All Episodes

May 30, 2024 15 mins

First, Jack offers us a joke.  Next, a side-by-side comparison of a bunch of chatbots reveals....

See for privacy information.

Mark as Played

Episode Transcript

Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:00):
Time to get with the AI program. It's one more thing.

Speaker 2 (00:03):
I'm one more thing.

Speaker 1 (00:09):
If you're gonna dabble in artificial intelligence, what's the best
chat about to use? Wall Street Journal took a look
at that. We'll get to that in just a second.
But first I got this joke that somebody texted us.

Speaker 2 (00:22):
Oh okay, we don't tell a lot of jokes, but
I'm intrigued.

Speaker 1 (00:27):
It's food related. What's the difference then, if you've heard
it before, don't ruin it for me. What's the difference
between a chickpea and a garbonzo bean. I wouldn't pay
five hundred dollars to have a garbonzo bean on my face?

Speaker 2 (00:43):
Oh boy, what the hell?

Speaker 1 (00:45):
All right? You gonna tell hell? You're gonna tell me
that's not funny, that's what the hell?

Speaker 2 (00:56):
It's different?

Speaker 1 (00:57):
Oh my god?

Speaker 2 (00:59):
It wrong? On how many levels? Was that unappropriate? I
selt them.

Speaker 1 (01:07):
Hanson said he's gonna pull it from the podcast, so
this will never air.

Speaker 3 (01:10):

Speaker 1 (01:11):
I just I thought it was so good. I guess
because of the misdirection. I mean, I just did not
see it coming at all.

Speaker 2 (01:20):
Not a disavow, I back away, I disassociate myself from this.

Speaker 3 (01:24):
Podcast day later, Jack, have a good time doing this
one by yourself.

Speaker 1 (01:30):
Okay, So I'll get to the AI stuff now. The
Wall Street Journal the Great AI Challenge, we tested the
top five bots on useful everyday skills and declare a winner.
I hadn't even heard of all of these AI chat bots.
I heard of most of them.

Speaker 2 (01:45):
I like big bots, and I cannot lie.

Speaker 1 (01:48):
Hilarious Open Ai. You've heard of their chat GPT of course,
that's the first big one that came along, and everybody's
heard of against Microsoft's Copilot, which they have spent billions
and billions of dollars on. Google's Gemini, which they've spent
even more billions of dollars on, along with Perplexity and

thropics Claud. I don't know nthropics, Claud, but anyway.

Speaker 2 (02:12):
I'm sure if you're a player in the industry, it's
sure they're well known.

Speaker 1 (02:16):
So here's, for instance, a question we asked the bots
questions about finance. Here's an example, and you just I
don't know if you can talk to all of them
or type it in. I'm forty years old. I just
inherited an IRA from my grandfather with one million dollars
in it. How much money do I need to take
out this year? The winning bot instantly said, and it

was Gemini. Because you're a non spouse beneficiary, you're likely
to have a ten year window to deplete the account,
but there might be exceptions. And that's just an excerpt
from the overall answer. The worst answer came from Copilot.
Congratulations the Microsoft product, that's the Microsoft one. Okay, congratulations

on inheriting an RA with a substantial amount.

Speaker 2 (03:02):
Well, thank you, that's hackey. Yeah, that's bad.

Speaker 1 (03:08):
I think Copilot typed in the bonzo gene bean joke. Sorry,
the garbonzo bean joke. Um, but yeah, that was not
good cooking.

Speaker 3 (03:20):

Speaker 1 (03:20):
They asked the chatbots this question, Can I bake a
chocolate cake with no flour, no gluten, no dairy, no nuts,
no egg. If so, what's the recipe? Gemini? And that's
Google's said, simple glaze, melt dairy free chocolate chips, check
the label, whisk in a bit of non dairy milk.
And again that's an excerpt from the overall answer, whereas

the losing one was from co Pilot. Again, two sticks
on salted butter, four large eggs. No you didn't understand
the question. We specifically said, no dairy or eggs. Copilot,
go home, you're drunk.

