May 29, 2024 • 48 mins
On May 14th 2024, Google introduced their Search Generative Experience, a service that uses hallucination-prone artificial intelligence to generate answers to queries rather than just presenting links, all so that they could Wall Street that they're innovative and future-forward. The result is an even-more-broken search experience, and in this episode, Ed Zitron walks you through the rotten state of Google, and speaks with Lily Ray, a 15-year veteran of the search engine optimization industry, about how Google abandoned the web.
LINKS:
Lily Ray: https://lilyray.nyc/ https://twitter.com/lilyraynyc
The Verge's Interview With Sundar Pichai: https://www.theverge.com/24158374/google-ceo-sundar-pichai-ai-search-gemini-future-of-the-internet-web-openai-decoder-interview
See omnystudio.com/listener for privacy information.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:02):
All Zone Media.
Speaker 2 (00:05):
Hello and welcome to Better Offline. I'm your host ed Zichron.
What Hey, You ever wonder who ruined Google? Well, the
answer is Prabagar Ragavan, Google's former head of ads, who
(00:26):
led a coup to run Google Search, which led to
websites to k Now, I'm not just recounting my own
work here. My source is Google Search. No, really really,
search who ruined Google on Google Search and it will
pop up with an AI generated summary of multiple articles
that cite a piece I wrote called the Man who
Killed Google Search? Why Because on May fourteenth, twenty twenty four,
(00:50):
Google decided, in their incredible wisdom to add generative artificial
intelligence to search results. These new AI generated results, dubbed
Google's Search Generative ex Experience, replace the initial links you'll
see on some pages with an aioverview, one generated by
scraping the text of other websites, so that Google can
answer your questions without you having to visit them and
(01:12):
consequently give them any ad revenue. It's very strange, but
also you may be thinking, isn't generate if AI the
thing that regularly gets things wrong and is kind of unreliable,
and everyone knows about that Google wouldn't do that right wrong.
McKinsey alumnus and Google CEO Sundar Pshai told The Verges
(01:33):
Nile Patel that he believes these results will actually help
the Web. I'd argue otherwise, these generative results will only
help centralize the Web's information around these big publishers that
Google likes, all while creating less of a need for
anybody to visit the rest of the Internet as in
any of the other links, even if they're presented along
an AI generated summary. As Lauren Good of Wyatt said
(01:56):
in a recent piece about Google's new shift to AI,
by choosing when and where these overviews appear, Google is
essentially deciding what is or is not a complex question,
and then they're making a judgment on what kind of
web content should inform its summaries, as well as what
content you're actually going to see or learn from. Yet,
these articles are also predicated on one very, very very
(02:17):
silly assumption. And I really respect Lauren and Nile, but seriously,
you're making one major mistake. You're assuming that Google actually
cares about building and maintaining a good search engine. Though
it's been fixed since I wrote this script, you used
to as in a few days ago, be able to
ask Google if there was a country in Africa that
began with the letter K, and it would have resulted
(02:40):
in the generative result that says as of September twenty
twenty one, no African country begins with the letter K. However,
Kenya is the closest African country to starting with a
K sound with two citations a forum post from twenty
twenty one that quoted a hallucination from chat GPT and
a website titled countries that start with the letter K
with the very sentence beginning with Kenya. Google's generative search
(03:04):
results have also recently said that one could use Elmer's
glue to keep cheese from sliding off of a pizza,
or that a dog played in the NBA, or that
something called a squad plug is a device that helps
keep the glutes tight while squatting, rather than a joke
scraped from a Reddit post about a very very specific
kind of sex toy. Do not look it up. You
do not need to know. These results, by the way,
(03:26):
have already been taken care of because of the massive
shit storm that Google is dealing with from the very
obvious mistake they've made and will continue to make. It's
very upsetting. And these results are a result of a
modified version of Google's Gemini Generative AI. They're equivalent of
chat GBT, generating answers based on the content of web pages.
(03:49):
And these web pages can include everything. I'm talking news outlets,
random blogs, posts from Reddit, and it's just this vile
new strategy where and I'm quoting the new head of Search,
Liz Read, You're allowing Google to do the googling for you. Now,
this is of course a really horrible idea for many
many reasons. As I've mentioned repeatedly in the past, GENERAFAI
(04:13):
has this tendency to hallucinate, where it authoritatively states something
that just is not true, mostly because these models don't
actually know anything. General IVAI, like Gemini or chat GBT
or Anthropics Claude or metas Lama. They're all responding to
your queries by predicting what the most likely to be
correct answer is based on parameters that they learned in
their training data, which means that Google's AI powered search
(04:36):
is googling for you with absolutely no comprehension or intellectual
understanding or really any read of the content itself. As
a result, all it's able to do is say, okay, ah,
all right. These phrases seem to make up what the
right answer could be based on mathematics, and these links
(04:56):
seem to include these phrases, So I guess that's fine,
which is why Google all briefly recommended that you eat
one rock a day because it was generating an answer
from the satirical newspaper The Onion. Google's response was to
tell the vergis Kylie Robison that these were generally very
uncommon queries that are not representative of most people's experience,
(05:17):
specifically responding to Robison's own query where Google told her
that the mammal with the most bones was a python
just to be clear of python is a reptile. They
do have bones though, fu fact. Now, as funny as
these results might be, Google's AI powered search is really
horrible for the Internet at large. By summarizing other people's
(05:37):
links to generate an answer, Google's effectively stolen the Internet,
picking and choosing what parts it deems worthy of showing you,
and then providing this kind of hackneyed, hallucination prone summary
of what they think the links mean. Google's already positioned
itself as a gatekeeper of the Internet. By the way,
they're meant to be the steward of what's important. And
important is the word that they used in the regional
(05:59):
paper to describe page rank, which was how they rank pages.
And the theoretical value exchange was that we get reliable,
safe results that actually answered our queries. But AI powered searches,
they kind of turned Google into a source of truth
and not a remotely reliable one. And they're using a
technology that is well known for getting things wrong. Now
(06:23):
you may think, well, surely Google would be both aware
of this and even if they even if they were
aware of it, they would they wouldn't put this hallucination
prone tech in deliberately. Right, Surely this must be some
kind of mistake. And you are wrong, because san Dar
Pashai told Nile Patel in an interview that happened earlier
(06:43):
last week that they're well aware of the problem of
hallucinations and that they were still unsolved and that they
were an inherent feature of large language models. Sundar, You're
completely correct, at which point the CEO of Google said
that large language models were good because of this, and
that in fact, it made them very creative, Sunda, what
(07:05):
the hell are you talking about, man, I don't need
Google to be creative. I need Google to answer my
bloody question. When you say someone's being creative with the truth,
you're describing someone lying. Oh my god, I feel like anyway. Anyway,
it's extremely obvious how dangerous this strategy is. Generative. AI
(07:26):
could hallucinate I don't know how to deal with a
chemical fire or a mental health issue, or authoritatively misinform
somebody that say, I don't know Barack Obama was a Muslim,
which is an actual generitive result that Google has since
removed billions of people, regular non technical people, people that
don't live on social media in the swill and constantly
(07:47):
see the new news. They rely on Google every day
to answer questions. And they're going to assume that a
multi trillion dollar tech company would not turn the most
well trafficked source of information in the bloody world over
to an aid. And Peter kafk go Over, a business insider,
actually put this really well. He said, it's like the
difference between being handed a map and being given directions
(08:10):
that will send your car barreling over a cliff. I
won't mince words, Google's AI powered searches a goddamn travesty,
both for the Internet and for society at large. By
choosing what queries to summarize and what sites to pull from,
Google both centralizes and polarizes the world's information, all while
depriving actual publishers of traffic, the actual human beings creating
(08:35):
the things that inform their search engine. And it's horrible.
That alone is such a horrible problem. Now. Of course,
Sandar Pashai told nil A Patel, who just went, yeah,
sure mate, okay, no pushback, man. He told them, no,
this is actually good for the web because people will
get the answers and then they'll click through. Yeah. Let
me give you an example of how wrong you are, Sandar.
(08:56):
Look at social media. Look at when people see clickbait.
Do you think people go, oh, I should check the link.
This could be used to get engagement for my tweet. No,
they tweet the thing and get mad about it. I've
done this within the last twenty four hours. I'm an idiot.
But still people are not going to be curious. They're
going to assume that Google Search is not full of nonsense,
(09:18):
that Google Search is not spitting out AI generated summaries
using hallucination pron AI I feel like and going crazy anyway,
while AI results cite their sources. It's just it's foolish
to believe that regular people are going to click through
to the results when Google itself has already answered the question,
especially on service driven queries that answer, say, whether a
(09:40):
food is safe or some sort of factual question. Publishers,
especially news outlets, already heavily dependent on Google Search traffic.
They're terrified that these summaries will further reduce traffic at
a time when social networks like Facebook, for example, are
no longer really sending them readers. And let's be honest,
Google search has already deteriorated drastically over the last few years,
(10:04):
and its results are plagued with AI generated spam and
search engine optimized nonsense. And these results overwhelm good, unique,
authoritative content to the point that major newspapers are put
out public service announcements like the Wall Street Journal, by
the way, about how to avoid scams in the very
first results you see on Google. Now, you'd think that
(10:25):
this deluge of bad press would make Google say, humh,
we should fix Google Search, we should make things better. No, No, no, no, no,
that's not what we do here at Google. No. Google
plugged in an already unreliable artificial intelligence into a product
that it's deliberately allowed to decay so that customers can
(10:46):
get quasi right answers from search results that Google itself
barely maintains the quality of. It's a sound a bit
frustrated because this is so strange. This is all so strange,
this multi trillion dollar company, with this search engine that
makes them tens of billions of dollars. They're just like,
(11:07):
you know, just put the II on it. They'll love it.
And it sent their stock ups, so I guess it
doesn't bloody matter. But back to the regular results just
for a moment. Google claims that in its recent Core
update that they've helped reduce low quality and unhelpful content
by forty five percent, and they're wrong. By the way,
search is still dominated by search engine optimized content full
(11:29):
of these links that are built to drive kickbacks to
publishers called affiliate links, as well as YouTube links and
Reddit links. And it's the very same content that informs
and by the way, misinforms Google's generative results Google Search.
