September 28, 2023 • 20 mins
Chinese-backed hackers are infiltrating computer systems by compromising routers. Some security researchers say that modern GPUs can steal pixels right off a web site. And Microsoft encounters yet more resistance regarding its acquisition of Activision Blizzard.
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 (00:04):
Welcome to Tech Stuff, a production from iHeartRadio. Hey thereon
Welcome to Tech Stuff. I'm your host, Jonathan Strickland. I'm
an executive producer with iHeartRadio. And how the tech are you.
I am still in Las Vegas, Nevada. You can probably
tell that my voice is going. I did a lot
(00:25):
of talking yesterday. I had to do some recordings of
the Restless Ones podcast, and I did a lot of
chatting with other folks, and yeah, I don't know if
I'm coming down with anything. I sure hope I'm not.
I've been masking every time I can, whenever I'm going
out in public, and I recommend that y'all who have
to go out in public do the same if you can,
(00:47):
because obviously COVID is having another surge, and I want
folks to be safe. And hopefully all I have is
just like wear and tear on the old vocal cords
and that's it. We'll obviously be very careful and keep checking.
But in the meantime, we have some tech news to
get to, so let's talk about that. Officials here in
(01:08):
the United States, as well as some in Japan, have
announced the discovery of a Chinese hacker attack that is
infiltrating organizations by compromising older routers. You can read up
on this in Ours Technica. There's an article by Dan Gooden.
The article is titled Backdoored firmware lets China State hackers
control routers with magic packets. The article goes into a
(01:31):
lot of technical detail. I'm not going to bother doing
that here because this is really just a news item.
But essentially this hacker group, which has lots of different names,
I'll just choose one of them, which is black Tech. First,
they have to get access to a system. Now how
they do that doesn't really matter. It's just they need
to get access. First, administrative access, like administrator level access
(01:54):
to be specific. They might get that through some compromise
login credentials, perhaps from another hacker group, maybe they use
social engineering, but whatever. First they get that access. With
that administrator level access to the system, they then push
an older version of firmware to certain routers, older routers
that don't have a protection against this kind of attack.
(02:16):
And typically they're targeting routers that are on the edge
of networks. So for example, they might target a branch
office of a larger company, partly because they're less likely
to be noticed, and that security won't be quite as
robust as it would be if it were at HQ. Now,
because these routers have a trusted relationship with all the
(02:39):
other computer systems that are connected to that company, the
hackers are then able to get widespread access to the
full organization. They target the edge routers, they infect those,
and then they snoop around Good and explains in his
piece that the more recent router hardware that has come
out over the last couple of years includes protections against
(03:01):
this type of attack, But obviously most organizations aren't updating
their routers regularly. They do so when they need to,
and determining if your system has been affected it seems
like it's not very easy or straightforward, particularly if the
hackers are actually being really careful. So it's another example
of how cyberwar is a real ongoing thing and that
(03:23):
state backed hackers continue to infiltrate both companies and public organizations.
As for the magic packets, those are little small packets
of data that the hackers are using essentially to open
or close backdoor access into these systems, and typically the
magic packets kind of blend in with overall network traffic
(03:44):
so that's kind of why they were called that. Back
in July, word got out that hackers had leveraged a
stolen Microsoft certificate to gain access to the email systems
belonging to really large organizations, including the United States State Department. Now,
the New York Times report that the hackers stole around
sixty thousand emails, primarily from ten email accounts. Further, nine
(04:07):
of those ten accounts belong to members of the State
Department who are working on projects that involve East Asian affairs. Now,
the State Department has not formally accused China of being
behind the attacks, but that seems to be the implication here.
The hack has sparked conversations within the US government on
how to harden systems against hacker attacks, including the possibility
(04:29):
of moving away from single vendor solutions like using Microsoft.
