57;26;11
One of the most important places
to from somebody who
00;57;27;19 - 00;57;29;26
is considering something
out of out of their context,
00;57;29;26 - 00;57;31;28
trying to make sense of it,
and comes up with a new way of seeing it.
00;57;31;28 - 00;57;35;20
And then now you have a new line of
business, or you avoid a terrible mistake.
00;57;37;15 - 00;57;39;05
Decision making looks
00;57;39;05 - 00;57;42;05
really different to its
00;57;42;27 - 00;57;45;27
it feels bad because it's a,
00;57;45;29 - 00;57;48;11
it's something that
unfortunately we need to do,
00;57;48;11 - 00;57;54;02
but it means reducing possibilities
and that's both that's scary.
00;57;54;02 - 00;57;57;29
And it also goes against
how generally humans have
00;57;58;13 - 00;58;01;23
thrived by increasing possibilities.
00;58;03;05 - 00;58;04;26
So so in short, I don't
00;58;04;26 - 00;58;07;26
know how would I know peace.
00;58;09;00 - 00;58;11;23
So there's it is really interesting.
00;58;11;23 - 00;58;15;13
And there's a sort of a,
I guess a paradoxical skill there because,
00;58;16;06 - 00;58;20;19
you know, coming back to Black swans,
you're saying like there's chaos.
00;58;20;19 - 00;58;24;05
There's no there's a skill in
keeping your options open,
00;58;24;05 - 00;58;27;06
but there's also a skill in filtering out
the noise.
00;58;27;08 - 00;58;31;17
Is that fair, or is there a better way
of thinking about that sort of paradox?
00;58;32;21 - 00;58;34;11
I it's basically right.
00;58;34;11 - 00;58;37;17
But instead of skill
which implies that skills belong
00;58;37;18 - 00;58;40;18
to single people, I would say there is a
00;58;40;26 - 00;58;42;19
I tend to think in terms of conversation.
00;58;42;19 - 00;58;45;10
So there are conversations,
but there are social organizations,
00;58;45;10 - 00;58;48;09
ways of organizing socially
so that many people are heard.
00;58;49;09 - 00;58;50;02
Leaving it up to
00;58;50;02 - 00;58;53;25
a single leader
just seems to me to be nuts.
00;58;54;08 - 00;58;56;13
I don't care how good they are.
00;58;56;13 - 00;58;59;13
I take my favorite president, I,
00;58;59;24 - 00;59;03;19
I don't want that person
to be making all of this.
00;59;03;19 - 00;59;07;09
I want that person to be a networker,
to be involved in conversations
00;59;07;09 - 00;59;10;19
and be happy to hear differences
and to have a group that is talking
00;59;10;19 - 00;59;13;28
not just to him or her someday,
00;59;14;15 - 00;59;17;15
but talking amongst themselves
and reaching out.
00;59;18;14 - 00;59;21;00
It's.
00;59;21;00 - 00;59;22;15
But in the days of the internet, we would,
00;59;22;15 - 00;59;26;21
you would say, accepting the network
which reaches out all over and forever.
00;59;26;26 - 00;59;29;26
Now, in days of AI,
I will say it's about accepting
00;59;29;26 - 00;59;34;24
the particular nature of life
and cherishing that,
00;59;35;04 - 00;59;36;27
cherishing the differences.
00;59;36;27 - 00;59;40;01
And this is not a novel idea
cherishing differences, he says.
00;59;40;01 - 00;59;43;21
For now, we have a machine that works
that way and it proves its efficacy.
00;59;44;11 - 00;59;47;25
No, but the model you're describing
in some ways of having this,
00;59;47;27 - 00;59;48;16
you know, network
00;59;48;16 - 00;59;52;10
or this socialization, it's actually
in some ways against the grain right now.
00;59;52;11 - 00;59;56;22
Like, I feel like we're in a moment
culturally where we sort of worship
00;59;57;00 - 01;00;01;00
the lone leader or founder
who makes the decisions
01;00;01;00 - 01;00;04;15
themselves versus
having a network of people doing it.