It's just this lazy ugly product, and it lacks any nuance,
it doesn't have ingenuity, and quite frankly, I'm beginning to
(11:50):
question whether it even has any utility anymore. Searching for
best laptop, and I mean this from Las Vegas, Nevada,
beautiful Las Vegas, Nevada brings up this affiliate mine market
funded tech site, and then a link to best Buy,
then another affiliate funded site, then another link to best Buy,
then a series of questions you can click through which
say things like which brand of laptop is best? And
(12:12):
when I click through the first one, by the way,
it led me to a site called New Indian Express
Deals and it's just full of affiliate marketing links, no
real journalism of any kind. It's all. It could be
AI generated. I have no idea, and it's tongues of
these links to MacBooks and then things like a Techno
megabook or the Honor Magic Book laptops that are not
(12:33):
available in America. Yeah, you know. One might imagine that
the point of a search company is that it searches
the web for you, surfacing things that you might find
interesting or informative that you might not find otherwise, highlighting
individual publishers that make great content that otherwise might get
lost in the churn of the Internet, which makes billions
(12:55):
and billions of pages so much to find the great
morass of the Web. That's what Google is meant to be.
Yet the reality is that Google wants to highlight the
same companies and the same publishers doing the same thing
in the same way, because Google, at its core is
a fundamentally lazy company that does not care about its customers.
As it related in The Man Who Killed Google Search.
(13:17):
Google's revenue arms are the ones pulling the levers at
the company right now, and they demand more queries, as
in more searches on the platform, even if doing so
makes the platform so much worse, because when you optimize
a platform for people to spend more time on it
and click on things more, you're not actually resolving a
query or a problem. I know, more queries are what
(13:40):
they're looking for, but they're not talking about resolving queries.
They're just looking for more of them so that they
can show you more ads. But when you look at
the state of Google right now, the current state of Google,
you can see how these decisions have poisoned it. Google
Search moves around its menus things like shopping images, news,
and videos. Based on what you're searching for. It pumps
(14:01):
results full of sponsored content YouTube videos. By the way,
of course, they own youtubes. That's good for them. And
then these arbitrarily chosen forum results, and then it deliberately
offer scates what might be a paid for link. It's
just this chaotically disorganized, nakedly manipulated piece of crap, let's
be honest, and the search engine optimization industry is eating
(14:23):
it alive. And it really isn't clear whether Google lacks
the ability to fix it or just doesn't really care
that much yet. The sorry state of Google Search did
a little digging here might also be the result of
Google distancing itself from the same search optimization community that
I've been kind of harsh on. While SEO is a
field that's inarguably harmed Google search results, it's also worth
(14:47):
considering the outside of technical standards like how fast the
site loads or whether it has a site map of
its links. Google also kind of hides things from the
SEO community. They offer scate what good looks like for
fear that optimize, we'll use these standards to manipulate its results, which,
by the way, happens anyway. So what Google does is,
instead of saying what good would be and I've kind
(15:09):
of been wrong here, they vaguely say, yeah, you should
kind of do this, but I'm not going to tell
you exactly how. And then you get these people who
still manipulate anyway. It's so strange, this company worth trillions
of dollars. They act like a plucky startup, but only
when it comes to actual quality control. But nevertheless, I
was quite interested in search optimization. So why you've got
(15:32):
in Lily Rays, a fifteen year SEO veterrun, and she
told me a really worrying story. I think that Google's
kind of failing everybody. They're failing publishers, they're failing SEO people,
and they're failing you and me. All right, so you've
(15:57):
been in se over about fifteen years. But what actually
is search engine optimization for a very basic listener who's
never dealt with it.
Speaker 1 (16:06):
Yeah, search engine optimization is the practice of trying to
make certain websites and the pages on those websites more
visible in search engines, by which I mean ranking in
the top positions on you know, Google Being or whatever
other search engine you're targeting. Because that's really where the
majority of people click on results.
Speaker 2 (16:27):
Isn't that a little bit manipulative of Google and these things?
Speaker 1 (16:31):
Is?
Speaker 2 (16:32):
How is that a good thing?
Speaker 1 (16:35):
So there's kind of a spectrum of approaches to SEO. There's,
first of all, Google's own guidance related to how to
do SEO in a way that I would argue is
actually great for the Internet. I think most of the
things that Google is recommending and that people like myself
who do SEO professionally day in and day out, what
we're doing is making websites more findable, more accessible, you know,
(16:58):
loading more quickly for people to find what they're looking for.
There's a lot of technical work that goes into that.
But of course there's a whole other end of the spectrum,
which is people trying to take advantage of the situation
and spam things and manipulate search engines. So it depends
which type of BESDEO you're talking about.
Speaker 2 (17:14):
So there are white hat SEO people, it's not just
all absolutely absolutely, So what do you do for your job?
What is the thing you make your customers do well?
Speaker 1 (17:25):
You know, companies come to us and they obviously want
to get more visibility, So we have to look at
all the different considerations and all the different factors that
influence how people are able to find that content. A
lot of it's technical, so literally a lot of our
work is like, you know, focusing on page speed, focusing
on overall accessibility. And you know, for example, if an
(17:48):
image is uploaded to a website and it doesn't have
image alt text, which is actually required for website accessibility,
search engines can't understand what's on that image traditionally, so
we have to go in there and kind of add
the right words in. So it's making that content more findable.
And again, a lot of that's technical, but a lot
of that is content oriented. So people might write a
certain article and in such a way that the way,
(18:09):
you know, when people go to Google and they're looking
for that article, they could never find it. So we're
helping companies to kind of structure their content in the
way where people can actually find what they're looking for.
Speaker 2 (18:19):
So you say that there are these signals, how much
of that comes from the search engines themselves.
Speaker 1 (18:25):
Well, you know, the SEO field, because search engines are
essentially a black box, and a lot of us are
working with you know, just third party data or working
the best we can with the analytics that Google any
other search engines give us. A lot of this is guesswork,
you know, we try to make informed guesses based on
the data itself, but Google and the search engines do
give us a lot of clues and a lot of
(18:46):
insights about what we should be focusing on, especially when
it comes to the technical thing, so you know, giving
us guidance about what we call like canonical tags. There
are different metadata that we have to add to a
website for them to really be able to parse crawl
through that website effectively. Those are direct signals that Google
gives us that we work with. But a lot of
it is kind of, you know, interpreting what we're seeing
(19:08):
as far as what the data shows and then making
informed guesses based on the history and on the analysis
that we do for different clients.
Speaker 2 (19:16):
So based on how Google describes it, they're fairly communicative
with web developers and sorry, web content creators, it doesn't
actually sound like that is true.
Speaker 1 (19:26):
I would say when it comes to the technical side
of things, Google's quite communicative. You know, it's really in
their best interest to have web you know, fast websites
websites that are accessible for users, websites that have a
great user experience. And Google does a lot of developer conferences,
a lot of communications on the technical side, and it's
quite easy to, you know, kind of get in touch
with them if you have a specific technical question, like
(19:48):
for example, John Vieler hosts Webmaster hangouts on Google often,
if not like several times per month where he answers
a lot of people's questions. So, yes, technically speaking, I
think they do provide a lot of guidance.
Speaker 2 (20:00):
About content because it seems like content, the actual things
on the web page is the thing that actually matters here.
Does Google give any guidance?
Speaker 1 (20:09):
They certainly do give guidance. They for example, they give
a lot of questionnaires that site owners can follow when
thinking about what Google considers helpful content. They also have
something called the Search Quality Radar Guidelines, which is a
document that you can find online. It's about maybe one
hundred and sixty plus pages, and that's specifically used to
inform Google's human Search Quality evaluator team as to what
(20:33):
good content or bad content kind of looks like. So
that's something that the SEO community looks at a lot
as well. But to your point, you know, getting into
granular questions about is this content good or bad for Google?
What does helpful content mean? That's become trickier in the
last few years because they can't tell us everything because
(20:54):
people spam everything that they can get their hands on.
Speaker 2 (20:57):
Right, So it almost feels like a kind of battle. Yes, absolutely,
but to this to this level, to this point you say,
they give this this advice and they're vague on the
content but direct on the technical. So how has Google
search quality got so much worse?
Speaker 1 (21:15):
Then? There's so many different factors at play, you know,
I think the reason why seos have a bad reputation
in general, And to your question earlier, you know, is
there such a thing as white hat seo. There's a
huge spectrum of approaches and beliefs within the SEO space.
For whatever reason, the people that spam Google the most
(21:36):
are the ones that have been kind of creating this
like the kind of the reputation that we have as
an industry lately. And I get that because they've been
quite successful. You know, a lot of people that are
spamming Google, especially with AI lately, have have really made
it into top positions, so they're getting a lot of
visibility with those tactics. But you know, there's a lot
of us that are doing work that again, it's like
(21:56):
kind of advised by search engines to do because it
really does make the Internet easier for everybody to use.
Uh So that's been a challenge lately. I think that
for whatever reason, you know, maybe because Google's shifted internally
so much of its focus to AI lately. Obviously there's
these big external challenges that Google's facing as a business lately,
(22:17):
maybe they're they're not haven't been as focused on search
quality for whatever reason. I don't actually have, you know,
any type of evidence of that other than what I'm seeing,
But I haven't seen as much spam and like SEO
manipulation as I've seen over the last year or so.
That being said, they're really cracking down on it in
the last few months, so sometimes it just takes them
(22:39):
a while to catch up.
Speaker 2 (22:41):
Something you've said publicly recently is real business is all
being affected by the most recent Google updates? Can you
kind of get into that, like who are the people
being affected here?
Speaker 1 (22:51):
Yeah? I mean that's that's nothing new. I think, you know, Google,
Google creates these algorithm updates multiple times big ones at
least multiple times per year every year. And we've seen
a lot of companies get caught up in those algorithm
updates before. In most cases they're doing something that goes
against Google's SEO or spam guidelines. The difference now is
(23:12):
that there's a lot of these new types of algorithm
updates that Google is launching that are affecting a lot
in a lot of people in the publishing space, so
both large and small publishers, especially smaller publishers over the
last couple of years. And these are, for example, people
that might have built a business doing travel blogging, doing
(23:32):
recipe blogging, product reviews, and so if you do consider
those people at business, and I think they would because
they often have staff, they have people going to the
places writing the travel reviews. Things like this, Google's dramatically
reduce the visibility of how those sites appear on Google
and the amount of traffic that they're getting.
Speaker 2 (23:52):
So when was the last major update to Google? Sometimes
in March right, or it's unclear when it actually started
and finish.