And while there's an element of closing the barn door
after the horse has already escaped going on here, I
do think putting all your trust in a single vendor
can lead to massive problems down the line. But honestly,
there aren't really any easy, fully reliable solutions here. Security
(04:51):
researchers revealed that all modern GPUs, that is graphics processing
units have their own way to compress data. But even
though they're each using a different method, that method can
be exploited to steal pixels from a website. Which sounds weird, right,
but essentially here's how it works. You get hackers and
they create a malicious website that has the same name
(05:14):
but a different domain as their target. So let's say
they've identified a target website that they don't have access
to directly, maybe it's behind some sort of protective measure,
So they create a different website has the same name
but a different domain, so instead of dot com, maybe
it's dot biz. Now in this malicious version of the website,
(05:34):
they include an iframe. The iframe serves as a spot
where you can embed other content into your website. So
one common use for iframes is to embed ads in them,
and that allows the site to swap ads out dynamically
and the iframe kind of holds its place. They then
rely on the GPU's data compression scheme to pull pixels
(05:55):
into a side channel, and they pull it from the
target website. You reconstruct what appears on your target site.
So you might have a corporate website that's usually behind
protection and the hackers create a malicious site named the
same thing, but on a different domain, and by stealing pixels,
they can recreate what was on that target website, potentially
(06:18):
gaining access to sensitive data, including things like you know,
log in credentials. The researchers were mostly showing that this
is possible. They didn't indicate that it's an active concern, Like,
they didn't say that this is something that they're seeing
actively deployed out there, but that in fact the hardware
across all GPU manufacturers allows for this, and since it
(06:40):
is possible, sooner or later someone will do it. Right,
So there are ways to block pixels stealing. Most of
those involve website administrators being proactive on the matter. They
can build stuff into the header of a web page
that will protect against that kind of thing, but that's
a lot of work. It is also probably, you know,
(07:01):
it's likely, I think that we're going to see companies
that make web browsers that facilitate this kind of attack
to address that in the future. The European Union issued
a report stating that after analyzing various social network platforms
out there, one in particular stands out when it comes
to the proliferation of disinformation. Do you have any guesses?
(07:23):
If you said X marks the spot, meaning the platform
formerly known as Twitter, you win a prize a virtual
donut enjoy Anyway, the EU previously established a voluntary code
of practice on disinformation, and tons of platforms signed to
(07:45):
pledge and that they would follow this voluntary code, so
that includes Google and YouTube, meta primarily Facebook and LinkedIn.
They all signed it, along with like forty other platforms.
It also used to include X, but at last May
Elon Musk decided that X was going to peace out
of that voluntary code. Now, researchers in the EU say
(08:07):
that disinformation is spreading like wildfire on X and that
the problem is only getting worse. Also, X has recently
disabled a feature that used to allow users to report
cases of misinformation about elections. So this means, assuming those
EU researchers are right that X is not just the
worst when it comes to preventing misinformation and disinformation from
(08:29):
propagating across their system, they are actively turning off features
that would do anything about it, which is a big
old yikes. An article in the LGBTQ Nation reveals that
TikTok is working with the government of Kenya to restrict
and remove LGBTQ plus content from the platform within Kenya's borders. Now.
(08:51):
This is because Kenya's government has banned same sex sexual relationships,
and currently the government is also considering an additional law
that would hunish homosexuality with a lifetime prison sentence. TikTok
has already deplatformed and demonetized TikTok users in Kenya who,
according to the country's laws, have posted restricted content. Apparently,
(09:13):
TikTok CEO has committed to not only being more active
to remove LGBTQ plus content from the platform within Kenya,
but to also launch a campaign to urge users to
instead post quote unquote positive content, which sounds disturbingly close
to conversion therapy. In my opinion, it's really ugly stuff
(09:34):
and my heart goes out to the people of Kenya.
