December 3, 2025 • 39 mins
Hour 1 of the show begins with Jon talking about the recent Tennessee elections. Then AI expert Daveed Gartenstein Ross joins the show to talk about AI, including why LLMs can sometime struggle with basic tasks.
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 (00:07):
And it's only the middle of the week, so much
has happened.
Speaker 2 (00:15):
We have a lot to talk about this morning on
Twin Cities News Talk Am eleven thirty one oh three
five FM. My name is John Justice, and I'm glad
that you're with the show this morning as we embark
on our broadcast from the six to five to one
carpet plus Next Day Install studios pushing the buttons in
(00:35):
the master control booth this morning.
Speaker 1 (00:37):
Is Devin. Nice to see him.
Speaker 2 (00:39):
Got a lot of guests on the show today, But
let's run down all of the major news items as
we dive in, dive into everything that needs to be
covered relating to the fact that Minnesota has become the
epicenter of the news cycle right now, much to the
chagrin of Tim Walls. Couldn't have happened to him, dude,
(01:00):
all right, So on the show today, as I had
been hearing, it has now become a reality. ICE is
launching extensive immigration raids here in Minnesota. We have a
lot of audio to share of local law enforcements, Minneapolis
man Baby mayors commenting their concern about ICE coming in
(01:23):
and doing their job. I mean, consider that for just
a moment, you have law enforcement criticizing that law enforcement
is going to come in and conduct law enforcement in it.
Make it make sense small business administration orders an investigation
into further COVID fraud in Minnesota.
Speaker 1 (01:42):
This is coming from the Trump administration.
Speaker 2 (01:44):
Walls does respond to all the latest of plaguing his administration,
all of the controversies, and we'll talk about that defeat
gartzan Stein. Ross will get into some AI conversations at
six point thirty. You can get your comments in early
via the iHeartRadio app Talkbacks. Representative Tom Emer will join
us at seven twenty this morning. Liz Colman with a
(02:07):
lot to cover relating to news at alphanews dot org.
That'll be at eight thirty as well. The Department of
Justice has also officially sued the state of Minnesota to
obtain the entire statewide voter roles, including all illegal aliens
living in Minnesota who may have been fraudulently registered to vote.
(02:33):
This is one of the things we're actually going to
talk about with David regarding AI because the potential of
AI actually going and speeding up these investigations. So we
got a lot of ground to cover today. Let's start
here though the election overnight that everybody was so worried about.
Speaker 1 (02:48):
Yeah, it wasn't that big of a concern.
Speaker 2 (02:53):
Republican Matt Van Epps fended off a challenge from the
far left whack a Doodle aften Bean on Tuesday in
the special election for the Tennessee seventh Congressional District. Ninety
five percent of the vote in the race was called
for Epps fifty four to forty five. This was a
nine point victory. And if you remember the commentary heading
(03:17):
into it, polls, oh, it's within two points, it's within
striking distance.
Speaker 3 (03:22):
It wasn't.
Speaker 4 (03:23):
It was not.
Speaker 1 (03:23):
He won by nine points. And this was after the Democrats.
Speaker 2 (03:28):
Poured millions of dollars into this race to try to
shore up support for this nutbar aften Bean. It was
forming Fox News this morning was framing this race as
a nailbier.
Speaker 1 (03:43):
It wasn't even close.
Speaker 2 (03:45):
Can we now have the commentary of the Democrats being
in disarray and how this election is a harbinger of
doom and gloom for Democrats in twenty twenty six. I
haven't seen any of that in the coverage relating to
this particular race. As a matter of fact, I have
his I see very little coverage relating to this race,
certainly when you compare it to the elections that just
(04:05):
took place last month, and the similarities are right there
in front of you. This was a safe Republican seat
that Democrats poured a ton of money into to try
to win, and they lost by nine points. You go
back and look at the November elections and you had
(04:27):
mostly blue races, less money because Republicans typically have less
money poured into some of those races, and Democrats won.
But see last month, it was all look at that.
Republicans better be worried. The Democrats just sweeping all these victories. Yeah,
they were blue areas.
Speaker 1 (04:50):
In case in point.
Speaker 2 (04:51):
Again, the commentary still this morning after a nine point.
Speaker 1 (04:55):
Victory was it was a nail bier. It wasn't a
nail bier.
Speaker 2 (05:00):
And this isn't a harbinger of doom and gloom for Democrats. No,
it was a special election scheduled after former Representative Mark
Green announced his resignation earlier this year. As the article
from Daily Wire covering this says, in recent weeks, money
poured into the race, as often being campaigned with Democrat.
Speaker 1 (05:23):
Favorites Al Gore and AOC.
Speaker 2 (05:27):
But this is no harbinger for Democrats heading in the
next year anymore than last year was a harbinger of
doom for Republicans in the other races. It just continues
to go and demonstrate the bias that we see every
single day. So it wasn't even I was worried about
it that Afton Bean was just a crazy person. The
(05:49):
thing about the crazies on the left, they typically don't
go away. She's actually raised quite a bit of a
name id now, which is hugely important when it comes
to Democrats. I don't think that she's going to be
going away anytime soon. So got a lot of ground
to cover. Let me give you one more item and
then we'll take a quick break here on Twin Cities
News Talk. As we get into Trump wanting to end
(06:10):
the federal income tax. Yesterday, Trump did go and make
it official. He followed through with this pledge announcing that
he's completely terminated every action signed by former President Joe Biden,
using the autopen and declaring those measures null and void.
So legal experts say the auto pen signed presidential actions
(06:31):
remain valid as long as the president authorized them, regardless
of who physically operated the device.
Speaker 1 (06:39):
But if the.
Speaker 2 (06:39):
President truly had no knowledge of the device's use or
did not consent to the signature, some scholars argue those
actions could be challengeable. The gop led House Oversight Committee
report questioned the legitimacy of numerous executive actions on the autopen,
alleging sometimes there was clear proof of Biden's personal approval.
(07:01):
The committee has urged the US Department of Justice to
investigate the potential misuse of presidential authority. So my guess
on what's taking place here, this seems to be more
of a means to the end, and Trump's been doing
this a lot lately. Challenges will ensue on the issue
of the auto pen, but those would likely end up
(07:21):
exposing the reality of how much, or more importantly, how
little was Biden involved, not just in the signing off
on these various issues, but also in his presidency. So again,
this seems like more of a means to an end
to bring about clarity on the issue of who was
really running the Biden administration more so than it is
(07:43):
nullifying what was signed by the auto pen. It's very
similar to what Trump did with the temporary protected status, which,
by the way, though that being said, because I am
convinced that Trump didn't make that move in order to
bring national attention to the issue of the fraud here
in Minnesota. That being said, that TPS will more than
(08:06):
likely be coming into play as ICE ramps up their
enforcement here in Minnesota. So we'll be getting into all
that throughout the show this morning. And of course I
want to hear from you phone number eight four four
nine four six five eight five five. You can email
Justice at iHeartRadio dot com. And of course, if you're
listening on the iHeartRadio app, your talkbacks are brought to
(08:27):
you by Lyndahl Realty You.
Speaker 1 (08:28):
We will get to those next here.
Speaker 2 (08:30):
On twin Citay's News Talk Am eleven thirty and one
oh three five FM.
Speaker 4 (08:33):
Good morning, and I love your show, John.
Speaker 3 (08:39):
I know nine points is a lot. However, when someone
is that crazy and that extreme, nine points isn't nearly enough.
Speaker 2 (08:49):
Well, when you consider how much attention Democrats gave the
millions that were poured into that race.
Speaker 1 (08:56):
Yeah, I'll take the nine points.
Speaker 2 (08:58):
And as I mentioned a moment go, the polls were
all way off.
Speaker 1 (09:03):
The media wanted to.
Speaker 2 (09:04):
Make this look a heck of a lot closer than
what it really was. You know, I said in the
earlier segment that the polls had the election down to
like two points, and.
Speaker 1 (09:16):
It wasn't even It wasn't even close.
Speaker 2 (09:20):
I want to see that commentary of how the Democrats
are in so much trouble next year.
Speaker 1 (09:23):
I just find the whole thing amusing.
Speaker 4 (09:25):
Morning John.
Speaker 5 (09:26):
I think, honestly, the scariest part about that Nashville election
is it forty five percent of the people or that
many people voted for somebody who literally hates their area
and hates their state.
Speaker 2 (09:42):
It's actually worse than that. So this was an election
in Tennessee. Nashville was the most prominent area of the
district as a whole. So Nashville is just a part
of the district itself, which spanned it, you know, beyond
the region of Nashville. In Nashville often had seventy seven
(10:04):
percent of the vote.
