May 22, 2026 • 47 mins
What does it mean to be at the “foothills of the singularity”? That’s how DeepMind CEO Demis Hassabis ended his speech at Google I/O, prompting questions and scratched heads. Oz and Reed Albergotti (Semafor) attempt to dissect the meaning behind Hassabis’s confounding statement. They also discuss why so many commencement speakers are getting booed by college graduates after bringing up AI, and what it means for SpaceX, Anthropic, and OpenAI to all be heading towards an IPO.
Then, Oz sits down with David Webster, Head of UX at Google Labs, for a deeper look at the products Google unveiled at their annual developer conference of the year.
Additional Reading:
- DeepMind founder Demis Hassabis on what Google AI products say about ‘singularity’ | Semafor
- A Guide to Commencement | Semafor
- SpaceX, Anthropic and OpenAI’s Sprint to Go Public Defines the AI Boom’s Big Day - WSJ
- Former Google CEO Eric Schmidt was booed | Strait Times IG
- Subscriber Q&A: Live @ Google I/O - by Alex Heath - Sources
Download SAILY in your app store and use our code techstuff at checkout to get an exclusive 15% off your first purchase! For further details go tohttps://saily.com/techstuff
See omnystudio.com/listener for privacy information.
Listen
Watch
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:11):
Readel Brigotti. Are we living in the foothills of the singularity?
Speaker 2 (00:15):
You know? So we're at Google Io and it's like
two hours of you know, and now we're adding AI
to this and now this is getting better and blah
blah blah, and it's like you're like, okay, cool, Like
you know, there's a million different things Google's doing. It's
very impressive and overwhelming. And then and then Demiss Sabist
closes it out and he's like, and by the way,
you know, we're in the foothills of the singularity. I
(00:37):
don't exactly know what that means. I'm like, well, I'm
going to get some time with him later. I'm going
to ask him, like what he exactly means about that.
I don't know if I have a better idea after
talking to him, like a more specific idea, because I
think he's been very specific about like AGI in five
years or ten years or whatever, and in this in
this case, it's like we're sort of in the realm
(01:00):
of this singularity thing, which no one can really define.
But I but I think I thought about it, and
I was like, you know it it is actually a
helpful comment in a sense because you can get sort
of bogged down in all in the weeds of all
these different product announcements, and it's like okay, like yeah,
AI is getting better and these and every Google product
(01:22):
now is being sort of you know, tuned or enhanced
with with everything that deep Mind is is working on.
But actually there is this like north star for the company,
and they are still like behind the scenes they see
deep Mind, you know, sees its goal as getting to
this you know, this this place of whatever you want
(01:44):
to call it, the singularity or or a g I
a s I you know, and I think, I mean
it's any means guests whether whether that will happen, whether
it's even possible, but it is. It does help to
sort of think, okay, yeah, this is like, this is
what's what's really happening the scenes.
Speaker 1 (02:04):
Let's get into it. Welcome to Tech Stuff. I'm Oz
Volosh and this is the Week in Tech Today, something
a little bit different. For the first part of our show,
we'll be joined by our regular contributor, Read Abricotti, tech
editor for TEMBI four, to discuss the big headlines you
need to know this week. And for the second part,
we've got a special interview. I'm excited to share with
you with David Webster, who is the head of UX
at Google Labs. He's essentially the guy who's responsible for
(02:27):
turning the AI research coming out of Deep Mind and
other parts of Google into consumer products that we're all using.
It's been a crazy week. I mean, the lead story
in the Wall Street Journal today is SpaceX Anthropic and
Open AI's sprint to go public defines the AI booms
big Day, potentially three trillion dollar IPOs this year. At
(02:48):
the same time, commencement speakers around the country are being
booed offstage for mentioning AI. He's Eric Schmidt, the former
CEO of Google at the University of Arizona. So the
question is whether you will help shape artificial intelligence.
Speaker 3 (03:04):
We do not know.
Speaker 1 (03:07):
We do not know.
Speaker 3 (03:08):
I know what many of you are feeling about that.
Speaker 2 (03:11):
I can hear you.
Speaker 3 (03:13):
There is a fear if you don't care about science.
Speaker 2 (03:20):
That's okay, because AI is going to touch everything else
as well.
Speaker 3 (03:24):
Whatever path you too, we are become more.
Speaker 2 (03:29):
Kind of been his favorite day of the year. No,
I mean, it is interesting though. I wrote this this
funny sort of newsletter intro yesterday about you know, my
guide to giving commencement speeches, because this has been happening
all over the country. If people give these speeches, and
of course they want to talk about AI, college students
don't want to hear it. I think they've been sort
(03:51):
of told probably honestly, I think inaccurately that AI is
going to mean no jobs for them. I think I
think actually the job market, it and and and a
lot of these tech layoffs are really have a lot
more to do with other, you know, a complex set
of issues, and AI is not really not really one
of them, maybe indirectly, but but no, I I wrote
(04:14):
this funny this funny piece which you should you should
check out at the seminar.
Speaker 1 (04:17):
Deadline where Sunscreen right, which was a termination.
Speaker 2 (04:20):
Everyone right, everyone everyone is uh, you know, was always
mentioned the where sunscreen you know, commencement. It was actually
I guess like a call a newspaper column commencement speech.
But but I was like, you know, you could go
that you don't talk about AI, but if you do,
go the Sarah Connor route, you know who who was
a college student right? And in Terminator one fictional person
(04:43):
if you're if you're too old for if you're too young, sorry,
I'm old, you know, but but you know she was
a college student who I who I said, you know,
became an AI safety advocate and as a as a
robot was chasing her and she had this great line
and terminator too, which is, you know, if you're not
wearing two million sun block, you're going to have a
really bad day.
Speaker 3 (05:02):
Get it. So it's a great, great wine for a
commencement speech. So that's my to.
Speaker 1 (05:10):
Wonder though, because you know, Schmid's commencement speech was after
this one. We talked about last week Gloria Callfield, who
gave the commencement speech at a Flora University and also
got bood talking about AI. Jensen Hawaian gave a Connie
hey Meling commencement of speech where he got a heroes
welcome talking about AI. But I guess, I guess it's
a robotics I mean it's a famous robotics university. So
(05:31):
the warmer crowds, maybe Erishman is looking at It's like,
am I more like Gloria Corefield and more like Jensen?
Speaker 3 (05:36):
Why?
Speaker 1 (05:36):
It's like I'm more like Jensen waning? I can I
can pull off talking about AI and not get a
bood off stage? But I'm just like, which students are
I going to bed on? H It's an interesting question.
So I actually asked David Webster a little bit of
plug for the second half of this episode for his
diagnosis of the tech clash and how the industry should respond,
and they'll talk about that later in this episode. But read,
(05:58):
I mean talk about this easy week from business news perspective.
I mean, the Will Street Journal was literally was giddy.
I opened the app when it was like jumping jumping
off the screen to me about this. Uh, you know,
the allies trillion dollar evaluations, these IPOs have been kind
of cooking for a while, but like, what is it
about this week that has has kind of been so insane?
Speaker 2 (06:18):
Yeah, I mean, you know, you've got You've got SpaceX,
you know filing, You've got open Ai, Anthropic. You know,
they all want they all want to go public this year.
They're all going to be the biggest I pos in history.
You know, SpaceX will probably take the crown there and
you know, I just it's like, Wow, there's a lot.
Speaker 3 (06:39):
Of money in the world. If all these companies can
go public in the same year.
Speaker 2 (06:45):
I don't know, I don't know what percentage they'll float,
you know, I think it'll be that'll be interesting and
we'll see what happens with like the lock up.
Speaker 3 (06:51):
There'll be some interesting stuff.
Speaker 2 (06:53):
Liz Hoffin's written about this at some and four, you know,
the interesting ideas around how to deal.
Speaker 3 (06:58):
With the lock up period.
Speaker 2 (06:59):
But it is I mean, look like so much wealth
and so much value is being created right now out
here in this like small small area that I live
in in Silicon Valley. I am not one of these
people who's become a billionaire in the last year, but
it's just like an incredible amount of wealth creation. And
I think it is I mean, there is something to
(07:21):
be excited about because you know, when these companies go
public now, they they'll be in you know four one
ks and and you know in Trump accounts and things
like that. And I think the next generation we'll be
able to you know, profit off of of all this
wealth creation.
Speaker 1 (07:35):
But you pointed out in your newsletter, which that was interesting,
was like for the last fifteen years, the luckiest students
who are in the cohort, who get jobs at you know,
Metro and Google and other places, basically had a life
of you know, free snacks and stock options. And you
know that didn't have never apply to the majority of students,
but applied to like a you know, somewhat of a
reasonable sliver. And now there's this crazy wealth creation going on,
(07:59):
but in the hands very few, and it's a stra
It's a very strange cultural moment. How much do you
think this threatens, Like if you're a tech post and
you like, on the one hand, my house is being
fly at boomed. On the other time, I'm about to
ip on my company for Triton dollos, Like, which are
you more focused on? Probably the Tritian dollos.
Speaker 2 (08:17):
I would hope these people are not that focused on,
you know, the the AI backlash. I mean, it's it's ultimately,
you know, it's a popular thing to say. I was
actually at my kid's hockey practice last night and their
ten year olds are in there talking about how much
they hate AI, and I was like, oh, that's interesting.
Like I asked one of them, like, what do you
(08:38):
what do you not like about AI? And he said, well,
it's it's wrong. You know, it's wrong, and like so like, yeah,
fifty percent of the time it's wrong. And I'm like,
your dad is like an AI executive, So I you know,
it's interesting that you're saying that, But no, I mean,
I don't know how much this this backlash. I mean,
certainly people don't want data centers, and certainly I think
(09:00):
AI is a convenient scapegoat. There there are bigger structural
issues and we don't have time to get into those
on why. You know, you have these periods of massive
wealth creation like we did, you know in the in
the the last couple of tech booms, and regular people
don't benefit from that. But I would if I were
a college student today, I'd be super excited because there's
(09:21):
more opportunity right now than ever. This technology is very
easy to learn, and I'd love to go just try
to start some companies. And I mean, I think it's
it's an incredible, incredible opportunity. I think this moment, I
think we'll look back and we'll say college students, who
people were coming out of college today, we look back
ten years from now, had like one of the best
(09:42):
opportunities to become wealthy in history.
