Video summary
Google is introducing a new feature designed to help news publishers regain visibility in search results and AI-generated responses by making it significantly easier for users to designate preferred sources. Previously, adding a publisher required visiting a specific Google tool, but the new interface allows users to click a button directly on a publisher's website while reading an article. This streamlined process ensures that trusted news outlets appear more frequently in top stories, Google Discover feeds, and AI overviews like Gemini responses. Additionally, Google is enhancing its discovery tools with natural language controls, allowing users to curate their content by asking for specific topics or requesting less coverage of certain subjects, while also improving the ability to link directly to full articles from audio briefings.
Despite these improvements, the hosts express a mix of optimism and skepticism regarding the feature's real-world impact, noting that many users have already moved away from traditional news sites toward social media platforms like Threads, Instagram, and Facebook for their daily information needs. While the ability to manually select preferred sources is valuable for tech-savvy individuals who actively manage their RSS feeds or specific settings, the average mobile user often prefers a passive experience where content is simply served up without manual intervention. The hosts worry that this update might be "too little, too late" to reverse the trend of news consumption shifting to social media, though they acknowledge that leveraging AI's natural language processing could eventually make it easier for non-technical users to maintain control over their information diet.
Beyond the main story, the episode covers a variety of other significant tech updates, including reviews of the new Pixel Watch 5 which offers improved health tracking features like Health Guardian and on-device AI models but still struggles with battery life compared to dedicated fitness trackers. Other headlines include Apple Music making AI-generated labels visible to users, Microsoft preparing to release an expensive 25th-anniversary Xbox Series X25, and Deepseek releasing a multimodal version of its model that rivals Anthropic's capabilities. The show also touches on regulatory and legal developments, such as Nevada approving permits for autonomous taxi services, Apple laying off staff from its VR division, and the ongoing trial of former Google engineer Linus Ding, who was convicted of stealing trade secrets but not espionage.
The broadcast concludes with a discussion on emerging technologies like satellite IoT for low-power devices in China and a prototype data center in Singapore powered by living human neurons, alongside a look at the shifting landscape of tech jobs where New York now hosts more workers than San Francisco. The hosts also address listener questions about the feasibility of facial recognition in smart glasses, citing studies showing that high-accuracy identification is possible even with very low-resolution images, which underscores the growing privacy concerns surrounding ubiquitous surveillance technology. Ultimately, the episode emphasizes the need to focus AI's strengths on specific tasks rather than trying to use it as a "golden hammer" for every problem, while encouraging listeners to stay informed about these rapid changes in the tech ecosystem.
Read the full video transcript
[music]
This is the daily tech news for Friday,
August 21st, 2026. We tell you what you
need to know, give you important
context, and try to help each other
understand.
>> Today, Google makes it easier for news
publishers to get you to add them to
preferred sources. This will make them
show up higher in search and more often
in AI overviews.
>> A little bone. Little bone. Maybe a big
bone. I don't know. Well, we'll talk
about it. Uh, but it is being thrown at
the publishers. I'm Tom Barrett.
>> I worked it down.
>> Let's start with what you need to know
with that big story.
>> Yeah. So, Google launched an easier way
to add preferred sources to your search,
Google News, or that Google discover
feed on mobile. Uh, if you didn't
remember, preferred sources launched in
May. I had kind of forgot and had to go
back and look, did I add any preferred
sources? Turned out I added eight of
them back in May and totally forgot I
did it. So, hey, here's a reminder. Uh,
you might have some preferred sources
settings. Uh, the preferred sources
button is the new thing, though. This
will show up if you're on the publishers
website. So, you're reading a news
website. Uh, and you'll see that
preferred sources button and you'll say,
"Gosh, I really like this news website.
I would like it to be one of my
preferred sources on Google." And then
you click that button and it will add it
to your preferred sources list in your
Google account. Right now, obviously you
have to be logged into Google, so if
you're using Chrome, that's going to be
easier. Um, but it is meant to make it
easier. Publishers had a badge before
that they could use, but it would send
you to Google's source preferences tool
and then you had to add the publisher
yourself. So, this just makes it a lot
faster and easier uh and let you stay on
the publishers's website. It's not going
to kick you out to anything. You just
press that button, keep reading your
story. uh preferred sources show up more
often in top stories and AI generated
responses. I think that's the other
thing to emphasize is when you are
getting a Gemini response, an AI
overview or even in in in just the the
the AI tab on search, your preferred
sources will be used more often if you
have them. In addition, uh Google's also
adding a feature to Google Discover that
when you select the three dot menu, or
sometimes it'll show you a chat prompt.