Speaker 1 (10:08):
That was over the GOP at like forty four.
Speaker 2 (10:11):
So again it tells you the dynamics in play, and
a lot like those core blue areas here in Minnesota
that we talk about so often on Twin Cities News
talk that drive elections here in the state. No, Nashville
is much in the same way, and even with winning
a massive majority in the key area of Nashville, often
(10:33):
still ended up losing by nine points. Continuing our show
here from the sixty five to one Carpet plus Next
Day Install Studios.
Speaker 6 (10:42):
Hey, John, I just wanted to say with that Nashville
election that we should definitely take the win. Sure, it
was a blowout, but I like how we kind of
think of stuff as being super close. It turns out
the vote a lot more.
Speaker 2 (10:59):
I feel, actually that's not a bad idea, and I
don't mind it either. I would rather have these And
here's the thing. On a comment like that, you're going
to get your wish because this is just what's going
to happen. Every election is likely you just can't trust
the polling. Sometimes you'll get more accurate polling, especially on
(11:20):
the national elections, but in situations like we have in
this race, where there aren't any other races to focus on,
the media is looking for things to talk about as
we head into the holidays, which is part of the
reason why Minnesota is getting so much attention right now.
That's not to say if we were outside of the holidays.
The level of fraud that we've been talking about for
(11:41):
years on the show would not have received exposure. However,
right now it is amplified because the media is looking
for something to talk about. They just ended up with
something really incredibly juicy to discuss. One thing I failed
to mention in my rundown of all the different issues
that Minnesota has faced. Also the Walls administration being in
(12:04):
the federal spotlight at risk of losing millions of dollars
in funding.
Speaker 1 (12:10):
For granting illegal.
Speaker 2 (12:12):
Driver's licenses to non domiciled residents those here illegally.
Speaker 1 (12:17):
But authorized to work.
Speaker 2 (12:18):
That's another one of the big issues that Minnesota is
currently facing right now. I failed to mention that in
the run down at the start of the show.
Speaker 5 (12:27):
And John, I do have one question for you and
the Republican Party. Why has it not been all over
the news Republican people screaming for Walls step down and
people to be held accountable. All I ever see is
Walls lying his face off on TV.
Speaker 2 (12:49):
I don't know about individuals calling for him to resign
elected officials at least, but I knew no. Plenty of
Republicans have spoken out. We'll talk to represent of Tom
Emmer in about an hour from now. He's been making
the news media rounds. He's not going to end up
on places like MSNBC and CNN, and a lot of
(13:10):
times local media doesn't go and talk to Republicans. So
I'm hesitant to go and do too much criticism over
Republicans apparently not speaking out on any one of these issues,
because a lot of times it simply is more an
aspect of the media not giving them proper attention. There
are plenty of individuals like myself going back to last
(13:32):
week when this all blew up after that City Journal
report calling for Walls to resign, and I have seen
those calls grow, albeit I haven't seen it coming from
elected officials, even though I have seen many members of
Congress that have been out speaking regarding the issues that
Minnesota is facing under Walls.
Speaker 1 (13:49):
And one of those will be Emma joining us next hour.
Speaker 2 (13:52):
So Trump says Americans may soon pay no income tax,
as White House explores alternative or revenue streams, and there's
a bit of a slow roll out on this. I've
been kind of keeping tabs on it. So last week
Trump made this comment talking with the media.
Speaker 7 (14:10):
Over the next couple of years, I think will substantially
be cutting and maybe cutting out completely, but we'll be
cutting income tax, could be almost completely cutting it because
the money we're taken in and is going to be
so large.
Speaker 1 (14:24):
And yet other.
Speaker 7 (14:26):
Countries who have been ripping us off for many, many
years and many years, they've just been ripping us as shreds.
Speaker 2 (14:31):
Now I think I actually reversed these. So that was yesterday.
That was Trump talking about cutting the income tax during
that lengthy cabinet meeting that took place off which we
have a lot of audio to share. I believe this
is the clip going back to last week.
Speaker 1 (14:45):
And I believe, yeah, Curial.
Speaker 7 (14:47):
That at some point in the not too distant future,
you won't even have income tax to pay because the
money would take it in is so great, it's so
enormous that you're not going to have income tax to pay.
Whether you get rid of it or just keep it
around for fun, or have it really low, much lower
than it is now, but you won't be paying income tax.
We've slashed one triggua down.
Speaker 2 (15:09):
What this is though for Trump, and he's done this
a lot when it comes to the things that he
wants to do, it's been a bit of a slow
ramp up to the actual action taking place, and I
had it right the first time. That clip you just
heard was from yesterday. The one prior was from last week.
And it's been just a slow rollout preparing everybody for
(15:29):
the potential of dramatically reducing the federal income tax or
eliminating it all together. And I'm of the opinion that
this is one of the keys that the Trump administration
is looking at utilizing in hopes of one turning the
economy around, because it's a much different mentality attached to
the federal income tax going away and how people go
(15:51):
and spend their money, because that's money that you made
and earned that the government take is taking from you,
as opposed to the government going and just arbitrarily cutting
you a stimulus check, which carries with it a different
mindset which brings about inflation. Cutting the federal income tax
is a much better way to drive that money into
the economy without driving up the cost of inflation. And
(16:11):
so I'm convinced that one Trump is going to be
rolling this out after the beginning of the year in
hopes of shoring up the economy, driving down prices, but
also helping Republicans in the mid terms later in the
year and certainly an issue that we'll be talking about
as we get more details on it in the near future.
(16:32):
But it is one that I've been watching for a while.
I'm incredibly excited over. I've heard the arguments from individuals saying, well,
if it happens and here in Minnesota, you're just going
to have the government go in taxes more at the
state level. Listen, it's nothing until it's something. I'm not
going to downplay my desire to see the federal income
(16:54):
tax be removed over concerns of what Democrats might attempt
to do here in Minnesota to offset that or take
advantage of it.
Speaker 1 (17:02):
We'll deal with that when the time comes.
Speaker 2 (17:05):
All right, coming up, David Gartenstein Ross, our AI expert,
will be joining us. Got a couple of different issues,
plus your questions for d V Those are already rolling
in on the iHeartRadio app. Scammers using AI deep fakes
to trick holiday shoppers, and almost half of gen Z
wants AI to run the government. Why you should probably
be terrified about that. We will talk with the v'd
(17:26):
next on Twin Cities News Talk Trinsday's News Talk Am
eleven thirty one three five FM from the sixty five
to one carpet Next Day Install Studios, John Justice and
(17:50):
joining me once.
Speaker 1 (17:51):
Again this week CEO.
Speaker 2 (17:54):
At Expert Theory Analysts regarding artificial Intelligence.
Speaker 1 (17:59):
Friend of the Show. So, David Carts and Sein Ross.
Speaker 2 (18:02):
David, good morning, and I hope that you and yours
had a fantastic Thanksgiving.
Speaker 1 (18:06):
Thanks for joining us.
Speaker 3 (18:08):
We most certainly did, John, and it's great to join you.
I hope you and yours had a great Thanksgiving too,
We really did.
Speaker 2 (18:14):
It was a nice little precursor to the holiday break
coming up. I always I try to use the Thanksgiving
breaks or to set the stage of how I'm going
to do with my lengthier.
Speaker 1 (18:24):
Christmas Christmas break, so it's like a trial and error.
But I'm just weird that way.
Speaker 2 (18:27):
So we have some talkbacks that I've already rolled in,
also an email that came in. So I want to
start off with the listener questions for you first divide
before we dive into some of the articles that we
pulled for today's show. If that works for you, perfect job,
all right, So first off, let's go and Jr. Friend
of the Show. He writes in and says question for
(18:49):
DGR I tested chat GPT with a mock am track
trip to Houston and Phoenix. It confused Amtrak codes with
airport codes, it missed to critical timing and leg requirement details,
and defaulted to check the website. In addition, despite repeated
requests for concise answers, when I speak to it, it
(19:11):
never improves.
Speaker 1 (19:12):
It keeps asking if it.
Speaker 2 (19:14):
Can help, on and on with let me know, if
I can help, let me know if what we need
to dig into when we can dig into it.
Speaker 1 (19:23):
So he then goes on to say it raises a
larger issue.
Speaker 2 (19:25):
If free chat GPT isn't reliable for basic tasks other
than Google searches, why should anyone trust to llms for
important questions? Now he has more to this, but I'll
go ahead and stop there and give you the opportunity
to respond to vi'd.
