KIT Episode 209 – AI Risk Repository, Google Monopoly, and Quantum Cryptography. Or Not.

WEBVTT

00:00:00.000 –> 00:00:06.000
From
somewhere deep in the cloud and the

00:00:06.025 –> 00:00:12.200
corners of the earth, this is the Killing
It podcast, with a focus on helping you

00:00:12.225 –> 00:00:15.660
make sense and dollars of all things IT.

00:00:15.685 –> 00:00:21.140
With your hosts, Dave Sobel,
Ryan Morris, and Karl Palachuk.

00:00:21.165 –> 00:00:29.100
Welcome, everybody, to the brand new
episode 209 of the Killing It, Killing It.

00:00:29.125 –> 00:00:31.375
Podcast.
You You guys are actually in sync.

00:00:31.400 –> 00:00:33.060
I’m not going to have
to do anything to that.

00:00:33.085 –> 00:00:34.640
This is instinctual at this point.

00:00:34.665 –> 00:00:36.100
We do it so well.

00:00:36.125 –> 00:00:37.420
Smooth professionals.

00:00:37.445 –> 00:00:40.460
Probably did it in your sleep last night.

00:00:40.485 –> 00:00:43.440
Guys, I’m going to just dive right in
because I’m excited to talk

00:00:43.465 –> 00:00:46.775
to you about our topics.
I want to warm us up a little bit.

00:00:46.800 –> 00:00:47.940
We’re coming off the Olympics.

00:00:47.965 –> 00:00:53.220
If you could actually compete at an
Olympic level, what athletic

00:00:53.245 –> 00:00:54.400
event would it be?

00:00:55.240 –> 00:00:56.980
I am super non-athletic.

00:00:57.005 –> 00:01:01.800
I’ve never actually been
able to run a block.

00:01:02.160 –> 00:01:06.700
But I think if I were in the Olympics, I
would want to do archery

00:01:06.725 –> 00:01:11.050
just because it would be fun to
actually be good at something like that.

00:01:11.080 –> 00:01:16.050
I can shoot an arrow,
I can’t hit anything.

00:01:16.080 –> 00:01:20.770
They always say, If you miss the
target, it’s not the target’s fault.

00:01:20.800 –> 00:01:24.380
That defines my business.

00:01:24.405 –> 00:01:31.620
See, and if one of the options is not
becoming a professional Olympic Olympic

00:01:31.645 –> 00:01:36.460
watcher, which, by the way, I did in the
last couple of weeks, it

00:01:36.485 –> 00:01:38.480
was very compelling stuff.

00:01:38.505 –> 00:01:42.975
And by the way, super kudos to Paris for
the way that they did the locations

00:01:43.000 –> 00:01:44.240
and the venues and everything.

00:01:44.265 –> 00:01:48.360
Seriously, that was cooler
than any other Olympics.

00:01:48.385 –> 00:01:53.140
But I would still go back to where I
am from, which is the soccer stuff.

00:01:53.165 –> 00:01:58.340
I was actually, as a teen, coming up in a
development program, Olympic track

00:01:58.365 –> 00:02:01.495
for the U23 Men’s Soccer Tournament.

00:02:01.520 –> 00:02:03.260
Now, I missed it because
I wasn’t good enough.

00:02:03.285 –> 00:02:09.880
But if I could overcome that in my youth,
one only can’t imagine what

00:02:09.905 –> 00:02:11.780
the trajectory would have been.

00:02:11.805 –> 00:02:14.700
So the Men’s Soccer team,
they need some help.

00:02:14.725 –> 00:02:17.255
Let’s just admit it.
They need some help right now.

00:02:17.280 –> 00:02:19.120
They need to take some lessons
from the women’s soccer.

00:02:19.145 –> 00:02:21.560
You might actually be the closest to being
able to compete at that level

00:02:21.585 –> 00:02:24.170
than many of the other sports.
I’m with you, Ryan.

00:02:24.200 –> 00:02:28.360
Gold Zone was the greatest
invention ever for peak.

00:02:28.385 –> 00:02:32.890
I got it like the ability to watch
match all the sport all day long.

00:02:32.920 –> 00:02:37.055
As they moved quickly around, and here’s
this one, and here’s this one, and here’s

00:02:37.080 –> 00:02:38.860
gold over here, and here’s gold over here.

00:02:38.885 –> 00:02:41.500
I was addicted, and it was too fun.

00:02:41.525 –> 00:02:45.360
But if you were actually- Oh, by the
way, it’s in front of the Eiffel Tower.

00:02:45.385 –> 00:02:46.280
It was pretty cool.

00:02:46.305 –> 00:02:50.400
When I’m picking a sport, I’m going to go
with distance cycling because

00:02:50.425 –> 00:02:52.180
that feels the most useful.

00:02:52.205 –> 00:02:58.040
Because not only could I actually do it at
that level, I could then also commute

00:02:58.065 –> 00:02:59.860
locally in really useful ways.

00:02:59.885 –> 00:03:02.695
It feels like it would
be a really cool sport.

00:03:02.720 –> 00:03:02.920
Right.

00:03:02.945 –> 00:03:05.820
You can get back to
pedaling your electric bike.

00:03:05.845 –> 00:03:07.320
I know.
Exactly.

00:03:07.345 –> 00:03:10.360
I want to take time off E-bike.

00:03:10.640 –> 00:03:15.000
And I’m pretty sure Carl’s skill is also
tactically applicable.

00:03:15.025 –> 00:03:20.460
So in a world, I’m thinking being
able to play soccer, that’s cute.

00:03:20.485 –> 00:03:23.240
But archery would get you some kudos.

00:03:23.265 –> 00:03:26.520
On occasion, you really need that skill.

00:03:26.920 –> 00:03:28.080
All righty.

00:03:28.105 –> 00:03:31.500
Well, let’s dive into we
have three topics As usual.