01;00;04;15 - 01;00;09;26
And it gets kind of,
you know, it's talking of
01;00;09;27 - 01;00;12;28
in a kind of a denigrated way is like,
you know, decision by committee.
01;00;12;29 - 01;00;15;29
You know, it's so slow and so burdensome
01;00;16;01 - 01;00;19;01
versus
having one person who just cuts the not.
01;00;19;04 - 01;00;22;17
So how do you respond to that concern?
01;00;23;17 - 01;00;25;26
It can be,
01;00;25;26 - 01;00;28;10
a pointless, time wasting circus.
01;00;28;10 - 01;00;31;10
It can be a gesture without meaning
01;00;31;17 - 01;00;34;03
somebody, some group,
01;00;34;03 - 01;00;37;03
somebody needs to be empowered
to make the decision.
01;00;37;13 - 01;00;42;05
However, to say, to put, to think of this
in terms of one supreme leader
01;00;42;05 - 01;00;46;06
who has miraculous powers to encompass
it all and come up with a good decision
01;00;46;10 - 01;00;48;03
is insanely risky.
01;00;48;03 - 01;00;52;20
And I understand
that's how to often businesses have worked
01;00;52;28 - 01;00;58;18
and governments at times is a deep change,
but it's one I think we're ready for.
01;00;58;18 - 01;01;02;27
We have the technological infrastructure
for doing it, which is the webbing of
01;01;03;26 - 01;01;06;26
we know that people want to connect
with one another.
01;01;06;29 - 01;01;09;29
We don't have good processes for,
01;01;10;16 - 01;01;14;09
taking that sort of input
and turning it into a decision,
01;01;15;01 - 01;01;17;26
but that's something
we could do better at.
01;01;17;26 - 01;01;21;10
And someday
we will figure out how actually to do.
01;01;21;16 - 01;01;23;16
Not going to be by voting.
01;01;23;16 - 01;01;28;11
But if you're not listening
to the conversations people are having,
01;01;30;02 - 01;01;33;02
then you're depriving yourselves of the
01;01;33;03 - 01;01;37;08
the best information you can have about
how your company is doing.
01;01;37;09 - 01;01;40;05
We're failing. We're succeeding.
01;01;40;05 - 01;01;43;14
The upcoming issues that are I don't know,
I think maybe there's going to be
01;01;43;15 - 01;01;46;28
and nobody wants to raise that as an issue
formally until the issue happens.
01;01;48;05 - 01;01;48;28
That's not quite true.
01;01;48;28 - 01;01;51;21
We all stand by for the woman.
01;01;51;21 - 01;01;54;15
The hierarchical nature of these systems,
01;01;54;15 - 01;01;57;15
which is, I guess, required,
01;01;58;07 - 01;02;01;20
is such a power organization,
which it doesn't have to be,
01;02;01;20 - 01;02;05;13
but it's so infested with implicit
01;02;05;15 - 01;02;08;13
threat to your livelihood
01;02;08;13 - 01;02;11;04
that it squashes
01;02;11;04 - 01;02;15;22
real information, real values,
real possibilities.
01;02;15;22 - 01;02;21;15
It narrows the possibility
space and good decisions are not made
01;02;21;15 - 01;02;25;20
in small possibility spaces that are made
where possibilities are rich,
01;02;26;18 - 01;02;30;01
a possibility
space has lots of possibilities.
01;02;30;06 - 01;02;32;25
They are related to one another,
and they are changed
01;02;32;25 - 01;02;35;26
by every actualization,
by every decision or whatever.
01;02;35;27 - 01;02;37;07
The whole thing can change.
01;02;38;18 - 01;02;41;01
That's because that's real.
01;02;41;01 - 01;02;43;19
So optimizing chords towards that
01;02;43;19 - 01;02;46;19
I think would be really healthy.
01;02;48;08 - 01;02;49;15
and you kind of touched on this,
01;02;49;15 - 01;02;53;14
but it seems like the enemy of that
in some ways is human ego, right?
01;02;53;15 - 01;02;54;08
It's power.
01;02;54;08 - 01;02;59;07
It's the notion of people trying to
hold on to that authority themselves.