Speaker 4 (19:42):
It's a really wonderful question, and the answer is sorry.
Speaker 3 (19:51):
He gets into like this core principle when we get this,
when if I get this answer right, it will be
really useful.
Speaker 1 (19:57):
Okay, you better get a people often.
Speaker 4 (20:02):
So yeah, and I apologize.
Speaker 3 (20:03):
I often don't take a second to really formulate how
to say something.
Speaker 4 (20:07):
But here's the way I would look at it.
Speaker 3 (20:11):
Lms are an assistant, and sometimes they're an amazing assistant,
and sometimes they're a really subbar assistant. So, as your
listeners know, I'm a huge fan of using large language models.
But what the listener described as JR right, is a
(20:32):
listener what JR describes happens every once in a while
to me where there's a really basic task and my
LLLM just fails at it, and I show it specifically
how to do that task, and even when corrected, it
will either fail again immediately or fail again after another
(20:56):
couple of tries, and then a different task which is
much more complex it will do almost perfectly at So
I would step back, and here's how I would analogize it. Okay,
most professionals have worked with employees or assistance of some kind,
(21:18):
and sometimes there'll be an employee who is very good
and simply fails at a certain kind of task, and
you wouldn't fire them, and you wouldn't say that they
don't have value, but there's just something where you understand
if you give it to them, they're not going to
(21:39):
do it well. And they might not do it well.
They might not do it well at one day and
they'll do it well on a week later or two
weeks later, not because they've been improved, but because they've
had a bad day. So people have to be mindful
of areas where the LM is just going to fail you.
Speaker 4 (22:02):
And so in areas where you're unfamiliar.
Speaker 3 (22:06):
With a topic, that's the area where you can get
you know, you can get toasted by over relying on
an LLM. In Jr's case, he understood that it was
mixing up Amtrak codes and airplane codes. He was trying
to correct the LM couldn't get it right and said, okay,
just go ahead and check the website because it was you,
(22:26):
quote unquote frustrated. Right, they don't really have personalities or emotions,
but they kind of approximate them.
Speaker 4 (22:32):
And that's what happened.
Speaker 3 (22:33):
It didn't get it right, it failed, and it just
defaulted to, yeah, well I'm not getting it right. But
there's a way you can get it right, which is
doing the work yourself. And so this is an aspect
of llm's John and JR.
Speaker 2 (22:46):
This fits in with what he asked at the very
end of his of his question, he basically says he
gets into some other technical things of what he does.
But then he says, for non technical users, what is
the practical value of llms beyond curiosity or Google search alternatives?
And why should we contribute our questions to train any
(23:08):
AI platform?
Speaker 4 (23:12):
The answer is.
Speaker 3 (23:15):
Similar to why you would do Google searches right, and
Google now uses AI and it search answers as well.
So llms are almost anywhere that you're searching for information.
It's that you'll get back information that is useful. Here's
my practical tip to avoid the problem that JR Is
(23:39):
putting a finger on. It's asked you, when you formulate
the question, ask the LM to provide you with links
substantiating its answer, and I think either on both the
paid and premium versions of Chat, GPT and others, it
(23:59):
will you with links. And this allows you to, as
they say, do your own research. It's especially important with llms,
which can be black boxes in terms of how they
think something through. If you make it provide you with
web with web links to substantiate its answer, it at
least allows you to go back over it and the
(24:20):
answer is as to why should we trust it and
why should we give it? You grace it with giving
us these training querias. The answer would be, the proof
is in the value you get back to me. I
get back got value even understanding that there are sometimes
that the LM fails me. But there's just an important
(24:42):
lesson here for everybody that you should be skeptical and
you should understand that it is going to get some
things wrong. So unless you're sufficiently specific and double checking,
it can sometimes lead you astray.
Speaker 2 (24:58):
There's an interesting common thing read with all of this,
where wherein AI in and of itself is really not
in every facet. Obviously, it has introduced a lot of
new aspects in terms of technology beyond that which we're
familiar with. That being said, AI does do differently a
(25:20):
lot of things that we were already dealing with. For example,
you mentioned you know whether or not we should trust AI.
You know, I think the same thing would ring true
when it comes to doing a search online prior to
AI or by example. Let me get into this piece
really quick, because I think it's along the same lines.
Speaker 1 (25:38):
There's a story out of Fox nine.
Speaker 2 (25:40):
Locally, US consumers lost twelve point five billion to fraud
in twenty twenty four, with online shopping being the second
most common scam type.
Speaker 1 (25:49):
AI powered fraud.
Speaker 2 (25:50):
Is becoming increasingly common, with scammers using technology to create
fake retailer websites, phishing emails, deep fake videos that impersonate
trusted brands or influencers.
Speaker 1 (26:00):
And again, so many of the.
Speaker 2 (26:01):
Concerns around AI aren't so much new concerns. They're just
an enhanced method of, in this case, previous scams that
we were already dealing with, but now we have the
element of AI being able to assist in this case
the bad guys with the scams they are trying to perform.
Speaker 4 (26:19):
Yeah, I agree with that.
Speaker 3 (26:22):
It's similar to problems that we faced before, and I
think the difference is in speed, synthesis, and what we
could call accuracy. Right, Like, lms are way way faster,
which allows scams like the one you're talking about, the
(26:43):
fake holiday shopping websites, which everyone should be able to
look out for. It allows them to erect it more quickly,
It allows you know, synthesis. This is the area where
I think LMS are different than anything we've had before,
or the ability to synthesize across large amounts of information,
(27:05):
even understanding that sometimes they're just going to get that
synthesis wrong, and then the final point, I called it accuracy.
It's that the Fox Fine article that you're putting your
figure on talks about it. Quotes Larry Zelvin, who's the
head of security advisory at BMO at the baker Botreal
(27:27):
who said that AI has made scams more convincing and
harder to detect. Fraustters can bimic trusts brands and voices
with alarming accuracy. And that's that's correct, right for anyone,
for people doing a scam. I won't get into how
someone could train an LM to do that, but it's
not hard at all, and it puts togures that you
(27:48):
and I have talked about John in other contexts. We've
talked about how deep fake videos it's. It can be
hard to determine, like whether the video is actually Trump
or not until he starts talking about how he's going
to deploy velociraptors at the border. It's pretty convincing and
it looks a lot like Donald Trump. And the same
(28:09):
thing is true of holiday shopping websites. It's easier than
ever for scammers to make the holiday shopping website look
a lot more like a legitimate website.
Speaker 4 (28:23):
It erodes the distinction between the two, and.
Speaker 3 (28:25):
There's a lot of danger where now it's not just
you know, old people or people who aren't Internet savvy
who could potentially fall for scams.
Speaker 4 (28:35):
They look more real.
Speaker 2 (28:36):
Talking with David garten Stein Ross, a AI analyst, a
CEO at Expert Theory, these aren't the questions. I just
want to add to the discussion here a couple of
a couple of different things. First off, it's really interesting.
You know, I've got two boys. Kyle's going to be
nineteen here in about a week and Logan's twenty three,
and you know, they're on their phones. We all enjoy
looking at the memes. We've talked about these a lot.
(28:58):
But it's interesting now because whenever these videos pop up
and we're sharing them with one another, there is a
debate that takes place between us if it's not labeled
as such.
Speaker 1 (29:09):
You know, I wonder if that's AI. I think it's A.
I don't think it's A.
Speaker 2 (29:12):
It's funny how it's entered into you know that aspect
of our enjoyment of just entertainment items like that. It's
just now a part of the of the conversation. I
have one more thing I want to add, is there
anything that you would like to mention to or you
want to say to that to that particular observation.
Speaker 3 (29:30):
The the interesting thing like that observation and the main
thing I thought about is how the way you deterribine
if a video or a beam is AI or not
is similar to what people should do to keep themselves
safe well, trying to do holiday line and try to
avoid falling for scams, which is double check, yeah right,
(29:54):
like just like you could research is a beam legit?
I always before using a shopping service to haven't used before,
I'm going to double and triple check to make sure
it's LEDGECT.
Speaker 2 (30:07):
I'm gonna send you a link to a video once
I got some time this morning. There's a there's a
YouTube channel. It's a special effects house run by a
group of millennials in southern California.
Speaker 1 (30:19):
It's called the Corridor Crew. I don't know if I've
ever heard of them before, but I happened.
Speaker 2 (30:24):
They analyze movies and CGI in movies and things of
this nature, and they've been on top of AI. They
did a video recently and again I'll share with you
wherein they went online and they ordered products that could
be purchased at an incredibly cheap price, but they looked
like high price items. So like a Carhart jacket. We've
(30:45):
all seen the images of these cable knit sweaters with
these expansive designs on them like you've never seen before.