Uh you can tell it in natural language,
hey, I want to see more uh uh K-pop
topics, or I'd like to see less about uh
enraged people, and it'll try to show
you more or less of that. You can also
curate your daily audio briefing in the
Google News app if you use that. uh you
can tell it to choose specific topics
and it will also now link out to full
articles more often. Uh the big one that
the industry is focusing on is the
preferred sources because that is a boon
to publishers. I don't know how many
people are going to use this, but I
guess enough people can use it that it
might make a little bit of a difference.
I that's what I was kind of curious and
you mentioned throwing the publishers a
bone and of course given that you know
AI summarization AI overviews and
everything like that are I I I assume I
don't know the numbers on this
disrupting you know kind of like the
traditional relationship between
>> definitely disrupting
>> definitely disrupting um I I was
actually as you were speaking
introspecting on how I use these things
I actually got a new phone recently and
I elected to not because it's a Samsung
I elected to not put Google Discover in
that like lefth hand side so I even more
removed myself from that just because I
and and I I'm doing that for a very
specific reason. I'm doing that because
I'm a little bit tired of the thing that
to be honest this the preferred
preferred sources is supposed to be
helping me with and that is to get more
results of things that I actually trust
and and care about and less of the what
I perceive to be not necessarily slot
but just things I really didn't care
about and don't need in my eyeballs in
the morning. So I I don't know whether
that makes me an outlier or whether that
makes me maybe part of the problem.
But I do agree with you that um as and
there's I don't think there's anything
inherently wrong with this feature. This
is a great feature to be honest with you
like giving people choice and allowing
people to be specific about what they
want to use. But that's kind of always
the problem with features like this is
that there's a lot of good intention
there. Discoverability and even just the
entropy of users to use something which
normally is like a very helpful feature.
there's like not really a downside. I I
guess one could say nominally to using
this, but a lot of it is I get entropy
and discoverability. Are we Is this like
too little too late? I don't know. But
it I would I would actually be a little
pessimistic about whether this kind of
brings people back to, you know, viewing
their prefer even their preferred
sources on these different surfaces
since we're all kind of been inundated
with this stuff for a while. I don't
know. What do you think? Would you
>> So you're you're on the small bone side.
This is
>> small bone. Tiny bone like like one of
the bones in your ear kind of like
>> Yeah.
>> It's [clears throat] like a hammer
anvil.
>> Anvil and a sturrup. Yeah.
>> Yeah. [laughter] Uh okay. I I think it
might be a little bigger than that, but
I'm still going no bigger than a neck
bone. Like this is this is not a huge
bone. Um
>> so not a femur. Not not a
>> Yeah. This is the kind of feature that
everyone says they want.
>> Yeah.
>> And 10% of you actually use.
>> Yes. No. Yeah. Like just give me
control. And then the Allison Sheridans
of the world actually use that control
and the 90% of other people out there
are like, "Oh, right. I forgot. Oh, but
that's so much work." These sorts of
things work better when you just do it
automatically. And that's the thing that
I I come to this and I'm like, "Okay, I
guess I should do this." And then I
looked and I'm like, "Oh, I already did
it. Okay, good. Good on me. Do I need to
update it? Do I need to add anything to
it?" But I had forgotten to maintain it
right now. This publisher button is a
good way to maintain it, but how often
do you actually go to a publishers's
website and read an article?
>> Like not never,
>> but rarely, right?
>> Like a lot of times you're reading it
somewhere else. Maybe it's Apple News,
maybe it's that discover section. Uh you
may be reading it in a way that gives
revenue to the source, but you're not on
the publishers's website. So making sure
that this button shows up every time you
read an article, maybe it's in an RSS
reader and stuff would go farther. I
don't think they're going to go that
extent though. Um, so yeah, small bone.
I'm with you.
>> Small bone. Small bone. Not quite as
small as not quite as small as a stirup.
>> My bone might be a little bigger than
yours. There's
I regretted that as soon as it came out
of my mouth. The size of the Never mind.
Uh, yes. I'm a slightly more optimistic
but still pessimistic, put it that way.