This week, Meta announced an update to its line of
smart glasses, previously known as Stories, but to be really
formal about they are now known as ray Ban Meta
smart Glasses. The glasses have a pair of cameras in
them when at the top of each corner of the frames,
(09:55):
and they contain five microphones that are meant to provide
better sound pickup when you use your glasses, you know,
take a phone call or something. They also have speakers
that the company says will make it easy for the
wearer to hear what's playing on their glasses, but it won't,
you know, irritate the living heck out of everybody else
who happens to be nearby. The glasses can take about
five hundred photos at twelve megapixel resolution or around one
(10:19):
hundred and thirty second long video clips at ten ADP
before you end up using up all the storage on
the device and it's time to offload stuff. The battery
reportedly is good for like four to six hours of use,
and they come in a case that can also serve
as a charger, and the case itself can hold enough
battery juice to recharge the glasses several times before the
case itself needs to be recharged. You can also live
(10:42):
stream while wearing these glasses too. The controls include both
a touch system and the stems of the eyeglasses and
voice commands, and essentially these things do some of, but
not all, of the stuff that Google Glass did years ago,
but they do look a lot better. They do not
as far as I can tell, create any sort of
ar view of the world around you, which Google Glass
(11:04):
could sort of do. And already folks are worried about
privacy issues with these glasses. They're concerned that you could
be on video or have someone snapping pictures of you
without your knowledge or consent. Now I get that concern.
It's a legit concern, don't get me wrong, But y'all,
that's already happening. There are so many cameras out there,
whether they're in phones or in security systems like in
(11:24):
door bells or in cars, So I pretty much guarantee
that you're on camera frequently throughout an average day. But
I do understand how it gets creepy when you're talking
about someone wearing the camera on their face. Anyway, Meta
is taking pre orders on these things, which start at
around two hundred and ninety nine US dollars, and they'll
start shipping in mid October. So maybe your Halloween costume
(11:48):
can be invasive surveillance. That's a fun one. All right,
We're gonna take a quick break and we're gonna come
back with more news in just a moment. We're back.
Reddit indicated this week that it will remove the opt
(12:10):
out feature for personalized ads, or at least some Reddit
users around the world. So the new policy says that
Reddit will take your activity on the platform and then
use that to decide which ads to display to you.
So let's say you spend a lot of time on
subreddits that are dedicated to gaming, Well, then you're more
(12:30):
likely to see ads related to gaming while on Reddit. Right.
It says this is for your benefit, But there are
a lot of users who are upset because the message
that they're picking up on is Reddit is determined to
track their behavior across the site and there's no way
to tell Reddit to knock it off. And while Reddit
says users all around the world will be able to
opt out of personalized advertising based off of quote unquote
(12:54):
information and activity from our partners, only users in select
locations will actually be able to opt out of personalized
ads based off their activity on Reddit itself. I'm guessing
those select locations will be places like the European Union,
which has some pretty strict rules in place when it
comes to user privacy, and any place that doesn't have
(13:15):
those kind of rules, Redd, it's going to be tracking
you like crazy. And now, the latest in the long
arduous story of Microsoft's plan to acquire Activision Blizzard. When
last we left our tail, Microsoft was working hard to
convince regulators in the UK that the deal would not
result in an anti competitive situation within the world of
(13:37):
console gaming in general and cloud based streaming gaming in particular.
Well last week that regulatory agency gave provisional approval to
the acquisition, so that roadblock was removed. But now the
United States Federal Trade Commission has once again objected to
this deal. Now, they previously attempted to secure an injunction
(13:58):
against the deal closing, but a US District Court judge
denied that request and said that the FTC had failed
to produce evidence that this deal would actually be harmful
toward competition. The FTC then appealed that ruling, but then
withdrew the complaint not that long afterward, and it turns
(14:19):
out that was temporary because now they have submitted the
appeal again to the Ninth Circuit Court of Appeals. Once
the regulators in the UK gave their provisional approval, and
then once the court actually gives a decision one way
or the other, the FTC plans on an evidentiary hearing
on the matter. Now This on the surface sounds bad
(14:42):
for Microsoft and Activision Blizzard, but they could still go
through with their deal because there's no injunction against it,
and that deal is scheduled to close on October eighteenth.