Speaker 1 (30:53):
They looked three D.
Speaker 2 (30:55):
So they went and ordered all these products to see
how closely it matched the obvious AI creation as it
was being presented online. And of course the result is
what you would expect the result to be. These were
incredibly cheap, flimsy items that were just knockoff products that
ended up ultimately looking.
Speaker 1 (31:14):
Similarly enough online.
Speaker 2 (31:16):
But nothing close to the quality that you would expect
it was. It's incredibly entertaining and I'll uh, I'll share
it with you when I get a chance here to v.
Speaker 3 (31:24):
That's so interesting, right, Like it's interesting because they're clearly
online scams of fake items and the knockoff items. And
what's interesting is that you know they were able to
purchase them and they were shipped and received. Yeah, right, Like,
you know, I typically what I think is that if
if I'm buying a knockoff, if I'm buying it online,
(31:48):
I'm not going to expect people to ship it to me,
it's really interesting that they shipped.
Speaker 2 (31:52):
Yeah, they got the products and they're close proximities, but
they are not. I mean, essentially, they end up like
the clothing ends up being of the quality that you
would get it like a halloween like a cheap Halloween costume.
Speaker 1 (32:02):
That's what the That's what they ended up being. So
I'll share that with you.
Speaker 2 (32:04):
Let's get back to a couple of the questions that
the audience has with for you. I actually had one
that came in from front of the show Rose mount Rick.
It's a really quick one. He just says, David, do
you talk to GROC every day? Do you try to
teach Grock?
Speaker 4 (32:18):
No?
Speaker 3 (32:19):
So I use I use chat GPT, and there's no
particular reason other than that's the one that I first
started using. I used CROC once in a while, but
chat GPT is the system I most hang out in.
Speaker 2 (32:34):
All right, let's go to the iHeart Radio app this
morning as we continue our discussion with AI analyst de
Vid Gartenstein Ross.
Speaker 5 (32:41):
Good morning, John.
Speaker 8 (32:42):
Can you please ask d g R about what the
differences between co pilot or chat GPT and an agent
created within co pilot or chat GPT.
Speaker 1 (32:53):
Thanks DVD.
Speaker 3 (32:56):
It's a it's a wonderful question. The quick answer is
that you know Copilot or chat Gypt. The main purpose
of it is to have a conversation with you. An
agent will actually perform tasks for you, and for that,
Copilot is much much better than chat Gypt because it's
(33:18):
integrated into the Microsoft Office suite, and so if you,
if you're a Microsoft person, Copilot can integrate any sort
of agentic function. Well, I don't use agents that often
in AI. I imagine though, that you know, a year
or eighteen months from now, I will use agents AI
(33:42):
agents multiple times a day. They're an important tool that
most people will be using within the next twelve to
eighteen months.
Speaker 2 (33:51):
All right, let's go here and talk with friend of
the show, Raquel, who has a question for David garten
Stein Ross.
Speaker 9 (33:57):
I have a question for deb Garden Santa Ross Ross.
A few weeks ago, maybe one or two weeks ago,
we played a piece of music, and you said that
you could tell it was written by AI.
Speaker 4 (34:09):
My question is about comic book art.
Speaker 9 (34:11):
My husband is a comic book artist, and comic books
can take weeks to complete. My husband's fear is that
AI will completely replace all comic book artists, because you know,
I could do it in a matter of minutes. I
would just like your opinion. Thank you Dad, Thanks Keel.
That's a it's a huge concern.
Speaker 3 (34:34):
AI does comic book art really well, and just most
kinds of art really well. Right now in certain industries,
like I work in the games industry as as folks know,
and for major board games, there's this stigma against using
AI and a preference for human artists.
Speaker 4 (34:54):
But that stigma is eroding.
Speaker 3 (35:00):
The concern is a real one, and so I'll provide
what I would do as an artist understanding that threat,
and that is I would actually it to hedge against that.
I would start using AI once in a while, not
(35:21):
in my art for customers, but to play around with
creating my own comic book art, because yeah, like what
Raquel is outlining is real, Like it's a huge problem.
I would definitely feel threat there. But there's going to
be things that comic book artists just know implicitly that
they can do more quickly with AI. That's actually, in
(35:45):
my view where job the security lies. And we can
talk about that at some length. But like Raquel's husband
definitely has worried to be concerned there.
Speaker 4 (35:55):
I definitely share the concern that she has Davdi.
Speaker 1 (35:58):
Guards and the Stein Ross.
Speaker 2 (35:59):
I'm going to hang on to the other articles that
we had planned to talk about today, but I really
enjoyed the conversation and the fact that many of the
listeners had a chance to get their questions answered. And always,
I appreciated the time that you give us every single week.
If you have further questions for David, you can always
email me Justice at iHeartRadio dot com. I'm happy to
forward those over to DVD David. If there's an email
(36:21):
that you would like to share or any place that
you would like to go and direct the listener before
I let you go this morning.
Speaker 4 (36:26):
Absolutely, My email is dav'd.
Speaker 3 (36:28):
That's dav Eed at Expert Theory dot com.
Speaker 2 (36:34):
David Gartinstein Ross. As always, thanks for the time, my friend.
I look forward to hopefully doing it again next week.
Speaker 4 (36:39):
Likewise, John, so do I all right?
Speaker 2 (36:42):
Cabinet meeting yesterday a lot took place with President Donald Trump.
Lengthy cabinet meeting, incredibly transparent this administration is. During that
cabinet meeting, one of the things that was talked about
was snap Secretary of Agriculture Brook Rawlins deliveredy sweeping update
on the nation's SNAP program, revealing what she called rampant
(37:04):
fraud and announcing a dramatic escalation in the administration's effort
to force state compliance in federal oversight. You may remember
that Minnesota was one of the states that was pushing
back on this. Rowland said that under the president's tenure,
about eight hundred thousand Americans have already moved off of
SNAP and the program is serving forty two million people.
She argued the trend underscores the need to restore integrity
(37:28):
to the system that she's said has been abused for
far too long. With that, before we head into next hour,
let me share this with you. This is USDA Secretary
Rawlins saying that she will be moving to halt federal
funding two states who refuse to turn over the SNAP
data to help root out the fraud, like Minnesota. And
we'll get to more of your talkbacks coming up in
(37:49):
just a moment here on Twin Cities News Talk. Here
is the USDA Secretary Ice.
Speaker 10 (37:55):
We had a couple of people receiving benefits in six
states in February of this year. We asked for all
the states for the first time to turn over their
data to the federal government to let the USDA partner
with them to root out this fraud, to make sure
that those who really need food stamps are getting them,
but also to ensure that the American taxpayer is protected.
(38:15):
Twenty one states said yes. Twenty nine states said yes,
not surprisingly the Red states, and that's where all of
that data that fraud comes from. But twenty one states,
including California, New York, and Minnesota, the Blue states, continue
to say no. So as of next week, we have
begun and will begin to stop moving federal funds into
(38:38):
those states until they comply and they tell us and
allow us to partner with them to root out this
fraud and to protect the American taxpayer.
Speaker 4 (38:46):
As Joe Biden was.
Speaker 10 (38:47):
Working to buy an election a year ago, he increased
food stamp program funding by forty percent. So now as
we continue to roll that back, so.
Speaker 2 (38:58):
Keep in mind, just really quick, keep in mind that
not only has billions of your dollars taxpayer money gone
up and smoke with fraud taking place under the Walls administration,
Walls now has put the state in a position, whether
it's because of what was mentioned there by the Agriculture
(39:21):
Secretary of losing funding for the lack of willingness to
send over snap data.
Speaker 1 (39:27):
Millions of dollars in funding could be lost.
Speaker 2 (39:29):
There some thirty million dollars in funding over drivers' licenses
for those who are here illegally. We could also lose
that funding as well. And it's all happening under Governor
Tim Walls's watch.
Speaker 1 (39:43):
He needs to.
Speaker 2 (39:43):
Resign, and as I keep saying, I do not expect
that he is going to be the nominee for governor
heading into next year. More coming up on Twin City's
News Talk AM eleven thirty and one to three five FM.
Speaker 4 (39:53):
I'm ready to get beat.
And it's only the middle of the week, so much
has happened.
Speaker 2 (00:15):
We have a lot to talk about this morning on
Twin Cities News Talk Am eleven thirty one oh three
five FM. My name is John Justice, and I'm glad
that you're with the show this morning as we embark
on our broadcast from the six to five to one
carpet plus Next Day Install studios pushing the buttons in
(00:35):
the master control booth this morning.