>> Uh I mean is there is there any in your
opinion especially someone who you know
I I know that you know how many people
are still using this is not an insult
like do like how do people like do we
know how people are really consuming
news because I know that you use an RSS
reader but that's kind of almost like
part of your day job and part of your
habit right for all these doing
reporting. I started to use an RSS
reader when I started uh when I subbed
in for writing this show. And I do find
it, oh wow, this is kind of cool. Like
if I'm like very forward and and I
really do care about my sources, this is
a great way to do it. But also, I'm a
I'm a lazy mobile user. I just want it
kind of served up to my eyeballs. Do we
have like a any sense at all about I
don't know how greatly people's
newsreading habits and their news
reading mechanisms have changed from say
like even like 10 years ago, 5 years
ago.
>> Yeah. Everything I see is that it's
social media. People get their sadly get
their news from at least it's not just
Twitter anymore. Uh you know it's some
of it's still X but a lot of it's
threads, Instagram, Facebook.
>> Uh that's where people see news.
>> Unless you're in a country where
Facebook has nerfed news. Uh you know,
and then you're getting it from some
other social media platform probably.
>> Does this feel like polishing the
silverware on the Titanic or is this do
you think this is going to change? Does
this have any momentum? I guess this is
this is maybe back to like the relative
size of our optimisms uh optimism or
never mind optimism
optimism buckets whatever this anyway
does
>> bone density
>> bone density there we go [laughter]
does this do you do you feel like do you
have do you feel optimistic about this
or do you think like like again is this
>> I am very pessimistic uh on where we are
with news I am optimistic that we will
eventually come out of this cycle but I
don't know when I don't think that this
moves the needle much if to make I'm
getting away from bone metaphors just
for me.
>> Yeah. Yeah. Um, I will say just to be
positive, I do think that one of the
powers of gener of of Gen AI LLM in
particular is that natural language
processing. And I would say that I do
think for maybe non professionals or non
non-professionals that using natural
language to say hey I don't want to see
this and I don't want to see that is
promising and that is an [clears throat]
improvement over I would only like to
see you know X Y and Z preferred surface
preferred sources because I assume that
people that aren't tech
>> enthusiasts like we are are probably
more in that gray. So I will say in in
in in a positive light that I would like
to see more of that and leveraging that
natural language to make it easier for
people and maybe once we make it easy
enough they will actually like if we
make the the the level effort low enough
people will actually do it.
>> It has it has to be low or no effort uh
to improve things for you. And honestly
some of my optimism does lie in the
LLM's uh being able to sort through and
say like this is a more reliable source.
Let me go with that. They didn't used to
be good at that at all. Uh, and I' I've
seen that they have gotten better. And
not just on Gemini, but but chat GPT as
well.
>> Uh, one thing I said today when I was
researching and I got an AI overview was
like, uh, can you give me a source that
isn't a YouTube video? And it was like,
oh, uh, sure. Yeah, here you go. That
that's an action I would take that most
people would not take, though. Uh, so
there's, you know, the mixed results,
but I I think it's actually better than
than search engine optimized gamed
results that we've had in the past.
>> That's a really good point, though, is
is maybe also exercising that muscle
because I should do that. I when you
said that just now, I feel like, well,
I'm a techsavvy person. I never thought
to do that. So, I should Yeah, maybe
just again exercising that muscle of we
can kind of go back and just tell it our
choices rather than doing it manually.
So,
>> yeah, because I I have found LLMs to be
useful in finding sources. I can get to
a reliable source faster. I'm not
relying on it
>> to be the source. I'm saying I need a
source that says this and it's like, oh,
here's the Department of Justice page on
this. I'm like, exactly what I want,
right? The actual primary source, not
somebody's clickbait post that doesn't
even mention the primary source.
>> This is one of the things I think is so
interesting about generative AI in
particular because it's so good at
certain things and I know that everyone
I know the temptation is there to use it
in everything. I think that it's a term
called a AI washing. I understand. But,
you know, like if we could just refocus
refocus it on the things it is really
good good at but then leave room for you
know not destroying that's very dramatic
but yeah like not disrupting you know
publisher relationships and and kind of
like website like kind of the website
economy. I would be like I think 30 to
40% more positive on it like and and
like because there are dollops of really
good stuff in here. It's just like okay
we just focus focus our powers for good
>> and then you can have like I don't know.
I'm about to go back to the weird
metaphors, but yeah. No, I like that.
>> Yeah, it's it's just get narrowing down.