Speaker 1 (09:46):
So we may not we may we may or may
not be in the foothills of the singularity, but we
are in the foothills of the of the of the
gold Rush as far as you're consent.
Speaker 2 (09:55):
I don't know what the singularity is. I'm still very
confused about that. I always thought it was we merge
with the with the machines or something, right, Yeah, and I
but but like what I do what I do see
is is that like when we employ AI and at Semaphore,
when we when we use these products, it creates more
work for people, for humans, Like we're going to have
(10:15):
to hire people because of AI because now we can
do these products like we launched Semaphore Intelligence. I think
we talked about this the other day, and it's like,
I mean, I think that's these giant tech companies that
we're paying people to do practically no work and give
them free snacks all day for stock options and a
good salary. Like, I mean, yeah, they're gonna have to
(10:36):
get lean. And there's a lot of reasons for that, right.
AI indirectly is one of them, because because there's a
lot of capex that needs to to go into this.
But you know, I I don't think this is like
going to kill jobs at all.
Speaker 1 (10:52):
Let's talk about those three IPOs and a bit more
detailed though, because everyone's many people have seen the headlines.
But that's that's kind of three different different stories. Space
that is actually filed this week, so the most information
which was previously you know probably but is no public
about that company. You know, three things struck me. You know,
one how profitable Starlink is and how big stall inkins,
(11:13):
and two how lols making SpaceX is and how lols
making x x AI is. What did you take away
from that fighting.
Speaker 3 (11:22):
Yeah, I mean, first of all, Starlink is amazing.
Speaker 2 (11:25):
I mean I remember when I first came out to
cover the tech industry, people told me that this this
idea of satellite internet was like pie in the sky,
like no one it was impossible, no one was ever
going to do it. I've I think I bought three
Starlink satellites in my in my in my life now
because you know, dishes, because like I live in California
where there's like outside of actually even in the cities,
(11:48):
like there's basically no cell reception. Great, and I like
to be outdoors and also have to work a lot.
So you know, like, I think this technology is incredible,
done surprises me at all that is making a ton
of revenue. But also you know, like Xai for instance, yeah,
I mean losing a ton of money. I was talking
with someone the other day at Google Io who was saying, oh,
(12:11):
I think like Xai took themselves off the board because
they they sold their compute to Anthropic and Anthropics, you
know in the lead.
Speaker 3 (12:20):
I mean I don't. I don't buy that at all.
Speaker 1 (12:22):
I think that's we love pank one point two billion
dollars a month for the next four or five years
to Xai and.
Speaker 2 (12:30):
And look Anthropics profitable, Like the Wall Street Journal had
to scoop the other day that Anthropic is now profitable.
I was joking around about this, like, oops, they're profitable,
like that that's a big mistake. But it's actually true,
Like they did not want to be profitable, Like that
is not a good thing in this in a in
a fast growing, high capex environment, to be profitable like
(12:51):
you you want to be if you're positible.
Speaker 1 (12:53):
What you're saying is they've underinvested.
Speaker 3 (12:55):
They've underinvested. I mean they may.
Speaker 2 (12:57):
I think it's pretty clear now, even though like all
the smartest financial journalists and tech journalists last year were
telling you that this was a bubble and that Opening
Eye was irresponsible, that it was all going to come
crashing down. It's pretty clear that was totally not true.
It was complete bs And if people had actually used
these products, they would know that because they'd, like me,
(13:17):
they'd be like at two in the morning running out
of tokens and freaking out signs.
Speaker 1 (13:21):
Throping one hundred and thirty percent revenue growth didn't surprise
me that much. First profitable quarter, I let your hot
take that's a bad thing. But the compute cost, basically
the cost of delivering the services went down from seventy
one cents per dollar of revenue to fifty six percent
fifty six cents per dollar revenue. That slightly niche factoid
was the one that I found most interesting. I mean,
(13:42):
that's a huge decline in the cost of the cost
of goods sold for them.
Speaker 3 (13:47):
Yeah, I know.
Speaker 2 (13:48):
I think I think that needs to be like unpacked
a little bit because I'm like, well, okay, what does
that actually mean? Like when you there's there's a question
I have, which is like if I pay a subscription
fee Tomanthropic or any of these companies, now, like they
are incentivized to provide as few tokens to me as
as possible. So I don't know where where all of
(14:09):
these efficiencies are coming. But generally, yeah, I mean I
I honestly, I'm I'm like a conspiracy theorist there probably,
but but like honestly, like the compute costs just come down,
Like that's the one thing. I mean, like if you're
if you're like there's a lot of companies where it's
like yeah, I mean like Uber was one of them.
It's like, are you really going to bring the cost
(14:29):
of driving a car down? Like not really, Like ultimately
you kind of just got this big cost that's never
going to change. Compute costs always come down, Like people
just find a way to make this cheaper. Right now,
it's like piling GPUs and denser and denser racks into
data centers and you know, liquid cooling and all this
optical connections, et cetera. And it's it's a huge industrial effort.
(14:52):
But like there will be breakthroughs that will be you know,
there's so much money in this. Like I just don't
think these compute costs are gonna get you know, they're
going to be like they're going to stay the same,
They're just going to continue to drop. So like be
unprofitable today, you will be extremely profitable tomorrow when the
costs come down and the demand just continues to go up.
We haven't even really seen penetration in the market yet.
(15:16):
I mean, this is just a tiny niche of people
who are like using all these tokens.
Speaker 1 (15:20):
So let's talk about open AI in exactly that respect.
Because today's Thursday, we're recording. The episode is going to
come up on Friday. And again the Journal reported that
maybe on Friday open Ai I would actually file, So
we don't know yet, but suddenly it's coming soon. And
one of the big issues about open AI, which you've
reported on, is the kind of conflict between Similtman and
the CFO. The Simultan wanting to basically invest more and
(15:43):
more ahead of revenue and compute in the CFO trying
to hold him back a little bit, like what what
can we expect from the IPO filing. In terms of
this debate.
Speaker 2 (15:52):
I have not written that there's a huge conflict between
Altman and the CFO. I mean, other publications Wall Street
Journal has said like there's this huge rift and the CFO.
Actually I think it was the information for me before
that we should give them a credit. But you know,
like the that Sarah friar Is doesn't want to IPO
(16:12):
and Sam Altman does, and the numbers are horrible. The
growth is slowing down, they can't get enough users in
chat gpt. They've painted a really bleak picture. So I
think we'll now get like a better sense, and then
when Nthropic files, it'll be interesting to compare because right now,
like there's all these comparisons on Entropics revenue and Opening
Eyes revenue and people are basically a lot of I
(16:36):
think it's like conventional wisdom that Anthropic is like fast
has faster growing revenue. Well, I think we'll hopefully we'll
get more of an apples to Apple's comparison. But Opening
as interesting because it's like chat gpt is a consumer business,
and then where all the tokens are being consumed is
kind of this like prosumer slash enterprise part of the
(16:57):
business with Codex, for like all these odors are using
you know, using these these harnesses. So that's like Opening
Eyes really two things and Anthropic is more one thing.
And then I think the way I look at the
way I look at it is like there are two
companies that are actually competing in the long term for
(17:17):
AI dominance, and that's Google and the other one doesn't
exist yet because it's going to be SpaceX and Tesla combined,
which is going to happen. And if you look at
it like those will be the only two companies in
the world that have leading frontier models. If you if
you agree that GROC is a leading frontier model, I
think it's I think it is. I think it gets
(17:39):
a bad ride. It's not as far behind as some
people say. And also have robotics deployed in the world
at scale, like advanced robotics. There's really because they have
there's Google has way Mo, and then Tesla has you know,
the most probably the most advanced robots in the world
out you know, out on the roads. And then and
(18:00):
then they're both kind of dabbling in space. I mean,
I don't know. Google doesn't have rockets obviously, but they're
they're an investor in SpaceX, so they'll make some they'll
make some money.
Speaker 1 (18:09):
Well that's I mean, that's why I want to come
back to Google just before we go to the break
and then come back to David Webster from Google. But
I mean that the sense in which all these companies
are intertwined is kind of extraordinary, right, Like you know,
anthropic of paying SpaceX Elon just lost his lawsuit against Semmeltman,
which is removed an impediment to open a eyes IPO.
(18:30):
Like I mean, I guess give give a reflection on
the conclusion of that trial. But also like is it
a rising tide will lift all these boats or is
Aaron sense going to be Is it going to be
more of a winner takes it all model?
Speaker 2 (18:42):
Do you think, Oh no, no, I don't mean it's
winner takes all at all? Like I actually think the
rising tide the demand is so huge, like all of
these companies are going to be hugely successful, Like there's not,
I don't I there's very little chance that like these
companies are just going to fail or something. Right, there's
so like I said earlier, like the demand, it's not
(19:03):
there's not even really market penetration yet and we're already
seeing supply.
Speaker 3 (19:06):
Can't you keep up with demand?
Speaker 4 (19:07):
Right?
Speaker 3 (19:08):
But but I think.
Speaker 1 (19:09):
Respect into twined nature of these I mean, if we
had a bit of fun with the musk Oltenland trial,
like how consequential was it that musk lost and and
and the that that does that allow for the IPO.
Is that not really standing in the way of of
OPA's ambitions.
Speaker 2 (19:23):
I don't think that was standing in the way at all,
because there's no way, there's no way the court's ultimately
on appeal eventually, Like there's just there's just no way
the courts were going to say, yeah, like we're going
to just dissolve Open a Eye, the most valuable company
and you know, created in a long time, and and
like the whole the whole idea behind the lawsuit was
like we have to protect this nonprofit. It's like, well,
(19:45):
dissolving Open a Eye just kills the nonprofit, whereas having
a for profit arm with shares and then and where
the nonprofit it shares in the for profit, like it's
going to be one of the most well funded nonprofits
in the world now, so especially when they IPO. So
it's like none of that may ever made any sense
at all. And it was just like as I said,
(20:06):
of those newsletter like great entertainment, love all the drama,
love the discovery. I mean, it's so fun to like
read all these juicy texts and emails. But like, ultimately
that was not a that was just a non issue
when it comes to the companies, but they are intertwined,
and I think the most interesting way they're intertwined in
addition to just they're they're constantly like investing in their
(20:27):
own customers in this circular economy. But like the token,
the AI token is becoming like this commodity that is
like transferable in general purpose and has utility. It's a
new financial instrument and we're actually going to see there.