00:03:31.525 –> 00:03:35.775
So the first topic we’re going
to dig into today is MIT.

00:03:35.800 –> 00:03:37.340
We’re going to put a
link in the show notes.

00:03:37.365 –> 00:03:43.100
But MIT has released an AI Risk
Repository, which is

00:03:43.125 –> 00:03:50.480
literally 700 articles and published
papers on risks that you need to be

00:03:50.505 –> 00:03:53.140
aware of with artificial intelligence.

00:03:53.165 –> 00:03:59.600
The cool thing about this is if you go to
the link that we give you, it’s at MIT at

00:03:59.625 –> 00:04:06.610
the university, and they allow you to
download this as an Excel spreadsheet.

00:04:06.640 –> 00:04:13.780
So if you want to do your own analysis,
sort, filter, whatever, do counts on

00:04:13.805 –> 00:04:17.020
keywords, anything you want to do.

00:04:17.045 –> 00:04:24.160
What I love about this is it’s an academic
view of something that is what we in

00:04:24.185 –> 00:04:26.620
the nerd world would call open source.

00:04:26.645 –> 00:04:32.800
So here you have this open database,
and you can decide for yourself

00:04:32.825 –> 00:04:36.020
whether or not it’s useful.
Now, who is it useful to?

00:04:36.045 –> 00:04:41.660
Well, it is 10% of this
audience and 0% of your clients.

00:04:41.685 –> 00:04:43.160
Oh, I disagree completely.

00:04:43.185 –> 00:04:46.060
Okay, this is good when we disagree.

00:04:46.085 –> 00:04:53.440
We’re in the golden age of NIST
and academic frameworks that I feel

00:04:53.465 –> 00:04:56.420
are incredibly useful in business.

00:04:56.445 –> 00:05:02.580
Again, I’m going to fall back on my basic
concept of the true value of IT services

00:05:02.605 –> 00:05:05.860
organizations
is in their expertise in guiding

00:05:05.885 –> 00:05:07.700
customers to the correct technologies.

00:05:07.725 –> 00:05:14.400
You have more than any time ever,
academics and industry professionals who

00:05:14.425 –> 00:05:19.260
spend their entire lives
thinking about this in a

00:05:19.285 –> 00:05:24.940
conceptual sense and give you all of the
resources to be incredibly smart

00:05:24.965 –> 00:05:27.420
very quickly on a lot of topics.

00:05:27.445 –> 00:05:32.700
I look at a database like this and say,
this is exactly the useful framework

00:05:32.725 –> 00:05:36.895
that I can then take to customers
and apply when I need to consider

00:05:36.920 –> 00:05:38.060
and answer their questions.

00:05:38.085 –> 00:05:41.900
I don’t have to talk in theoretical
terms about what AI risk is.

00:05:41.925 –> 00:05:46.460
I have exact models and use cases to
compare against to understand where the

00:05:46.485 –> 00:05:51.000
smartest people around who just spit their
time thinking about this have given

00:05:51.025 –> 00:05:52.800
me all of the resources needed.

00:05:52.825 –> 00:06:01.620
You can get up to speed on practices
at an incredibly rapid rate and leverage

00:06:01.645 –> 00:06:04.140
incredible intelligence
quickly with your customers.

00:06:04.165 –> 00:06:09.080
What I want people to think is, I’m not
saying I think everyone needs to go

00:06:09.105 –> 00:06:12.860
and learn all 700 and some scenarios.

00:06:12.885 –> 00:06:17.060
It’s more the, you need to be a really
good person

00:06:17.085 –> 00:06:23.740
at matching information to your customers,
and you’re given a searchable, indexable,

00:06:23.765 –> 00:06:27.580
sortable resource that
you can use very quickly.

00:06:27.605 –> 00:06:31.580
I look at this and say, your job
is knowing all these resources.

00:06:31.605 –> 00:06:35.700
These resources are available to you, and
leveraging them makes you exceptional.

00:06:35.725 –> 00:06:38.940
To your point, Carl, is
only 10% will use it?

00:06:38.965 –> 00:06:40.880
Yeah, those are the best 10%.

00:06:42.360 –> 00:06:48.400
Well, and those are the 10% who are going
to get paid for their professional

00:06:48.425 –> 00:06:50.820
services and consulting advice.

00:06:50.845 –> 00:06:56.000
I have three comments about this
that I think it is very interesting.

00:06:56.025 –> 00:07:01.300
Number one, thank you very much, MIT, for
quantifying what has up until now been

00:07:01.325 –> 00:07:04.580
just a general sense of foreboding doom.

00:07:04.605 –> 00:07:08.420
Everybody has said, AI equals risk.
What risk?

00:07:08.445 –> 00:07:12.100
Well, one, and two, and three, and
holy cow, you can’t imagine how many.

00:07:12.125 –> 00:07:16.880
Yes, we can because we’re scientists and
Because in a professional world, you can

00:07:16.905 –> 00:07:23.500
say there are not just a lot,
there are a number, and that’s very good.

00:07:23.525 –> 00:07:27.440
Number two, and by the way, there
are at least 700 risks with AI.

00:07:27.465 –> 00:07:31.540
We all know that intuitively, but
we need a number to work against.

00:07:31.565 –> 00:07:36.800
Number two, exactly to your point, Carl,
very few customers are going to go into

00:07:36.825 –> 00:07:40.760
that and try to solve for these problems
because it is the paralysis

00:07:40.785 –> 00:07:42.980
by analysis paradox.

00:07:43.005 –> 00:07:45.800
It’s literally, I can be responsible.

00:07:45.825 –> 00:07:49.295
Holy cow, there’s too many
things to be responsible about.

00:07:49.320 –> 00:07:51.400
I’m just going to do what I was going to
do anyway, and we’re just going to

00:07:51.425 –> 00:07:53.660
hope that everything works out okay.

00:07:53.685 –> 00:07:58.100
Customers are not going to go in
there and use this information.