Speaker 1 (00:37):
Is Devin. Nice to see him.
Speaker 2 (00:39):
Got a lot of guests on the show today, But
let's run down all of the major news items as
we dive in, dive into everything that needs to be
covered relating to the fact that Minnesota has become the
epicenter of the news cycle right now, much to the
chagrin of Tim Walls. Couldn't have happened to him, dude,
(01:00):
all right, So on the show today, as I had
been hearing, it has now become a reality. ICE is
launching extensive immigration raids here in Minnesota. We have a
lot of audio to share of local law enforcements, Minneapolis
man Baby mayors commenting their concern about ICE coming in
(01:23):
and doing their job. I mean, consider that for just
a moment, you have law enforcement criticizing that law enforcement
is going to come in and conduct law enforcement in it.
Make it make sense small business administration orders an investigation
into further COVID fraud in Minnesota.
Speaker 1 (01:42):
This is coming from the Trump administration.
Speaker 2 (01:44):
Walls does respond to all the latest of plaguing his administration,
all of the controversies, and we'll talk about that defeat
gartzan Stein. Ross will get into some AI conversations at
six point thirty. You can get your comments in early
via the iHeartRadio app Talkbacks. Representative Tom Emer will join
us at seven twenty this morning. Liz Colman with a
(02:07):
lot to cover relating to news at alphanews dot org.
That'll be at eight thirty as well. The Department of
Justice has also officially sued the state of Minnesota to
obtain the entire statewide voter roles, including all illegal aliens
living in Minnesota who may have been fraudulently registered to vote.
(02:33):
This is one of the things we're actually going to
talk about with David regarding AI because the potential of
AI actually going and speeding up these investigations. So we
got a lot of ground to cover today. Let's start
here though the election overnight that everybody was so worried about.
Speaker 1 (02:48):
Yeah, it wasn't that big of a concern.
Speaker 2 (02:53):
Republican Matt Van Epps fended off a challenge from the
far left whack a Doodle aften Bean on Tuesday in
the special election for the Tennessee seventh Congressional District. Ninety
five percent of the vote in the race was called
for Epps fifty four to forty five. This was a
nine point victory. And if you remember the commentary heading
(03:17):
into it, polls, oh, it's within two points, it's within
striking distance.
Speaker 3 (03:22):
It wasn't.
Speaker 4 (03:23):
It was not.
Speaker 1 (03:23):
He won by nine points. And this was after the Democrats.
Speaker 2 (03:28):
Poured millions of dollars into this race to try to
shore up support for this nutbar aften Bean. It was
forming Fox News this morning was framing this race as
a nailbier.
Speaker 1 (03:43):
It wasn't even close.
Speaker 2 (03:45):
Can we now have the commentary of the Democrats being
in disarray and how this election is a harbinger of
doom and gloom for Democrats in twenty twenty six. I
haven't seen any of that in the coverage relating to
this particular race. As a matter of fact, I have
his I see very little coverage relating to this race,
certainly when you compare it to the elections that just
(04:05):
took place last month, and the similarities are right there
in front of you. This was a safe Republican seat
that Democrats poured a ton of money into to try
to win, and they lost by nine points. You go
back and look at the November elections and you had
(04:27):
mostly blue races, less money because Republicans typically have less
money poured into some of those races, and Democrats won.
But see last month, it was all look at that.
Republicans better be worried. The Democrats just sweeping all these victories. Yeah,
they were blue areas.
Speaker 1 (04:50):
In case in point.
Speaker 2 (04:51):
Again, the commentary still this morning after a nine point.
Speaker 1 (04:55):
Victory was it was a nail bier. It wasn't a
nail bier.
Speaker 2 (05:00):
And this isn't a harbinger of doom and gloom for Democrats. No,
it was a special election scheduled after former Representative Mark
Green announced his resignation earlier this year. As the article
from Daily Wire covering this says, in recent weeks, money
poured into the race, as often being campaigned with Democrat.
Speaker 1 (05:23):
Favorites Al Gore and AOC.
Speaker 2 (05:27):
But this is no harbinger for Democrats heading in the
next year anymore than last year was a harbinger of
doom for Republicans in the other races. It just continues
to go and demonstrate the bias that we see every
single day. So it wasn't even I was worried about
it that Afton Bean was just a crazy person. The
(05:49):
thing about the crazies on the left, they typically don't
go away. She's actually raised quite a bit of a
name id now, which is hugely important when it comes
to Democrats. I don't think that she's going to be
going away anytime soon. So got a lot of ground
to cover. Let me give you one more item and
then we'll take a quick break here on Twin Cities
News Talk. As we get into Trump wanting to end
(06:10):
the federal income tax. Yesterday, Trump did go and make
it official. He followed through with this pledge announcing that
he's completely terminated every action signed by former President Joe Biden,
using the autopen and declaring those measures null and void.
So legal experts say the auto pen signed presidential actions
(06:31):
remain valid as long as the president authorized them, regardless
of who physically operated the device.
Speaker 1 (06:39):
But if the.
Speaker 2 (06:39):
President truly had no knowledge of the device's use or
did not consent to the signature, some scholars argue those
actions could be challengeable. The gop led House Oversight Committee
report questioned the legitimacy of numerous executive actions on the autopen,
alleging sometimes there was clear proof of Biden's personal approval.
(07:01):
The committee has urged the US Department of Justice to
investigate the potential misuse of presidential authority. So my guess
on what's taking place here, this seems to be more
of a means to the end, and Trump's been doing
this a lot lately. Challenges will ensue on the issue
of the auto pen, but those would likely end up
(07:21):
exposing the reality of how much, or more importantly, how
little was Biden involved, not just in the signing off
on these various issues, but also in his presidency. So again,
this seems like more of a means to an end
to bring about clarity on the issue of who was
really running the Biden administration more so than it is
(07:43):
nullifying what was signed by the auto pen. It's very
similar to what Trump did with the temporary protected status, which,
by the way, though that being said, because I am
convinced that Trump didn't make that move in order to
bring national attention to the issue of the fraud here
in Minnesota. That being said, that TPS will more than
(08:06):
likely be coming into play as ICE ramps up their
enforcement here in Minnesota. So we'll be getting into all
that throughout the show this morning. And of course I
want to hear from you phone number eight four four
nine four six five eight five five. You can email
Justice at iHeartRadio dot com. And of course, if you're
listening on the iHeartRadio app, your talkbacks are brought to
(08:27):
you by Lyndahl Realty You.
Speaker 1 (08:28):
We will get to those next here.
Speaker 2 (08:30):
On twin Citay's News Talk Am eleven thirty and one
oh three five FM.
Speaker 4 (08:33):
Good morning, and I love your show, John.
Speaker 3 (08:39):
I know nine points is a lot. However, when someone
is that crazy and that extreme, nine points isn't nearly enough.
Speaker 2 (08:49):
Well, when you consider how much attention Democrats gave the
millions that were poured into that race.
Speaker 1 (08:56):
Yeah, I'll take the nine points.
Speaker 2 (08:58):
And as I mentioned a moment go, the polls were
all way off.
Speaker 1 (09:03):
The media wanted to.
Speaker 2 (09:04):
Make this look a heck of a lot closer than
what it really was. You know, I said in the
earlier segment that the polls had the election down to
like two points, and.
Speaker 1 (09:16):
It wasn't even It wasn't even close.
Speaker 2 (09:20):
I want to see that commentary of how the Democrats
are in so much trouble next year.
Speaker 1 (09:23):
I just find the whole thing amusing.
Speaker 4 (09:25):
Morning John.
Speaker 5 (09:26):
I think, honestly, the scariest part about that Nashville election
is it forty five percent of the people or that
many people voted for somebody who literally hates their area
and hates their state.
Speaker 2 (09:42):
It's actually worse than that. So this was an election
in Tennessee. Nashville was the most prominent area of the
district as a whole. So Nashville is just a part
of the district itself, which spanned it, you know, beyond
the region of Nashville. In Nashville often had seventy seven
(10:04):
percent of the vote.
Speaker 1 (10:08):
That was over the GOP at like forty four.
Speaker 2 (10:11):
So again it tells you the dynamics in play, and
a lot like those core blue areas here in Minnesota
that we talk about so often on Twin Cities News
talk that drive elections here in the state. No, Nashville
is much in the same way, and even with winning
a massive majority in the key area of Nashville, often
(10:33):
still ended up losing by nine points. Continuing our show
here from the sixty five to one Carpet plus Next
Day Install Studios.