They are like startups that are that are in stealth
right now. They're going to be launching products in you know,
(20:49):
token futures and things like that right now. Like of
course there's not all tokens are the same. You've got Nvidio,
GPUs that run on Kuda, You've got AMD You've got
you know, Cerebris and TPUs and all these different companies,
but like they're moving in even to be more transferable,
more general purpose, and all the labs can use like
any different GPU now. So it's basically like they've they've
(21:14):
got they've all got this resource that they can that
they can trade and share. And that's what this this
deal with between XAI and Thropic was about. And Thropic
underinvested as we talked about earlier, and xa Xai was
slow on the harnesses, so they didn't have the as
much of the token demands as the labs, so they
sold a bunch of They sold basically their the use
(21:36):
of their gigantic data center in Memphis called Colossus to Anthropic,
and I just thought that was so fascinating. I mean,
just the way that the way that these tokens are
now like underpinning the economy and are just are almost
like a like a cross between like cash and a commodity.
Speaker 3 (21:54):
I don't even know.
Speaker 2 (21:55):
We're gonna have to come up with like a new
definition for what they are.
Speaker 1 (21:58):
We're going to take a short break now when we
can back, We're going to hear from David Webster of
Google Labs all about how Google is continuing to turn
AI research into consumer products, and also to get David's
view on the tech lash and what companies can do
to respond what their role in all of this is.
Before we go to the break, though, read as we
(22:19):
talk about the future of AI and the future of
these companies and stuff, I mean, you talk about it,
It's ultimately coming down to agi kind of battle between
Google and the soon to come SpaceX Tesla conglomerate how
consequential will it be that Google doesn't need to raise
external money to do this that Google does so SpaceX,
(22:40):
SpaceX tester will I mean, SpaceX is burning huge amounts
of cash right and Tesla's business is not growing in
the way that it was, whereas Google is. You know,
their ambition is a city on top of what continues
to be just an extraordinary cash generating.
Speaker 2 (22:54):
Machine that they have this cash cow it you know, yeah,
I mean, it is an advantage for Google, but it's
also like this albatross because they are disrupting their cash cow.
So it's it's very it's unclear exactly how the ads economy,
which is still like eighty percent of their revenue, is
gonna is going to exist in the AI world. I
(23:16):
think they're doing a good job. I think they're doing
their best to trying to transition there. But that's gonna
that's gonna make them slower. And so I don't Elon
Elon world. You know, if you look at his companies,
like he so far has not had any trouble raising
money from investors. People are willing to give him insane multiples,
you know, betting on the future. I mean, look at
(23:37):
his his pay package at Tesla and shareholders approving that.
I mean, his advantage has and always probably will be
just that he moves fast, right, and that's and that's
what Google isn't good at. And so I think it
kind of evens out. It makes like a like a
really interesting competition, like if this, if this were the
the finals of the World Series of AI. Like it's
(23:59):
it's a two really good good teams with like different
strengths and weaknesses. The Elon world has Elon, Like that's
another advantage.
Speaker 3 (24:06):
It's just there.
Speaker 2 (24:07):
There's something about that, you know, that sort of founder
CEO who's kind of crazy and you know, willing to
just like turn on a dime and try new things
and throw out like I'm just not going to make
these cars anymore. We're going to devote it to robots
and do these crazy things. And then like you know,
Google doesn't have that, but they have you know, they
(24:28):
have Demisissabists, and they have Jeff Dean and Soundar. I
think is actually has actually been a good shepherd in
this in this transition, and so they have like a
lot of talent on the executive front as well, but
it's more spread out. It's a bigger ship, it's it's
harder to turn it around, so you know it'll be
like like I I think of it if as maybe
(24:49):
because I'm a former sports reporter of it just a
great a great matchup between two opponents.
Speaker 1 (24:55):
Well when we when we come back, we're going to
hear from inside Google from David Webster, So stay with us.
When you're traveling abroad, Wi Fi is always an issue.
Maybe you can't find it, or maybe you can, but
the Wi Fi network that's available looks a little bit
(25:16):
sketchy and makes you think it might not be safe
to connect to. I've certainly had that feeling, and it's
why having a local SIM card on your phone can
make all the difference. But you don't have to literally
buy one. That's why sale comes in an e sim
service that is as simple as downloading an app, and
that gives you instant Internet access wherever you're traveling. That
(25:37):
is safe and secure Internet with built in cybersecurity features
like web protection and ad blockers, and no need for
local Wi Fi. All you have to do is download
the salely app on your mobile device and choose the
plan that works best for you. Once you install the
e sim sale will be instantly activated, so you can
start your trip off right. Download salely in your app
store and use our code tech Stuff at checkout to
(25:59):
get an extra elusive fifteen percent off your first purchase.
See details in the podcast episode description box. We're joined
now by David Webster, head of UX at Google Labs.
David was on tech Stuff last year for a profile
conversation where we spoke about everything from his training in
product design in Japan, where he learned about the concept
(26:20):
of human machine soul energy, to his role today as
head of UX at Google Labs, where he contributed to
building notebook LM, the viral tool to make personalized podcasts
with Ai. David, you haven't put me out of business yet.
Speaker 4 (26:33):
Not yet. It's really nice to be back on. Nice
to see you again.
Speaker 1 (26:38):
Oz. It's great to see you too. Welcome back to
tech Stuff. We're recording this conversation on Wednesday, May twentieth,
which is day two of Google's annual developer conference Google Io.
And David, you've been quite busy getting ready for this event.
Speaker 4 (26:53):
Yes, yes, indeed, things reach fever pitch the in the
weeks and days before. Iowe it's this entirely self imposed deadline.
That has a great effect on people. Everyone gets really excited.
Everyone surgeries. There's this kind of genuine sort of electricity
in the air. And then we have the big day
(27:14):
where we show our wares to the world, and that
was yesterday. It went well.
Speaker 1 (27:19):
So the headline Axios was Google reinvented and they went
on to say Google is betting that AI is breaking
free from the chatbot and we'll find a home in
virtually everything it makes, including new types of hardware such
as wearables. Is that how it feels from the inside?
Speaker 4 (27:35):
It does? I like that. Actually, it certainly feels like
we're coming out of the of the constrained evolutionary phase
that was the chatbot, and it really for me, it
feels like we're in this moment where it's very very exciting.
I mean, you mentioned in the intro that I get
(27:57):
up in the morning obsessed with this idea of human
machine energy, and it feels like we're in this really
humanizing moment that the technology has got to the point
where now the computer can come to us on our
terms like never before, you know, after decades of us
having to contort to go to the computer and its terms,
(28:17):
it can morph and take the form of a collaborative partner,
which is another way of saying agents. You will have
heard the word agents a lot for me.
Speaker 1 (28:27):
I never had an agent.
Speaker 4 (28:31):
For me. What an agent is this collaborative partner that
meets you where you're at in what feels like a
very organic, very human way. And that's a massive unlock
I think for the relationship between human and technology.
Speaker 1 (28:47):
Now, the computer can come to us is a great phrase.
And that's been your mission I think since before joining Google.
But I mean you joined labs in twenty twenty two.
How does the mission of lab ladder up to and
deliver frankly on the wider company goals of Google, like,
what do you do?
Speaker 4 (29:06):
So what we do in labs is we we The
way we put it is we discover and deliver a
new AI products that serve Google's mission. So, you know,
we look at the new capabilities that are brewing in GDM,
in particular in terms of the model.
Speaker 1 (29:27):
Capabilities GDM is Google deep Mind, Google Deep.
Speaker 4 (29:30):
Mind, and then we from a very user centric perspective,
look for opportunities to bring those capabilities to bear in
ways that are that serve Google's mission. Meaning in ways
that are useful in people's lives. And it does seem
(29:51):
that we're we're in this moment now where the capabilities
are at the point where we can really usefully enhance
a lot of activities the humans value. And I like,
I like thinking about it that way around Googles. Google's
uniquely amongst the tech companies being woven into people's lives
(30:12):
for a long time and and distrusted with all sorts
of important things. And how do we bring the new
affordances of AI to bear and enhancing the kind of
experiences that the people value. It is a is a
great endeavor for for Google, I think.
Speaker 1 (30:31):
And that's what sort of xos for pointing out with
this the idea of breaking free from the chatbot I
guess in their story. But as you mentioned, I mean
Google Io, it's sort of like a it's a developer conference,
but it's sort of like all of the all of
the senior Google that's getting ready for a kind of
annual wedding essentially, right, I mean, I have to the
date is not changing, but but but you and the
product are to get ready for the day.
Speaker 4 (30:53):
Definitely feels like that we don't have too many braidzillas.
But you know we have our moment.
Speaker 1 (30:59):
So but what you I mean, what were you most
proud of and most excited about delivering it? This on
behalf of Google Labs and for the company.
Speaker 4 (31:06):
If I zoom in a click from what I was
saying about, you know, enhancing uniquely sort of human experiences
and bringing the technology, formatting the technology to enable that.
I think the headliner of those experiences is the experience
of creativity. You know, I'm a creative professional. I think
(31:30):
the world would be a better place if as many
people as possible we're able to relish the creative act.
And I think we've got a number of our products
that we've been working on in Labs that do just that.
They enable people to actually bring this technology to bear
and service of creativity. And specifically within that is this
(31:55):
idea of kind of iterating like instead of the take turns,
give a prompt to a chatbot and wait and see
what it comes back with. We're now at the point
with our new omni model in particular, where you can
the experience is much more about you create something in
(32:16):
a very natural way, giving assets that are kind of
indicators of your intention, and then you're able to edit
and sculpt and hone the work in progress in a
way that feels much more natural, much more much more
like sculpting, much more like, much much more familiar to
a creative person.