01;02;59;27 - 01;03;03;29
So and I'm curious
if you've worked in organizations
01;03;04;07 - 01;03;08;10
or seen organizations
that are actually a good example of this,
01;03;08;12 - 01;03;12;28
and you know what they look like
and how they kind of operate,
01;03;14;03 - 01;03;17;03
like optimize toward this, I guess.
01;03;17;08 - 01;03;18;26
Yeah.
01;03;18;26 - 01;03;21;26
I have
01;03;22;08 - 01;03;23;24
the very first business I ever worked
01;03;23;24 - 01;03;26;25
in, which was in the mid
90s called interleave.
01;03;27;17 - 01;03;30;06
It was an engineering
company was making the world's first word
01;03;30;06 - 01;03;32;11
processor and integrated graphical system.
01;03;32;11 - 01;03;33;25
It was fantastic software.
01;03;33;25 - 01;03;39;11
It was pre web, but very based
on the same ideas of network.
01;03;39;18 - 01;03;42;05
Fantastic engineering group.
01;03;43;17 - 01;03;44;02
Just.
01;03;44;02 - 01;03;48;05
And they had back then 1986 I guess,
01;03;49;07 - 01;03;52;07
a internet network
01;03;52;07 - 01;03;55;07
for discussion and the engineers,
01;03;57;06 - 01;03;58;18
it was open to any conversation.
01;03;58;18 - 01;04;03;06
And one of one of the some that worked was
for gamers, very early computer gamers,
01;04;04;19 - 01;04;07;19
and it was pretty active,
and a bunch of the engineers were on it.
01;04;07;25 - 01;04;10;25
And one day the company,
01;04;11;02 - 01;04;14;20
one of the things that it manufactured,
which was a faster image
01;04;16;01 - 01;04;19;01
thingy processor for printers.
01;04;19;25 - 01;04;21;20
It stopped working in some situation.
01;04;21;20 - 01;04;26;04
Nobody could figure it out
until the people in the gaming
01;04;27;20 - 01;04;30;17
discussion group, they knew a lot about
01;04;30;17 - 01;04;33;27
graphic, graphic hardware
because they were gamers.
01;04;33;29 - 01;04;34;22
And they.
01;04;34;22 - 01;04;36;14
That was their hobby.
01;04;36;14 - 01;04;38;23
They figured it out instantly.
01;04;38;23 - 01;04;41;03
This was because of side conversation.
01;04;41;03 - 01;04;42;29
Who would have predicted.
That's the point.
01;04;42;29 - 01;04;45;23
Nobody would have.
But this group solved the problem.
01;04;45;23 - 01;04;48;20
So that's a concrete, simple example
that's from 40 years ago.
01;04;48;20 - 01;04;50;22
But I've seen this over
and over and over again.
01;04;50;22 - 01;04;52;14
I'll just jump ahead.
01;04;52;14 - 01;04;55;03
I was a writer in residence at Google,
01;04;56;02 - 01;04;59;00
for four years until a couple of years ago
when I was in there.
01;04;59;00 - 01;05;02;22
People part of the time, at the time
I was in there, people and AI research
01;05;02;22 - 01;05;06;00
group, which was a lovely, lovely
01;05;06;27 - 01;05;09;27
group of people working on,
01;05;10;12 - 01;05;13;24
concrete issues of humans living with AI,
01;05;14;15 - 01;05;17;14
a very humanistic,
01;05;17;19 - 01;05;19;19
really, really smart
01;05;19;19 - 01;05;22;19
engineers, designers and others.
01;05;23;14 - 01;05;25;29
One of the reasons that group was
so this is completely obvious.
01;05;25;29 - 01;05;28;21
One of the reasons that group is
so, was so prolific is that it was
01;05;28;21 - 01;05;32;10
a completely open discussion space,
which made it a completely open
01;05;33;04 - 01;05;36;09
possibility space bounded by Google. But,
01;05;37;10 - 01;05;40;11
anything could be talked
was not ever viewed as a waste of time.
01;05;40;14 - 01;05;42;04
If you're talking about something very,
01;05;42;04 - 01;05;44;03
in my point of view,
very enlightened management.