Then once the deal is closed, they can worry about
any ongoing legal issues they might face after the fact,
but by then the deal will already be done. So
what I'm seeing is that most analysts think this deal
(15:04):
will finally close next month, and now for a few
AI stories, it's always going to be that. In a
tech Stuff News episode, an SEO consultant named Gagan Gotra
brought some concerning information about Google Bard conversations. So it
turns out if someone has a conversation with google Bard,
(15:27):
which I'm just going to remind you that's Google's AI
powered chatbot. It's similar to Open AI's chat GPT, specifically,
as it is integrated with bing, well, if you then
share a link of that conversation with someone else, then
Google will actually index that conversation, which means that conversation
(15:51):
with google Bard can pop up in future search results.
So let's say that you and a coworker are using
google Bard to help develop a business plan and this
is something that you do not want to share outside
of your organization. When you share that conversational link to
your coworker, the link itself becomes indexable, and then Google's
(16:13):
web craller will index the conversation, and if someone else
uses the right search query, that conversation can potentially pop
up in those search results, and the information that you
intended to be private has now been made public. Now,
obviously there are all sorts of situations where indexing the
barred conversation could turn out to be a really bad thing.
(16:33):
It's not that different from when open Ai had an
issue with chat GPT, where the chatbot would occasionally give
users access to other users chat histories with the chatbot. Now,
in the Google case, we're not talking about a bug,
we're talking about a feature. Because Gotra had an exchange
with Google research scientist Peter Lieu, and Peter pointed out
(16:58):
that the search engine will only end conversations if someone
has clicked on the share button, and Goscher's reply was
the same one I would have made. Namely, I think
most users would just assume that share just means you've
elected to share that conversation with someone in particular not
the whole world in general, but that appears to be
(17:19):
the case. Once upon a time, when chat GPT first
splashed on the scene, the chatbot could only draw information
from before September twenty twenty one. It didn't have access
to the real time web. It could not crawl the
web for current information. So you could not ask it
about breaking news or anything like that because it literally
(17:42):
just could not access those information sources. Now, back in
July of this year, OpenAI created a subscription based tier
of service that would let users rely on chat GPT
through bing searches to access current information. But that feature
soon went away. Why because users figured out that they
could use this particular feature to bypass paywalls and get
(18:05):
to content that normally you would have to pay a
subscription to access. That's not exactly something that either open
ai or Microsoft wants to deal with, so the feature
went offline. But now it's come back. OpenAI announced yesterday
that through bing, it will let users access real time
information using the chatbot, and it calls this authoritative. Now.
(18:28):
I personally find that designation questionable given how chatbots like
chat GPT are prone to producing hallucinations or confabulations, which
just means sometimes they make stuff up. So I'm not
sure how authoritative you can actually claim to be. If
you've been following the union strikes in Hollywood, you likely
know that the Writer's Guild of America or WGA, has
(18:50):
reached a provisional agreement with the Alliance of Motion Pictures
and Television Producers aka the AMPTP. A lot of the
WGA's concerns relate to tech, ranging from how streaming companies
determine payouts like residuals, to the role of artificial intelligence
should play in film and television. The agreement, which WGA
(19:10):
members will have to ratify in a vote, essentially says
that the studios will not be able to compel writers
to use AI, that any content that does involve generative
AI has to be labeled as such, and that AI
is not eligible to get a writing credit. Now that's
important when it comes to things like payments and residuals
and so on, Considering the US courts have recently found
that AI generated material is not eligible for copyright, I
(19:33):
think these changes are ultimately of benefit to both the
writers and the producers. Meanwhile, the actor strike still goes on. Okay,
that's it for the news for today Thursday, September twenty eighth,
twenty twenty three. Next week, I'll be back in Atlanta
and episodes will return as normal. They'll probably sound more
(19:54):
like the older episodes and not weird because I'm recording
in some hotel room in Las Vegas. So I hope
you are all well, and I'll talk to you again
really soon. Tech Stuff is an iHeartRadio production. For more
podcasts from iHeartRadio, visit the iHeartRadio app, Apple Podcasts, or
(20:18):
wherever you listen to your favorite shows.