Speaker 1 (32:36):
The tagline for Omni is create anything from any input.
And this is sort of like shelving the idea of
the prompt engineer and focusing the idea that the technology
you will are building at Google and building on top
of I guess to work at Google deep Mind should
be responsive to human creativity rather than requiring a specialized
(32:59):
set of behaviors to get the most out of.
Speaker 4 (33:02):
Yeah, the trajectory that Omni's on is anything in anything
out is I think the way the demis said it yesterday,
and that is particularly experienceable right now in the context
of Flow, the Flow Suite our surface for media creation.
We've got Flow which is a for movie making. We've
got Flow Music, which is for music and associated media making.
(33:27):
And you can really go in there and it feels
like a very natural, very dimensional collaboration with a partner.
Speaker 3 (33:38):
Right.
Speaker 4 (33:38):
You go in there and you kind of, I don't know,
throw in some images, throw in some like loosely worded
conversation about your intention, get a first instance of the
thing that you want to create, and then with the
capabilities of Omni, you can say, okay, change the camera angle,
change the lighting, change the character, what keeping it consistent
(34:01):
across all the scenes, and you can really start to
craft the character of the of the new Ascid that
you're creating. And it's very it's very satisfying experience.
Speaker 1 (34:12):
It was also the annoucement or Gemini Spark, which somebody
called open floor for normies. For us, what does that mean?
Speaker 4 (34:23):
I love that open club for enormies. It means it's interesting.
It feels like an evolution. The last time we spoke,
we spoke about Mariner, which was this experiment that we
did where you could hand over control of the computer
to an agent, which we would then go and do
(34:46):
things that you wanted it to do. And Spark is
kind of the evolution of of that and incorporation of
that into the Gemini app that they experience. And so
Spark will work ase synchronously. You can, you know, you
can essentially delegate tasks and actions and topics that you're
(35:08):
interested into it, and it will go and work away
in those and then you can check back in and
it will have completed them and have them ready for
your review and approval. And it's across surfaces. It can
be on your mobile device, or it can be in
your computer. It's really it's really nice to see the
early signal from users that the Marina experiment gave us
(35:32):
kind of roll forward into this into this big product
within Gemini.
Speaker 1 (35:36):
So this is this is this is a gentic AI
for people who maybe have never even heard that term,
where if they have heard it, have no idea what
it means.
Speaker 4 (35:45):
Yeah, or that's what the goal is. I have no
idea what it means. Or I was going to say,
maybe roll their eyes at it. Yeah, you don't need
to understand the jargon to experience the upside. It's matured
to the point where you can go and actually just
experience it.
Speaker 1 (36:04):
How do you see the future, because there's a lot
of names flying around like Flow and Omni and Gemini
and Spark and Mariner, but it sort of feels like
from the outside, the direction of all of this is
a one interface for consumers and a AI system that
(36:27):
goes interprets what you want and then goes and does
it without you having to You know choose Am I
going to use Flow or Spark or whatever it make
or notebook, whatever it may be. I mean is it
will will success look like all of these subsidiary products
disappearing and just us reverting to Google being Google.
Speaker 4 (36:47):
It's a good question. I would say maybe one side
of a both, and the version of success looks like that.
I think you're right if I see a common pattern
across a lot of the products that we're working on.
We're getting to this point where the affordancies of the
(37:07):
models and the agents enable everything to be on everything
machine is to some extent, and then the question is that, okay,
do we access that all through one surface? And I
think the way to think about it is you can
certainly provide a door into the room through a surface
(37:28):
like Gemini, where yes, you can go and make videos
through Gemini, or you can go and build a website
or do all the other kind of agent enhanced collaborative
partner sort of experiences through Gemini. But I think there's
also value in having some task specific doors into the room.
(37:48):
It's maybe a consumer prosumer thing, so you can access
a version of Flow that allows you to make movie
assets via Gemini but you can also do it through
the flow surface, maybe in a little bit more of
a detailed or under the hood way, and that's that's
something that we're doing with Notebook LM for example. Right now.
Speaker 1 (38:09):
Demis from from Deep Mind in his in his keynote,
talked about potentially being in the foothills of the singularity. Yeah,
what does that mean to you?
Speaker 4 (38:21):
I mean it's exciting because it's Demis. There's a there's
certain things you can say with U and people listen
if you have a Nobel prize, right, And like Demis,
I find very inspiring because he consistently for his entire
(38:43):
career has been enacting this strategy to bring about a
GI artificial general intelligence, and is now telling people consistently
that we are close in terms of the the capability
of the model. And you know, I think he's one
(39:03):
of the world's best strategists. You know, he's a chess champion,
and so I kind of believe it. And so he's
in he's saying we're getting close. We're starting to have
experiences that feel like glimpses of the potential of AGI,
And for me, I'm very excited that we're at a
(39:26):
moment where those experiences, those glimpses are reassuringly human enhancing.
Speaker 1 (39:36):
Demi's obviously one his Nobel Prize in chemistry for the
work on alpha fold and you know protein simulation, which
is kind of a holy grail of medicine, illness, drugs,
interaction between drugs and molecules, or is it. We haven't
seen any of those drugs come near to market, and
(39:56):
part of it has to do of course with regulations
and FDN all those things. But one of the announcements
that came out of IO was Science Sweet. I wonder
if you can talk about that and also just give
us a prediction when when does Science Suite and the
foothills of the Singularity lead to radically better outcomes in
(40:17):
terms of health potentially.
Speaker 4 (40:19):
So Science Suite is really exciting. Denis mentioned at the
very end of IO yesterday. So Science Suite is this
set of tools that Google Labs our group has been
working on in collaboration with Google Deep Mind, Google Research,
and Google Cloud that massively enhance scientist's ability to do
(40:41):
their kind of creativity, which is a sort of inquisitive
kind of creativity, and in particular within the Science Suite,
what we've been working on our three different applications of
the models. One is an hypothesis generation. This is this
incredibly involved process that can take scientists months. If you're
(41:04):
a research scientist, if you're a PhD. There's a huge
amount of effort that goes into just identifying the high
potential uncharted territory that is worth exploring, worth applying yourself to.
And it's incredibly involved thing. It involves. You know, you
(41:26):
have to review all the previous and current science and
research landscape. You have to do some sort of an
evaluation of the likely value of the yield of your
line of inquiry, and some sort of an environment an
evaluation of the odds of discovering something, and so hypothesis
(41:50):
generation is this place where we can apply AI in
a very dedicated way to confer a lot of advantage
to scientists in terms of where they they decided to
put their attention and energy.
Speaker 1 (42:02):
I was, I think we talked about this, but I was.
I got to do an interview on stage at the
Royal Institution a few weeks ago with me from oh
deep mind, Oh cool. But I was struck then with
I mean this idea, as we're both Brits so I
mean the Royal Institution was designed to, you know, promulgate
(42:23):
the wonder of science and make it accessible to the
British public and kind of inspire inspire people who what
the industrial Revolution could do for society and for them.
I was thinking about those scientists you know who are
dissecting mummies on stage and discovering sodium and stuff. And
if you could teleport them to now and they could
see talk about step change, I mean multiple multiple step
(42:46):
changes in where sciences, they would be blown away with wonder,
I would confidently predict. Yet most people living today in
the world feel the opposite, I mean not only despondent,
but also increasingly angry. What is like, what is your
diagnosis as to why the wonder that you experience on
(43:07):
a day to day basis, your team experiences, the breakthroughs
that you're seeing in some cases building are not only
failing to connect with society but actually enraging people.
Speaker 4 (43:21):
I would say it's.
Speaker 5 (43:24):
There's maybe something about the fact that it's been such
a sudden tech push that is off putting for people.
Meaning I think that that certainly, in my opinion, doing
what I do the maximally kind of human palatical way
(43:46):
for innovation to happen is to lead by human pool
rather than tech push.
Speaker 3 (43:54):
It's to.
Speaker 4 (43:56):
You know, get curious about what matters in people's life,
use that as the jumping off point for generating ideas
about how to format the technology, and and kind of
with a little bit of savor affare format the technology
so that the untie some nuts in people's lives. Right
that that that's the ideal. There's something about the the
(44:18):
speed at which this platform shift came about and the
how how aggressive the competition is to win in this space.
The means that the flavor of the platform shift is
a little bit more tech push than than human poll.
And and then and then it can be quite off
(44:41):
putting in a lot of ways. You know. You can
people get confronted by other people who are slinging jargon
around that doesn't really make sense to them. They're that
they they don't they hear very dramatic articulations of potential
upsides and downsides and don't quite know how to situate
(45:03):
themselves relative to that. And it can create this sort
of aimlaise that we hear about. And I think, what
do you do about that?
Speaker 1 (45:16):
Is you.
Speaker 4 (45:19):
Start from a pace you start from a place of
user curiosity, you start to respectfully understand how this is
feeling to people, and you start to format the technologies
and in ways where they experience the value without having
to hear all the exposition. I actually think Google's really
(45:42):
really well situated to do that. Like Google's already in
a bunch of different flavors giving billions and billions of
people value that they've woven into their existence quite happily.
And Google can enhance that using all of the new
(46:02):
bag of tricks without predicating it with a description of
the mega tricks. You know, but I think the antipathy
is real right.
Speaker 1 (46:11):
Now, the same thing we say in in podcasts show
don't tell.
Speaker 4 (46:17):
Yeah. Yeah, if you're explaining, you're losing, you know.
Speaker 3 (46:21):
Yeah.
Speaker 1 (46:22):
David Webster, thank you.
Speaker 4 (46:23):
You're very welcome. Nice to see you again.
Speaker 1 (46:25):
Oz for tech Stuff, I'm as Voloshin. This episode was
produced by Eliza Dennis and Melissa Slaughter. Is executive produced
by me Julian Nutter and Kate Osborne for Kaleidoscope and
Katrian Novel for iHeart Podcasts. Our Engineers, But he'd Fraser
(46:49):
jack Insley mixed this episode and Kyle Murdoch wrote our
theme song special thank You to Read Abrigotti and David Webster.