Speaker 1 (24:00):
Yeah, the last major update I would say was the
March core update, which took forty five days to complete.
But that started in early March and then ended in April.
But there's been a few different aspects of that update
in new spam policies that they've introduced as well, So
we're actually still kind of in the middle of some
of that rollout happening.
Speaker 2 (24:19):
Yet it's it's strange, though you and this is not
on you, This is on Google. You say there's technical things,
they say, there's quality things they say, but even recipe
sites look terrible. They're full of ads there they make
your phone so hot it burns you. Okay, maybe not
that hot, but it feels like Google doesn't actually seem
to be maintaining quality, and it seems like they almost
(24:42):
fight those who do. An SEO, just how do you
feel about your relationship with Google?
Speaker 1 (24:50):
It's hard to say right now. I think, you know,
with something that's been very tricky for me is that
I have a bit of a I've built a bit
of a name for myself as the person that goes
around and and really echoes a lot of what they
recommend site owners do. That's been kind of my brand
for the last several years. Google has different, you know,
policies and things that they recommend to site owners that
(25:13):
I have found, I would say, up until twenty twenty three,
I've found those recommendations to be very consistent with the
work that we're doing for our clients. So it's very
easy for me to go say, yeah, what Google says
to do is what works right, And we've worked with
a lot of different companies and found those tactics and
those recommendations to be very effective. The last few months,
(25:35):
that's changed a bit for me. I'm not well, I'm
not you know, my personal relationship with them is I
want to be very careful to say, like, you know,
the people that I know that that work at Google,
and they you know, speak on behalf of Google, Like,
this has nothing to do with them, right, This has
to do with what's happening in the search results. And
you know when I see ai spam ranking above my
(25:59):
clients for what we call top stories, so that's you know,
the news features on Google for example, or I see
that in Google's new aioverviews is getting everything so many
different things wrong, and there's spam and there's malware in there,
like things like this. It's just it makes us as
seos look bad because we then have to go defend
those results to our clients and try to make sense
(26:21):
of them. And that's been really challenging, very frustrating for
me lately, because it's very hard to defend the quality
of the search results when the search results are in
a state of very poor quality.
Speaker 2 (26:33):
So you and you said previously up until twenty twenty three,
was it what happened in twenty twenty four? Was it
the core up they or was there just something that
happened this year?
Speaker 1 (26:43):
Writ large, I think that probably the biggest change that
happened was honestly chat GPT. That's there's one of many
big changes. But the you know, not only chat Gypt
but the different tools like it empowered spammers and sea
alike to find new ways to speed up what they
were doing and to create a lot of terrible content
(27:05):
at scale. And that worked extremely well for a year
because guess what, it takes Google and the other search
engines a while to kind of figure out what they're
doing to develop new algorithms and new spam policies that
you know, devalue or demote that type of content. But
there was about a year where that content was succeeding
extremely well. So towards the end of twenty twenty three,
(27:28):
you would see a ton of AI generated spam and
the search results and people using AI tools in new
and different ways to just accelerate whatever spamming they were doing.
Speaker 2 (27:38):
Do you think Google actually has control of search or
has it just kind of got out of hand.
Speaker 1 (27:46):
I think they you know, they had these new set
of updates beginning in March twenty twenty four where I
think that they made loud and clear that they're not
messing around. You know, they really cracked down on a
lot of sites that were doing a lot of different
SEO tactics and spam tactics. So I do think they're
getting controlled back. I do actually also think they kind
(28:08):
of overstepped in some areas and there's a lot of
collateral damage there as well, which I know they're working on.
And that's where you hear a lot of people saying,
you know, this destroyed my business because there were a
lot of people that just had no idea that they
were going to get caught up in that. So that's
on the search side. I think, you know, Google's proven
time and time again that they can eventually get ahead
of it. But I think they're presented with a lot
(28:29):
of new and different challenges right now on the AI side.
You know, I'm sure eventually they'll figure it out, or
this AI overview thing will get better over time. But
it's been quite interesting to see publicly this week, so
many examples of Google's new AI overviews getting information blatantly wrong.
Speaker 2 (28:49):
I want to push back a little bit about the
control over search that Google has. The reason I asked
that is it's not been just chat GPT going back
to twenty twenty. Results have been bad for a long time,
and it feels that Google is fighting SEO, and SEO
I realized is a much wider umbrella because there's people
(29:11):
trying to do the web properly. There's people who were
just trying to mess with Google, and then there are
the people who just don't care and are spamming. It's
just it doesn't feel like they're able to make quick changes,
and I just don't understand it. It drives me insane, honestly.
Speaker 1 (29:24):
Yeah. Well, well, one of the other things that I
didn't mention that's been a big change to the quality
of search results, and one of the big complaints that
people are having is a lot of big sites with
a lot of authority on Google a lot of SEO history,
A lot of the big publishers that we all know,
you know, big brands that everybody's familiar with, found a
lot of SEO opportunities in the last couple of years
(29:46):
that have worked splendidly for them. So, for example, affiliate content.
You might have noticed that nearly every big publisher on
the Internet has ventured into product reviews in the last
couple of years, and that's because it's a huge money maker.
So one of the big complaints that people have is like,
it's not fair that X, y Z big publisher shows
up every time I'm typing best running shoes for women
(30:07):
or whatever the keyword is, and believe it or not,
you know this. I think this has been one of
the biggest complaints that people have had. But in the
last two or three months, Google is cracking down on
that in a very big way. I think it's just
taking them a long time to figure out specifically how
to target that.
Speaker 2 (30:25):
Why do you think it takes them so long? They're
worth like two trillion, they surely they have the people.
Speaker 1 (30:31):
I wonder the same things myself. Sometimes I'm not a
search ranking engineer, so I don't know why I also,
I think you know better than I do what's going
on internally at their company right now. I think there's
a lot of a lot of kind of pandemonium in
different ways. I don't know, but it seems that way.
Speaker 2 (30:47):
But that being yeah, go on, you go on.
Speaker 1 (30:50):
I was just gonna say, another reason why this is
so difficult for them to target is because not all
of this is easy to spot algorithmically. For example, if
a publisher decides to start working with third party contributors
who are contributing product review content, those relationships are not
always entirely clear to a point where an algorithm wouldn't
(31:13):
necessarily be able to figure that out. So Google has
to use a variety of different methods to demote that
content and learn about those business relationships, and you can't
just solve that all algorithmically.
Speaker 2 (31:27):
Do you think they want to solve it?
Speaker 1 (31:30):
I would imagine yeah, they want to do it algorithmically
instead of using their web spam team. But currently they're
using their web spam teams. So that's real human beings
that are sending out what they call manual actions or
penalties to sites that are engaged in different types of
spam policies, and they've created some new ones recently.
Speaker 2 (31:47):
It just feels like they could I say this not
as an SEO expert, but it feels like they could
do very obvious things. They could go and go with
dot Meredith, who has the met who you read, Hello
House Read is whole thing around how Meredith has been
basically spamming affiliate marketing content that's not particularly well done,
and you've written about this as well. It feels that
(32:09):
they could just do something to them immediately that why
don't they take these fast actions because it doesn't seem
like they're doing much anyway.
Speaker 1 (32:17):
Well, a lot of these publishers have content that a
lot of people really love, and you know, Google has
a lot of signals that indicate that these are, you know,
sites that produce extremely helpful content in a lot of
different categories. So it's not so easy to say this
category of the website is not okay by users, but
this one is. You know, it starts to become tricky
(32:39):
when they have to go after certain sections of a website.
You know, a lot of times with sites that are
just doing pure spam, it'll be the whole site that's
doing spam. But these are obviously not to mention these
are very very big brands. You can't you can't remove
a site from Google or demote a site from Google,
for if it's a brand that people are looking for
and a brand that people would expect to see, right,
(33:01):
because that may mean you can sort.
Speaker 2 (33:03):
Of well, it's kind of I think what I'm what
I'm getting at is what they judge is good for
the Internet doesn't necessarily seem so because there's a self
fulfilling prophecy there there is, Oh, this is popular with
a bunch of people. Well maybe it's popular because Google
made it popular. Maybe it's not. Actually it's just an
algorithmic choice. But I'll get I'll get onto something a
(33:25):
little more direct, which is, how do you feel about AI.
Speaker 1 (33:29):
Search, like Google's new AI overviews?
Speaker 2 (33:32):
Yes, yes, sge Yes, that's generative experience.
Speaker 1 (33:36):
It's like, obviously we're in a big transition moment where
you know, speaking about it now is not reflective of
what it's going to look like in six months. So
I always want to be cognizant of like this thing
changes every single day. My biggest issue with it, so
first of all, Google made it public, you know, a
week ago, at Google Io, and myself and many others
have been experimenting with it in Google Labs for a
(33:58):
year before that, and in theory, you know, with AI,
it should be improving and learning and getting better every day, right.
For a year, myself and many others were raising flags
about quality issues and also the fact that we just
didn't feel like it often provided that much of an
improvement to the existing search results. For example, Google already
(34:19):
had featured snippets, so if they want to take a
paragraph or a snippet from somebody's website and display that
directly in the search results, there was already a mechanism
for that. So for the past year, you know, I've
raised a million different examples of when it got things wrong.
When I think it was, you know, responding to some
type of super controversial question in a weird way, whatever
(34:40):
it was. And when they launched it, you know, literally
I can tell you because I spent my whole life
at SEO conferences. Most of us were like, there's no
way they can launch this thing in its current form,
like it's just so not ready, right, And they did,
And here we are a week later, and nearly every
you know, social media site has tons of examples of
people wanting to it off, the tons of examples of
(35:02):
it getting things wrong. So you know, I want to
be cognizant and respectful the fact that it will change,
it will get better. I'm just quite surprised that they
launched it in its current form when they did.
Speaker 2 (35:13):
One last question And don't take this the wrong way,
but why do you think they'll improve, because the last
few years have shown that they've got worse.
Speaker 1 (35:23):
Well, in theory, these AI lm should learn and improve
and get better and get retrained all the time. I'm
going off with that optimistic idea. But at the same time,
it's like we're doing, you know, offloading this work onto
Google searchers and relying on them to provide feedback and
everything is I guess it doesn't necessarily mean it's going
to go in the right direction. So it's a good point.
(35:45):
It's hard to say.