And I think we see signs that that is
happening where instead of being a do
everything. Anytime anything's a do
everything, it's not it's not at its
best. That was true of social media
platforms. They're actually better now
because they've been broken up into
different ones than they were when it
was like one platform for everything.
And that's true of LLMs. Uh when they
are tailored, when you get them like,
oh, they are focused and they're good on
this, they always perform better.
>> Yeah. Well, tech people like them a
golden hammer. Uh, but uh, you know, not
everything is a nail. So, there we go.
>> Well, we are made possible by you, the
listener, this year right here, DTNS. I
don't even know what to call it. There
was like a golden hammer metaphor there.
But, we want to be your news for
everything. Hopefully, we're not a
golden hammer. So, hopefully we're just
a golden I don't know.
>> We're just golden.
>> We're just golden. There you go. Got to
be. There you go. We cut it back to the
K-pop. I love that. Um, and you know, I
would also love to thank or we would
love to thank Ryan Marx, Michelle Serju,
Kirk Stephenson, and Miranda Janelle for
being our supporters.
>> There is more to talk about today. Let's
get to the briefs.
>> All right. Well, the Pixel Watch 5 is on
sale now and the reviews are out today
as well. So, here are just a few of the
common things that reviewers praise
about the Pixel Watch 5. it being
faster, health and wellness features
like health guardian, strength training
is now useful, accurate GPS and sleep
tracking, offline Gemini ondevice
models, double the storage as before, so
more room for your watches and uh and
other things like that or not watches,
your watch faces as well as a faster
charging battery. And with regards to
the cons, watch faces and I think in
particular things like you know AI
generated watch faces and like that and
also just like the the availability
battery life is still two days which is
kind of pedestrian compared to some uh
fitness trackers out there. Active band
the band that comes in the box with your
Pixel Watch 5 is a bit boring and has
been the same band for several
generations now. Too many settings just
lots of different toggles to toggle as
well as a unsurprising but still sad
price rise of $50. The Pixel Watch 5
costs $399, sorry, $399.99
for the 41 millimeter Bluetooth Wi-Fi
model and $429.99 for the 45 mm version.
>> I have to say from the outside looking
in, uh, the cons here are not terribly
damning. Like, didn't improve the
battery life much. You know, that
included active band is kind of boring.
Uh, yeah. Uh, the
>> too many settings, but
>> yeah, your watch faces are slop. Uh
yeah. Uh it could be worse. Like it does
seem like the reviews of the Pixel Watch
5 are are fairly favorable. People seem
to like it.
>> Yeah. I And I agree. Like there there
are other things too. Like I just like
everything else with this current
generation of Pixel devices, whether
that's the phone or not. It's a good
iterative. It's nothing. It's It's
generally just nothing but improvements.
I would say like yeah, nothing the
battery life didn't get worse. You know,
the bands didn't get worse. It's just
there's not a lot of improvements. I
will say the active band is I mean I
think the fact that the active band is
is quote boring is is not a is I I guess
that's kind of fair. Um especially since
the if you're not aware the active band
is this like solid silicone thing. So if
you tend to have skin issues or stuff
like that or you like to work out with
it and you want something more
breathable that is kind of a pita. But I
I I it just a lot of things seem great
you know. Um and and even like I'd say
even some of the good things are just
kind of almost unabashly good things. Of
course, we can't get away from Gemini
and you know AI and all of our things.
So having ondevice models is great like
and I think they did a good job. Um and
like people like to one thing that
people do like to do is attack battery
life on these devices. I So for do you
have an Apple Watch?
>> Yeah, I do. I the 11. Yeah.
>> What is your charge time? Like what is
um what is your battery
>> charge? Oh battery life I can go a day
and a half.
>> Okay. A day and a half. Um, and that
seems so reasonable given that our, you
know, these watches are, sorry, these
are not watches. These are fitness
trackers. I got two of them in this
hand. I got a Pixel watch in this.
[laughter] Um, it makes a lot of sense,
right? It's it's so funny to me still
because the negative with the battery
life still being at two days is is still
comparing it to again fitness trackers
like a Garmin, like you know, these kind
of bands or even like to be fair like a
tick watch uh something like that where
the functionality and the usage is quite
different. you know, your Apple Watch,
my Pixel Watch is not just a health
tracker. It's a bunch of other things,
including having Gemini on it. So, I
still find it interesting that maybe one
of the biggest cons is like, oh, well,
this watch that does a billion and a
half things and runs, you know, an LLM
on here or large language model on here
still only does two days of battery
life, which in my opinion is still
pretty good, uh, given the fact that
they've done the the fast charging easy.