Readel Brigotti. Are we living in the foothills of the singularity?
Speaker 2 (00:15):
You know? So we're at Google Io and it's like
two hours of you know, and now we're adding AI
to this and now this is getting better and blah
blah blah, and it's like you're like, okay, cool, Like
you know, there's a million different things Google's doing. It's
very impressive and overwhelming. And then and then Demiss Sabist
closes it out and he's like, and by the way,
you know, we're in the foothills of the singularity. I
(00:37):
don't exactly know what that means. I'm like, well, I'm
going to get some time with him later. I'm going
to ask him, like what he exactly means about that.
I don't know if I have a better idea after
talking to him, like a more specific idea, because I
think he's been very specific about like AGI in five
years or ten years or whatever, and in this in
this case, it's like we're sort of in the realm
(01:00):
of this singularity thing, which no one can really define.
But I but I think I thought about it, and
I was like, you know it it is actually a
helpful comment in a sense because you can get sort
of bogged down in all in the weeds of all
these different product announcements, and it's like okay, like yeah,
AI is getting better and these and every Google product
(01:22):
now is being sort of you know, tuned or enhanced
with with everything that deep Mind is is working on.
But actually there is this like north star for the company,
and they are still like behind the scenes they see
deep Mind, you know, sees its goal as getting to
this you know, this this place of whatever you want
(01:44):
to call it, the singularity or or a g I
a s I you know, and I think, I mean
it's any means guests whether whether that will happen, whether
it's even possible, but it is. It does help to
sort of think, okay, yeah, this is like, this is
what's what's really happening the scenes.
Speaker 1 (02:04):
Let's get into it. Welcome to Tech Stuff. I'm Oz
Volosh and this is the Week in Tech Today, something
a little bit different. For the first part of our show,
we'll be joined by our regular contributor, Read Abricotti, tech
editor for TEMBI four, to discuss the big headlines you
need to know this week. And for the second part,
we've got a special interview. I'm excited to share with
you with David Webster, who is the head of UX
at Google Labs. He's essentially the guy who's responsible for
(02:27):
turning the AI research coming out of Deep Mind and
other parts of Google into consumer products that we're all using.
It's been a crazy week. I mean, the lead story
in the Wall Street Journal today is SpaceX Anthropic and
Open AI's sprint to go public defines the AI booms
big Day, potentially three trillion dollar IPOs this year. At
(02:48):
the same time, commencement speakers around the country are being
booed offstage for mentioning AI. He's Eric Schmidt, the former
CEO of Google at the University of Arizona. So the
question is whether you will help shape artificial intelligence.
Speaker 3 (03:04):
We do not know.
Speaker 1 (03:07):
We do not know.
Speaker 3 (03:08):
I know what many of you are feeling about that.
Speaker 2 (03:11):
I can hear you.
Speaker 3 (03:13):
There is a fear if you don't care about science.
Speaker 2 (03:20):
That's okay, because AI is going to touch everything else
as well.
Speaker 3 (03:24):
Whatever path you too, we are become more.
Speaker 2 (03:29):
Kind of been his favorite day of the year. No,
I mean, it is interesting though. I wrote this this
funny sort of newsletter intro yesterday about you know, my
guide to giving commencement speeches, because this has been happening
all over the country. If people give these speeches, and
of course they want to talk about AI, college students
don't want to hear it. I think they've been sort
(03:51):
of told probably honestly, I think inaccurately that AI is
going to mean no jobs for them. I think I
think actually the job market, it and and and a
lot of these tech layoffs are really have a lot
more to do with other, you know, a complex set
of issues, and AI is not really not really one
of them, maybe indirectly, but but no, I I wrote
(04:14):
this funny this funny piece which you should you should
check out at the seminar.
Speaker 1 (04:17):
Deadline where Sunscreen right, which was a termination.
Speaker 2 (04:20):
Everyone right, everyone everyone is uh, you know, was always
mentioned the where sunscreen you know, commencement. It was actually
I guess like a call a newspaper column commencement speech.
But but I was like, you know, you could go
that you don't talk about AI, but if you do,
go the Sarah Connor route, you know who who was
a college student right? And in Terminator one fictional person
(04:43):
if you're if you're too old for if you're too young, sorry,
I'm old, you know, but but you know she was
a college student who I who I said, you know,
became an AI safety advocate and as a as a
robot was chasing her and she had this great line
and terminator too, which is, you know, if you're not
wearing two million sun block, you're going to have a
really bad day.
Speaker 3 (05:02):
Get it. So it's a great, great wine for a
commencement speech. So that's my to.
Speaker 1 (05:10):
Wonder though, because you know, Schmid's commencement speech was after
this one. We talked about last week Gloria Callfield, who
gave the commencement speech at a Flora University and also
got bood talking about AI. Jensen Hawaian gave a Connie
hey Meling commencement of speech where he got a heroes
welcome talking about AI. But I guess, I guess it's
a robotics I mean it's a famous robotics university. So
(05:31):
the warmer crowds, maybe Erishman is looking at It's like,
am I more like Gloria Corefield and more like Jensen?
Speaker 3 (05:36):
Why?
Speaker 1 (05:36):
It's like I'm more like Jensen waning? I can I
can pull off talking about AI and not get a
bood off stage? But I'm just like, which students are
I going to bed on? H It's an interesting question.
So I actually asked David Webster a little bit of
plug for the second half of this episode for his
diagnosis of the tech clash and how the industry should respond,
and they'll talk about that later in this episode. But read,
(05:58):
I mean talk about this easy week from business news perspective.
I mean, the Will Street Journal was literally was giddy.
I opened the app when it was like jumping jumping
off the screen to me about this. Uh, you know,
the allies trillion dollar evaluations, these IPOs have been kind
of cooking for a while, but like, what is it
about this week that has has kind of been so insane?
Speaker 2 (06:18):
Yeah, I mean, you know, you've got You've got SpaceX,
you know filing, You've got open Ai, Anthropic. You know,
they all want they all want to go public this year.
They're all going to be the biggest I pos in history.
You know, SpaceX will probably take the crown there and
you know, I just it's like, Wow, there's a lot.
Speaker 3 (06:39):
Of money in the world. If all these companies can
go public in the same year.
Speaker 2 (06:45):
I don't know, I don't know what percentage they'll float,
you know, I think it'll be that'll be interesting and
we'll see what happens with like the lock up.
Speaker 3 (06:51):
There'll be some interesting stuff.
Speaker 2 (06:53):
Liz Hoffin's written about this at some and four, you know,
the interesting ideas around how to deal.
Speaker 3 (06:58):
With the lock up period.
Speaker 2 (06:59):
But it is I mean, look like so much wealth
and so much value is being created right now out
here in this like small small area that I live
in in Silicon Valley. I am not one of these
people who's become a billionaire in the last year, but
it's just like an incredible amount of wealth creation. And
I think it is I mean, there is something to
(07:21):
be excited about because you know, when these companies go
public now, they they'll be in you know four one
ks and and you know in Trump accounts and things
like that. And I think the next generation we'll be
able to you know, profit off of of all this
wealth creation.
Speaker 1 (07:35):
But you pointed out in your newsletter, which that was interesting,
was like for the last fifteen years, the luckiest students
who are in the cohort, who get jobs at you know,
Metro and Google and other places, basically had a life
of you know, free snacks and stock options. And you
know that didn't have never apply to the majority of students,
but applied to like a you know, somewhat of a
reasonable sliver. And now there's this crazy wealth creation going on,
(07:59):
but in the hands very few, and it's a stra
It's a very strange cultural moment. How much do you
think this threatens, Like if you're a tech post and
you like, on the one hand, my house is being
fly at boomed. On the other time, I'm about to
ip on my company for Triton dollos, Like, which are
you more focused on? Probably the Tritian dollos.
Speaker 2 (08:17):
I would hope these people are not that focused on,
you know, the the AI backlash. I mean, it's it's ultimately,
you know, it's a popular thing to say. I was
actually at my kid's hockey practice last night and their
ten year olds are in there talking about how much
they hate AI, and I was like, oh, that's interesting.
Like I asked one of them, like, what do you
(08:38):
what do you not like about AI? And he said, well,
it's it's wrong. You know, it's wrong, and like so like, yeah,
fifty percent of the time it's wrong. And I'm like,
your dad is like an AI executive, So I you know,
it's interesting that you're saying that, But no, I mean,
I don't know how much this this backlash. I mean,
certainly people don't want data centers, and certainly I think
(09:00):
AI is a convenient scapegoat. There there are bigger structural
issues and we don't have time to get into those
on why. You know, you have these periods of massive
wealth creation like we did, you know in the in
the the last couple of tech booms, and regular people
don't benefit from that. But I would if I were
a college student today, I'd be super excited because there's
(09:21):
more opportunity right now than ever. This technology is very
easy to learn, and I'd love to go just try
to start some companies. And I mean, I think it's
it's an incredible, incredible opportunity. I think this moment, I
think we'll look back and we'll say college students, who
people were coming out of college today, we look back
ten years from now, had like one of the best
(09:42):
opportunities to become wealthy in history.
Speaker 1 (09:46):
So we may not we may we may or may
not be in the foothills of the singularity, but we
are in the foothills of the of the of the
gold Rush as far as you're consent.
Speaker 2 (09:55):
I don't know what the singularity is. I'm still very
confused about that. I always thought it was we merge
with the with the machines or something, right, Yeah, and I
but but like what I do what I do see
is is that like when we employ AI and at Semaphore,
when we when we use these products, it creates more
work for people, for humans, Like we're going to have
(10:15):
to hire people because of AI because now we can
do these products like we launched Semaphore Intelligence. I think
we talked about this the other day, and it's like,
I mean, I think that's these giant tech companies that
we're paying people to do practically no work and give
them free snacks all day for stock options and a
good salary. Like, I mean, yeah, they're gonna have to
(10:36):
get lean. And there's a lot of reasons for that, right.
AI indirectly is one of them, because because there's a
lot of capex that needs to to go into this.
But you know, I I don't think this is like
going to kill jobs at all.