Speaker 2 (35:47):
Yeah, it's just it feels almost like the AI side
is getting away from what search is meant to do,
because search is meant to help you find useful things,
and this doesn't even seem to be capable of helping
you find the right answer, let alone a good link.
Speaker 1 (36:01):
Yeah, I think The main issue that I currently have
with it is that once I know that it's twenty
or ten percent wrong or whatever whatever the number is,
we don't know the actual percentage of wrong answers. You know,
based on how many we've already seen in the last week,
it's quite a lot. So when people know that this
thing has a tendency to be wrong, will they trust it? Right?
(36:23):
And what does that do for trust and Google overall?
Because with featured snippets, at least you can say, Okay,
this comes from this website, you can blame that website
because they pull that data verbatim. When they're showing a
featured snippet with.
Speaker 2 (36:35):
Air or reviews.
Speaker 1 (36:35):
It's like there's a tendency that one of these sentences
is completely wrong, and it's pulled from an eleven year
old Breddit threat that was making a joke, and we
just don't know which sentence it is right.
Speaker 2 (36:46):
Right. It just it feels like they're almost moving against
the Internet. It's just a very very depressing thing to watch.
Speaker 1 (36:53):
Yeah, I agree. I mean I think I would have
expected this thing to be more fact, more evidence based
in its rollout. I would have expected it to show
for far fewer queries where getting things wrong is extremely dangerous.
So yeah, I mean, I like, our industry now is
(37:14):
digging into the data and we're finding that it seems
like health is actually one of the categories where AI
already is triggered the most health topics.
Speaker 2 (37:22):
That's genuinely scary. But okay, actual lost question here, What
is the thing that Google could change to get better?
Speaker 1 (37:32):
Huh? I mean, I would imagine they're they're working on
a lot of the feedback that they've gotten in the
last week with the AI over reviews and finding a
lot of examples where it gets things wrong. And if
they're going to continue with aio reviews, I would hope
that they dramatically improve the product as quickly as possible.
That's on the AI front. I mean, if if you
(37:52):
want to be if you want to.
Speaker 2 (37:53):
Completely in general with search, Yeah, with.
Speaker 1 (37:56):
With search, I think that they should reassess this notion
of helpful content because I think that they've inadvertently caused
a lot of sites that are truly producing helpful, experience
(38:17):
driven content to lose a lot of visibility in the
last year or so, and a lot of like you know,
we're still seeing a lot of big brands seemingly able
to show up for every possible query, and we're not
hearing enough from small publishers. We're not giving small publishers
enough of a chance for their content to compete, and
they're kind of At this point, Google's just seems to
(38:37):
be blasting Reddit everywhere for everything, maybe as a solution
to that, because yeah, you get a lot of firsthand
experience from Reddit, but I'm not sure that putting Reddit
in top positions for so many different queries is a
solution either. So personally, I'd like to see a broader
diversity of the types of websites that are able to compete.
Speaker 2 (39:05):
The problem, it seems, is that this multi trillion dollar
tech company, one that has a monopoly over the Internet's information,
is somehow we can we can power us when faced
with people that create websites specifically to trick it, Google,
a company that sells to militaries and governments and professional
sports leagues that makes tens of billions of dollars a
quarter and employs nearly two hundred thousand people, is incapable
(39:28):
of thwarting a hodgepodge industry full of nakedly craven bad actors,
or they just un willing to intervene because it Google
doesn't really care about anything other than how many times
people search and how many ads they see, And they
don't really mind if maybe this big mess of crap
makes the whole thing worse. This problem, it's been going
(39:50):
on for years, and I feel crazy every time I
talk about Google Search as a product and has a
business that's such an incredibly profitable and ubiquitous product can
be it's so utterly shoddy and regularly unfit for the task,
I feel insane. I feel crazy when I see article
after article about Sandhar Pushai and Google talking about the
(40:10):
remarkable success of Google, yet never the fact that they've
done so much to damage this search engine, and they've
so blatantly destroyed their core product. And what really upsets
me is the sheer contempt that Google seems to have
for its customers. Google Search is awful. It's a terrible,
(40:31):
unreliable product, and now they've made it even less reliable
with the introduction of Generatifai. And yet sun Dar Pushai,
the CEO of Google, can stride into interviews all piss
and vinegar and ramble about how evalues independent sources and
more authentic voices as his service continually fails the surface
unique interesting or useful content. Nil Patel of The Verge
(40:56):
got a lot of credit for giving Peshai a tough interview,
but I'm gonna be bloody honest, paper Tiger. It's just
another opportunity for a Google executive to vaguely suggest that
they care about search without really being held accountable. The
question is, sun Dar, why is Google so bad? Now?
And he will give you some vague pr spiel. Show
him results. Nille did show him Chromebook results. The thing
(41:19):
to do would be come with a few results and
show him in front of him and say why is
this so bad? Why? Sun Dar? And if you can
get him to say on the record, actually, I think
it's good. That says everything. What it says is he's
a liar and he doesn't really care. But I guess
the interview did that anyway. Look, the problem, other than
(41:39):
the fact that Google should be broken up and that
search should be a public utility or a nonprofit, is
that the media seems to continually lack the ability to
simplify the problem. Google Search is bad, and it's been
bad for a while. Results are worse. They're full of
spam and low quality content. Google does a terrible job
of maintaining the content on Google Search, and it's now
introduced an AI to search through these results. That makes
(42:02):
things worse by generating answers based on search results from
a search engine that's rotting in real time, and then
it hallucinates the results. Anyway, Eh God, I apologize them
getting a little bit head up, but this whole thing
just it kills me. It makes me feel a little
bit nuts because if any of us, anyone listening to this,
(42:24):
went into our jobs and our trousers were down and
our our bits were out, and we were walking in,
and we go in and we half do our job
and the job's terribly done, would we get fired? We
get fired. I think we'd get fired, maybe if the
crap work we did made money, but we'd still be
in trouble for our willies hanging out. And that's the thing.
None of these tech executives, even when faced with the media,
(42:45):
seemed to actually be held accountable. And Google acts without integrity,
and they just act with this stunning contempt for their customers.
I've said it before, but it bears repeating. Suned up
a shite is a goddamn management concern, a former McKinsey
free with little interest in providing users with any kind
of usable service, and his cronies people like Probagara, Ragavan
(43:06):
and Liz Read. By the way, Probagara has been upgraded
to head of a bunch of things now and Liz
Read is now the head of Search. It's so good
he picked his successor, and it's so good that this
person loves Ai. And these people they've continually helped Google
put out bad software, all in pursuit of perpetual growth.
And I believe they're part of a larger problem in
(43:29):
the tech industry. Tech firms are no longer building things
for customers. Big tech and honestly, far too many startups.
They're creating these products that are more like symbolic capital
for investors so that they can place bets on them
and say, ah, maybe red war here, no, maybe black
this time, maybe double zero, And these things they only
(43:49):
exist to express meteoric growth ten x, twenty x, thirty
x returns things that sound satisfying for people to chirp
on CNBC without ever actually really having to provide the service.
They're discussing I'm not to be honest, it's breaking my arm.
As I said in my last episode, I'm not a pessimist.
I'm more of a broken heared romantic. There was a time,
(44:12):
and it wasn't super long ago, that tech was exciting
and fun and interesting. When things came out that subtly
or revertedly, they changed our lives, They made things better,
they maybe connected us, or they made our time on
our own better and more substantive, and things felt like
they were progressing. Yet it feels like so much of
today's tech industry is focused on shipping products that nobody
(44:33):
asked for to fix problems that don't exist, all while
demanding applause for a future that they've yet to make
a reality. Cryptocurrency, the metaverse, now generative AI. They're all
trends that built this kind of pseudo religious movement to
carry water for products that didn't do what they said.
They would guilt tripping those that said otherwise for not
being optimistic enough about an industry there's continually failed to
(44:56):
give us hope for the future, and it's disgraceful. It's disgraceful.
Look at Facebook from my last episodes. Look at that website,
that multi billion dollar website. Look how bad it is.
Look how full it's of AI, spam and junk it is.
Look how unsafe it is. Look at Google right now.
Google Search is falling. Apart of the seams. It kind
(45:19):
of works and sometimes it does, don't get me wrong,
but for the most part you just can't trust it
any more. And that was before they plugged generative AI
into it. Now the world's probably the world's most profitable
tech product is just languishing. It's so much worse. And
what's crazy is I don't even know how they're gonna
make money from it because these AI results they don't
(45:41):
seem to connect to ads or anything. They don't seem
to make Google money. But because Sandha Peshei needs to
convince Wall Street that Google will grow forever, well, you
gotta put AI in it, man, You gotta make the
AI happen. Except regular people using this, they're gonna get misinformed.
Someone's gonna die. And I really sound dramatic, I know,
(46:01):
but they are because people use Google Search and rely
on Google Search. They rely on Google Search to give
them real answers to real queries. And we can laugh
about the fact it said that someone had to eat
a rock a day. But what if they're trying to
put out a chemical fire. What if they're depressed then
they ask Google something and Google tells them, I don't know,
(46:22):
some sort of weird CBD treatment because Google got poisoned
by an AI spam result. There is, by the way,
a fake result about Google telling someone to jump off
the gong Gate Bridge. It's very fake, done for for it.
But seriously, though, something like that's going to happen. And yeah,
I'm sure someone will say, oh, it's just an edge case.
I don't give a crap an edge case where someone
(46:42):
dies or is harmed or is misinformed. That's not worth it.
It's not worth it. And these companies must learn. And
I think what's going to happen is far simpler. I
think they believe that people are just going to keep
using it because they have these horrible monopolies. I think
at some point people are just not going to use
search at all. I think that is the actual end
(47:05):
result here. I think you see open AI create search
engine and it kind of is worse than this. I
think you're going to see a proliferation of this tech
as people desperately try and catch up with Google and
it's so distressing, it's so insulting to customers, turns my
bloody stomach. I realized this wasn't a very positive podcast,
(47:26):
but hope it filled in the gaps for you. Thanks
for listening, Thank you for listening to Better Offline. The
editor and composer of the Better Offline theme song is Metasowski.
You can check out more of his music and audio
projects at Matasowski dot com, M A T.
Speaker 1 (47:48):
T O.
Speaker 2 (47:49):
S O w Ski dot com. You can email me
at easy at Better Offline dot com or visit Better
Offline dot com to find more podcast links and of course,
my newsletter. I also really recommend you go to chat
dot Where's youread dot at to visit the discord, and
go to our slash Better Offline to check out I'll Reddit.