Speaker 6 (10:42):
Hey, John, I just wanted to say with that Nashville
election that we should definitely take the win. Sure, it
was a blowout, but I like how we kind of
think of stuff as being super close. It turns out
the vote a lot more.
Speaker 2 (10:59):
I feel, actually that's not a bad idea, and I
don't mind it either. I would rather have these And
here's the thing. On a comment like that, you're going
to get your wish because this is just what's going
to happen. Every election is likely you just can't trust
the polling. Sometimes you'll get more accurate polling, especially on
(11:20):
the national elections, but in situations like we have in
this race, where there aren't any other races to focus on,
the media is looking for things to talk about as
we head into the holidays, which is part of the
reason why Minnesota is getting so much attention right now.
That's not to say if we were outside of the holidays.
The level of fraud that we've been talking about for
(11:41):
years on the show would not have received exposure. However,
right now it is amplified because the media is looking
for something to talk about. They just ended up with
something really incredibly juicy to discuss. One thing I failed
to mention in my rundown of all the different issues
that Minnesota has faced. Also the Walls administration being in
(12:04):
the federal spotlight at risk of losing millions of dollars
in funding.
Speaker 1 (12:10):
For granting illegal.
Speaker 2 (12:12):
Driver's licenses to non domiciled residents those here illegally.
Speaker 1 (12:17):
But authorized to work.
Speaker 2 (12:18):
That's another one of the big issues that Minnesota is
currently facing right now. I failed to mention that in
the run down at the start of the show.
Speaker 5 (12:27):
And John, I do have one question for you and
the Republican Party. Why has it not been all over
the news Republican people screaming for Walls step down and
people to be held accountable. All I ever see is
Walls lying his face off on TV.
Speaker 2 (12:49):
I don't know about individuals calling for him to resign
elected officials at least, but I knew no. Plenty of
Republicans have spoken out. We'll talk to represent of Tom
Emmer in about an hour from now. He's been making
the news media rounds. He's not going to end up
on places like MSNBC and CNN, and a lot of
(13:10):
times local media doesn't go and talk to Republicans. So
I'm hesitant to go and do too much criticism over
Republicans apparently not speaking out on any one of these issues,
because a lot of times it simply is more an
aspect of the media not giving them proper attention. There
are plenty of individuals like myself going back to last
(13:32):
week when this all blew up after that City Journal
report calling for Walls to resign, and I have seen
those calls grow, albeit I haven't seen it coming from
elected officials, even though I have seen many members of
Congress that have been out speaking regarding the issues that
Minnesota is facing under Walls.
Speaker 1 (13:49):
And one of those will be Emma joining us next hour.
Speaker 2 (13:52):
So Trump says Americans may soon pay no income tax,
as White House explores alternative or revenue streams, and there's
a bit of a slow roll out on this. I've
been kind of keeping tabs on it. So last week
Trump made this comment talking with the media.
Speaker 7 (14:10):
Over the next couple of years, I think will substantially
be cutting and maybe cutting out completely, but we'll be
cutting income tax, could be almost completely cutting it because
the money we're taken in and is going to be
so large.
Speaker 1 (14:24):
And yet other.
Speaker 7 (14:26):
Countries who have been ripping us off for many, many
years and many years, they've just been ripping us as shreds.
Speaker 2 (14:31):
Now I think I actually reversed these. So that was yesterday.
That was Trump talking about cutting the income tax during
that lengthy cabinet meeting that took place off which we
have a lot of audio to share. I believe this
is the clip going back to last week.
Speaker 1 (14:45):
And I believe, yeah, Curial.
Speaker 7 (14:47):
That at some point in the not too distant future,
you won't even have income tax to pay because the
money would take it in is so great, it's so
enormous that you're not going to have income tax to pay.
Whether you get rid of it or just keep it
around for fun, or have it really low, much lower
than it is now, but you won't be paying income tax.
We've slashed one triggua down.
Speaker 2 (15:09):
What this is though for Trump, and he's done this
a lot when it comes to the things that he
wants to do, it's been a bit of a slow
ramp up to the actual action taking place, and I
had it right the first time. That clip you just
heard was from yesterday. The one prior was from last week.
And it's been just a slow rollout preparing everybody for
(15:29):
the potential of dramatically reducing the federal income tax or
eliminating it all together. And I'm of the opinion that
this is one of the keys that the Trump administration
is looking at utilizing in hopes of one turning the
economy around, because it's a much different mentality attached to
the federal income tax going away and how people go
(15:51):
and spend their money, because that's money that you made
and earned that the government take is taking from you,
as opposed to the government going and just arbitrarily cutting
you a stimulus check, which carries with it a different
mindset which brings about inflation. Cutting the federal income tax
is a much better way to drive that money into
the economy without driving up the cost of inflation. And
(16:11):
so I'm convinced that one Trump is going to be
rolling this out after the beginning of the year in
hopes of shoring up the economy, driving down prices, but
also helping Republicans in the mid terms later in the
year and certainly an issue that we'll be talking about
as we get more details on it in the near future.
(16:32):
But it is one that I've been watching for a while.
I'm incredibly excited over. I've heard the arguments from individuals saying, well,
if it happens and here in Minnesota, you're just going
to have the government go in taxes more at the
state level. Listen, it's nothing until it's something. I'm not
going to downplay my desire to see the federal income
(16:54):
tax be removed over concerns of what Democrats might attempt
to do here in Minnesota to offset that or take
advantage of it.
Speaker 1 (17:02):
We'll deal with that when the time comes.
Speaker 2 (17:05):
All right, coming up, David Gartenstein Ross, our AI expert,
will be joining us. Got a couple of different issues,
plus your questions for d V Those are already rolling
in on the iHeartRadio app. Scammers using AI deep fakes
to trick holiday shoppers, and almost half of gen Z
wants AI to run the government. Why you should probably
be terrified about that. We will talk with the v'd
(17:26):
next on Twin Cities News Talk Trinsday's News Talk Am
eleven thirty one three five FM from the sixty five
to one carpet Next Day Install Studios, John Justice and
(17:50):
joining me once.
Speaker 1 (17:51):
Again this week CEO.
Speaker 2 (17:54):
At Expert Theory Analysts regarding artificial Intelligence.
Speaker 1 (17:59):
Friend of the Show. So, David Carts and Sein Ross.
Speaker 2 (18:02):
David, good morning, and I hope that you and yours
had a fantastic Thanksgiving.
Speaker 1 (18:06):
Thanks for joining us.
Speaker 3 (18:08):
We most certainly did, John, and it's great to join you.
I hope you and yours had a great Thanksgiving too,
We really did.
Speaker 2 (18:14):
It was a nice little precursor to the holiday break
coming up. I always I try to use the Thanksgiving
breaks or to set the stage of how I'm going
to do with my lengthier.
Speaker 1 (18:24):
Christmas Christmas break, so it's like a trial and error.
But I'm just weird that way.
Speaker 2 (18:27):
So we have some talkbacks that I've already rolled in,
also an email that came in. So I want to
start off with the listener questions for you first divide
before we dive into some of the articles that we
pulled for today's show. If that works for you, perfect job,
all right, So first off, let's go and Jr. Friend
of the Show. He writes in and says question for
(18:49):
DGR I tested chat GPT with a mock am track
trip to Houston and Phoenix. It confused Amtrak codes with
airport codes, it missed to critical timing and leg requirement details,
and defaulted to check the website. In addition, despite repeated
requests for concise answers, when I speak to it, it
(19:11):
never improves.
Speaker 1 (19:12):
It keeps asking if it.
Speaker 2 (19:14):
Can help, on and on with let me know, if
I can help, let me know if what we need
to dig into when we can dig into it.
Speaker 1 (19:23):
So he then goes on to say it raises a
larger issue.
Speaker 2 (19:25):
If free chat GPT isn't reliable for basic tasks other
than Google searches, why should anyone trust to llms for
important questions? Now he has more to this, but I'll
go ahead and stop there and give you the opportunity
to respond to vi'd.
Speaker 4 (19:42):
It's a really wonderful question, and the answer is sorry.
Speaker 3 (19:51):
He gets into like this core principle when we get this,
when if I get this answer right, it will be
really useful.
Speaker 1 (19:57):
Okay, you better get a people often.
Speaker 4 (20:02):
So yeah, and I apologize.
Speaker 3 (20:03):
I often don't take a second to really formulate how
to say something.
Speaker 4 (20:07):
But here's the way I would look at it.