Speaker 1 (10:52):
Let's talk about those three IPOs and a bit more
detailed though, because everyone's many people have seen the headlines.
But that's that's kind of three different different stories. Space
that is actually filed this week, so the most information
which was previously you know probably but is no public
about that company. You know, three things struck me. You know,
one how profitable Starlink is and how big stall inkins,
(11:13):
and two how lols making SpaceX is and how lols
making x x AI is. What did you take away
from that fighting.
Speaker 3 (11:22):
Yeah, I mean, first of all, Starlink is amazing.
Speaker 2 (11:25):
I mean I remember when I first came out to
cover the tech industry, people told me that this this
idea of satellite internet was like pie in the sky,
like no one it was impossible, no one was ever
going to do it. I've I think I bought three
Starlink satellites in my in my in my life now
because you know, dishes, because like I live in California
where there's like outside of actually even in the cities,
(11:48):
like there's basically no cell reception. Great, and I like
to be outdoors and also have to work a lot.
So you know, like, I think this technology is incredible,
done surprises me at all that is making a ton
of revenue. But also you know, like Xai for instance, yeah,
I mean losing a ton of money. I was talking
with someone the other day at Google Io who was saying, oh,
(12:11):
I think like Xai took themselves off the board because
they they sold their compute to Anthropic and Anthropics, you
know in the lead.
Speaker 3 (12:20):
I mean I don't. I don't buy that at all.
Speaker 1 (12:22):
I think that's we love pank one point two billion
dollars a month for the next four or five years
to Xai and.
Speaker 2 (12:30):
And look Anthropics profitable, Like the Wall Street Journal had
to scoop the other day that Anthropic is now profitable.
I was joking around about this, like, oops, they're profitable,
like that that's a big mistake. But it's actually true,
Like they did not want to be profitable, Like that
is not a good thing in this in a in
a fast growing, high capex environment, to be profitable like
(12:51):
you you want to be if you're positible.
Speaker 1 (12:53):
What you're saying is they've underinvested.
Speaker 3 (12:55):
They've underinvested. I mean they may.
Speaker 2 (12:57):
I think it's pretty clear now, even though like all
the smartest financial journalists and tech journalists last year were
telling you that this was a bubble and that Opening
Eye was irresponsible, that it was all going to come
crashing down. It's pretty clear that was totally not true.
It was complete bs And if people had actually used
these products, they would know that because they'd, like me,
(13:17):
they'd be like at two in the morning running out
of tokens and freaking out signs.
Speaker 1 (13:21):
Throping one hundred and thirty percent revenue growth didn't surprise
me that much. First profitable quarter, I let your hot
take that's a bad thing. But the compute cost, basically
the cost of delivering the services went down from seventy
one cents per dollar of revenue to fifty six percent
fifty six cents per dollar revenue. That slightly niche factoid
was the one that I found most interesting. I mean,
(13:42):
that's a huge decline in the cost of the cost
of goods sold for them.
Speaker 3 (13:47):
Yeah, I know.
Speaker 2 (13:48):
I think I think that needs to be like unpacked
a little bit because I'm like, well, okay, what does
that actually mean? Like when you there's there's a question
I have, which is like if I pay a subscription
fee Tomanthropic or any of these companies, now, like they
are incentivized to provide as few tokens to me as
as possible. So I don't know where where all of
(14:09):
these efficiencies are coming. But generally, yeah, I mean I
I honestly, I'm I'm like a conspiracy theorist there probably,
but but like honestly, like the compute costs just come down,
Like that's the one thing. I mean, like if you're
if you're like there's a lot of companies where it's
like yeah, I mean like Uber was one of them.
It's like, are you really going to bring the cost
(14:29):
of driving a car down? Like not really, Like ultimately
you kind of just got this big cost that's never
going to change. Compute costs always come down, Like people
just find a way to make this cheaper. Right now,
it's like piling GPUs and denser and denser racks into
data centers and you know, liquid cooling and all this
optical connections, et cetera. And it's it's a huge industrial effort.
(14:52):
But like there will be breakthroughs that will be you know,
there's so much money in this. Like I just don't
think these compute costs are gonna get you know, they're
going to be like they're going to stay the same,
They're just going to continue to drop. So like be
unprofitable today, you will be extremely profitable tomorrow when the
costs come down and the demand just continues to go up.
We haven't even really seen penetration in the market yet.
(15:16):
I mean, this is just a tiny niche of people
who are like using all these tokens.
Speaker 1 (15:20):
So let's talk about open AI in exactly that respect.
Because today's Thursday, we're recording. The episode is going to
come up on Friday. And again the Journal reported that
maybe on Friday open Ai I would actually file, So
we don't know yet, but suddenly it's coming soon. And
one of the big issues about open AI, which you've
reported on, is the kind of conflict between Similtman and
the CFO. The Simultan wanting to basically invest more and
(15:43):
more ahead of revenue and compute in the CFO trying
to hold him back a little bit, like what what
can we expect from the IPO filing. In terms of
this debate.
Speaker 2 (15:52):
I have not written that there's a huge conflict between
Altman and the CFO. I mean, other publications Wall Street
Journal has said like there's this huge rift and the CFO.
Actually I think it was the information for me before
that we should give them a credit. But you know,
like the that Sarah friar Is doesn't want to IPO
(16:12):
and Sam Altman does, and the numbers are horrible. The
growth is slowing down, they can't get enough users in
chat gpt. They've painted a really bleak picture. So I
think we'll now get like a better sense, and then
when Nthropic files, it'll be interesting to compare because right now,
like there's all these comparisons on Entropics revenue and Opening
Eyes revenue and people are basically a lot of I
(16:36):
think it's like conventional wisdom that Anthropic is like fast
has faster growing revenue. Well, I think we'll hopefully we'll
get more of an apples to Apple's comparison. But Opening
as interesting because it's like chat gpt is a consumer business,
and then where all the tokens are being consumed is
kind of this like prosumer slash enterprise part of the
(16:57):
business with Codex, for like all these odors are using
you know, using these these harnesses. So that's like Opening
Eyes really two things and Anthropic is more one thing.
And then I think the way I look at the
way I look at it is like there are two
companies that are actually competing in the long term for
(17:17):
AI dominance, and that's Google and the other one doesn't
exist yet because it's going to be SpaceX and Tesla combined,
which is going to happen. And if you look at
it like those will be the only two companies in
the world that have leading frontier models. If you if
you agree that GROC is a leading frontier model, I
think it's I think it is. I think it gets
(17:39):
a bad ride. It's not as far behind as some
people say. And also have robotics deployed in the world
at scale, like advanced robotics. There's really because they have
there's Google has way Mo, and then Tesla has you know,
the most probably the most advanced robots in the world
out you know, out on the roads. And then and
(18:00):
then they're both kind of dabbling in space. I mean,
I don't know. Google doesn't have rockets obviously, but they're
they're an investor in SpaceX, so they'll make some they'll
make some money.
Speaker 1 (18:09):
Well that's I mean, that's why I want to come
back to Google just before we go to the break
and then come back to David Webster from Google. But
I mean that the sense in which all these companies
are intertwined is kind of extraordinary, right, Like you know,
anthropic of paying SpaceX Elon just lost his lawsuit against Semmeltman,
which is removed an impediment to open a eyes IPO.
(18:30):
Like I mean, I guess give give a reflection on
the conclusion of that trial. But also like is it
a rising tide will lift all these boats or is
Aaron sense going to be Is it going to be
more of a winner takes it all model?
Speaker 2 (18:42):
Do you think, Oh no, no, I don't mean it's
winner takes all at all? Like I actually think the
rising tide the demand is so huge, like all of
these companies are going to be hugely successful, Like there's not,
I don't I there's very little chance that like these
companies are just going to fail or something. Right, there's
so like I said earlier, like the demand, it's not
(19:03):
there's not even really market penetration yet and we're already
seeing supply.
Speaker 3 (19:06):
Can't you keep up with demand?
Speaker 4 (19:07):
Right?
Speaker 3 (19:08):
But but I think.
Speaker 1 (19:09):
Respect into twined nature of these I mean, if we
had a bit of fun with the musk Oltenland trial,
like how consequential was it that musk lost and and
and the that that does that allow for the IPO.
Is that not really standing in the way of of
OPA's ambitions.
Speaker 2 (19:23):
I don't think that was standing in the way at all,
because there's no way, there's no way the court's ultimately
on appeal eventually, Like there's just there's just no way
the courts were going to say, yeah, like we're going
to just dissolve Open a Eye, the most valuable company
and you know, created in a long time, and and
like the whole the whole idea behind the lawsuit was
like we have to protect this nonprofit. It's like, well,
(19:45):
dissolving Open a Eye just kills the nonprofit, whereas having
a for profit arm with shares and then and where
the nonprofit it shares in the for profit, like it's
going to be one of the most well funded nonprofits
in the world now, so especially when they IPO. So
it's like none of that may ever made any sense
at all. And it was just like as I said,
(20:06):
of those newsletter like great entertainment, love all the drama,
love the discovery. I mean, it's so fun to like
read all these juicy texts and emails. But like, ultimately
that was not a that was just a non issue
when it comes to the companies, but they are intertwined,
and I think the most interesting way they're intertwined in
addition to just they're they're constantly like investing in their
(20:27):
own customers in this circular economy. But like the token,
the AI token is becoming like this commodity that is
like transferable in general purpose and has utility. It's a
new financial instrument and we're actually going to see there.
They are like startups that are that are in stealth
right now. They're going to be launching products in you know,
(20:49):
token futures and things like that right now. Like of
course there's not all tokens are the same. You've got Nvidio,
GPUs that run on Kuda, You've got AMD You've got
you know, Cerebris and TPUs and all these different companies,
but like they're moving in even to be more transferable,
more general purpose, and all the labs can use like
any different GPU now. So it's basically like they've they've
(21:14):
got they've all got this resource that they can that
they can trade and share. And that's what this this
deal with between XAI and Thropic was about. And Thropic
underinvested as we talked about earlier, and xa Xai was
slow on the harnesses, so they didn't have the as
much of the token demands as the labs, so they
sold a bunch of They sold basically their the use
(21:36):
of their gigantic data center in Memphis called Colossus to Anthropic,
and I just thought that was so fascinating. I mean,
just the way that the way that these tokens are
now like underpinning the economy and are just are almost
like a like a cross between like cash and a commodity.