00:07:58.125 –> 00:08:02.520
But point number three, I think, and if
anybody would like to know how to do this,

00:08:02.545 –> 00:08:04.500
we would be happy to do this for you.

00:08:04.525 –> 00:08:12.040
You can package a service offering by
industry, by customer size that says an

00:08:12.065 –> 00:08:16.220
AI risk assessment for your organization.

00:08:16.245 –> 00:08:21.020
I will examine the following X number
of categories and specific use cases.

00:08:21.045 –> 00:08:27.020
I will identify actual behavior within
your organization, a profile of what your

00:08:27.045 –> 00:08:32.080
business processes are, what is actually
going on in the real world, too, because I

00:08:32.105 –> 00:08:35.400
have network tools that allow me to
identify whether these tools are

00:08:35.425 –> 00:08:37.380
being used in your environment.

00:08:37.405 –> 00:08:41.000
And I can give you an action oriented
report that says, here are your

00:08:41.025 –> 00:08:44.720
vulnerabilities, here are the action item
recommendations, and here’s exactly

00:08:44.745 –> 00:08:46.580
what you should do about that stuff.

00:08:46.605 –> 00:08:51.920
Oh, and by the way, if it’s in the SMB
world, that costs 5 to $10,000 for that

00:08:51.945 –> 00:08:55.140
level of work, and it’s going
to be highly templatized.

00:08:55.165 –> 00:09:00.335
If I’m in the mid-market,
I’m talking 25 to 35, and it’s

00:09:00.360 –> 00:09:01.740
going to require some interviews.

00:09:01.765 –> 00:09:07.280
If you happen to want to get involved with
some divisions of larger enterprises in

00:09:07.305 –> 00:09:11.960
your local neighborhood, this is a great
way to get your foot in the door and get a

00:09:11.985 –> 00:09:19.720
50 to 100K services gig that is, A, very
timely, and B, once you’ve done it a few

00:09:19.745 –> 00:09:23.860
times and you’ve figured out what the
templates are, it is oh, so infinitely

00:09:23.885 –> 00:09:28.060
repeatable that your margin
shouldn’t be 50% on that thing.

00:09:28.085 –> 00:09:31.220
They should be 75% on that thing.

00:09:31.245 –> 00:09:32.060
This is great.

00:09:32.085 –> 00:09:36.380
That, to me, makes the point that 10% of
our audience

00:09:36.405 –> 00:09:41.740
is actually interested in going to
all that work or do that a deep dive.

00:09:41.765 –> 00:09:48.280
I do think this is a rich, rich area
to create a spectacular presentation on

00:09:48.305 –> 00:09:54.540
the dangers of AI that you can give in 20
minutes, not a super crazy deep dive, but

00:09:54.565 –> 00:09:58.120
just literally browse through this thing
and say, Well, what are

00:09:58.145 –> 00:09:59.695
the concerns about privacy?

00:09:59.720 –> 00:10:01.040
What are the What are the
concerns about misinformation?

00:10:01.065 –> 00:10:03.380
What are the concerns about bad actors?

00:10:03.405 –> 00:10:09.695
And get a few things where you can point
to the references that are outlined

00:10:09.720 –> 00:10:12.300
because this is all referencing articles,
referencing articles,

00:10:12.325 –> 00:10:14.000
referencing articles.

00:10:14.560 –> 00:10:19.920
You could repeat that to the Kauanis, the
Rotary, the BNI, the Morning

00:10:19.945 –> 00:10:21.300
Chambers of Commerce.

00:10:21.325 –> 00:10:25.160
You can give this presentation every day
for the next month, and you

00:10:25.185 –> 00:10:26.735
might get some clients out of it.

00:10:26.760 –> 00:10:27.760
Most people are not going to do it.

00:10:27.785 –> 00:10:28.500
Yeah, but we can say
that about everything.

00:10:28.525 –> 00:10:30.095
We can say that about everything.

00:10:30.120 –> 00:10:32.380
They can say that everything.
Most people will not execute.

00:10:32.405 –> 00:10:35.935
And that’s why, in a certain degree, I
don’t mind talking about it

00:10:35.960 –> 00:10:36.540
because it’s the…

00:10:36.565 –> 00:10:38.300
Look, I know a lot of people
aren’t going to do it.

00:10:38.325 –> 00:10:40.015
You don’t want to be that person.

00:10:40.040 –> 00:10:42.180
You want to be the person
that actually executes.

00:10:42.205 –> 00:10:46.900
And that’s where the value is.

00:10:46.925 –> 00:10:50.720
Well, and to your point, Carl, if you’re
not going to package it the way I

00:10:50.745 –> 00:10:55.680
described and sell it as a professional
service, please use it as a promotional

00:10:55.705 –> 00:11:00.895
vehicle to get into a conversation that
then says, Oh, and by the way, we

00:11:00.920 –> 00:11:02.380
could monitor your network for you.

00:11:02.405 –> 00:11:04.340
You should sign a managed
services contract.

00:11:04.365 –> 00:11:11.580
That, to me, is a very topical
reality that almost everybody can use.

00:11:11.605 –> 00:11:14.560
But we’re done giving
you advice on that topic.

00:11:14.585 –> 00:11:18.100
Let’s move on to our next one, guys.

00:11:18.125 –> 00:11:21.400
Ripped from the headlines of the most
obvious news that any of us have

00:11:21.425 –> 00:11:25.420
encountered recently,
Google has a monopoly.

00:11:25.445 –> 00:11:26.940
What?

00:11:26.965 –> 00:11:28.820
Say it in a sense.

00:11:28.845 –> 00:11:32.360
And not just a monopoly,
but a monopoly in search.

00:11:32.385 –> 00:11:38.080
So the news, if you have not been keeping
up with this, this is not immediate.