Speaker 3 (20:11):
Lms are an assistant, and sometimes they're an amazing assistant,
and sometimes they're a really subbar assistant. So, as your
listeners know, I'm a huge fan of using large language models.
But what the listener described as JR right, is a
(20:32):
listener what JR describes happens every once in a while
to me where there's a really basic task and my
LLLM just fails at it, and I show it specifically
how to do that task, and even when corrected, it
will either fail again immediately or fail again after another
(20:56):
couple of tries, and then a different task which is
much more complex it will do almost perfectly at So
I would step back, and here's how I would analogize it. Okay,
most professionals have worked with employees or assistance of some kind,
(21:18):
and sometimes there'll be an employee who is very good
and simply fails at a certain kind of task, and
you wouldn't fire them, and you wouldn't say that they
don't have value, but there's just something where you understand
if you give it to them, they're not going to
(21:39):
do it well. And they might not do it well.
They might not do it well at one day and
they'll do it well on a week later or two
weeks later, not because they've been improved, but because they've
had a bad day. So people have to be mindful
of areas where the LM is just going to fail you.
Speaker 4 (22:02):
And so in areas where you're unfamiliar.
Speaker 3 (22:06):
With a topic, that's the area where you can get
you know, you can get toasted by over relying on
an LLM. In Jr's case, he understood that it was
mixing up Amtrak codes and airplane codes. He was trying
to correct the LM couldn't get it right and said, okay,
just go ahead and check the website because it was you,
(22:26):
quote unquote frustrated. Right, they don't really have personalities or emotions,
but they kind of approximate them.
Speaker 4 (22:32):
And that's what happened.
Speaker 3 (22:33):
It didn't get it right, it failed, and it just
defaulted to, yeah, well I'm not getting it right. But
there's a way you can get it right, which is
doing the work yourself. And so this is an aspect
of llm's John and JR.
Speaker 2 (22:46):
This fits in with what he asked at the very
end of his of his question, he basically says he
gets into some other technical things of what he does.
But then he says, for non technical users, what is
the practical value of llms beyond curiosity or Google search alternatives?
And why should we contribute our questions to train any
(23:08):
AI platform?
Speaker 4 (23:12):
The answer is.
Speaker 3 (23:15):
Similar to why you would do Google searches right, and
Google now uses AI and it search answers as well.
So llms are almost anywhere that you're searching for information.
It's that you'll get back information that is useful. Here's
my practical tip to avoid the problem that JR Is
(23:39):
putting a finger on. It's asked you, when you formulate
the question, ask the LM to provide you with links
substantiating its answer, and I think either on both the
paid and premium versions of Chat, GPT and others, it
(23:59):
will you with links. And this allows you to, as
they say, do your own research. It's especially important with llms,
which can be black boxes in terms of how they
think something through. If you make it provide you with
web with web links to substantiate its answer, it at
least allows you to go back over it and the
(24:20):
answer is as to why should we trust it and
why should we give it? You grace it with giving
us these training querias. The answer would be, the proof
is in the value you get back to me. I
get back got value even understanding that there are sometimes
that the LM fails me. But there's just an important
(24:42):
lesson here for everybody that you should be skeptical and
you should understand that it is going to get some
things wrong. So unless you're sufficiently specific and double checking,
it can sometimes lead you astray.
Speaker 2 (24:58):
There's an interesting common thing read with all of this,
where wherein AI in and of itself is really not
in every facet. Obviously, it has introduced a lot of
new aspects in terms of technology beyond that which we're
familiar with. That being said, AI does do differently a
(25:20):
lot of things that we were already dealing with. For example,
you mentioned you know whether or not we should trust AI.
You know, I think the same thing would ring true
when it comes to doing a search online prior to
AI or by example. Let me get into this piece
really quick, because I think it's along the same lines.
Speaker 1 (25:38):
There's a story out of Fox nine.
Speaker 2 (25:40):
Locally, US consumers lost twelve point five billion to fraud
in twenty twenty four, with online shopping being the second
most common scam type.
Speaker 1 (25:49):
AI powered fraud.
Speaker 2 (25:50):
Is becoming increasingly common, with scammers using technology to create
fake retailer websites, phishing emails, deep fake videos that impersonate
trusted brands or influencers.
Speaker 1 (26:00):
And again, so many of the.
Speaker 2 (26:01):
Concerns around AI aren't so much new concerns. They're just
an enhanced method of, in this case, previous scams that
we were already dealing with, but now we have the
element of AI being able to assist in this case
the bad guys with the scams they are trying to perform.
Speaker 4 (26:19):
Yeah, I agree with that.
Speaker 3 (26:22):
It's similar to problems that we faced before, and I
think the difference is in speed, synthesis, and what we
could call accuracy. Right, Like, lms are way way faster,
which allows scams like the one you're talking about, the
(26:43):
fake holiday shopping websites, which everyone should be able to
look out for. It allows them to erect it more quickly,
It allows you know, synthesis. This is the area where
I think LMS are different than anything we've had before,
or the ability to synthesize across large amounts of information,
(27:05):
even understanding that sometimes they're just going to get that
synthesis wrong, and then the final point, I called it accuracy.
It's that the Fox Fine article that you're putting your
figure on talks about it. Quotes Larry Zelvin, who's the
head of security advisory at BMO at the baker Botreal
(27:27):
who said that AI has made scams more convincing and
harder to detect. Fraustters can bimic trusts brands and voices
with alarming accuracy. And that's that's correct, right for anyone,
for people doing a scam. I won't get into how
someone could train an LM to do that, but it's
not hard at all, and it puts togures that you
(27:48):
and I have talked about John in other contexts. We've
talked about how deep fake videos it's. It can be
hard to determine, like whether the video is actually Trump
or not until he starts talking about how he's going
to deploy velociraptors at the border. It's pretty convincing and
it looks a lot like Donald Trump. And the same
(28:09):
thing is true of holiday shopping websites. It's easier than
ever for scammers to make the holiday shopping website look
a lot more like a legitimate website.
Speaker 4 (28:23):
It erodes the distinction between the two, and.
Speaker 3 (28:25):
There's a lot of danger where now it's not just
you know, old people or people who aren't Internet savvy
who could potentially fall for scams.
Speaker 4 (28:35):
They look more real.
Speaker 2 (28:36):
Talking with David garten Stein Ross, a AI analyst, a
CEO at Expert Theory, these aren't the questions. I just
want to add to the discussion here a couple of
a couple of different things. First off, it's really interesting.
You know, I've got two boys. Kyle's going to be
nineteen here in about a week and Logan's twenty three,
and you know, they're on their phones. We all enjoy
looking at the memes. We've talked about these a lot.
(28:58):
But it's interesting now because whenever these videos pop up
and we're sharing them with one another, there is a
debate that takes place between us if it's not labeled
as such.
Speaker 1 (29:09):
You know, I wonder if that's AI. I think it's A.
I don't think it's A.
Speaker 2 (29:12):
It's funny how it's entered into you know that aspect
of our enjoyment of just entertainment items like that. It's
just now a part of the of the conversation. I
have one more thing I want to add, is there
anything that you would like to mention to or you
want to say to that to that particular observation.
Speaker 3 (29:30):
The the interesting thing like that observation and the main
thing I thought about is how the way you deterribine
if a video or a beam is AI or not
is similar to what people should do to keep themselves
safe well, trying to do holiday line and try to
avoid falling for scams, which is double check, yeah right,
(29:54):
like just like you could research is a beam legit?
I always before using a shopping service to haven't used before,
I'm going to double and triple check to make sure
it's LEDGECT.
Speaker 2 (30:07):
I'm gonna send you a link to a video once
I got some time this morning. There's a there's a
YouTube channel. It's a special effects house run by a
group of millennials in southern California.
Speaker 1 (30:19):
It's called the Corridor Crew. I don't know if I've
ever heard of them before, but I happened.
Speaker 2 (30:24):
They analyze movies and CGI in movies and things of
this nature, and they've been on top of AI. They
did a video recently and again I'll share with you
wherein they went online and they ordered products that could
be purchased at an incredibly cheap price, but they looked
like high price items. So like a Carhart jacket. We've
(30:45):
all seen the images of these cable knit sweaters with
these expansive designs on them like you've never seen before.
Speaker 1 (30:53):
They looked three D.
Speaker 2 (30:55):
So they went and ordered all these products to see
how closely it matched the obvious AI creation as it
was being presented online. And of course the result is
what you would expect the result to be. These were
incredibly cheap, flimsy items that were just knockoff products that
ended up ultimately looking.
Speaker 1 (31:14):
Similarly enough online.