Speaker 3 (21:54):
I don't even know.
Speaker 2 (21:55):
We're gonna have to come up with like a new
definition for what they are.
Speaker 1 (21:58):
We're going to take a short break now when we
can back, We're going to hear from David Webster of
Google Labs all about how Google is continuing to turn
AI research into consumer products, and also to get David's
view on the tech lash and what companies can do
to respond what their role in all of this is.
Before we go to the break, though, read as we
(22:19):
talk about the future of AI and the future of
these companies and stuff, I mean, you talk about it,
It's ultimately coming down to agi kind of battle between
Google and the soon to come SpaceX Tesla conglomerate how
consequential will it be that Google doesn't need to raise
external money to do this that Google does so SpaceX,
(22:40):
SpaceX tester will I mean, SpaceX is burning huge amounts
of cash right and Tesla's business is not growing in
the way that it was, whereas Google is. You know,
their ambition is a city on top of what continues
to be just an extraordinary cash generating.
Speaker 2 (22:54):
Machine that they have this cash cow it you know, yeah,
I mean, it is an advantage for Google, but it's
also like this albatross because they are disrupting their cash cow.
So it's it's very it's unclear exactly how the ads economy,
which is still like eighty percent of their revenue, is
gonna is going to exist in the AI world. I
(23:16):
think they're doing a good job. I think they're doing
their best to trying to transition there. But that's gonna
that's gonna make them slower. And so I don't Elon
Elon world. You know, if you look at his companies,
like he so far has not had any trouble raising
money from investors. People are willing to give him insane multiples,
you know, betting on the future. I mean, look at
(23:37):
his his pay package at Tesla and shareholders approving that.
I mean, his advantage has and always probably will be
just that he moves fast, right, and that's and that's
what Google isn't good at. And so I think it
kind of evens out. It makes like a like a
really interesting competition, like if this, if this were the
the finals of the World Series of AI. Like it's
(23:59):
it's a two really good good teams with like different
strengths and weaknesses. The Elon world has Elon, Like that's
another advantage.
Speaker 3 (24:06):
It's just there.
Speaker 2 (24:07):
There's something about that, you know, that sort of founder
CEO who's kind of crazy and you know, willing to
just like turn on a dime and try new things
and throw out like I'm just not going to make
these cars anymore. We're going to devote it to robots
and do these crazy things. And then like you know,
Google doesn't have that, but they have you know, they
(24:28):
have Demisissabists, and they have Jeff Dean and Soundar. I
think is actually has actually been a good shepherd in
this in this transition, and so they have like a
lot of talent on the executive front as well, but
it's more spread out. It's a bigger ship, it's it's
harder to turn it around, so you know it'll be
like like I I think of it if as maybe
(24:49):
because I'm a former sports reporter of it just a
great a great matchup between two opponents.
Speaker 1 (24:55):
Well when we when we come back, we're going to
hear from inside Google from David Webster, So stay with us.
When you're traveling abroad, Wi Fi is always an issue.
Maybe you can't find it, or maybe you can, but
the Wi Fi network that's available looks a little bit
(25:16):
sketchy and makes you think it might not be safe
to connect to. I've certainly had that feeling, and it's
why having a local SIM card on your phone can
make all the difference. But you don't have to literally
buy one. That's why sale comes in an e sim
service that is as simple as downloading an app, and
that gives you instant Internet access wherever you're traveling. That
(25:37):
is safe and secure Internet with built in cybersecurity features
like web protection and ad blockers, and no need for
local Wi Fi. All you have to do is download
the salely app on your mobile device and choose the
plan that works best for you. Once you install the
e sim sale will be instantly activated, so you can
start your trip off right. Download salely in your app
store and use our code tech Stuff at checkout to
(25:59):
get an extra elusive fifteen percent off your first purchase.
See details in the podcast episode description box. We're joined
now by David Webster, head of UX at Google Labs.
David was on tech Stuff last year for a profile
conversation where we spoke about everything from his training in
product design in Japan, where he learned about the concept
(26:20):
of human machine soul energy, to his role today as
head of UX at Google Labs, where he contributed to
building notebook LM, the viral tool to make personalized podcasts
with Ai. David, you haven't put me out of business yet.
Speaker 4 (26:33):
Not yet. It's really nice to be back on. Nice
to see you again.
Speaker 1 (26:38):
Oz. It's great to see you too. Welcome back to
tech Stuff. We're recording this conversation on Wednesday, May twentieth,
which is day two of Google's annual developer conference Google Io.
And David, you've been quite busy getting ready for this event.
Speaker 4 (26:53):
Yes, yes, indeed, things reach fever pitch the in the
weeks and days before. Iowe it's this entirely self imposed deadline.
That has a great effect on people. Everyone gets really excited.
Everyone surgeries. There's this kind of genuine sort of electricity
in the air. And then we have the big day
(27:14):
where we show our wares to the world, and that
was yesterday. It went well.
Speaker 1 (27:19):
So the headline Axios was Google reinvented and they went
on to say Google is betting that AI is breaking
free from the chatbot and we'll find a home in
virtually everything it makes, including new types of hardware such
as wearables. Is that how it feels from the inside?
Speaker 4 (27:35):
It does? I like that. Actually, it certainly feels like
we're coming out of the of the constrained evolutionary phase
that was the chatbot, and it really for me, it
feels like we're in this moment where it's very very exciting.
I mean, you mentioned in the intro that I get
(27:57):
up in the morning obsessed with this idea of human
machine energy, and it feels like we're in this really
humanizing moment that the technology has got to the point
where now the computer can come to us on our
terms like never before, you know, after decades of us
having to contort to go to the computer and its terms,
(28:17):
it can morph and take the form of a collaborative partner,
which is another way of saying agents. You will have
heard the word agents a lot for me.
Speaker 1 (28:27):
I never had an agent.
Speaker 4 (28:31):
For me. What an agent is this collaborative partner that
meets you where you're at in what feels like a
very organic, very human way. And that's a massive unlock
I think for the relationship between human and technology.
Speaker 1 (28:47):
Now, the computer can come to us is a great phrase.
And that's been your mission I think since before joining Google.
But I mean you joined labs in twenty twenty two.
How does the mission of lab ladder up to and
deliver frankly on the wider company goals of Google, like,
what do you do?
Speaker 4 (29:06):
So what we do in labs is we we The
way we put it is we discover and deliver a
new AI products that serve Google's mission. So, you know,
we look at the new capabilities that are brewing in GDM,
in particular in terms of the model.
Speaker 1 (29:27):
Capabilities GDM is Google deep Mind, Google Deep.
Speaker 4 (29:30):
Mind, and then we from a very user centric perspective,
look for opportunities to bring those capabilities to bear in
ways that are that serve Google's mission. Meaning in ways
that are useful in people's lives. And it does seem
(29:51):
that we're we're in this moment now where the capabilities
are at the point where we can really usefully enhance
a lot of activities the humans value. And I like,
I like thinking about it that way around Googles. Google's
uniquely amongst the tech companies being woven into people's lives
(30:12):
for a long time and and distrusted with all sorts
of important things. And how do we bring the new
affordances of AI to bear and enhancing the kind of
experiences that the people value. It is a is a
great endeavor for for Google, I think.
Speaker 1 (30:31):
And that's what sort of xos for pointing out with
this the idea of breaking free from the chatbot I
guess in their story. But as you mentioned, I mean
Google Io, it's sort of like a it's a developer conference,
but it's sort of like all of the all of
the senior Google that's getting ready for a kind of
annual wedding essentially, right, I mean, I have to the
date is not changing, but but but you and the
product are to get ready for the day.
Speaker 4 (30:53):
Definitely feels like that we don't have too many braidzillas.
But you know we have our moment.
Speaker 1 (30:59):
So but what you I mean, what were you most
proud of and most excited about delivering it? This on
behalf of Google Labs and for the company.
Speaker 4 (31:06):
If I zoom in a click from what I was
saying about, you know, enhancing uniquely sort of human experiences
and bringing the technology, formatting the technology to enable that.
I think the headliner of those experiences is the experience
of creativity. You know, I'm a creative professional. I think
(31:30):
the world would be a better place if as many
people as possible we're able to relish the creative act.
And I think we've got a number of our products
that we've been working on in Labs that do just that.
They enable people to actually bring this technology to bear
and service of creativity. And specifically within that is this
(31:55):
idea of kind of iterating like instead of the take turns,
give a prompt to a chatbot and wait and see
what it comes back with. We're now at the point
with our new omni model in particular, where you can
the experience is much more about you create something in
(32:16):
a very natural way, giving assets that are kind of
indicators of your intention, and then you're able to edit
and sculpt and hone the work in progress in a
way that feels much more natural, much more much more
like sculpting, much more like, much much more familiar to
a creative person.
Speaker 1 (32:36):
The tagline for Omni is create anything from any input.
And this is sort of like shelving the idea of
the prompt engineer and focusing the idea that the technology
you will are building at Google and building on top
of I guess to work at Google deep Mind should
be responsive to human creativity rather than requiring a specialized
(32:59):
set of behaviors to get the most out of.
Speaker 4 (33:02):
Yeah, the trajectory that Omni's on is anything in anything
out is I think the way the demis said it yesterday,
and that is particularly experienceable right now in the context
of Flow, the Flow Suite our surface for media creation.
We've got Flow which is a for movie making. We've
got Flow Music, which is for music and associated media making.
(33:27):
And you can really go in there and it feels
like a very natural, very dimensional collaboration with a partner.
Speaker 3 (33:38):
Right.
Speaker 4 (33:38):
You go in there and you kind of, I don't know,
throw in some images, throw in some like loosely worded
conversation about your intention, get a first instance of the
thing that you want to create, and then with the
capabilities of Omni, you can say, okay, change the camera angle,
change the lighting, change the character, what keeping it consistent
(34:01):
across all the scenes, and you can really start to
craft the character of the of the new Ascid that
you're creating. And it's very it's very satisfying experience.