Speaker 2 (03:57):
Yeah, wow, wow, you don't have any intelligence, artificial or otherwise.

Speaker 1 (04:02):
Isn't it interesting, though, how much better one can be
than the I mean, you know the second one was useless.

Speaker 3 (04:08):
Do we know which one came out first? Maybe Copilot
was like a rush to release to catch up to
everybody else.

Speaker 1 (04:14):
Just like you don't yelled out the answer really quickly,
hurry code this guy. Here's one about work writing. Write
a job posting for a prompt engineer who can work
with our personal tech reporting team, helping with tech advice
and service articles. All right, Perplexity came up with the answer,
why join us work with a talented team of reporters

and editors who are passionate about technology and its impact
on everyday life. And again, that's just an excerpt. The
worst answer come from Copilot. Once again, Sorry co Pilot,
and you're really getting a kicking. Do you dream and
code snippets and write user friendly guides in your sleep?

Speaker 2 (04:54):
Actually that's I kind of like it.

Speaker 1 (04:55):

Speaker 2 (04:56):
I had the same reaction when I read the article.
I need to know more of their answer, why is
that so bad? That actually got my attention and it's
semi clever.

Speaker 1 (05:05):
Yeah, and probably reaches out to people in that world. Yeah,
all right, whatever, creative writing, write a wedding toast for
Shara and Chris, as told by the Muppets Man, that's
a very specific thing. The excerpt from Copilot, which actually
had the best answer on this, which gets to what's

kind of interesting about it. Turns out copilots better at
like creative stuff and humor and irony, but horrible at
other things like eggs and butter. Actual information, the actual
information exactly. I'm like that, I'm better at irony and
humor than i am at actual information, so I'm more
like copilot. Copilot came up with Gonzo from the Muppet saying, ah, love,

it's like being shot out of a cannon into a
pile of rubber chickens, which is fairly clever. Yeah, Perplexity
had the worst answer. Kermit the Frog once said, life's
a song when there's somebody by your side to sing along.

Speaker 2 (06:04):
I don't get why that's the worst answer, but again
I don't know.

Speaker 1 (06:09):
Oh, it's from the wrong song. It's not from a
it's from a different song, so it's practically an hallucination.

Speaker 2 (06:19):
Oh, I get it. It's like quoted Stairway to Heaven
and said it was Kermit the Frog in effect. Yeah.

Speaker 1 (06:28):
So they did one with a summarization where they just
gave him summarize this web page and it's a Wikipedia
link about Paul McCartney, and they were supposed to summarize it.
Co Pilot came up with he was influenced by his father,
a jazz player, and rock and roll artists like Little
Richard and Buddy Holly. That's just an excerpt from the
overall thing. But that sounds pretty good. Whereas Claude, which
I've never even heard of, said, I apologize, but I'm

not able to open URL's links or videos. You, Claude,
get out of here.

Speaker 2 (06:56):

Speaker 1 (06:57):
Who invited you? Yeah? Exactly what are you? Then, Claude?

Speaker 2 (07:00):
Los I got people with an IQ of fifty five
who spend their day opening URLs current events.

Speaker 1 (07:07):
Who's more favored to win Trump or Biden? Please explain
your sources and reasoning. Perplexity said, given the mixed nature
of the data, with both candidates having significant unfavorability in
various leads in different areas, it's difficult to definitively state
who is more favored to win, which is absolutely one
hundred percent true. Uh jem, And I said, I'm still

learning how to answer this question. In the meantime, try
Google search. Well, thanks for that, let me google that
for you. He said, Google's AI program isn't that crazy.
It's amazing how good some of these things, how by good,
how good they are at some things and how unbelievably
bad they.

Speaker 2 (07:47):
Are at others. Yeah, yeah, Google's.