So, yeah, yeah. So, I I got one. if you
can fast charge it up during a shower,
you know, or or or something like
because I always charge mine when I'm
asleep, but I don't use mine for sleep
tracking. So, obviously, if you're using
it for sleep tracking, you don't want to
charge it while you're asleep. Uh so,
you need a short period in the day uh to
charge things up. So, fast charging, I
think, is an is an important thing here.
The Apple Watch does not charge very
fast in my experience. So,
>> yeah, I I have no complaints. I think
she think they closed like I think even
the Pixel Watch 3 they had already
closed the loop on in my opinion on the
charging story because within like a
shower's length like say 10 to 20
minutes you have at least a whole day.
>> In fact, I'm the kind of person who's a
little bit uh not battery anxiety on
running it out but battery health
anxiety where I try not to charge this
up to past 90%. And if anything it
charges too fast I end up at 100% when
I'm trying to save the battery. So I I
will say like and and that's kind of the
thing is like coming from again like
more of a Garmin, more of a tick watch,
you're you're like I don't have to
charge it for two weeks.
>> It kind kind of re refraraming it as
like okay this is a this is like a smart
device not just a health tracker. So
yeah um if we can just change our habits
>> it's more capable both these watches are
more capable than the original iPhone or
the the original Nexus. like, [laughter]
you know, like it's crazy to think
you're just wearing that you can wear a
large language model on on device on
your wrist.
>> Yeah,
>> that's that's kind of insane. Um, I will
say just just just because I'm the
fitness weirdo here that Yeah. Um, if
you're curious, Health Guardian is kind
of a new thing. And I I do find this
really interesting because we've seen a
lot of, you know, of these health
wearables, including like those by
Samsung, like the Galaxy Watch and
stuff. Try to like get into that more,
not just like fitness health, but actual
general health, of course, like, you
know, digital well-being or like kind of
overall well-being is like a big theme
for both like Samsung, Apple, and and
Google. And so, HealthGuardian is kind
of along the lines of like, you know,
how people want to do glucose tracking
and like blood pressure monitoring. And
it kind of gets into a weird space with
those specific things. And I I looked it
up. HealthGardian basically tries to
correlate your sleep uh and and kind of
like your activity and other things like
that into a metric that can correlate to
your metabolic health, which then ties
into glucose. What I do think is
fascinating is that there is still some
interesting legal loopholes or regional
like like laws around, you know, saying
something is a blood pressure monitor or
not. And that's like I think does Apple
Watch have the blood pressure monitoring
yet or is it
>> Yes, it does. But but it's an estimate,
right? It's not going to give you your
120 over 70, right?
>> Yeah. Um and so similarly,
HealthGuardian is a good feature to kind
of augment, you know, any treatments or
kind of like, you know, uh like I guess
medical advice that you're getting from
your uh primary care physician in
regards to any kind of uh metabolic
conditions that you have like diabetes,
but there's always this asterisk of like
it's not a direct measurement, which is
how they get around the laws of, you
know, uh kind of like get around the
very specific laws around like um
medical like medical grade monitoring.
Um so just to say that they're really
great features. always consult your
primary care physician when using
strength training, when using Health
Guardian. It's a really good augment.
Just make sure you also, you know, close
the loop and and see your doctor as
well. So,
>> I mean, the the reviews were universally
praising HealthGuardian. Uh, but that
that is a good caveat of like this won't
replace accurate medical tools if that's
what you need, right? This is just kind
of giving you a generalized indication
if you're somebody who doesn't need to
use those accurate tools to let you know
like, oh, you're headed this direction
or that direction. Basically,
>> absolutely.
>> Yeah. Uh, ChatGpt's latest update to its
Mac app adds a plugin for Apple
Messages, which is usable on all the
plans. Uh, it can read and search your
messages if you enable the integration
and can send messages, but by default,
it requires user approval before each
send. OpenAI notes that there are risks
to disabling this approval, though you
can, and you can even disable it on a
per conversation basis. say, "You know
what? I'm okay here, but then once we
start a new conversation, come and ask
me again." Uh, it also doesn't let you
control your agent remotely through
Apple messages, which I think a lot of
people were hoping they'd be able to do,
like, oh, I can just send a text message
to my agent. Uh, because you can do
stuff like that with OpenClaw. Uh, it
doesn't work in regular chat GPT chats
either. Uh, it only works in the chat
GPT app chats uh on the Mac, at least
for now.