Speaker 1 (34:12):
It was also the annoucement or Gemini Spark, which somebody
called open floor for normies. For us, what does that mean?
Speaker 4 (34:23):
I love that open club for enormies. It means it's interesting.
It feels like an evolution. The last time we spoke,
we spoke about Mariner, which was this experiment that we
did where you could hand over control of the computer
to an agent, which we would then go and do
(34:46):
things that you wanted it to do. And Spark is
kind of the evolution of of that and incorporation of
that into the Gemini app that they experience. And so
Spark will work ase synchronously. You can, you know, you
can essentially delegate tasks and actions and topics that you're
(35:08):
interested into it, and it will go and work away
in those and then you can check back in and
it will have completed them and have them ready for
your review and approval. And it's across surfaces. It can
be on your mobile device, or it can be in
your computer. It's really it's really nice to see the
early signal from users that the Marina experiment gave us
(35:32):
kind of roll forward into this into this big product
within Gemini.
Speaker 1 (35:36):
So this is this is this is a gentic AI
for people who maybe have never even heard that term,
where if they have heard it, have no idea what
it means.
Speaker 4 (35:45):
Yeah, or that's what the goal is. I have no
idea what it means. Or I was going to say,
maybe roll their eyes at it. Yeah, you don't need
to understand the jargon to experience the upside. It's matured
to the point where you can go and actually just
experience it.
Speaker 1 (36:04):
How do you see the future, because there's a lot
of names flying around like Flow and Omni and Gemini
and Spark and Mariner, but it sort of feels like
from the outside, the direction of all of this is
a one interface for consumers and a AI system that
(36:27):
goes interprets what you want and then goes and does
it without you having to You know choose Am I
going to use Flow or Spark or whatever it make
or notebook, whatever it may be. I mean is it
will will success look like all of these subsidiary products
disappearing and just us reverting to Google being Google.
Speaker 4 (36:47):
It's a good question. I would say maybe one side
of a both, and the version of success looks like that.
I think you're right if I see a common pattern
across a lot of the products that we're working on.
We're getting to this point where the affordancies of the
(37:07):
models and the agents enable everything to be on everything
machine is to some extent, and then the question is that, okay,
do we access that all through one surface? And I
think the way to think about it is you can
certainly provide a door into the room through a surface
(37:28):
like Gemini, where yes, you can go and make videos
through Gemini, or you can go and build a website
or do all the other kind of agent enhanced collaborative
partner sort of experiences through Gemini. But I think there's
also value in having some task specific doors into the room.
(37:48):
It's maybe a consumer prosumer thing, so you can access
a version of Flow that allows you to make movie
assets via Gemini but you can also do it through
the flow surface, maybe in a little bit more of
a detailed or under the hood way, and that's that's
something that we're doing with Notebook LM for example. Right now.
Speaker 1 (38:09):
Demis from from Deep Mind in his in his keynote,
talked about potentially being in the foothills of the singularity. Yeah,
what does that mean to you?
Speaker 4 (38:21):
I mean it's exciting because it's Demis. There's a there's
certain things you can say with U and people listen
if you have a Nobel prize, right, And like Demis,
I find very inspiring because he consistently for his entire
(38:43):
career has been enacting this strategy to bring about a
GI artificial general intelligence, and is now telling people consistently
that we are close in terms of the the capability
of the model. And you know, I think he's one
(39:03):
of the world's best strategists. You know, he's a chess champion,
and so I kind of believe it. And so he's
in he's saying we're getting close. We're starting to have
experiences that feel like glimpses of the potential of AGI,
And for me, I'm very excited that we're at a
(39:26):
moment where those experiences, those glimpses are reassuringly human enhancing.
Speaker 1 (39:36):
Demi's obviously one his Nobel Prize in chemistry for the
work on alpha fold and you know protein simulation, which
is kind of a holy grail of medicine, illness, drugs,
interaction between drugs and molecules, or is it. We haven't
seen any of those drugs come near to market, and
(39:56):
part of it has to do of course with regulations
and FDN all those things. But one of the announcements
that came out of IO was Science Sweet. I wonder
if you can talk about that and also just give
us a prediction when when does Science Suite and the
foothills of the Singularity lead to radically better outcomes in
(40:17):
terms of health potentially.
Speaker 4 (40:19):
So Science Suite is really exciting. Denis mentioned at the
very end of IO yesterday. So Science Suite is this
set of tools that Google Labs our group has been
working on in collaboration with Google Deep Mind, Google Research,
and Google Cloud that massively enhance scientist's ability to do
(40:41):
their kind of creativity, which is a sort of inquisitive
kind of creativity, and in particular within the Science Suite,
what we've been working on our three different applications of
the models. One is an hypothesis generation. This is this
incredibly involved process that can take scientists months. If you're
(41:04):
a research scientist, if you're a PhD. There's a huge
amount of effort that goes into just identifying the high
potential uncharted territory that is worth exploring, worth applying yourself to.
And it's incredibly involved thing. It involves. You know, you
(41:26):
have to review all the previous and current science and
research landscape. You have to do some sort of an
evaluation of the likely value of the yield of your
line of inquiry, and some sort of an environment an
evaluation of the odds of discovering something, and so hypothesis
(41:50):
generation is this place where we can apply AI in
a very dedicated way to confer a lot of advantage
to scientists in terms of where they they decided to
put their attention and energy.
Speaker 1 (42:02):
I was, I think we talked about this, but I was.
I got to do an interview on stage at the
Royal Institution a few weeks ago with me from oh
deep mind, Oh cool. But I was struck then with
I mean this idea, as we're both Brits so I
mean the Royal Institution was designed to, you know, promulgate
(42:23):
the wonder of science and make it accessible to the
British public and kind of inspire inspire people who what
the industrial Revolution could do for society and for them.
I was thinking about those scientists you know who are
dissecting mummies on stage and discovering sodium and stuff. And
if you could teleport them to now and they could
see talk about step change, I mean multiple multiple step
(42:46):
changes in where sciences, they would be blown away with wonder,
I would confidently predict. Yet most people living today in
the world feel the opposite, I mean not only despondent,
but also increasingly angry. What is like, what is your
diagnosis as to why the wonder that you experience on
(43:07):
a day to day basis, your team experiences, the breakthroughs
that you're seeing in some cases building are not only
failing to connect with society but actually enraging people.
Speaker 4 (43:21):
I would say it's.
Speaker 5 (43:24):
There's maybe something about the fact that it's been such
a sudden tech push that is off putting for people.
Meaning I think that that certainly, in my opinion, doing
what I do the maximally kind of human palatical way
(43:46):
for innovation to happen is to lead by human pool
rather than tech push.
Speaker 3 (43:54):
It's to.
Speaker 4 (43:56):
You know, get curious about what matters in people's life,
use that as the jumping off point for generating ideas
about how to format the technology, and and kind of
with a little bit of savor affare format the technology
so that the untie some nuts in people's lives. Right
that that that's the ideal. There's something about the the
(44:18):
speed at which this platform shift came about and the
how how aggressive the competition is to win in this space.
The means that the flavor of the platform shift is
a little bit more tech push than than human poll.
And and then and then it can be quite off
(44:41):
putting in a lot of ways. You know. You can
people get confronted by other people who are slinging jargon
around that doesn't really make sense to them. They're that
they they don't they hear very dramatic articulations of potential
upsides and downsides and don't quite know how to situate
(45:03):
themselves relative to that. And it can create this sort
of aimlaise that we hear about. And I think, what
do you do about that?
Speaker 1 (45:16):
Is you.
Speaker 4 (45:19):
Start from a pace you start from a place of
user curiosity, you start to respectfully understand how this is
feeling to people, and you start to format the technologies
and in ways where they experience the value without having
to hear all the exposition. I actually think Google's really
(45:42):
really well situated to do that. Like Google's already in
a bunch of different flavors giving billions and billions of
people value that they've woven into their existence quite happily.
And Google can enhance that using all of the new
(46:02):
bag of tricks without predicating it with a description of
the mega tricks. You know, but I think the antipathy
is real right.
Speaker 1 (46:11):
Now, the same thing we say in in podcasts show
don't tell.
Speaker 4 (46:17):
Yeah. Yeah, if you're explaining, you're losing, you know.
Speaker 3 (46:21):
Yeah.
Speaker 1 (46:22):
David Webster, thank you.
Speaker 4 (46:23):
You're very welcome. Nice to see you again.
Speaker 1 (46:25):
Oz for tech Stuff, I'm as Voloshin. This episode was
produced by Eliza Dennis and Melissa Slaughter. Is executive produced
by me Julian Nutter and Kate Osborne for Kaleidoscope and
Katrian Novel for iHeart Podcasts. Our Engineers, But he'd Fraser
(46:49):
jack Insley mixed this episode and Kyle Murdoch wrote our
theme song special thank You to Read Abrigotti and David Webster.
TechStuff News
Hosts And Creators
Oz Woloshyn
Karah Preiss
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Hey Jonas!
Hey Jonas! The official Jonas Brothers podcast. Hosted by Kevin, Joe, and Nick Jonas. It’s the Jonas Brothers you know... musicians, actors, and well, yes, brothers. Now, they’re sharing another side of themselves in the playful, intimate, and irreverent way only they can. Spend time with the Jonas Brothers here and stay a little bit longer for deep conversations like never before.
Crime Junkie
Does hearing about a true crime case always leave you scouring the internet for the truth behind the story? Dive into your next mystery with Crime Junkie. Every Monday, join your host Ashley Flowers as she unravels all the details of infamous and underreported true crime cases with her best friend Brit Prawat. From cold cases to missing persons and heroes in our community who seek justice, Crime Junkie is your destination for theories and stories you won’t hear anywhere else. Whether you're a seasoned true crime enthusiast or new to the genre, you'll find yourself on the edge of your seat awaiting a new episode every Monday. If you can never get enough true crime... Congratulations, you’ve found your people. Follow to join a community of Crime Junkies! Crime Junkie is presented by Audiochuck Media Company.