Welcome to Tech Stuff, a production from iHeartRadio. Hey thereon
Welcome to Tech Stuff. I'm your host, Jonathan Strickland. I'm
an executive producer with iHeartRadio. And how the tech are you.
I am still in Las Vegas, Nevada. You can probably
tell that my voice is going. I did a lot
(00:25):
of talking yesterday. I had to do some recordings of
the Restless Ones podcast, and I did a lot of
chatting with other folks, and yeah, I don't know if
I'm coming down with anything. I sure hope I'm not.
I've been masking every time I can, whenever I'm going
out in public, and I recommend that y'all who have
to go out in public do the same if you can,
(00:47):
because obviously COVID is having another surge, and I want
folks to be safe. And hopefully all I have is
just like wear and tear on the old vocal cords
and that's it. We'll obviously be very careful and keep checking.
But in the meantime, we have some tech news to
get to, so let's talk about that. Officials here in
(01:08):
the United States, as well as some in Japan, have
announced the discovery of a Chinese hacker attack that is
infiltrating organizations by compromising older routers. You can read up
on this in Ours Technica. There's an article by Dan Gooden.
The article is titled Backdoored firmware lets China State hackers
control routers with magic packets. The article goes into a
(01:31):
lot of technical detail. I'm not going to bother doing
that here because this is really just a news item.
But essentially this hacker group, which has lots of different names,
I'll just choose one of them, which is black Tech. First,
they have to get access to a system. Now how
they do that doesn't really matter. It's just they need
to get access. First, administrative access, like administrator level access
(01:54):
to be specific. They might get that through some compromise
login credentials, perhaps from another hacker group, maybe they use
social engineering, but whatever. First they get that access. With
that administrator level access to the system, they then push
an older version of firmware to certain routers, older routers
that don't have a protection against this kind of attack.
(02:16):
And typically they're targeting routers that are on the edge
of networks. So for example, they might target a branch
office of a larger company, partly because they're less likely
to be noticed, and that security won't be quite as
robust as it would be if it were at HQ. Now,
because these routers have a trusted relationship with all the
(02:39):
other computer systems that are connected to that company, the
hackers are then able to get widespread access to the
full organization. They target the edge routers, they infect those,
and then they snoop around Good and explains in his
piece that the more recent router hardware that has come
out over the last couple of years includes protections against
(03:01):
this type of attack, But obviously most organizations aren't updating
their routers regularly. They do so when they need to,
and determining if your system has been affected it seems
like it's not very easy or straightforward, particularly if the
hackers are actually being really careful. So it's another example
of how cyberwar is a real ongoing thing and that
(03:23):
state backed hackers continue to infiltrate both companies and public organizations.
As for the magic packets, those are little small packets
of data that the hackers are using essentially to open
or close backdoor access into these systems, and typically the
magic packets kind of blend in with overall network traffic
(03:44):
so that's kind of why they were called that. Back
in July, word got out that hackers had leveraged a
stolen Microsoft certificate to gain access to the email systems
belonging to really large organizations, including the United States State Department. Now,
the New York Times report that the hackers stole around
sixty thousand emails, primarily from ten email accounts. Further, nine
(04:07):
of those ten accounts belong to members of the State
Department who are working on projects that involve East Asian affairs. Now,
the State Department has not formally accused China of being
behind the attacks, but that seems to be the implication here.
The hack has sparked conversations within the US government on
how to harden systems against hacker attacks, including the possibility
(04:29):
of moving away from single vendor solutions like using Microsoft.