Thank you so much for listening. Better Offline is a
(48:11):
production of cool Zone Media. For more from cool Zone Media,
visit our website Coolzonemedia dot com, or check us out
on the iHeartRadio app, Apple Podcasts, or wherever you get
your podcasts.
All Zone Media.
Speaker 2 (00:05):
Hello and welcome to Better Offline. I'm your host ed Zichron.
What Hey, You ever wonder who ruined Google? Well, the
answer is Prabagar Ragavan, Google's former head of ads, who
(00:26):
led a coup to run Google Search, which led to
websites to k Now, I'm not just recounting my own
work here. My source is Google Search. No, really really,
search who ruined Google on Google Search and it will
pop up with an AI generated summary of multiple articles
that cite a piece I wrote called the Man who
Killed Google Search? Why Because on May fourteenth, twenty twenty four,
(00:50):
Google decided, in their incredible wisdom to add generative artificial
intelligence to search results. These new AI generated results, dubbed
Google's Search Generative ex Experience, replace the initial links you'll
see on some pages with an aioverview, one generated by
scraping the text of other websites, so that Google can
answer your questions without you having to visit them and
(01:12):
consequently give them any ad revenue. It's very strange, but
also you may be thinking, isn't generate if AI the
thing that regularly gets things wrong and is kind of unreliable,
and everyone knows about that Google wouldn't do that right wrong.
McKinsey alumnus and Google CEO Sundar Pshai told The Verges
(01:33):
Nile Patel that he believes these results will actually help
the Web. I'd argue otherwise, these generative results will only
help centralize the Web's information around these big publishers that
Google likes, all while creating less of a need for
anybody to visit the rest of the Internet as in
any of the other links, even if they're presented along
an AI generated summary. As Lauren Good of Wyatt said
(01:56):
in a recent piece about Google's new shift to AI,
by choosing when and where these overviews appear, Google is
essentially deciding what is or is not a complex question,
and then they're making a judgment on what kind of
web content should inform its summaries, as well as what
content you're actually going to see or learn from. Yet,
these articles are also predicated on one very, very very
(02:17):
silly assumption. And I really respect Lauren and Nile, but seriously,
you're making one major mistake. You're assuming that Google actually
cares about building and maintaining a good search engine. Though
it's been fixed since I wrote this script, you used
to as in a few days ago, be able to
ask Google if there was a country in Africa that
began with the letter K, and it would have resulted
(02:40):
in the generative result that says as of September twenty
twenty one, no African country begins with the letter K. However,
Kenya is the closest African country to starting with a
K sound with two citations a forum post from twenty
twenty one that quoted a hallucination from chat GPT and
a website titled countries that start with the letter K
with the very sentence beginning with Kenya. Google's generative search
(03:04):
results have also recently said that one could use Elmer's
glue to keep cheese from sliding off of a pizza,
or that a dog played in the NBA, or that
something called a squad plug is a device that helps
keep the glutes tight while squatting, rather than a joke
scraped from a Reddit post about a very very specific
kind of sex toy. Do not look it up. You
do not need to know. These results, by the way,
(03:26):
have already been taken care of because of the massive
shit storm that Google is dealing with from the very
obvious mistake they've made and will continue to make. It's
very upsetting. And these results are a result of a
modified version of Google's Gemini Generative AI. They're equivalent of
chat GBT, generating answers based on the content of web pages.
(03:49):
And these web pages can include everything. I'm talking news outlets,
random blogs, posts from Reddit, and it's just this vile
new strategy where and I'm quoting the new head of Search,
Liz Read, You're allowing Google to do the googling for you. Now,
this is of course a really horrible idea for many
many reasons. As I've mentioned repeatedly in the past, GENERAFAI
(04:13):
has this tendency to hallucinate, where it authoritatively states something
that just is not true, mostly because these models don't
actually know anything. General IVAI, like Gemini or chat GBT
or Anthropics Claude or metas Lama. They're all responding to
your queries by predicting what the most likely to be
correct answer is based on parameters that they learned in
their training data, which means that Google's AI powered search
(04:36):
is googling for you with absolutely no comprehension or intellectual
understanding or really any read of the content itself. As
a result, all it's able to do is say, okay, ah,
all right. These phrases seem to make up what the
right answer could be based on mathematics, and these links
(04:56):
seem to include these phrases, So I guess that's fine,
which is why Google all briefly recommended that you eat
one rock a day because it was generating an answer
from the satirical newspaper The Onion. Google's response was to
tell the vergis Kylie Robison that these were generally very
uncommon queries that are not representative of most people's experience,
(05:17):
specifically responding to Robison's own query where Google told her
that the mammal with the most bones was a python
just to be clear of python is a reptile. They
do have bones though, fu fact. Now, as funny as
these results might be, Google's AI powered search is really
horrible for the Internet at large. By summarizing other people's
(05:37):
links to generate an answer, Google's effectively stolen the Internet,
picking and choosing what parts it deems worthy of showing you,
and then providing this kind of hackneyed, hallucination prone summary
of what they think the links mean. Google's already positioned
itself as a gatekeeper of the Internet. By the way,
they're meant to be the steward of what's important. And
important is the word that they used in the regional
(05:59):
paper to describe page rank, which was how they rank pages.
And the theoretical value exchange was that we get reliable,
safe results that actually answered our queries. But AI powered searches,
they kind of turned Google into a source of truth
and not a remotely reliable one. And they're using a
technology that is well known for getting things wrong. Now
(06:23):
you may think, well, surely Google would be both aware
of this and even if they even if they were
aware of it, they would they wouldn't put this hallucination
prone tech in deliberately. Right, Surely this must be some
kind of mistake. And you are wrong, because san Dar
Pashai told Nile Patel in an interview that happened earlier
(06:43):
last week that they're well aware of the problem of
hallucinations and that they were still unsolved and that they
were an inherent feature of large language models. Sundar, You're
completely correct, at which point the CEO of Google said
that large language models were good because of this, and
that in fact, it made them very creative, Sunda, what
(07:05):
the hell are you talking about, man, I don't need
Google to be creative. I need Google to answer my
bloody question. When you say someone's being creative with the truth,
you're describing someone lying. Oh my god, I feel like anyway. Anyway,
it's extremely obvious how dangerous this strategy is. Generative. AI
(07:26):
could hallucinate I don't know how to deal with a
chemical fire or a mental health issue, or authoritatively misinform
somebody that say, I don't know Barack Obama was a Muslim,
which is an actual generitive result that Google has since
removed billions of people, regular non technical people, people that
don't live on social media in the swill and constantly
(07:47):
see the new news. They rely on Google every day
to answer questions. And they're going to assume that a
multi trillion dollar tech company would not turn the most
well trafficked source of information in the bloody world over
to an aid. And Peter kafk go Over, a business insider,
actually put this really well. He said, it's like the
difference between being handed a map and being given directions
(08:10):
that will send your car barreling over a cliff. I
won't mince words, Google's AI powered searches a goddamn travesty,
both for the Internet and for society at large. By
choosing what queries to summarize and what sites to pull from,
Google both centralizes and polarizes the world's information, all while
depriving actual publishers of traffic, the actual human beings creating
(08:35):
the things that inform their search engine. And it's horrible.
That alone is such a horrible problem. Now. Of course,
Sandar Pashai told nil A Patel, who just went, yeah,
sure mate, okay, no pushback, man. He told them, no,
this is actually good for the web because people will
get the answers and then they'll click through. Yeah. Let
me give you an example of how wrong you are, Sandar.
(08:56):
Look at social media. Look at when people see clickbait.
Do you think people go, oh, I should check the link.
This could be used to get engagement for my tweet. No,
they tweet the thing and get mad about it. I've
done this within the last twenty four hours. I'm an idiot.
But still people are not going to be curious. They're
going to assume that Google Search is not full of nonsense,
(09:18):
that Google Search is not spitting out AI generated summaries
using hallucination pron AI I feel like and going crazy anyway,
while AI results cite their sources. It's just it's foolish
to believe that regular people are going to click through
to the results when Google itself has already answered the question,
especially on service driven queries that answer, say, whether a
(09:40):
food is safe or some sort of factual question. Publishers,
especially news outlets, already heavily dependent on Google Search traffic.
They're terrified that these summaries will further reduce traffic at
a time when social networks like Facebook, for example, are
no longer really sending them readers. And let's be honest,
Google search has already deteriorated drastically over the last few years,
(10:04):
and its results are plagued with AI generated spam and
search engine optimized nonsense. And these results overwhelm good, unique,
authoritative content to the point that major newspapers are put
out public service announcements like the Wall Street Journal, by
the way, about how to avoid scams in the very
first results you see on Google. Now, you'd think that
(10:25):
this deluge of bad press would make Google say, humh,
we should fix Google Search, we should make things better. No, No, no, no, no,
that's not what we do here at Google. No. Google
plugged in an already unreliable artificial intelligence into a product
that it's deliberately allowed to decay so that customers can
(10:46):
get quasi right answers from search results that Google itself
barely maintains the quality of. It's a sound a bit
frustrated because this is so strange. This is all so strange,
this multi trillion dollar company, with this search engine that
makes them tens of billions of dollars. They're just like,
(11:07):
you know, just put the II on it. They'll love it.
And it sent their stock ups, so I guess it
doesn't bloody matter. But back to the regular results just
for a moment. Google claims that in its recent Core
update that they've helped reduce low quality and unhelpful content
by forty five percent, and they're wrong. By the way,
search is still dominated by search engine optimized content full
(11:29):
of these links that are built to drive kickbacks to
publishers called affiliate links, as well as YouTube links and
Reddit links. And it's the very same content that informs
and by the way, misinforms Google's generative results Google Search.
It's just this lazy ugly product, and it lacks any nuance,
it doesn't have ingenuity, and quite frankly, I'm beginning to
(11:50):
question whether it even has any utility anymore. Searching for
best laptop, and I mean this from Las Vegas, Nevada,
beautiful Las Vegas, Nevada brings up this affiliate mine market
funded tech site, and then a link to best Buy,
then another affiliate funded site, then another link to best Buy,
then a series of questions you can click through which
say things like which brand of laptop is best? And
(12:12):
when I click through the first one, by the way,
it led me to a site called New Indian Express
Deals and it's just full of affiliate marketing links, no
real journalism of any kind. It's all. It could be
AI generated. I have no idea, and it's tongues of
these links to MacBooks and then things like a Techno
megabook or the Honor Magic Book laptops that are not
(12:33):
available in America. Yeah, you know. One might imagine that
the point of a search company is that it searches
the web for you, surfacing things that you might find
interesting or informative that you might not find otherwise, highlighting
individual publishers that make great content that otherwise might get
lost in the churn of the Internet, which makes billions
(12:55):
and billions of pages so much to find the great
morass of the Web. That's what Google is meant to be.