01;05;47;22 - 01;05;50;22
This is slightly off topic, but
01;05;50;29 - 01;05;52;06
it's brief. Don't worry.
01;05;52;06 - 01;05;55;06
So I was there,
01;05;56;10 - 01;05;56;29
Hanging out,
01;05;56;29 - 01;05;59;26
passing by in a hallway,
and one of the developers
01;05;59;26 - 01;06;02;20
was calling another one over to
01;06;02;20 - 01;06;05;26
his computer, and he said to her,
he got to come over here and see this.
01;06;06;07 - 01;06;10;11
It was they were working on
embedding spaces, which is key to,
01;06;11;24 - 01;06;15;05
the modern language, AI and
01;06;18;13 - 01;06;21;21
Kind of having an old age name loss.
01;06;21;24 - 01;06;24;00
So I'll make alter.
01;06;24;00 - 01;06;26;16
You came over and he said
01;06;26;16 - 01;06;29;09
it just it just figured out prefixes.
01;06;29;09 - 01;06;32;26
I didn't, and they both started
laughing and happiness.
01;06;32;27 - 01;06;35;27
I holy they figured out prefixes.
01;06;35;28 - 01;06;38;28
I didn't tell it anything about prefixes.
01;06;39;12 - 01;06;41;15
That was a sense of joy,
01;06;41;15 - 01;06;44;15
of discovery, that is.
01;06;45;07 - 01;06;47;10
As stayed with me.
01;06;47;10 - 01;06;49;29
That's the type of that's
where I've always wanted to work.
01;06;49;29 - 01;06;55;07
I don't say that everybody's like me,
of course, but I think there is something.
01;06;55;12 - 01;06;56;28
I think there is something pretty.
01;06;56;28 - 01;06;58;15
Oh, I'm about to generalize wildly.
01;06;58;15 - 01;07;03;06
I think Youmans depend upon
and get deep pleasure
01;07;03;06 - 01;07;07;03
from living in possibility spaces
where new things happen.
01;07;08;00 - 01;07;10;04
Things are put into a new light like that.
01;07;10;04 - 01;07;12;13
Suddenly
you never thought about it that way.
01;07;13;27 - 01;07;16;06
I like that live.
01;07;16;06 - 01;07;19;17
So, David, you know, just,
you know, in the spirit of all of this,
01;07;19;21 - 01;07;20;00
you know,
01;07;20;00 - 01;07;24;07
one of the things I like to kind of wind
these conversations down on is
01;07;24;07 - 01;07;28;26
just given everything we've talked about,
which is immense, right?
01;07;28;28 - 01;07;31;23
That there's there's, you know,
there's the chaos.
01;07;31;23 - 01;07;34;19
There's
there's the need to work collaboratively.
01;07;34;19 - 01;07;36;27
There's the benefit of us all
working together.
01;07;36;27 - 01;07;39;09
There's AI in the mix.
01;07;39;09 - 01;07;43;02
You know,
if you have any sort of parting advice
01;07;43;14 - 01;07;47;23
for business and tech leaders
who are facing these challenges, you know,
01;07;47;25 - 01;07;51;06
how they should be thinking or what
they should be doing differently to,
01;07;51;24 - 01;07;56;27
you know, de-stress and to end, to thrive
in this sort of age of uncertainty.
01;07;58;24 - 01;08;01;24
other than day drinking.
01;08;01;24 - 01;08;03;16
Bring back the three martini lunch.
01;08;05;01 - 01;08;05;18
Yeah.
01;08;05;18 - 01;08;08;16
Boy, I miss them. miss them.
01;08;08;16 - 01;08;09;29
As before my time.
01;08;09;29 - 01;08;10;11
Yeah.
01;08;10;11 - 01;08;13;11
I was never part of my time.
01;08;14;18 - 01;08;16;02
machine learning, AI works.
01;08;16;02 - 01;08;17;15
I think we're all convinced of that.
01;08;17;15 - 01;08;21;07
We know it has terrible problems, and it's
often inaccurate because probabilistic.