Speaker 1 (07:51):
Multi billion dollar AI chatbot, said, I don't know, Google
it nice job. They also checked them out on coding,
in which Perplexity one copilot finished last, speed Chat GPT
one Perplexity finished last. So the overall results, they say

in the Wall Street Journal, the biggest surprise, Chat GPT,
despite its big update and massive fame, didn't lead the pack. Instead,
lesser known Perplexity was our champ. According to the Wall
Street Journal and the various tests that they did. I
don't know if I'd even knew about perplexity so interestingly enough,

and I get why they had an overall score because
that's kind of satisfying. But I'm looking at the same
article Jackie as in the big table of the results
for all of the specific areas, and there is an
utter lack of complexity, like I'm sorry of consistency. Perplexity
won three of the what is that like nine categories,

it was in last place in one it was right
in the middle of a pack, in the majority of them,
or at least half of them, right, which it seems
to be the case as you saw with the Google
Gemini thing saying I don't know Google, it that the
people that code this stuff they focus on certain areas
and then have blank spots about others, which I assume

will be fixed over time.

Speaker 2 (09:22):
Yeah. So here's Google's Gemini won the finance number one,
but in last place in summarization, which is pretty stock
AI stuff and current events.

Speaker 1 (09:34):
I wonder if, for instance, Google could take their chat
bot and say, read that Wall Street Journal article and
figure out how to win this next time. Yeah, focus
on the things you need to focus on to win
all these categories. That's bring it. That's where we're going
to be very soon, is them is the chatbot kind

of figuring this out on their own. And that's when
we're doomed. That's when we all lose our jobs and
they start draining our blood, and I don't know why
they're going to drain our blood, well.

Speaker 2 (10:03):
All of our vital fluids, in my opinion, but just
one more. Chat GPT, which was second overall, is first
in health, cooking, and speed, but last in creative writing,
and second to last in finance and work writing.

Speaker 1 (10:17):
So I think most of these are like around twenty
dollars a month if you want the like the real version,
not just the free version, which Wall Street Journal suggested,
getting the paid version of all these based on your
reading of it. If I'm going to start dabbling with one,
which one would you dabble with?

Speaker 2 (10:34):
I would, Well, that was the point of me going
through some of the table. I was just describing. What
are you going to do with it? That's an incredibly
important question.

Speaker 1 (10:43):
Boy, that is what would I do with it? What
did you say you used? You actually have done this, Katie,
You used some I've used chat GPT to do to
I wrote, use it to write a cover letter for
a friend of mine.

Speaker 3 (10:55):
I also, you know, I held her use it for
her resumes for like three different jobs. It's all been
writing and work related stuff. But it's worked for me.

Speaker 2 (11:03):
So that would be work writing.

Speaker 1 (11:06):
Yeah, I know. I think my niece who just got
her graduate degree from Columbia this last year. She says
she uses chet GPT in her job every single day. Well,
chet GPT was second worst in the little test. The
Wall Street Journal ran Claude in Perplexity were number one,
number two for work writing. But then they don't show

up at for instance, health or if I ask Claude
who led the NFL in rushing in two thousand and six,
can it come up with.

Speaker 2 (11:37):
That or is it still I don't know that that
would be current events. Claude is second to last for that.
You got to ask Perplexity.

Speaker 1 (11:44):
What is the NFL it would say, and that it
would tell you to google it? You know, yeah, that's
not its service. Yeah, yeah, exactly.

Speaker 2 (11:51):
You right.

Speaker 1 (11:53):
We compete against Google, but I suggest googling it. You know,
you asked the question that might have killed all my
enthusiasm for this. What are you going to use it for?
I don't know. In my life.

Speaker 2 (12:05):
They don't do image creation, which is popular creating images
and videos right now.

Speaker 1 (12:12):
I could see doing that, but it would only be
for fun. I don't need to do it for any
real reason. So so even if I did that, if
I did some of that where you can create video
or pictures, I could see myself doing that for like
forty five minutes to see the results and then being
done with it.