Speaker 2 (31:16):
But nothing close to the quality that you would expect
it was. It's incredibly entertaining and I'll uh, I'll share
it with you when I get a chance here to v.
Speaker 3 (31:24):
That's so interesting, right, Like it's interesting because they're clearly
online scams of fake items and the knockoff items. And
what's interesting is that you know they were able to
purchase them and they were shipped and received. Yeah, right, Like,
you know, I typically what I think is that if
if I'm buying a knockoff, if I'm buying it online,
(31:48):
I'm not going to expect people to ship it to me,
it's really interesting that they shipped.
Speaker 2 (31:52):
Yeah, they got the products and they're close proximities, but
they are not. I mean, essentially, they end up like
the clothing ends up being of the quality that you
would get it like a halloween like a cheap Halloween costume.
Speaker 1 (32:02):
That's what the That's what they ended up being. So
I'll share that with you.
Speaker 2 (32:04):
Let's get back to a couple of the questions that
the audience has with for you. I actually had one
that came in from front of the show Rose mount Rick.
It's a really quick one. He just says, David, do
you talk to GROC every day? Do you try to
teach Grock?
Speaker 4 (32:18):
No?
Speaker 3 (32:19):
So I use I use chat GPT, and there's no
particular reason other than that's the one that I first
started using. I used CROC once in a while, but
chat GPT is the system I most hang out in.
Speaker 2 (32:34):
All right, let's go to the iHeart Radio app this
morning as we continue our discussion with AI analyst de
Vid Gartenstein Ross.
Speaker 5 (32:41):
Good morning, John.
Speaker 8 (32:42):
Can you please ask d g R about what the
differences between co pilot or chat GPT and an agent
created within co pilot or chat GPT.
Speaker 1 (32:53):
Thanks DVD.
Speaker 3 (32:56):
It's a it's a wonderful question. The quick answer is
that you know Copilot or chat Gypt. The main purpose
of it is to have a conversation with you. An
agent will actually perform tasks for you, and for that,
Copilot is much much better than chat Gypt because it's
(33:18):
integrated into the Microsoft Office suite, and so if you,
if you're a Microsoft person, Copilot can integrate any sort
of agentic function. Well, I don't use agents that often
in AI. I imagine though, that you know, a year
or eighteen months from now, I will use agents AI
(33:42):
agents multiple times a day. They're an important tool that
most people will be using within the next twelve to
eighteen months.
Speaker 2 (33:51):
All right, let's go here and talk with friend of
the show, Raquel, who has a question for David garten
Stein Ross.
Speaker 9 (33:57):
I have a question for deb Garden Santa Ross Ross.
A few weeks ago, maybe one or two weeks ago,
we played a piece of music, and you said that
you could tell it was written by AI.
Speaker 4 (34:09):
My question is about comic book art.
Speaker 9 (34:11):
My husband is a comic book artist, and comic books
can take weeks to complete. My husband's fear is that
AI will completely replace all comic book artists, because you know,
I could do it in a matter of minutes. I
would just like your opinion. Thank you Dad, Thanks Keel.
That's a it's a huge concern.
Speaker 3 (34:34):
AI does comic book art really well, and just most
kinds of art really well. Right now in certain industries,
like I work in the games industry as as folks know,
and for major board games, there's this stigma against using
AI and a preference for human artists.
Speaker 4 (34:54):
But that stigma is eroding.
Speaker 3 (35:00):
The concern is a real one, and so I'll provide
what I would do as an artist understanding that threat,
and that is I would actually it to hedge against that.
I would start using AI once in a while, not
(35:21):
in my art for customers, but to play around with
creating my own comic book art, because yeah, like what
Raquel is outlining is real, Like it's a huge problem.
I would definitely feel threat there. But there's going to
be things that comic book artists just know implicitly that
they can do more quickly with AI. That's actually, in
(35:45):
my view where job the security lies. And we can
talk about that at some length. But like Raquel's husband
definitely has worried to be concerned there.
Speaker 4 (35:55):
I definitely share the concern that she has Davdi.
Speaker 1 (35:58):
Guards and the Stein Ross.
Speaker 2 (35:59):
I'm going to hang on to the other articles that
we had planned to talk about today, but I really
enjoyed the conversation and the fact that many of the
listeners had a chance to get their questions answered. And always,
I appreciated the time that you give us every single week.
If you have further questions for David, you can always
email me Justice at iHeartRadio dot com. I'm happy to
forward those over to DVD David. If there's an email
(36:21):
that you would like to share or any place that
you would like to go and direct the listener before
I let you go this morning.
Speaker 4 (36:26):
Absolutely, My email is dav'd.
Speaker 3 (36:28):
That's dav Eed at Expert Theory dot com.
Speaker 2 (36:34):
David Gartinstein Ross. As always, thanks for the time, my friend.
I look forward to hopefully doing it again next week.
Speaker 4 (36:39):
Likewise, John, so do I all right?
Speaker 2 (36:42):
Cabinet meeting yesterday a lot took place with President Donald Trump.
Lengthy cabinet meeting, incredibly transparent this administration is. During that
cabinet meeting, one of the things that was talked about
was snap Secretary of Agriculture Brook Rawlins deliveredy sweeping update
on the nation's SNAP program, revealing what she called rampant
(37:04):
fraud and announcing a dramatic escalation in the administration's effort
to force state compliance in federal oversight. You may remember
that Minnesota was one of the states that was pushing
back on this. Rowland said that under the president's tenure,
about eight hundred thousand Americans have already moved off of
SNAP and the program is serving forty two million people.
She argued the trend underscores the need to restore integrity
(37:28):
to the system that she's said has been abused for
far too long. With that, before we head into next hour,
let me share this with you. This is USDA Secretary
Rawlins saying that she will be moving to halt federal
funding two states who refuse to turn over the SNAP
data to help root out the fraud, like Minnesota. And
we'll get to more of your talkbacks coming up in
(37:49):
just a moment here on Twin Cities News Talk. Here
is the USDA Secretary Ice.
Speaker 10 (37:55):
We had a couple of people receiving benefits in six
states in February of this year. We asked for all
the states for the first time to turn over their
data to the federal government to let the USDA partner
with them to root out this fraud, to make sure
that those who really need food stamps are getting them,
but also to ensure that the American taxpayer is protected.
(38:15):
Twenty one states said yes. Twenty nine states said yes,
not surprisingly the Red states, and that's where all of
that data that fraud comes from. But twenty one states,
including California, New York, and Minnesota, the Blue states, continue
to say no. So as of next week, we have
begun and will begin to stop moving federal funds into
(38:38):
those states until they comply and they tell us and
allow us to partner with them to root out this
fraud and to protect the American taxpayer.
Speaker 4 (38:46):
As Joe Biden was.
Speaker 10 (38:47):
Working to buy an election a year ago, he increased
food stamp program funding by forty percent. So now as
we continue to roll that back, so.
Speaker 2 (38:58):
Keep in mind, just really quick, keep in mind that
not only has billions of your dollars taxpayer money gone
up and smoke with fraud taking place under the Walls administration,
Walls now has put the state in a position, whether
it's because of what was mentioned there by the Agriculture
(39:21):
Secretary of losing funding for the lack of willingness to
send over snap data.
Speaker 1 (39:27):
Millions of dollars in funding could be lost.
Speaker 2 (39:29):
There some thirty million dollars in funding over drivers' licenses
for those who are here illegally. We could also lose
that funding as well. And it's all happening under Governor
Tim Walls's watch.
Speaker 1 (39:43):
He needs to.
Speaker 2 (39:43):
Resign, and as I keep saying, I do not expect
that he is going to be the nominee for governor
heading into next year. More coming up on Twin City's
News Talk AM eleven thirty and one to three five FM.
Speaker 4 (39:53):
I'm ready to get beat.
Popular Podcasts
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
Are You A Charlotte?
In 1997, actress Kristin Davis’ life was forever changed when she took on the role of Charlotte York in Sex and the City. As we watched Carrie, Samantha, Miranda and Charlotte navigate relationships in NYC, the show helped push once unacceptable conversation topics out of the shadows and altered the narrative around women and sex. We all saw ourselves in them as they searched for fulfillment in life, sex and friendships. Now, Kristin Davis wants to connect with you, the fans, and share untold stories and all the behind the scenes. Together, with Kristin and special guests, what will begin with Sex and the City will evolve into talks about themes that are still so relevant today. "Are you a Charlotte?" is much more than just rewatching this beloved show, it brings the past and the present together as we talk with heart, humor and of course some optimism.
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.