00:11:38.105 –> 00:11:41.360
This is not just a short term assessment
where somebody woke up one day

00:11:41.385 –> 00:11:43.000
and decided to point fingers.

00:11:43.025 –> 00:11:49.660
This is years of analysis and legal
exploration going into a scenario to

00:11:49.685 –> 00:11:54.300
examine the business behavior, not the
prevalence of technology,

00:11:54.325 –> 00:11:59.660
but the business practices around the
application, requiring it to be positioned

00:11:59.685 –> 00:12:06.640
as default, requiring compatibilities and
embedded capabilities with their search

00:12:06.665 –> 00:12:09.900
engine in other people’s software.

00:12:09.925 –> 00:12:11.740
It’s not the technology.

00:12:11.765 –> 00:12:15.820
It is the business behavior around the
technology that

00:12:15.845 –> 00:12:18.060
apparently has risen to a level.

00:12:18.085 –> 00:12:23.720
Now, we’ve been talking about
some things that we believe are obvious

00:12:23.745 –> 00:12:28.220
monopolies in many dimensions of our
industries for a number of years.

00:12:28.245 –> 00:12:33.940
This is the first time anybody with any
authority actually said something about it

00:12:33.965 –> 00:12:35.860
since Microsoft.

00:12:35.885 –> 00:12:37.540
That’s a million years ago.

00:12:37.565 –> 00:12:41.660
What do you guys think is
actually going to come from this?

00:12:41.685 –> 00:12:45.840
And do you believe this is going to change
anything about their behavior?

00:12:45.865 –> 00:12:48.820
I want to make predictions.
I’m totally on board for this.

00:12:48.845 –> 00:12:53.280
So I’m going to say I don’t
think Google will get broken up.

00:12:53.305 –> 00:12:58.980
I think that is one of those moves that
the government doesn’t generally like to

00:12:59.005 –> 00:13:03.420
do if they don’t have I’m
not sure it’s necessarily as clean.

00:13:03.445 –> 00:13:07.560
I could see a very clear argument for you
could split main Google

00:13:07.585 –> 00:13:09.500
search with YouTube.

00:13:09.525 –> 00:13:13.020
It makes sense to me because you end
up with a world of two search engines.

00:13:13.045 –> 00:13:14.400
But I don’t actually…

00:13:14.425 –> 00:13:18.420
Because first, you immediate move would
to make themselves search the other thing.

00:13:18.445 –> 00:13:22.260
It makes a lot of sense if you
want to create a search engine bit.

00:13:22.285 –> 00:13:25.460
But I actually think they’re going
to be a little bit more cautious.

00:13:25.485 –> 00:13:30.460
I think the main thing I’m going to see
out of this is a ban on

00:13:30.485 –> 00:13:36.060
these contracts where they’re buying
off someone else to not get in there.

00:13:36.085 –> 00:13:38.260
The obvious one is Apple.

00:13:38.285 –> 00:13:43.240
Google just writes a check to Apple so
that they don’t want this for placement as

00:13:43.265 –> 00:13:45.895
a search engine, but it also
disincentivizes Apple

00:13:45.920 –> 00:13:46.980
from ever looking at that.

00:13:47.005 –> 00:13:51.720
I think they’re going to clearly say, When
you are at some measurable level of the

00:13:51.745 –> 00:13:55.980
dominant player, you cannot
buy your way to hold on that.

00:13:56.005 –> 00:13:57.340
You’ve got to compete.

00:13:57.365 –> 00:14:01.380
I think that’s going to be the obvious
remedy that we’ll see out of this.

00:14:01.405 –> 00:14:05.980
Whether or not that will inspire players
like Apple to get into search

00:14:06.005 –> 00:14:09.240
will be an interesting play.

00:14:09.520 –> 00:14:16.820
I think the one thing that I’m looking for
is a remedy that actually does create

00:14:16.845 –> 00:14:19.300
the opportunity to create new markets.

00:14:19.325 –> 00:14:24.460
And I’ll be interested in, and by the way,
just slowing Google down might do that.

00:14:24.485 –> 00:14:28.060
I think it’s argument that their search
results have been

00:14:28.085 –> 00:14:33.140
garbage for a while and that they are
not delivering on the best product.

00:14:33.165 –> 00:14:39.680
If they’re distracted by having to handle
a bunch of legal stuff, it might open the

00:14:39.705 –> 00:14:42.135
opportunity for somebody
else to do something there.

00:14:42.160 –> 00:14:45.200
But I think the obvious bit is we’re going
to say, No, you can’t buy your

00:14:45.225 –> 00:14:47.340
way to the top of the market?

00:14:47.365 –> 00:14:48.700
A couple of things.

00:14:48.725 –> 00:14:56.900
First of all, thank goodness, the EU
has done a better job of being effective

00:14:56.925 –> 00:15:02.060
with managing the big tech giants
than the United States.

00:15:02.085 –> 00:15:06.500
I think that basically saying, No, you
can’t do that,

00:15:06.525 –> 00:15:09.760
has already set Google on the path to
figuring out how they’re

00:15:09.785 –> 00:15:10.740
going to handle this.

00:15:10.765 –> 00:15:14.280
I don’t know what their answer is, but I
know that Google knows

00:15:14.305 –> 00:15:15.960
what their answer is.

00:15:15.985 –> 00:15:17.360
I do think it’s interesting.

00:15:17.385 –> 00:15:20.420
It’s not illegal to have a monopoly.

00:15:20.445 –> 00:15:25.360
It is illegal to behave like a monopoly.

00:15:25.385 –> 00:15:27.620
It’s like, Okay, here’s a question.

00:15:27.645 –> 00:15:31.380
If you could choose
any search engine in the world, which

00:15:31.405 –> 00:15:34.660
you can, would you choose Google?

00:15:34.685 –> 00:15:39.740
Today and day after day after day, I do.

00:15:39.765 –> 00:15:43.040
Not that it is the greatest
I could possibly imagine.