Privacy-wise, it runs locally on your
machine. also uh and it doesn't create
an index of all someone's messages uh
which you know I would hope it doesn't
but I'm glad they specified that that
was the case.
>> Okay. As a green bubble is this
something that you blue bubbles are
excited about? [laughter]
>> Sorry.
>> Um let me put my my pixel filled away
and pick up my iPhone. [laughter] Um
>> I guess I think yes. Uh if I am on a Mac
and I have Apple messages uh and that is
something where I'm I'm I'm wanting an
agent to be like oh and can you tell
when this right I could send you a text
message through iMes uh it will be green
not blue but it will go through because
it can do SMS and it can do RCS as well.
Um so yeah you can you can set this up
on your Mac. Uh and yes you do have to
have access to Apple messages. So, if
you have a Mac and you use just an
Android phone, then you're not going to
be able to do that, right?
>> Um I I don't use agents that way. So,
personally, I'm I'm not that excited,
but I can see where some people would be
like, "Oh, this is going to be a great
help because I do all of my coordination
over text message." And I and my agents
can now just send notes on my behalf,
which is fantastic.
>> I'm sorry. I miss I I miss mobile over
operating system. Do you? Uh
>> I'm by operating system. [laughter]
I call it bios.
>> Oh, I see what you did there. [laughter]
That's the best. Okay.
>> Uh, yeah. So, I Yeah, this is I I think
this is uh one of those things that
people are going to be scared about. I
think OpenAI did a good job of saying
no, we put in the controls and also we
are actively telling you don't take off
the controls unless you're really sure
you want to assume the risk.
>> Yeah.
>> For that.
>> No complaints there.
>> Now, when if you're like me, you're more
than half water.
>> Yes. [laughter] But that water keeps
evaporating every time you breathe. And
stopping breathing can lead to more
serious problems. So instead, if
stopping breathing, we have a solution.
The DTNS water bottle. It replenishes
that water. You just carry the water
bottle around, drink the water, and you
can keep breathing and still be your
full self. 55 to 60% water. [laughter]
Get a DTNS water bottle, folks.
Dailytechneow.com/store.
Do it for your health.
All right, let's get to some quick
headlines that are just good to know.
Might make you look smarter in the
future.
>> All right. Well, Apple Music sent an
email to music industry partners saying
that later this year it will make AI
labels on songs visible to users. Apple
earlier announced it requires content
providers to disclose when AI is used in
a quote material way.
>> Yeah. So, this is telling the labels
like, "Remember when we asked you to do
that thing? People are going to be able
to see that you did that thing. We're
just letting you know." Uh, Microsoft
will release a 25th anniversary Xbox
this autumn called the Xbox Series X25.
They announced that, I think, back in
June. And now, Deal Labs Bill Coun says
that their sources say that the price is
going to be 900 bucks. Now, Microsoft
hasn't announced that, but Bill Kun is
pretty reliable with these sorts of
things. Uh he also thinks it's going to
launch on November 17th. The day or I'm
sorry, 27th. 27th. Yeah, 27th. The day
after US Thanksgiving, whatever day that
is. So, Black Friday. It's going to la
he's the or Bill Kun thinks it's going
to launch on Black Friday.
>> It is not going to be a cheap device. Uh
well, deep.
>> It's $100 more than the current one
though, right? Like, isn't that wild to
think like we are now at an $800 Xbox
Series X? I suppose it's been a while
since I had an Xbox. So, yes, you know,
my my expectations are a little bit
>> your expectations. Oh, my Xbox
expectations. All right. Well, Deepseek
announced the release of an experimental
version of its flagship textonly V4
flash mode that can accept visual
prompts. Deepseek says it it says it
Deepseek says it is close to as capable
as Anthropic's Opus 4.8 eight mode on
multimodal agentic capability. I
>> you know what I I can't help it but I
read the story as deepseek still behind
anthropic [laughter] like you know
deepseat's very capable and it's very
very it's not as cheap as it used to be
to run uh but it's it's still not at
par. Uh the US state of Nevada's
transportation authority has approved
three permits Thursday to let Tesla,
Uber, and Whimo charge people for
autonomous taxi rides. So, commercial
service in Clark County, Nevada, aka the
home county of Las Vegas. Permits allow
for up to 8,000 autonomous taxis to be
deployed over the next year. And Uber
works with Zuks, which already has a
permit. So, that's adding to the Zuks uh
thing that you'll see out there, too.