01;08;21;28 - 01;08;24;10
Let's assume
01;08;24;10 - 01;08;28;06
we have enough evidence in our own lives
that this stuff works.
01;08;28;06 - 01;08;30;16
It's doing
stuff was never possibly before.
01;08;30;16 - 01;08;35;01
And in doing that, I think it
it tells us something about the world.
01;08;35;09 - 01;08;38;02
Because if we're not lined up
with the world in some way,
01;08;38;02 - 01;08;39;05
it wouldn't be worrying.
01;08;39;05 - 01;08;44;16
So it's telling us something
about the world itself, and it works
01;08;44;16 - 01;08;49;17
because it's able
it doesn't know about the rules,
01;08;49;27 - 01;08;54;00
but it's about to work
because it doesn't know about the rules,
01;08;54;00 - 01;08;59;15
but it does know about particulars,
impossibly large numbers of particulars
01;08;59;15 - 01;09;03;12
computed at enormous expense
and environmental damage
01;09;03;20 - 01;09;09;05
in order to get a machine
that gives us probabilistic answers
01;09;09;13 - 01;09;12;10
that are useful by
01;09;12;10 - 01;09;16;28
considering only the details,
only the particulars,
01;09;16;28 - 01;09;20;15
only the data, that here's
a metric for this thing. Now,
01;09;22;14 - 01;09;23;17
maybe in this or that
01;09;23;17 - 01;09;26;18
condition, it's all particulars.
01;09;27;14 - 01;09;30;03
It can talk to us and give us knowledge.
01;09;30;03 - 01;09;33;26
Because of that, as well as all
the nonlinguistic stuff that it does.
01;09;33;26 - 01;09;38;21
And that tells us that
particulars are in many ways constitutive
01;09;39;00 - 01;09;42;14
of the world, particulars
in their differences.
01;09;43;03 - 01;09;43;10
Right.
01;09;43;10 - 01;09;45;15
When you say, oh, no,
I'd like that particular donut.
01;09;45;15 - 01;09;49;06
What you're saying is that particular
donut is different in some meaningful way
01;09;49;06 - 01;09;49;25
from the other ones.
01;09;49;25 - 01;09;54;11
That's the one that I want particularly
is exist in their differences.
01;09;54;29 - 01;09;58;05
That is a really different way
of thinking,
01;09;58;11 - 01;10;02;12
explicitly thinking about the world, but
it seems to be a truth about the world.
01;10;02;17 - 01;10;04;17
The laws are there,
the generalizations are there.
01;10;04;17 - 01;10;07;16
We need them as well, of course.
01;10;08;16 - 01;10;11;08
But it opens our eyes to
01;10;12;08 - 01;10;15;16
when I think is magnificent particularity.
01;10;15;17 - 01;10;18;12
It helps clarify for us
01;10;18;12 - 01;10;21;03
our personal relationships, why
01;10;21;03 - 01;10;24;10
we like some music, some particular
01;10;25;16 - 01;10;29;00
song, and one particular change
in that song where you just sort of
01;10;29;01 - 01;10;31;14
get lifted up,
even though all that happened
01;10;31;14 - 01;10;34;14
was they played an unexpected
particular note.
01;10;34;20 - 01;10;40;00
This happens all the time, and it helps
explain humor, where it's the difference
01;10;40;00 - 01;10;43;15
between two things,
between the setup and the punchline.
01;10;43;27 - 01;10;46;07
If they're too close,
they're too obvious. It's not funny.
01;10;46;07 - 01;10;47;04
It's a bad joke.
01;10;47;04 - 01;10;48;22
It's too far, it's absurd.
01;10;48;22 - 01;10;51;22
IST if it's far
enough, gets to be pretty funny.
01;10;51;23 - 01;10;54;19
It takes you that second say, well,
oh, I get it.
01;10;54;19 - 01;10;58;14
Like getting it
is your love for difference.
01;10;59;15 - 01;11;02;08
It's liberating,
01;11;02;08 - 01;11;03;23
I think. I think it's a hell.
01;11;03;23 - 01;11;07;17
And being relieved of the idea
that we are the one creature
01;11;07;17 - 01;11;09;23
in the universe who can understand it.