Yet the reality is that Google wants to highlight the
same companies and the same publishers doing the same thing
in the same way, because Google, at its core is
a fundamentally lazy company that does not care about its customers.
As it related in The Man Who Killed Google Search.
(13:17):
Google's revenue arms are the ones pulling the levers at
the company right now, and they demand more queries, as
in more searches on the platform, even if doing so
makes the platform so much worse, because when you optimize
a platform for people to spend more time on it
and click on things more, you're not actually resolving a
query or a problem. I know, more queries are what
(13:40):
they're looking for, but they're not talking about resolving queries.
They're just looking for more of them so that they
can show you more ads. But when you look at
the state of Google right now, the current state of Google,
you can see how these decisions have poisoned it. Google
Search moves around its menus things like shopping images, news,
and videos. Based on what you're searching for. It pumps
(14:01):
results full of sponsored content YouTube videos. By the way,
of course, they own youtubes. That's good for them. And
then these arbitrarily chosen forum results, and then it deliberately
offer scates what might be a paid for link. It's
just this chaotically disorganized, nakedly manipulated piece of crap, let's
be honest, and the search engine optimization industry is eating
(14:23):
it alive. And it really isn't clear whether Google lacks
the ability to fix it or just doesn't really care
that much yet. The sorry state of Google Search did
a little digging here might also be the result of
Google distancing itself from the same search optimization community that
I've been kind of harsh on. While SEO is a
field that's inarguably harmed Google search results, it's also worth
(14:47):
considering the outside of technical standards like how fast the
site loads or whether it has a site map of
its links. Google also kind of hides things from the
SEO community. They offer scate what good looks like for
fear that optimize, we'll use these standards to manipulate its results, which,
by the way, happens anyway. So what Google does is,
instead of saying what good would be and I've kind
(15:09):
of been wrong here, they vaguely say, yeah, you should
kind of do this, but I'm not going to tell
you exactly how. And then you get these people who
still manipulate anyway. It's so strange, this company worth trillions
of dollars. They act like a plucky startup, but only
when it comes to actual quality control. But nevertheless, I
was quite interested in search optimization. So why you've got
(15:32):
in Lily Rays, a fifteen year SEO veterrun, and she
told me a really worrying story. I think that Google's
kind of failing everybody. They're failing publishers, they're failing SEO people,
and they're failing you and me. All right, so you've
(15:57):
been in se over about fifteen years. But what actually
is search engine optimization for a very basic listener who's
never dealt with it.
Speaker 1 (16:06):
Yeah, search engine optimization is the practice of trying to
make certain websites and the pages on those websites more
visible in search engines, by which I mean ranking in
the top positions on you know, Google Being or whatever
other search engine you're targeting. Because that's really where the
majority of people click on results.
Speaker 2 (16:27):
Isn't that a little bit manipulative of Google and these things?
Speaker 1 (16:31):
Is?
Speaker 2 (16:32):
How is that a good thing?
Speaker 1 (16:35):
So there's kind of a spectrum of approaches to SEO. There's,
first of all, Google's own guidance related to how to
do SEO in a way that I would argue is
actually great for the Internet. I think most of the
things that Google is recommending and that people like myself
who do SEO professionally day in and day out, what
we're doing is making websites more findable, more accessible, you know,
(16:58):
loading more quickly for people to find what they're looking for.
There's a lot of technical work that goes into that.
But of course there's a whole other end of the spectrum,
which is people trying to take advantage of the situation
and spam things and manipulate search engines. So it depends
which type of BESDEO you're talking about.
Speaker 2 (17:14):
So there are white hat SEO people, it's not just
all absolutely absolutely, So what do you do for your job?
What is the thing you make your customers do well?
Speaker 1 (17:25):
You know, companies come to us and they obviously want
to get more visibility, So we have to look at
all the different considerations and all the different factors that
influence how people are able to find that content. A
lot of it's technical, so literally a lot of our
work is like, you know, focusing on page speed, focusing
on overall accessibility. And you know, for example, if an
(17:48):
image is uploaded to a website and it doesn't have
image alt text, which is actually required for website accessibility,
search engines can't understand what's on that image traditionally, so
we have to go in there and kind of add
the right words in. So it's making that content more findable.
And again, a lot of that's technical, but a lot
of that is content oriented. So people might write a
certain article and in such a way that the way,
(18:09):
you know, when people go to Google and they're looking
for that article, they could never find it. So we're
helping companies to kind of structure their content in the
way where people can actually find what they're looking for.
Speaker 2 (18:19):
So you say that there are these signals, how much
of that comes from the search engines themselves.
Speaker 1 (18:25):
Well, you know, the SEO field, because search engines are
essentially a black box, and a lot of us are
working with you know, just third party data or working
the best we can with the analytics that Google any
other search engines give us. A lot of this is guesswork,
you know, we try to make informed guesses based on
the data itself, but Google and the search engines do
give us a lot of clues and a lot of
(18:46):
insights about what we should be focusing on, especially when
it comes to the technical thing, so you know, giving
us guidance about what we call like canonical tags. There
are different metadata that we have to add to a
website for them to really be able to parse crawl
through that website effectively. Those are direct signals that Google
gives us that we work with. But a lot of
it is kind of, you know, interpreting what we're seeing
(19:08):
as far as what the data shows and then making
informed guesses based on the history and on the analysis
that we do for different clients.
Speaker 2 (19:16):
So based on how Google describes it, they're fairly communicative
with web developers and sorry, web content creators, it doesn't
actually sound like that is true.
Speaker 1 (19:26):
I would say when it comes to the technical side
of things, Google's quite communicative. You know, it's really in
their best interest to have web you know, fast websites
websites that are accessible for users, websites that have a
great user experience. And Google does a lot of developer conferences,
a lot of communications on the technical side, and it's
quite easy to, you know, kind of get in touch
with them if you have a specific technical question, like
(19:48):
for example, John Vieler hosts Webmaster hangouts on Google often,
if not like several times per month where he answers
a lot of people's questions. So, yes, technically speaking, I
think they do provide a lot of guidance.
Speaker 2 (20:00):
About content because it seems like content, the actual things
on the web page is the thing that actually matters here.
Does Google give any guidance?
Speaker 1 (20:09):
They certainly do give guidance. They for example, they give
a lot of questionnaires that site owners can follow when
thinking about what Google considers helpful content. They also have
something called the Search Quality Radar Guidelines, which is a
document that you can find online. It's about maybe one
hundred and sixty plus pages, and that's specifically used to
inform Google's human Search Quality evaluator team as to what
(20:33):
good content or bad content kind of looks like. So
that's something that the SEO community looks at a lot
as well. But to your point, you know, getting into
granular questions about is this content good or bad for Google?
What does helpful content mean? That's become trickier in the
last few years because they can't tell us everything because
(20:54):
people spam everything that they can get their hands on.
Speaker 2 (20:57):
Right, So it almost feels like a kind of battle. Yes, absolutely,
but to this to this level, to this point you say,
they give this this advice and they're vague on the
content but direct on the technical. So how has Google
search quality got so much worse?
Speaker 1 (21:15):
Then? There's so many different factors at play, you know,
I think the reason why seos have a bad reputation
in general, And to your question earlier, you know, is
there such a thing as white hat seo. There's a
huge spectrum of approaches and beliefs within the SEO space.
For whatever reason, the people that spam Google the most
(21:36):
are the ones that have been kind of creating this
like the kind of the reputation that we have as
an industry lately. And I get that because they've been
quite successful. You know, a lot of people that are
spamming Google, especially with AI lately, have have really made
it into top positions, so they're getting a lot of
visibility with those tactics. But you know, there's a lot
of us that are doing work that again, it's like
(21:56):
kind of advised by search engines to do because it
really does make the Internet easier for everybody to use.
Uh So that's been a challenge lately. I think that
for whatever reason, you know, maybe because Google's shifted internally
so much of its focus to AI lately. Obviously there's
these big external challenges that Google's facing as a business lately,
(22:17):
maybe they're they're not haven't been as focused on search
quality for whatever reason. I don't actually have, you know,
any type of evidence of that other than what I'm seeing,
But I haven't seen as much spam and like SEO
manipulation as I've seen over the last year or so.
That being said, they're really cracking down on it in
the last few months, so sometimes it just takes them
(22:39):
a while to catch up.
Speaker 2 (22:41):
Something you've said publicly recently is real business is all
being affected by the most recent Google updates? Can you
kind of get into that, like who are the people
being affected here?
Speaker 1 (22:51):
Yeah? I mean that's that's nothing new. I think, you know, Google,
Google creates these algorithm updates multiple times big ones at
least multiple times per year every year. And we've seen
a lot of companies get caught up in those algorithm
updates before. In most cases they're doing something that goes
against Google's SEO or spam guidelines. The difference now is
(23:12):
that there's a lot of these new types of algorithm
updates that Google is launching that are affecting a lot
in a lot of people in the publishing space, so
both large and small publishers, especially smaller publishers over the
last couple of years. And these are, for example, people
that might have built a business doing travel blogging, doing
(23:32):
recipe blogging, product reviews, and so if you do consider
those people at business, and I think they would because
they often have staff, they have people going to the
places writing the travel reviews. Things like this, Google's dramatically
reduce the visibility of how those sites appear on Google
and the amount of traffic that they're getting.
Speaker 2 (23:52):
So when was the last major update to Google? Sometimes
in March right, or it's unclear when it actually started
and finish.
Speaker 1 (24:00):
Yeah, the last major update I would say was the
March core update, which took forty five days to complete.
But that started in early March and then ended in April.
But there's been a few different aspects of that update
in new spam policies that they've introduced as well, So
we're actually still kind of in the middle of some
of that rollout happening.