00:15:43.065 –> 00:15:44.300
It’s like the government.

00:15:44.325 –> 00:15:47.580
It’s not the best
democracy that can be made.

00:15:47.605 –> 00:15:50.060
It’s the best that we will accept.

00:15:50.085 –> 00:15:54.260
I would like Google to
be better in many ways.

00:15:54.285 –> 00:16:00.220
I think the interesting part is the look
at Microsoft is probably a good example

00:16:00.245 –> 00:16:04.100
that what they really did is they looked
at little things like,

00:16:04.125 –> 00:16:08.560
Well, you can’t make a deal with IBM
so that IBM has to pay you for an

00:16:08.585 –> 00:16:11.300
operating system, whether
they install it or not.

00:16:11.325 –> 00:16:13.615
Like, holy crap.

00:16:13.640 –> 00:16:15.800
That’s a genius It’s not a
great move if you’re selling.

00:16:15.825 –> 00:16:19.500
It’s not a great move if you’re buying.

00:16:19.525 –> 00:16:23.380
And maybe they will figure out how to
open it up and have some real competition.

00:16:23.405 –> 00:16:27.740
Apple would be smart to never
get into the search business.

00:16:27.765 –> 00:16:32.680
So they’re always going to buy a partner
It’s just a matter of, Okay, can somebody

00:16:32.705 –> 00:16:36.780
else offer them more money than Google or
a better deal or something that fits

00:16:36.805 –> 00:16:42.940
better with their future AI offering
or fits better with their ecosystem?

00:16:42.965 –> 00:16:48.660
What’s the searching that works best
for them, separate from everybody else?

00:16:48.685 –> 00:16:52.040
I would love to think that there’s a
future where we each have

00:16:52.065 –> 00:16:54.300
these bespoke search engines.

00:16:54.325 –> 00:16:58.300
I have what works for my business,
you have what works for your business.

00:16:58.325 –> 00:17:02.720
There are lots of specialties search
engines, but they’re really hard to

00:17:02.745 –> 00:17:04.900
use unless you’re in that specialty.

00:17:04.925 –> 00:17:09.575
Maybe we’re going to force some
innovation to other people as well.

00:17:09.600 –> 00:17:11.280
Before I let Ryan, I’m going
to just make a quick comment.

00:17:11.305 –> 00:17:13.380
You asked a really good question
like, would you choose something else?

00:17:13.405 –> 00:17:17.040
I’m finding that for at least 50% of
my searches, I’m not choosing Google.

00:17:17.065 –> 00:17:18.960
I’m, in fact, choosing ChatGPT.

00:17:19.400 –> 00:17:23.420
The reason is I have a focus question.

00:17:23.445 –> 00:17:28.050
I generally know that it will be in the
vast, broad, settled amount of knowledge.

00:17:28.080 –> 00:17:32.135
It’s not something that this is just
that is going to be variable or timely.

00:17:32.160 –> 00:17:34.360
And by the way, the fact that it gives me
a couple of sources means I

00:17:34.385 –> 00:17:35.810
can double-check its work.

00:17:35.840 –> 00:17:39.770
It’s an established thing that
hasn’t been known for 20 years.

00:17:39.800 –> 00:17:40.720
That’s a great way of getting it.

00:17:40.745 –> 00:17:43.760
I’m 50% of my searches, I’m
not choosing it.

00:17:43.785 –> 00:17:48.380
Well, see, and this is where
if you go back to the fundamentals of

00:17:48.405 –> 00:17:52.090
why is monopoly behavior frowned upon?

00:17:52.120 –> 00:17:54.220
Why do we actually care about that?

00:17:54.245 –> 00:17:59.740
Well, it’s because you use market position
and money to substitute

00:17:59.765 –> 00:18:02.010
competition and innovation.

00:18:02.040 –> 00:18:08.810
If you have the very best product and you
achieve a dominant market position because

00:18:08.840 –> 00:18:12.570
it is bigger, better, faster than anything
else that’s out there,

00:18:12.600 –> 00:18:17.160
congratulations to you and congratulations
to me as the user because you’ve made my

00:18:17.185 –> 00:18:22.570
life better with a product that is
demonstrably better than anything else.

00:18:22.600 –> 00:18:26.770
When you get to that position, and to your
point, Carl, you begin to

00:18:26.800 –> 00:18:29.290
behave in a monopolistic way.

00:18:29.320 –> 00:18:33.480
What happens is you buy your market
dominance instead of earn your market

00:18:33.505 –> 00:18:38.180
dominance, meaning you take your foot off
the gas pedal of innovation and you cease

00:18:38.205 –> 00:18:43.360
to ship the best technology, which
punishes not only the buyers, but

00:18:43.385 –> 00:18:45.900
also the actual customers, the users.

00:18:45.925 –> 00:18:48.700
That’s exactly where
Dave’s example is going.

00:18:48.725 –> 00:18:58.000
The capabilities of Gen AI added to search
makes the potential for that service, for

00:18:58.025 –> 00:19:02.240
that basic technology, dramatic radically
better than anything we have been

00:19:02.265 –> 00:19:05.010
accustomed to in the last 20 years.

00:19:05.040 –> 00:19:08.250
Google has not done well with that.

00:19:08.280 –> 00:19:11.740
They have stumbled into a lot of the
research that I’m reading in

00:19:11.765 –> 00:19:13.250
my own personal experience.

00:19:13.280 –> 00:19:16.740
There are many times I will ask a question
in Google, and everything

00:19:16.765 –> 00:19:18.940
I get back is an ad.

00:19:18.965 –> 00:19:21.960
Literally everything I get back is an ad.
Not cool.

00:19:21.985 –> 00:19:23.860
That’s not what I’m looking for.

00:19:23.885 –> 00:19:28.320
Chatgpt has some reliability and
hallucination problems, but it

00:19:28.345 –> 00:19:30.770
does give me a more robust answer.