And while there's an element of closing the barn door
after the horse has already escaped going on here, I
do think putting all your trust in a single vendor
can lead to massive problems down the line. But honestly,
there aren't really any easy, fully reliable solutions here. Security
(04:51):
researchers revealed that all modern GPUs, that is graphics processing
units have their own way to compress data. But even
though they're each using a different method, that method can
be exploited to steal pixels from a website. Which sounds weird, right,
but essentially here's how it works. You get hackers and
they create a malicious website that has the same name
(05:14):
but a different domain as their target. So let's say
they've identified a target website that they don't have access
to directly, maybe it's behind some sort of protective measure,
So they create a different website has the same name
but a different domain, so instead of dot com, maybe
it's dot biz. Now in this malicious version of the website,
(05:34):
they include an iframe. The iframe serves as a spot
where you can embed other content into your website. So
one common use for iframes is to embed ads in them,
and that allows the site to swap ads out dynamically
and the iframe kind of holds its place. They then
rely on the GPU's data compression scheme to pull pixels
(05:55):
into a side channel, and they pull it from the
target website. You reconstruct what appears on your target site.
So you might have a corporate website that's usually behind
protection and the hackers create a malicious site named the
same thing, but on a different domain, and by stealing pixels,
they can recreate what was on that target website, potentially
(06:18):
gaining access to sensitive data, including things like you know,
log in credentials. The researchers were mostly showing that this
is possible. They didn't indicate that it's an active concern, Like,
they didn't say that this is something that they're seeing
actively deployed out there, but that in fact the hardware
across all GPU manufacturers allows for this, and since it
(06:40):
is possible, sooner or later someone will do it. Right,
So there are ways to block pixels stealing. Most of
those involve website administrators being proactive on the matter. They
can build stuff into the header of a web page
that will protect against that kind of thing, but that's
a lot of work. It is also probably, you know,
(07:01):
it's likely, I think that we're going to see companies
that make web browsers that facilitate this kind of attack
to address that in the future. The European Union issued
a report stating that after analyzing various social network platforms
out there, one in particular stands out when it comes
to the proliferation of disinformation. Do you have any guesses?
(07:23):
If you said X marks the spot, meaning the platform
formerly known as Twitter, you win a prize a virtual
donut enjoy Anyway, the EU previously established a voluntary code
of practice on disinformation, and tons of platforms signed to
(07:45):
pledge and that they would follow this voluntary code, so
that includes Google and YouTube, meta primarily Facebook and LinkedIn.
They all signed it, along with like forty other platforms.
It also used to include X, but at last May
Elon Musk decided that X was going to peace out
of that voluntary code. Now, researchers in the EU say
(08:07):
that disinformation is spreading like wildfire on X and that
the problem is only getting worse. Also, X has recently
disabled a feature that used to allow users to report
cases of misinformation about elections. So this means, assuming those
EU researchers are right that X is not just the
worst when it comes to preventing misinformation and disinformation from
(08:29):
propagating across their system, they are actively turning off features
that would do anything about it, which is a big
old yikes. An article in the LGBTQ Nation reveals that
TikTok is working with the government of Kenya to restrict
and remove LGBTQ plus content from the platform within Kenya's borders. Now.
(08:51):
This is because Kenya's government has banned same sex sexual relationships,
and currently the government is also considering an additional law
that would hunish homosexuality with a lifetime prison sentence. TikTok
has already deplatformed and demonetized TikTok users in Kenya who,
according to the country's laws, have posted restricted content. Apparently,
(09:13):
TikTok CEO has committed to not only being more active
to remove LGBTQ plus content from the platform within Kenya,
but to also launch a campaign to urge users to
instead post quote unquote positive content, which sounds disturbingly close
to conversion therapy. In my opinion, it's really ugly stuff
(09:34):
and my heart goes out to the people of Kenya.