>> There's got to be like a Pixar's Cars
like Las Vegas, you know, autonomous
taxi edition. Maybe that's just
something that only I want, but
>> Oh my gosh. Yes. A new Cars character
would be amazing. I love that. Well,
Apple Insider reports that Apple has
laid off about 60 people from its VR
development division, indicating a
further shift from away from the Apple
Vision Pro and towards smart glasses.
>> The government of China has issued a
permit to J's Gpace to offer satellite
internet of things service for small
transport, marine fisheries, energy and
water infrastructure, stuff like that.
Satellite IoT uses low power connections
for a large number of devices rather
than the high-speed wide coverage you
need for satellite internet. So this is
less Starlink and more like let's let's
just make sure we get coverage for these
these low power devices. Uh and a small
but but growing
industry aka buzzword. I think we're
going to see more satellite IoT stories
over the next year or so.
>> Really cool applications though with the
small transport the fisheries and the
internet like the infrastructure where
you don't need that high speed. That's
actually cool. Exactly.
>> All right. Well, US District Court Judge
Vince Chabria in San Francisco ruled
Thursday that there was not enough
evidence that former Google engineer
Linuay Ding intended or knew his conduct
would benefit the government of China.
So he so the judge did not convict him
of espionage. He was however found
guilty of theft of trade secrets.
>> So [laughter] So yeah, it's like it's
not like he got off scot-free, but he's
like, "Oh yeah, you you're not a spy.
[laughter]
you just you're just a thief.
So, [gasps] uh 22year-old researcher who
goes by zeroistic devised a way to
enroll a Linux-based machine into
Apple's Find My Network. Now, this
doesn't get you other people's devices.
It just let him add a non-Apple device
to his own Find My list, but still
pretty cool. And register has all the
details about it if you want to read it.
>> Cool. Well, N US Medicine data center
operator day1 and Melbourne's Cortical
Labs have announced they are running a
data data center prototype in Singapore
built from a 20 unit rack of
neuronpowered CL1 computers. Yep, that
is 20 computers running on living human
neurons. No benchmarks or anything, but
it is running.
>> What could go wrong? [laughter]
>> [gasps]
>> No, that that is a a significant
deployment of neuronal storage, you
know, and individual neurons are not
aware. So, it's fine. It's fine.
>> It's fine for now.
>> Uh, US company Robo Store can no longer
distribute humanoid robots from China
after a US ban on doing so. So, it's
decided to make its own. Uh, it's going
to start making humanoid robots from a
facility in Long Island, New York. And
I'm sure no one will ever accuse it of
having taken any of the intellectual
property of the humanoid robots it used
to distribute.
>> Well, attention Walmart shoppers.
Walmart now accepts Apple Pay and Google
Pay at its Walmart and Sam's Club
locations.
>> Oh my gosh. Cats, the dogs sleeping
together. Massive. [laughter]
uh this is going to significantly
increase the number of mobile payment
users especially in those older demos
and I'm not not trying to be snide when
I say that it it's like 89% amongst
younger demos but it's I think in 65
plus it's less than 30%. Uh and this
this is going to boost it because it's
going to make it much more widespread.
>> Absolutely.
>> Put your blazers back on tech workers. A
report from CBRE found there are now
394,300
tech jobs in New York compared to
375,730
in the San Francisco Bay area. If you
weren't paying attention, 394 is bigger
than 375. That makes New York home to
more tech workers than San Francisco for
the first time since they started
tracking this.
>> Wow, that is impressive. And I don't
think I've unfortunately had a Blazer in
many, many years. I would be in trouble.
Um well, Brazil's National Fund for
Scientific and Technological Development
has split a supercomp supercomputer
project into two, building half with
Huawei and the other half with Nvidia in
order to reduce reliance on either China
or the US.
>> Yes, this is the way most of the world
lives. Like, oh [laughter] god. Okay,
fine. You know, Brazil's genius. Like,
we'll just do half one, half the other.
That Yeah, that way, you know, we're in
trouble with both.
>> [laughter]
>> Uh the US now accounts for 15 of the
world's 20 largest hypers scale data
center markets and 12
of global capacity comes out of Northern
Virginia alone uh where a lot of data
centers are. So uh you know if if the
world is going to run on data centers
the US is kind of cornering the market
on that.