Speaker 2 (24:19):
Yet it's it's strange, though you and this is not
on you, This is on Google. You say there's technical things,
they say, there's quality things they say, but even recipe
sites look terrible. They're full of ads there they make
your phone so hot it burns you. Okay, maybe not
that hot, but it feels like Google doesn't actually seem
to be maintaining quality, and it seems like they almost
(24:42):
fight those who do. An SEO, just how do you
feel about your relationship with Google?
Speaker 1 (24:50):
It's hard to say right now. I think, you know,
with something that's been very tricky for me is that
I have a bit of a I've built a bit
of a name for myself as the person that goes
around and and really echoes a lot of what they
recommend site owners do. That's been kind of my brand
for the last several years. Google has different, you know,
policies and things that they recommend to site owners that
(25:13):
I have found, I would say, up until twenty twenty three,
I've found those recommendations to be very consistent with the
work that we're doing for our clients. So it's very
easy for me to go say, yeah, what Google says
to do is what works right, And we've worked with
a lot of different companies and found those tactics and
those recommendations to be very effective. The last few months,
(25:35):
that's changed a bit for me. I'm not well, I'm
not you know, my personal relationship with them is I
want to be very careful to say, like, you know,
the people that I know that that work at Google,
and they you know, speak on behalf of Google, Like,
this has nothing to do with them, right, This has
to do with what's happening in the search results. And
you know when I see ai spam ranking above my
(25:59):
clients for what we call top stories, so that's you know,
the news features on Google for example, or I see
that in Google's new aioverviews is getting everything so many
different things wrong, and there's spam and there's malware in there,
like things like this. It's just it makes us as
seos look bad because we then have to go defend
those results to our clients and try to make sense
(26:21):
of them. And that's been really challenging, very frustrating for
me lately, because it's very hard to defend the quality
of the search results when the search results are in
a state of very poor quality.
Speaker 2 (26:33):
So you and you said previously up until twenty twenty three,
was it what happened in twenty twenty four? Was it
the core up they or was there just something that
happened this year?
Speaker 1 (26:43):
Writ large, I think that probably the biggest change that
happened was honestly chat GPT. That's there's one of many
big changes. But the you know, not only chat Gypt
but the different tools like it empowered spammers and sea
alike to find new ways to speed up what they
were doing and to create a lot of terrible content
(27:05):
at scale. And that worked extremely well for a year
because guess what, it takes Google and the other search
engines a while to kind of figure out what they're
doing to develop new algorithms and new spam policies that
you know, devalue or demote that type of content. But
there was about a year where that content was succeeding
extremely well. So towards the end of twenty twenty three,
(27:28):
you would see a ton of AI generated spam and
the search results and people using AI tools in new
and different ways to just accelerate whatever spamming they were doing.
Speaker 2 (27:38):
Do you think Google actually has control of search or
has it just kind of got out of hand.
Speaker 1 (27:46):
I think they you know, they had these new set
of updates beginning in March twenty twenty four where I
think that they made loud and clear that they're not
messing around. You know, they really cracked down on a
lot of sites that were doing a lot of different
SEO tactics and spam tactics. So I do think they're
getting controlled back. I do actually also think they kind
(28:08):
of overstepped in some areas and there's a lot of
collateral damage there as well, which I know they're working on.
And that's where you hear a lot of people saying,
you know, this destroyed my business because there were a
lot of people that just had no idea that they
were going to get caught up in that. So that's
on the search side. I think, you know, Google's proven
time and time again that they can eventually get ahead
of it. But I think they're presented with a lot
(28:29):
of new and different challenges right now on the AI side.
You know, I'm sure eventually they'll figure it out, or
this AI overview thing will get better over time. But
it's been quite interesting to see publicly this week, so
many examples of Google's new AI overviews getting information blatantly wrong.
Speaker 2 (28:49):
I want to push back a little bit about the
control over search that Google has. The reason I asked
that is it's not been just chat GPT going back
to twenty twenty. Results have been bad for a long time,
and it feels that Google is fighting SEO, and SEO
I realized is a much wider umbrella because there's people
(29:11):
trying to do the web properly. There's people who were
just trying to mess with Google, and then there are
the people who just don't care and are spamming. It's
just it doesn't feel like they're able to make quick changes,
and I just don't understand it. It drives me insane, honestly.
Speaker 1 (29:24):
Yeah. Well, well, one of the other things that I
didn't mention that's been a big change to the quality
of search results, and one of the big complaints that
people are having is a lot of big sites with
a lot of authority on Google a lot of SEO history,
A lot of the big publishers that we all know,
you know, big brands that everybody's familiar with, found a
lot of SEO opportunities in the last couple of years
(29:46):
that have worked splendidly for them. So, for example, affiliate content.
You might have noticed that nearly every big publisher on
the Internet has ventured into product reviews in the last
couple of years, and that's because it's a huge money maker.
So one of the big complaints that people have is like,
it's not fair that X, y Z big publisher shows
up every time I'm typing best running shoes for women
(30:07):
or whatever the keyword is, and believe it or not,
you know this. I think this has been one of
the biggest complaints that people have had. But in the
last two or three months, Google is cracking down on
that in a very big way. I think it's just
taking them a long time to figure out specifically how
to target that.
Speaker 2 (30:25):
Why do you think it takes them so long? They're
worth like two trillion, they surely they have the people.
Speaker 1 (30:31):
I wonder the same things myself. Sometimes I'm not a
search ranking engineer, so I don't know why I also,
I think you know better than I do what's going
on internally at their company right now. I think there's
a lot of a lot of kind of pandemonium in
different ways. I don't know, but it seems that way.
Speaker 2 (30:47):
But that being yeah, go on, you go on.
Speaker 1 (30:50):
I was just gonna say, another reason why this is
so difficult for them to target is because not all
of this is easy to spot algorithmically. For example, if
a publisher decides to start working with third party contributors
who are contributing product review content, those relationships are not
always entirely clear to a point where an algorithm wouldn't
(31:13):
necessarily be able to figure that out. So Google has
to use a variety of different methods to demote that
content and learn about those business relationships, and you can't
just solve that all algorithmically.
Speaker 2 (31:27):
Do you think they want to solve it?
Speaker 1 (31:30):
I would imagine yeah, they want to do it algorithmically
instead of using their web spam team. But currently they're
using their web spam teams. So that's real human beings
that are sending out what they call manual actions or
penalties to sites that are engaged in different types of
spam policies, and they've created some new ones recently.
Speaker 2 (31:47):
It just feels like they could I say this not
as an SEO expert, but it feels like they could
do very obvious things. They could go and go with
dot Meredith, who has the met who you read, Hello
House Read is whole thing around how Meredith has been
basically spamming affiliate marketing content that's not particularly well done,
and you've written about this as well. It feels that
(32:09):
they could just do something to them immediately that why
don't they take these fast actions because it doesn't seem
like they're doing much anyway.
Speaker 1 (32:17):
Well, a lot of these publishers have content that a
lot of people really love, and you know, Google has
a lot of signals that indicate that these are, you know,
sites that produce extremely helpful content in a lot of
different categories. So it's not so easy to say this
category of the website is not okay by users, but
this one is. You know, it starts to become tricky
(32:39):
when they have to go after certain sections of a website.
You know, a lot of times with sites that are
just doing pure spam, it'll be the whole site that's
doing spam. But these are obviously not to mention these
are very very big brands. You can't you can't remove
a site from Google or demote a site from Google,
for if it's a brand that people are looking for
and a brand that people would expect to see, right,
(33:01):
because that may mean you can sort.
Speaker 2 (33:03):
Of well, it's kind of I think what I'm what
I'm getting at is what they judge is good for
the Internet doesn't necessarily seem so because there's a self
fulfilling prophecy there there is, Oh, this is popular with
a bunch of people. Well maybe it's popular because Google
made it popular. Maybe it's not. Actually it's just an
algorithmic choice. But I'll get I'll get onto something a
(33:25):
little more direct, which is, how do you feel about AI.
Speaker 1 (33:29):
Search, like Google's new AI overviews?
Speaker 2 (33:32):
Yes, yes, sge Yes, that's generative experience.
Speaker 1 (33:36):
It's like, obviously we're in a big transition moment where
you know, speaking about it now is not reflective of
what it's going to look like in six months. So
I always want to be cognizant of like this thing
changes every single day. My biggest issue with it, so
first of all, Google made it public, you know, a
week ago, at Google Io, and myself and many others
have been experimenting with it in Google Labs for a
(33:58):
year before that, and in theory, you know, with AI,
it should be improving and learning and getting better every day, right.
For a year, myself and many others were raising flags
about quality issues and also the fact that we just
didn't feel like it often provided that much of an
improvement to the existing search results. For example, Google already
(34:19):
had featured snippets, so if they want to take a
paragraph or a snippet from somebody's website and display that
directly in the search results, there was already a mechanism
for that. So for the past year, you know, I've
raised a million different examples of when it got things wrong.
When I think it was, you know, responding to some
type of super controversial question in a weird way, whatever
(34:40):
it was. And when they launched it, you know, literally
I can tell you because I spent my whole life
at SEO conferences. Most of us were like, there's no
way they can launch this thing in its current form,
like it's just so not ready, right, And they did,
And here we are a week later, and nearly every
you know, social media site has tons of examples of
people wanting to it off, the tons of examples of
(35:02):
it getting things wrong. So you know, I want to
be cognizant and respectful the fact that it will change,
it will get better. I'm just quite surprised that they
launched it in its current form when they did.
Speaker 2 (35:13):
One last question And don't take this the wrong way,
but why do you think they'll improve, because the last
few years have shown that they've got worse.
Speaker 1 (35:23):
Well, in theory, these AI lm should learn and improve
and get better and get retrained all the time. I'm
going off with that optimistic idea. But at the same time,
it's like we're doing, you know, offloading this work onto
Google searchers and relying on them to provide feedback and
everything is I guess it doesn't necessarily mean it's going
to go in the right direction. So it's a good point.
(35:45):
It's hard to say.
Speaker 2 (35:47):
Yeah, it's just it feels almost like the AI side
is getting away from what search is meant to do,
because search is meant to help you find useful things,
and this doesn't even seem to be capable of helping
you find the right answer, let alone a good link.
Speaker 1 (36:01):
Yeah, I think The main issue that I currently have
with it is that once I know that it's twenty
or ten percent wrong or whatever whatever the number is,
we don't know the actual percentage of wrong answers. You know,
based on how many we've already seen in the last week,
it's quite a lot. So when people know that this
thing has a tendency to be wrong, will they trust it? Right?