00:19:30.800 –> 00:19:36.090
But Google still is the A number one by a
long way in the marketplace because they

00:19:36.120 –> 00:19:41.250
forced it through monopolistic behavior
rather than earned it through innovation.

00:19:41.280 –> 00:19:47.380
If this has that simple impact on the
search market, it solves a problem.

00:19:47.405 –> 00:19:48.380
I’m all for it.

00:19:48.405 –> 00:19:53.180
If it sends a warning sign to others in
other marketplaces to

00:19:53.205 –> 00:19:57.810
stop it with the buying dominance and go
back to a world where you actually

00:19:57.840 –> 00:20:01.290
innovate and earn,
that That would be a pipe dream.

00:20:01.320 –> 00:20:03.010
I don’t think we’re
going to get there yet.

00:20:03.040 –> 00:20:06.500
It’ll take some more wraps on the knuckle
before anybody else

00:20:06.525 –> 00:20:08.380
actually changes behavior.

00:20:08.405 –> 00:20:10.400
But this is a good start.

00:20:10.720 –> 00:20:13.810
I would say, just for a note,
Google has made a pretty good case.

00:20:13.840 –> 00:20:19.000
They do continue to innovate, and my
number two search engine is YouTube.

00:20:20.600 –> 00:20:21.980
Maybe they will get programed.

00:20:22.005 –> 00:20:25.215
But I’m going to move us on to the last
topic, and I want to take an

00:20:25.240 –> 00:20:26.160
interesting angle on this.

00:20:26.185 –> 00:20:32.220
Nist just released their guidance for
cryptography for quantum computing.

00:20:32.245 –> 00:20:36.480
I will say, look, it’s an interesting set
of framework for those that

00:20:36.505 –> 00:20:38.140
are in the encryption space.

00:20:38.165 –> 00:20:40.620
Get to work, guys, because
you’ve got new data.

00:20:40.645 –> 00:20:45.320
But what I wanted to do is I actually
rejected this story for the business of

00:20:45.345 –> 00:20:49.560
tech, and I wanted to bring it to you guys
instead because I looked at this and said,

00:20:49.585 –> 00:20:52.500
okay, there’s the obvious element of
upgrade your cryptography

00:20:52.525 –> 00:20:54.290
when there’s a better version.

00:20:54.320 –> 00:20:56.640
That just always makes sense.

00:20:56.760 –> 00:21:02.280
But I feel like quantum computing as a
thing has always been, well, we’re within

00:21:02.305 –> 00:21:07.640
10 years of it, every three years, as
far as I can remember now at this point.

00:21:07.665 –> 00:21:12.600
I mean, it just feels like we’re so close
to quantum that I’ve come to the point

00:21:12.625 –> 00:21:15.290
where I think I just don’t believe them.

00:21:15.320 –> 00:21:20.050
I think I just don’t
think this is a thing.

00:21:20.080 –> 00:21:26.360
I will open a space for, sure, in the 27th
century, when the next version of Star

00:21:26.385 –> 00:21:29.050
Trek is out, we may
have quantum computing.

00:21:29.080 –> 00:21:33.700
But I don’t think in any
practical terms, this is a thing.

00:21:33.725 –> 00:21:36.600
Am I missing the boat?
I want to get check here.

00:21:36.625 –> 00:21:37.980
What’s your take on quantum?

00:21:38.005 –> 00:21:43.215
So much technology,
and we’ve seen this just recently in the

00:21:43.240 –> 00:21:45.920
pandemic where we’re saying, Oh, how
come robots aren’t taking over the world?

00:21:45.945 –> 00:21:51.455
Well, because in the real world,
people are saying, Well, I’ll work for a

00:21:51.480 –> 00:21:53.260
dollar less if you don’t take my job away.

00:21:53.285 –> 00:21:57.540
And so it’s delayed
the actual use of that technology.

00:21:57.565 –> 00:22:03.360
I think things like
the fact that we have GPUs and we have

00:22:03.385 –> 00:22:09.500
these processors that can just
put more horsepower on a problem, we don’t

00:22:09.525 –> 00:22:12.140
have a need for quantum computing.

00:22:12.165 –> 00:22:16.940
It’s not that it’s not real or it’s
not there or it’s not going to happen.

00:22:16.965 –> 00:22:22.020
If it doesn’t happen, it’ll be because we
don’t need it, because we can

00:22:22.045 –> 00:22:24.620
continually increase horsepower.

00:22:24.645 –> 00:22:28.700
Even though we don’t increase the
horsepower necessarily with one chip,

00:22:28.725 –> 00:22:34.860
having more and more processors, but
we can now buy NVIDIA chips

00:22:34.885 –> 00:22:36.920
by the gallon, right?

00:22:36.945 –> 00:22:41.980
We’ll just buy more horsepower, and then
they get smaller and smaller and smaller.

00:22:42.005 –> 00:22:45.940
See, and I will go to
the next order of impact.

00:22:45.965 –> 00:22:50.960
Because what Carl is saying is,
if I have an alternative technology that I

00:22:50.985 –> 00:22:56.090
can use to solve the future problem, it’s
existing, it’s tested, it’s understood.

00:22:56.120 –> 00:22:57.940
Let’s just do what’s familiar.

00:22:57.965 –> 00:23:03.050
I’ll go to the next level and say that I I
think the thing that will prevent rapid

00:23:03.080 –> 00:23:07.800
deployment of quantum is
the environmental impact.

00:23:07.825 –> 00:23:11.055
And I don’t mean just on trees and
waterways and cute

00:23:11.080 –> 00:23:11.900
little bunnies in nature.

00:23:11.925 –> 00:23:17.090
What I mean is the literal environment
in which these systems play.

00:23:17.120 –> 00:23:22.975
The power consumption requirements for
quantum computing are projected because

00:23:23.000 –> 00:23:24.800
nobody really has one of these things yet.