01;11;11;01 - 01;11;14;01
I'm not saying AI can understand it,
it doesn't understand anything.
01;11;14;04 - 01;11;17;26
But that notion is so it's such a burden.
01;11;18;15 - 01;11;22;21
And if we can give up on that,
then we'll see the world more clearly.
01;11;22;21 - 01;11;24;24
I think
01;11;24;24 - 01;11;27;08
that's what I'm hopeful for.
01;11;27;08 - 01;11;30;08
If it doesn't tell us first.
01;11;30;14 - 01;11;32;16
It's a it's a really interesting idea.
01;11;32;16 - 01;11;36;18
It's it's simple,
but it's really profound,
01;11;36;28 - 01;11;40;23
I think and you know, certainly gives me
a lot to chew on in terms of
01;11;41;11 - 01;11;46;01
how I'm looking at things and,
you know, all the areas in my life,
01;11;46;01 - 01;11;49;02
both personal and professional,
where I can bring that, that lens.
01;11;49;05 - 01;11;52;06
I want to say a big thank you
for coming on to the program today.
01;11;52;06 - 01;11;55;00
You've given me a lot
to chew on, a lot to think about.
01;11;55;00 - 01;11;56;18
And I really appreciate the,
01;11;56;18 - 01;12;00;20
the interesting,
heady discussion and all your insight.
01;12;02;19 - 01;12;03;06
Thank you.
01;12;03;06 - 01;12;05;10
And I
01;12;05;10 - 01;12;06;25
appreciate the opportunity
01;12;06;25 - 01;12;11;21
to have this sort of conversation
and with you as you.
01;12;12;02 - 01;12;14;00
Those are really great questions.
01;12;14;00 - 01;12;17;00
So I have a bunch of things to go
think about.
01;12;18;06 - 01;12;22;10
Most viewers don't know this, but Digital
Disruption is developed by Infotech
01;12;22;10 - 01;12;26;14
Research Group, a leading advisor
to technology leaders around the world.
01;12;26;22 - 01;12;29;00
If that's not you, you don't need to care.
01;12;29;00 - 01;12;31;08
So skip ahead and enjoy our content.
01;12;31;08 - 01;12;33;29
But if you are a technology
leader, Infotech helps
01;12;33;29 - 01;12;37;29
IT teams get projects done faster, better,
and at a lower cost.
01;12;38;14 - 01;12;42;13
Infotech provides unlimited access
to practical tools and expert guidance
01;12;42;13 - 01;12;46;18
you need to execute at a fraction
of the cost of traditional consulting.
01;12;46;29 - 01;12;51;26
No matter the project, from AI strategy
to cybersecurity to vendor negotiation.
01;12;52;02 - 01;12;53;28
Infotech has you covered.
01;12;53;28 - 01;12;57;09
Check it out at the link below
and don't forget to like and subscribe!
The Next Industrial Revolution Is Already Here
Digital Disruption is where leaders and experts share their insights on using technology to build the organizations of the future. As intelligent technologies reshape our lives and our livelihoods, we speak with the thinkers and the doers who will help us predict and harness this disruption.
Our Guest Chris Lattner Discusses
The AI Coding Revolution Has a Dirty Secret
Chris Lattner, creator of Swift and LLVM and co-founder of Modular, joins this week’s episode to talk about coding, vibe coding, and why more code doesn’t always mean better software.
We Discuss
The Jensen Huang Scandal Is Bigger Than You Think
What happens when Big Tech depends too much on one person, and we depend too much on AI? This week we’re covering the Jensen Huang controversy, the backlash against AI data centers, and how AI coding tools could weaken critical thinking.
Our Guest David Weinberger Discusses
AI Has Exposed the Biggest Lie About Humanity
Author, philosopher, and technologist David Weinberger joins Geoff Nielson to explore how artificial intelligence challenges our assumptions about knowledge, human judgment, and business leadership.
Our Guest Jeremy Roberts Discusses
AI Data Centers in Space: Stupid or Genius
Are AI data centers in space a practical solution for the future or just an expensive fantasy?