(36:23):
And what does that do for trust and Google overall?
Because with featured snippets, at least you can say, Okay,
this comes from this website, you can blame that website
because they pull that data verbatim. When they're showing a
featured snippet with.
Speaker 2 (36:35):
Air or reviews.
Speaker 1 (36:35):
It's like there's a tendency that one of these sentences
is completely wrong, and it's pulled from an eleven year
old Breddit threat that was making a joke, and we
just don't know which sentence it is right.
Speaker 2 (36:46):
Right. It just it feels like they're almost moving against
the Internet. It's just a very very depressing thing to watch.
Speaker 1 (36:53):
Yeah, I agree. I mean I think I would have
expected this thing to be more fact, more evidence based
in its rollout. I would have expected it to show
for far fewer queries where getting things wrong is extremely dangerous.
So yeah, I mean, I like, our industry now is
(37:14):
digging into the data and we're finding that it seems
like health is actually one of the categories where AI
already is triggered the most health topics.
Speaker 2 (37:22):
That's genuinely scary. But okay, actual lost question here, What
is the thing that Google could change to get better?
Speaker 1 (37:32):
Huh? I mean, I would imagine they're they're working on
a lot of the feedback that they've gotten in the
last week with the AI over reviews and finding a
lot of examples where it gets things wrong. And if
they're going to continue with aio reviews, I would hope
that they dramatically improve the product as quickly as possible.
That's on the AI front. I mean, if if you
(37:52):
want to be if you want to.
Speaker 2 (37:53):
Completely in general with search, Yeah, with.
Speaker 1 (37:56):
With search, I think that they should reassess this notion
of helpful content because I think that they've inadvertently caused
a lot of sites that are truly producing helpful, experience
(38:17):
driven content to lose a lot of visibility in the
last year or so, and a lot of like you know,
we're still seeing a lot of big brands seemingly able
to show up for every possible query, and we're not
hearing enough from small publishers. We're not giving small publishers
enough of a chance for their content to compete, and
they're kind of At this point, Google's just seems to
(38:37):
be blasting Reddit everywhere for everything, maybe as a solution
to that, because yeah, you get a lot of firsthand
experience from Reddit, but I'm not sure that putting Reddit
in top positions for so many different queries is a
solution either. So personally, I'd like to see a broader
diversity of the types of websites that are able to compete.
Speaker 2 (39:05):
The problem, it seems, is that this multi trillion dollar
tech company, one that has a monopoly over the Internet's information,
is somehow we can we can power us when faced
with people that create websites specifically to trick it, Google,
a company that sells to militaries and governments and professional
sports leagues that makes tens of billions of dollars a
quarter and employs nearly two hundred thousand people, is incapable
(39:28):
of thwarting a hodgepodge industry full of nakedly craven bad actors,
or they just un willing to intervene because it Google
doesn't really care about anything other than how many times
people search and how many ads they see, And they
don't really mind if maybe this big mess of crap
makes the whole thing worse. This problem, it's been going
(39:50):
on for years, and I feel crazy every time I
talk about Google Search as a product and has a
business that's such an incredibly profitable and ubiquitous product can
be it's so utterly shoddy and regularly unfit for the task,
I feel insane. I feel crazy when I see article
after article about Sandhar Pushai and Google talking about the
(40:10):
remarkable success of Google, yet never the fact that they've
done so much to damage this search engine, and they've
so blatantly destroyed their core product. And what really upsets
me is the sheer contempt that Google seems to have
for its customers. Google Search is awful. It's a terrible,
(40:31):
unreliable product, and now they've made it even less reliable
with the introduction of Generatifai. And yet sun Dar Pushai,
the CEO of Google, can stride into interviews all piss
and vinegar and ramble about how evalues independent sources and
more authentic voices as his service continually fails the surface
unique interesting or useful content. Nil Patel of The Verge
(40:56):
got a lot of credit for giving Peshai a tough interview,
but I'm gonna be bloody honest, paper Tiger. It's just
another opportunity for a Google executive to vaguely suggest that
they care about search without really being held accountable. The
question is, sun Dar, why is Google so bad? Now?
And he will give you some vague pr spiel. Show
him results. Nille did show him Chromebook results. The thing
(41:19):
to do would be come with a few results and
show him in front of him and say why is
this so bad? Why? Sun Dar? And if you can
get him to say on the record, actually, I think
it's good. That says everything. What it says is he's
a liar and he doesn't really care. But I guess
the interview did that anyway. Look, the problem, other than
(41:39):
the fact that Google should be broken up and that
search should be a public utility or a nonprofit, is
that the media seems to continually lack the ability to
simplify the problem. Google Search is bad, and it's been
bad for a while. Results are worse. They're full of
spam and low quality content. Google does a terrible job
of maintaining the content on Google Search, and it's now
introduced an AI to search through these results. That makes
(42:02):
things worse by generating answers based on search results from
a search engine that's rotting in real time, and then
it hallucinates the results. Anyway, Eh God, I apologize them
getting a little bit head up, but this whole thing
just it kills me. It makes me feel a little
bit nuts because if any of us, anyone listening to this,
(42:24):
went into our jobs and our trousers were down and
our our bits were out, and we were walking in,
and we go in and we half do our job
and the job's terribly done, would we get fired? We
get fired. I think we'd get fired, maybe if the
crap work we did made money, but we'd still be
in trouble for our willies hanging out. And that's the thing.
None of these tech executives, even when faced with the media,
(42:45):
seemed to actually be held accountable. And Google acts without integrity,
and they just act with this stunning contempt for their customers.
I've said it before, but it bears repeating. Suned up
a shite is a goddamn management concern, a former McKinsey
free with little interest in providing users with any kind
of usable service, and his cronies people like Probagara, Ragavan
(43:06):
and Liz Read. By the way, Probagara has been upgraded
to head of a bunch of things now and Liz
Read is now the head of Search. It's so good
he picked his successor, and it's so good that this
person loves Ai. And these people they've continually helped Google
put out bad software, all in pursuit of perpetual growth.
And I believe they're part of a larger problem in
(43:29):
the tech industry. Tech firms are no longer building things
for customers. Big tech and honestly, far too many startups.
They're creating these products that are more like symbolic capital
for investors so that they can place bets on them
and say, ah, maybe red war here, no, maybe black
this time, maybe double zero, And these things they only
(43:49):
exist to express meteoric growth ten x, twenty x, thirty
x returns things that sound satisfying for people to chirp
on CNBC without ever actually really having to provide the service.
They're discussing I'm not to be honest, it's breaking my arm.
As I said in my last episode, I'm not a pessimist.
I'm more of a broken heared romantic. There was a time,
(44:12):
and it wasn't super long ago, that tech was exciting
and fun and interesting. When things came out that subtly
or revertedly, they changed our lives, They made things better,
they maybe connected us, or they made our time on
our own better and more substantive, and things felt like
they were progressing. Yet it feels like so much of
today's tech industry is focused on shipping products that nobody
(44:33):
asked for to fix problems that don't exist, all while
demanding applause for a future that they've yet to make
a reality. Cryptocurrency, the metaverse, now generative AI. They're all
trends that built this kind of pseudo religious movement to
carry water for products that didn't do what they said.
They would guilt tripping those that said otherwise for not
being optimistic enough about an industry there's continually failed to
(44:56):
give us hope for the future, and it's disgraceful. It's disgraceful.
Look at Facebook from my last episodes. Look at that website,
that multi billion dollar website. Look how bad it is.
Look how full it's of AI, spam and junk it is.
Look how unsafe it is. Look at Google right now.
Google Search is falling. Apart of the seams. It kind
(45:19):
of works and sometimes it does, don't get me wrong,
but for the most part you just can't trust it
any more. And that was before they plugged generative AI
into it. Now the world's probably the world's most profitable
tech product is just languishing. It's so much worse. And
what's crazy is I don't even know how they're gonna
make money from it because these AI results they don't
(45:41):
seem to connect to ads or anything. They don't seem
to make Google money. But because Sandha Peshei needs to
convince Wall Street that Google will grow forever, well, you
gotta put AI in it, man, You gotta make the
AI happen. Except regular people using this, they're gonna get misinformed.
Someone's gonna die. And I really sound dramatic, I know,
(46:01):
but they are because people use Google Search and rely
on Google Search. They rely on Google Search to give
them real answers to real queries. And we can laugh
about the fact it said that someone had to eat
a rock a day. But what if they're trying to
put out a chemical fire. What if they're depressed then
they ask Google something and Google tells them, I don't know,
(46:22):
some sort of weird CBD treatment because Google got poisoned
by an AI spam result. There is, by the way,
a fake result about Google telling someone to jump off
the gong Gate Bridge. It's very fake, done for for it.
But seriously, though, something like that's going to happen. And yeah,
I'm sure someone will say, oh, it's just an edge case.
I don't give a crap an edge case where someone
(46:42):
dies or is harmed or is misinformed. That's not worth it.
It's not worth it. And these companies must learn. And
I think what's going to happen is far simpler. I
think they believe that people are just going to keep
using it because they have these horrible monopolies. I think
at some point people are just not going to use
search at all. I think that is the actual end
(47:05):
result here. I think you see open AI create search
engine and it kind of is worse than this. I
think you're going to see a proliferation of this tech
as people desperately try and catch up with Google and
it's so distressing, it's so insulting to customers, turns my
bloody stomach. I realized this wasn't a very positive podcast,
(47:26):
but hope it filled in the gaps for you. Thanks
for listening, Thank you for listening to Better Offline. The
editor and composer of the Better Offline theme song is Metasowski.
You can check out more of his music and audio
projects at Matasowski dot com, M A T.
Speaker 1 (47:48):
T O.
Speaker 2 (47:49):
S O w Ski dot com. You can email me
at easy at Better Offline dot com or visit Better
Offline dot com to find more podcast links and of course,
my newsletter. I also really recommend you go to chat
dot Where's youread dot at to visit the discord, and
go to our slash Better Offline to check out I'll Reddit.
Thank you so much for listening. Better Offline is a
(48:11):
production of cool Zone Media. For more from cool Zone Media,
visit our website Coolzonemedia dot com, or check us out
on the iHeartRadio app, Apple Podcasts, or wherever you get
your podcasts.
Host
Ed Zitron
Popular Podcasts
24/7 News: The Latest
Today’s Latest News In 4 Minutes. Updated Hourly.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations.
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.