00:23:24.825 –> 00:23:29.520
If you’re going to use one of these that
does what it is scientifically suggested

00:23:29.545 –> 00:23:35.280
it could do, it’s going to consume power
at a rate beyond even what these data

00:23:35.305 –> 00:23:39.920
centers are doing for AI, which we know is
an order of magnitude beyond

00:23:39.945 –> 00:23:42.050
what regular data centers did.

00:23:42.080 –> 00:23:47.050
We already covered, years ago, if you guys
remembered, a story about

00:23:47.080 –> 00:23:53.090
the health impacts for people living in
proximity of large format data centers.

00:23:53.120 –> 00:23:57.320
The hum that happens 24/7
leads to hearing problems.

00:23:57.345 –> 00:24:00.570
It creates sleeping problems, et cetera.

00:24:00.600 –> 00:24:06.050
Order of magnitude to AI, further order of
magnitude up to the consumption of

00:24:06.080 –> 00:24:08.700
what quantum computing is going to be.

00:24:08.725 –> 00:24:11.020
I don’t think we can’t
innovate out of that.

00:24:11.045 –> 00:24:17.260
I think that the world will look around
and go, I’m not willing to pay you the

00:24:17.285 –> 00:24:22.290
extra second order costs in
order to consume this technology.

00:24:22.320 –> 00:24:27.500
I will buy a quantum computer for X,
but then in order to cool

00:24:27.525 –> 00:24:32.440
it, to power it, in order for it to live
in a place where it doesn’t poison the

00:24:32.465 –> 00:24:36.330
local citizenry, it’s going
to cost me X times four.

00:24:36.360 –> 00:24:38.960
I’m not willing to pay the X times four.

00:24:38.985 –> 00:24:40.400
So thank you very much.

00:24:40.425 –> 00:24:42.570
I’ll just stick with
what I’m doing right now.

00:24:42.600 –> 00:24:46.680
I think the capability, the
potential, the stories we’ve been told

00:24:46.705 –> 00:24:53.620
about quantum, yes, they are fascinating,
but they are not yet economically viable.

00:24:53.645 –> 00:24:55.455
And that’s an engineering problem.

00:24:55.480 –> 00:24:56.780
That’s not a scientific problem.

00:24:56.805 –> 00:24:58.200
That’s an engineering problem.

00:24:58.225 –> 00:25:03.420
You need to figure out a way to deploy
power without consumption.

00:25:03.445 –> 00:25:08.015
That’s a single variable that you need to
control for in designing and

00:25:08.040 –> 00:25:09.500
deploying these kinds of systems.

00:25:09.525 –> 00:25:11.460
By the way, one last comment.

00:25:11.485 –> 00:25:18.330
If quantum does happen,
all your cybersecurity is grandma stuff.

00:25:18.360 –> 00:25:19.360
It is outdated.

00:25:19.385 –> 00:25:24.400
It is the olden times, and it
does not work in a quantum world.

00:25:24.425 –> 00:25:30.420
We do not presently possess
cryptography that can withstand

00:25:30.445 –> 00:25:34.570
quantum attacks for longer
than just a couple of minutes.

00:25:34.600 –> 00:25:37.570
We’re literally not using technology.

00:25:37.600 –> 00:25:41.920
So if and when quantum comes,
I hope you all are thinking about the

00:25:41.945 –> 00:25:44.720
cyber impacts because
it’s going to break all your tools.

00:25:44.745 –> 00:25:47.020
Hold on.
I want to break that sentence down.

00:25:47.045 –> 00:25:49.260
First off, when is the question?

00:25:49.285 –> 00:25:52.330
And I’m coming to the
conclusion of it isn’t.

00:25:52.360 –> 00:25:57.220
And if is doing a whole lot of heavy
lifting right there, Ryan, because

00:25:57.245 –> 00:26:01.420
if when is never, if doesn’t matter.

00:26:01.445 –> 00:26:04.940
That’s your sentence
diagramming by the way.

00:26:04.965 –> 00:26:08.940
I’m just getting break down to
the basics here and observing it.

00:26:08.965 –> 00:26:09.680
Okay, sure.

00:26:09.705 –> 00:26:13.920
I hear you on all of this stuff,
but I’m coming to the conclusion

00:26:13.945 –> 00:26:16.290
that this isn’t a thing.

00:26:16.320 –> 00:26:24.050
In fact, head cycle spent on it in
any level is just not worth my time.

00:26:24.080 –> 00:26:29.700
Now, I will give a space of,
Look, I think making things

00:26:29.725 –> 00:26:32.330
generally more secure, and I’m putting
that in big old air

00:26:32.360 –> 00:26:33.570
quotes, is a good thing.

00:26:33.600 –> 00:26:38.860
My statement is, Okay, if you’ve designed
a new set of encryption that is

00:26:38.885 –> 00:26:43.360
invulnerable to this theoretical thing
that is way more advanced than our regular

00:26:43.385 –> 00:26:47.800
stuff, well, then it should also be good
enough for our regular stuff, and I

00:26:47.825 –> 00:26:50.860
don’t see any reason to not do it.

00:26:50.885 –> 00:26:55.680
But if somebody says, Dave, I want you to
spend time and brain cycles on this thing

00:26:55.705 –> 00:26:58.940
that only benefits quantum computing, I
think my answer is,

00:26:58.965 –> 00:27:03.660
That’s a waste of my time,
and I’m not doing any work on it because

00:27:03.685 –> 00:27:06.500
you don’t even prove
you can make the thing.

00:27:06.525 –> 00:27:10.330
Like, literally, the basics of
making it have not even been proven.

00:27:10.360 –> 00:27:15.140
So to Ryan, I would say, literally, all of
your objections, old man,

00:27:15.165 –> 00:27:19.380
can be overcome and will be
overcome by time and technology.

00:27:19.405 –> 00:27:22.180
Everything starts out being too expensive.