This week, Meta announced an update to its line of
smart glasses, previously known as Stories, but to be really
formal about they are now known as ray Ban Meta
smart Glasses. The glasses have a pair of cameras in
them when at the top of each corner of the frames,
(09:55):
and they contain five microphones that are meant to provide
better sound pickup when you use your glasses, you know,
take a phone call or something. They also have speakers
that the company says will make it easy for the
wearer to hear what's playing on their glasses, but it won't,
you know, irritate the living heck out of everybody else
who happens to be nearby. The glasses can take about
five hundred photos at twelve megapixel resolution or around one
(10:19):
hundred and thirty second long video clips at ten ADP
before you end up using up all the storage on
the device and it's time to offload stuff. The battery
reportedly is good for like four to six hours of use,
and they come in a case that can also serve
as a charger, and the case itself can hold enough
battery juice to recharge the glasses several times before the
case itself needs to be recharged. You can also live
(10:42):
stream while wearing these glasses too. The controls include both
a touch system and the stems of the eyeglasses and
voice commands, and essentially these things do some of, but
not all, of the stuff that Google Glass did years ago,
but they do look a lot better. They do not
as far as I can tell, create any sort of
ar view of the world around you, which Google Glass
(11:04):
could sort of do. And already folks are worried about
privacy issues with these glasses. They're concerned that you could
be on video or have someone snapping pictures of you
without your knowledge or consent. Now I get that concern.
It's a legit concern, don't get me wrong, But y'all,
that's already happening. There are so many cameras out there,
whether they're in phones or in security systems like in
(11:24):
door bells or in cars, So I pretty much guarantee
that you're on camera frequently throughout an average day. But
I do understand how it gets creepy when you're talking
about someone wearing the camera on their face. Anyway, Meta
is taking pre orders on these things, which start at
around two hundred and ninety nine US dollars, and they'll
start shipping in mid October. So maybe your Halloween costume
(11:48):
can be invasive surveillance. That's a fun one. All right,
We're gonna take a quick break and we're gonna come
back with more news in just a moment. We're back.
Reddit indicated this week that it will remove the opt
(12:10):
out feature for personalized ads, or at least some Reddit
users around the world. So the new policy says that
Reddit will take your activity on the platform and then
use that to decide which ads to display to you.
So let's say you spend a lot of time on
subreddits that are dedicated to gaming, Well, then you're more
(12:30):
likely to see ads related to gaming while on Reddit. Right.
It says this is for your benefit, But there are
a lot of users who are upset because the message
that they're picking up on is Reddit is determined to
track their behavior across the site and there's no way
to tell Reddit to knock it off. And while Reddit
says users all around the world will be able to
opt out of personalized advertising based off of quote unquote
(12:54):
information and activity from our partners, only users in select
locations will actually be able to opt out of personalized
ads based off their activity on Reddit itself. I'm guessing
those select locations will be places like the European Union,
which has some pretty strict rules in place when it
comes to user privacy, and any place that doesn't have
(13:15):
those kind of rules, Redd, it's going to be tracking
you like crazy. And now, the latest in the long
arduous story of Microsoft's plan to acquire Activision Blizzard. When
last we left our tail, Microsoft was working hard to
convince regulators in the UK that the deal would not
result in an anti competitive situation within the world of
(13:37):
console gaming in general and cloud based streaming gaming in particular.
Well last week that regulatory agency gave provisional approval to
the acquisition, so that roadblock was removed. But now the
United States Federal Trade Commission has once again objected to
this deal. Now, they previously attempted to secure an injunction
(13:58):
against the deal closing, but a US District Court judge
denied that request and said that the FTC had failed
to produce evidence that this deal would actually be harmful
toward competition. The FTC then appealed that ruling, but then
withdrew the complaint not that long afterward, and it turns
(14:19):
out that was temporary because now they have submitted the
appeal again to the Ninth Circuit Court of Appeals. Once
the regulators in the UK gave their provisional approval, and
then once the court actually gives a decision one way
or the other, the FTC plans on an evidentiary hearing
on the matter. Now This on the surface sounds bad
(14:42):
for Microsoft and Activision Blizzard, but they could still go
through with their deal because there's no injunction against it,
and that deal is scheduled to close on October eighteenth.