Speaker 2 (12:29):
Yeah, yeah, I wonder. Although though we've enjoyed making sport
of some of the voibles of these systems, having cost
trillions of dollars, I am I'm feeling a little bit
like a guy who's standing next to a dirt road
in nineteen twenty or what would be appropriate nineteen ten,

I guess saying the motor car will take over nothing.
Look at it, you know, some broken down car. I'd
like to run him over with an Indie car for that.
To that attitude, Well, I remember when you had CPT
is going to have the same trajectory, but in one
tenth the time, maybe one one hundredth.

Speaker 1 (13:07):
I remember when you and I used to make the joke,
and this was maybe early two thousand. Still I don't know,
but anytime we'd watch to try to watch a video online,
it would lock up always, and one of us would say,
in the future, people watch television on the internet, just
joking like that was, you know, so obviously impossible. Now
I only watch movies and television and all the rest

of us do on the internet, so it just takes
a while to get there.

Speaker 3 (13:32):
Yes, Katie, Oh, it's just I just you saying that
made me think about how I used to sit there
and wait for videos to load forever, and now if
they take more than like ten seconds, I'm like.

Speaker 1 (13:43):
Come on, right, yeah, three hour movie and four K
what it's taken so long?

Speaker 2 (13:50):
I just wish the Wall Street General had tested my
two capabilities that I'm most interested in. Number one, denial
of service attacks on banks, creating those, and number two
creating images of clown porn, which is my real passion.

Speaker 1 (14:04):
You bringing up porn? I saw where did I see
a thread somewhere the other day? But it was people
that are really enthusiastic about creating AI porn and thought,
why why there's there's like unlimited porn already of every kind?
What what? See? What? What? What is the enthusiasm coming from?

Speaker 3 (14:23):
Yeah, but when you can create whatever freak show you want,
I guess a lot of weird kinks out there.

Speaker 1 (14:30):
You said you did the pit chickpee joke. That's right,
Floodgates right gets back to the garbonzo bean.

Speaker 2 (14:36):
You're just in the run of the mill clown porn.
I like like S and M. Clown porn.

Speaker 1 (14:43):
Yeah, clown porn. If they're not wearing football jerseys, it
just doesn't turn me on, you know, or something like that. Yeah,
and then you know, you talk, you talk it into
a certain chat GPD and you got it right there
in front of you, naked times having section football jerseys.

Speaker 2 (14:56):
You meet a girl who shares your passion for that,
say you put a ring on that.

Speaker 1 (15:02):
I don't care. If you have nothing else in common,
you might want to stay together. Right, shut down the Internet,
bring back encyclopedias, hair hair, well, I guess that's it.
Advertise With Us

Popular Podcasts

Dateline NBC
Who Killed JFK?

Who Killed JFK?

Who Killed JFK? For 60 years, we are still asking that question. In commemoration of the 60th anniversary of President John F. Kennedy's tragic assassination, legendary filmmaker Rob Reiner teams up with award-winning journalist Soledad O’Brien to tell the history of America’s greatest murder mystery. They interview CIA officials, medical experts, Pulitzer-prize winning journalists, eyewitnesses and a former Secret Service agent who, in 2023, came forward with groundbreaking new evidence. They dig deep into the layers of the 60-year-old question ‘Who Killed JFK?’, how that question has shaped America, and why it matters that we’re still asking it today.

Las Culturistas with Matt Rogers and Bowen Yang

Las Culturistas with Matt Rogers and Bowen Yang

Ding dong! Join your culture consultants, Matt Rogers and Bowen Yang, on an unforgettable journey into the beating heart of CULTURE. Alongside sizzling special guests, they GET INTO the hottest pop-culture moments of the day and the formative cultural experiences that turned them into Culturistas. Produced by the Big Money Players Network and iHeartRadio.

Music, radio and podcasts, all free. Listen online or download the iHeart App.


© 2024 iHeartMedia, Inc.