>> All right well we end every episode of
DTNS with some shared perspectives.
Today, Peter has a question about using
smart glasses for facial recognition.
>> Yes, thank you, Peter, for writing in.
Uh, he says, "Good day, DTNS group." A
followup about the email from the
professor from show 5335.
That was two days ago. Uh, I could see
allowing the glasses to tag a person's
face you record with the glasses with a
name, which would help the professor or
others with facial recognition
difficulties. But where would this
database of faces come from that is
causing all the outcry? Even if Meta
used all the public Facebook, Instagram,
and WhatsApp public profile pictures, a
large number wouldn't be a high enough
resolution for a facial match. It isn't
like the madeup facial recognition tech
that is in a lot of TV shows and movies.
Peter from just outside Toronto, Canada.
Uh, I hate to say this, Peter, but it is
good enough. Uh, I looked up uh just
just one example. There's a company
called Briefcam uh that can do pretty
high accuracy facial recognition with 40
by 40 pixel resolution. Uh and your
average Facebook photo is probably
around 300 uh by 300 maybe a little
more. Um there's also that study from
2024 where a couple of Harvard students
demonstrated a system using info from
PIMIS which is a thirdparty database uh
combined with an LLM and some metaray
bands and they could just identify
people at a train station and in their
demo they were not only identifying the
person but using the LLM to pull up
other data about them so they could have
a conversation as if they knew the
person. Um, so it was a very
disconcerting demonstration that they
were showing what's possible uh with the
amount of information that's out there
on the internet about you. But I I think
I think that database is good enough uh
you know and and so we can have a debate
over how you can access the that
database and what the rules should be.
But I I think the technology is there.
>> Um yeah, 40 by 40 pixels is is nothing.
Like that's that's like if you're audio
user and you can't see but I mean
literally that's that's kind of um
remarkable in a terrifying way. It's a
very small bone.
>> Yeah. [laughter]
Not quite as small as your syrup, your
hammer, your anvil, I don't think, but
pretty darn close. And I mean, I guess
this is just one of those things where
it feels like, you know, we've how long
has it been now? like a couple decades
now of of you know like Facebook,
Instagram, of social media being being
intact and there just like I I know we
all know this and I know we're all very
cognizant of it but at some point all of
this accumulated data I think I think
the I I totally get the the feeling
Peter of like you know we watch we see
things on on on in fictional things in
TVs and movies and it still feels very
very far out but um eventually we do
catch up to these things unfortunately.
Um,
>> I used to make fun of Zoom and enhance
>> and now you can literally enhance like
>> those were fun days when we could we can
just use that meme unabashardly and it
was fun, but now it's like, oh, so you
you did it.
>> You actually can. Oh, okay.
>> Okay.
>> Yeah.
>> All right. Well, what are you thinking
about? Do you have a question? Do you
have some insight? We would love to hear
it. Feedback atailytechnewshow.com.
>> Yeah. And thank you, Peter, for
contributing to today's show. Thank
everyone listening for being along for
Daily Techne Show. You're the folks who
keep us in business. Become a patron if
you haven't already. Uh we've got great
perks. We're going to be doing one of
them later today called the producers
room. Little behind the scenes. That and
more available at patreon.com/dtns.
[music]
This week's episodes of Daily Tech News
Show were created by the following
people. Host producer and writer Tom
Merritt. Host and writer Jason Hal.
Co-host Sarah Lane. Co-host Rob Dunwood.
Co-host Jen Cutter. Co-host Wento Dao.
Producer Anthony Lamos. Producer Roger
Chang, editor Hammond Chamberlain,
editor Victor Bognot, contributing
producers Kevin Tech, Noel Cow, and
Brandon Richards, science correspondent
Dr. Nikki Acrimans, social media
producer and moderator Zoe Deding. Our
mods, Beatmaster W. Scottiswan, Biocow,
Captain Kipper, Steve, Paul Reese,
Matthew J. Stevens, and JD Galloway. Mod
and video hosting from Dan Christensen.
Music provided by Martin Bell and Dan
Lutters. Art by Len Peralta. Acast ad
support from Alli Gallow, Patreon
support from Bobby [music] Wagner, and
our guest this week was science
communicator Bobby Frankenberger. Thanks
to all our patrons who make the show
possible.
[music]
The DTNS family of podcasts
helping each other understand.
>> Diamond Club hopes you have enjoyed this
broer. [laughter]