Then once the deal is closed, they can worry about
any ongoing legal issues they might face after the fact,
but by then the deal will already be done. So
what I'm seeing is that most analysts think this deal
(15:04):
will finally close next month, and now for a few
AI stories, it's always going to be that. In a
tech Stuff News episode, an SEO consultant named Gagan Gotra
brought some concerning information about Google Bard conversations. So it
turns out if someone has a conversation with google Bard,
(15:27):
which I'm just going to remind you that's Google's AI
powered chatbot. It's similar to Open AI's chat GPT, specifically,
as it is integrated with bing, well, if you then
share a link of that conversation with someone else, then
Google will actually index that conversation, which means that conversation
(15:51):
with google Bard can pop up in future search results.
So let's say that you and a coworker are using
google Bard to help develop a business plan and this
is something that you do not want to share outside
of your organization. When you share that conversational link to
your coworker, the link itself becomes indexable, and then Google's
(16:13):
web craller will index the conversation, and if someone else
uses the right search query, that conversation can potentially pop
up in those search results, and the information that you
intended to be private has now been made public. Now,
obviously there are all sorts of situations where indexing the
barred conversation could turn out to be a really bad thing.
(16:33):
It's not that different from when open Ai had an
issue with chat GPT, where the chatbot would occasionally give
users access to other users chat histories with the chatbot. Now,
in the Google case, we're not talking about a bug,
we're talking about a feature. Because Gotra had an exchange
with Google research scientist Peter Lieu, and Peter pointed out
(16:58):
that the search engine will only end conversations if someone
has clicked on the share button, and Goscher's reply was
the same one I would have made. Namely, I think
most users would just assume that share just means you've
elected to share that conversation with someone in particular not
the whole world in general, but that appears to be
(17:19):
the case. Once upon a time, when chat GPT first
splashed on the scene, the chatbot could only draw information
from before September twenty twenty one. It didn't have access
to the real time web. It could not crawl the
web for current information. So you could not ask it
about breaking news or anything like that because it literally
(17:42):
just could not access those information sources. Now, back in
July of this year, OpenAI created a subscription based tier
of service that would let users rely on chat GPT
through bing searches to access current information. But that feature
soon went away. Why because users figured out that they
could use this particular feature to bypass paywalls and get
(18:05):
to content that normally you would have to pay a
subscription to access. That's not exactly something that either open
ai or Microsoft wants to deal with, so the feature
went offline. But now it's come back. OpenAI announced yesterday
that through bing, it will let users access real time
information using the chatbot, and it calls this authoritative. Now.
(18:28):
I personally find that designation questionable given how chatbots like
chat GPT are prone to producing hallucinations or confabulations, which
just means sometimes they make stuff up. So I'm not
sure how authoritative you can actually claim to be. If
you've been following the union strikes in Hollywood, you likely
know that the Writer's Guild of America or WGA, has
(18:50):
reached a provisional agreement with the Alliance of Motion Pictures
and Television Producers aka the AMPTP. A lot of the
WGA's concerns relate to tech, ranging from how streaming companies
determine payouts like residuals, to the role of artificial intelligence
should play in film and television. The agreement, which WGA
(19:10):
members will have to ratify in a vote, essentially says
that the studios will not be able to compel writers
to use AI, that any content that does involve generative
AI has to be labeled as such, and that AI
is not eligible to get a writing credit. Now that's
important when it comes to things like payments and residuals
and so on, Considering the US courts have recently found
that AI generated material is not eligible for copyright, I
(19:33):
think these changes are ultimately of benefit to both the
writers and the producers. Meanwhile, the actor strike still goes on. Okay,
that's it for the news for today Thursday, September twenty eighth,
twenty twenty three. Next week, I'll be back in Atlanta
and episodes will return as normal. They'll probably sound more
(19:54):
like the older episodes and not weird because I'm recording
in some hotel room in Las Vegas. So I hope
you are all well, and I'll talk to you again
really soon. Tech Stuff is an iHeartRadio production. For more
podcasts from iHeartRadio, visit the iHeartRadio app, Apple Podcasts, or
(20:18):
wherever you listen to your favorite shows.
TechStuff News
Hosts And Creators
Oz Woloshyn
Karah Preiss
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.