00:27:22.205 –> 00:27:23.520
Everything’s impossible.

00:27:23.545 –> 00:27:24.780
Everything costs too much.

00:27:24.805 –> 00:27:31.695
If you think of how many BTUs did it take
to light a house 100 years ago, well, you

00:27:31.720 –> 00:27:33.120
got to get the kerosine, but you have to
get the kerosine to the

00:27:33.145 –> 00:27:35.900
house and like, holy smokes.

00:27:35.925 –> 00:27:39.420
Now you have LED lights that take
essentially nothing except

00:27:39.445 –> 00:27:44.260
static electricity.
So the future will take care of itself.

00:27:44.285 –> 00:27:51.200
Today, if I would say, I have been
hearing about AI since before I was born.

00:27:51.225 –> 00:27:56.260
Like, literally, you watch old black and
white science fiction movies and TV shows.

00:27:56.285 –> 00:28:00.090
They’ve been talking
about AI for 60 years.

00:28:00.120 –> 00:28:03.940
And then one day, it became reality.

00:28:03.965 –> 00:28:08.360
Everybody, five years ago, you could have
made the argument, They’ve been talking

00:28:08.385 –> 00:28:10.800
about AI for 50 years, and
it’s never going to happen.

00:28:10.825 –> 00:28:17.980
I think it’s going to be the exact
same thing with quantum computing.

00:28:18.005 –> 00:28:22.000
Part of it is with AI, we still haven’t
got, we haven’t figured

00:28:22.025 –> 00:28:23.810
out the killer app.

00:28:23.840 –> 00:28:30.280
We haven’t figured out a thing that’s
the email equivalent of the web.

00:28:30.305 –> 00:28:33.570
What is it that we would
do with quantum computing?

00:28:33.600 –> 00:28:35.520
What’s the killer wrap
of quantum computing?

00:28:35.545 –> 00:28:41.090
Well, if you’re not a quantum physicist or
you’re not a mathematician,

00:28:41.120 –> 00:28:43.975
I’m not sure there’s a use for
it outside of your paycheck.

00:28:44.000 –> 00:28:45.480
Well, and that’s the thing, right?

00:28:45.505 –> 00:28:46.760
There are use cases.

00:28:46.785 –> 00:28:50.220
They’re just not economically
viable yet, right?

00:28:50.245 –> 00:28:51.740
I agree with you, Carl.

00:28:51.765 –> 00:28:57.615
The ecosystem of technology will solve
for the problems that I’ve been outlining.

00:28:57.640 –> 00:28:59.840
That sounds like opportunity, and it
sounds like business

00:28:59.865 –> 00:29:01.520
development to me, right?

00:29:01.545 –> 00:29:03.420
From an industry perspective.

00:29:03.445 –> 00:29:08.280
My thing is, I do believe that
the future takes forever

00:29:08.305 –> 00:29:10.380
until it happens all at once.

00:29:10.405 –> 00:29:16.400
Exactly as Carl is describing, it will
come to Dave’s point, not this year.

00:29:16.425 –> 00:29:19.760
If you’re focused on this year, please
don’t spend any time building a

00:29:19.785 –> 00:29:21.900
quantum practice, but it will come.

00:29:21.925 –> 00:29:23.760
We just need to be ready for it.

00:29:23.785 –> 00:29:27.640
Well, I want to push back slightly there,
Carl, because while you’re right on the AI

00:29:27.665 –> 00:29:32.020
stuff, we have seen advancements
advancements along the way

00:29:32.045 –> 00:29:34.360
for a good portion of that.

00:29:34.385 –> 00:29:37.680
We could go back in time, let’s say 2015.

00:29:37.705 –> 00:29:42.320
I worked on projects that were advanced,
that were data science stuff, that was

00:29:42.345 –> 00:29:45.090
using machine learning,
that was a stretch to be AI.

00:29:45.120 –> 00:29:46.570
That had the glimmers of that.

00:29:46.600 –> 00:29:50.320
You could go back further than that and
you could see versions where we’re

00:29:50.345 –> 00:29:54.000
building the systems
that communicate that.

00:29:54.025 –> 00:29:58.240
I have this element of I saw a trend line
over time of advancements,

00:29:58.265 –> 00:29:59.215
and you’re right.

00:29:59.240 –> 00:30:01.810
All of a sudden, we see the
breakthrough and it crashes through.

00:30:01.840 –> 00:30:06.330
What I’m pushing back on is, they don’t
even have the basic bits over on quantum.

00:30:06.360 –> 00:30:10.290
It’s all whiteboard stuff
where nothing actually works.

00:30:10.320 –> 00:30:15.720
There’s a difference between seeing
elements of it and You guys are having

00:30:15.745 –> 00:30:17.810
fun on a whiteboard, and that’s all cool.

00:30:17.840 –> 00:30:22.050
But until you can actually build
any of it, stop wasting my time.

00:30:22.080 –> 00:30:27.280
Maybe AI will be the tool that
productises quantum.

00:30:27.305 –> 00:30:30.380
Perhaps.
That replaces itself with quantum.

00:30:30.405 –> 00:30:35.080
Well,
when the word theoretical is built into

00:30:35.105 –> 00:30:38.000
your definition, maybe you’ll
never show up in the real world.

00:30:38.025 –> 00:30:40.900
Maybe you’ll never show up.

00:30:40.925 –> 00:30:49.860
And with that happy note, we bring an
end to episode 209 of the Killing It…

00:30:49.885 –> 00:30:51.320
Podcast.

00:30:51.400 –> 00:30:54.380
Thanks for tuning in to
the Killing It podcast.

00:30:54.405 –> 00:30:59.620
Please share with your friends and tell
everyone to subscribe on iTunes,

00:30:59.645 –> 00:31:03.020
Stitcher, all the podcast places.

00:31:03.045 –> 00:31:07.160
Join us next week and help us keep
killing it in the technology business.