Hater Season: Cal Newport on AI Reporting

Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:02):
As media your witness to a great becoming. It's better
offline and I'm ed Zetron. Today we're joined by computer

(00:23):
science professor and tech writer Caln Newport for Hater Season.
Cal thank you for joining me.

Speaker 2 (00:29):
Always happy to do some heding.

Speaker 1 (00:30):
I suppose, well, you had an excellent YouTube that I'll
be linking in the show notes about the mistakes in
AI reporting. Though I would my hater in me says,
I don't think these are mistakes, but you really you touched.
One of the best videos I've seen on AI report
or AI in general was what you did. But it's
basically this thing of like the digital iic that these

(00:51):
stories are meant to make you feel uncomfortable, and of
course this full astonishment thing. You know what, Just run
the tape. Tell me a little bit about what the
bits that you found, because I watched it going yeah, yeah, yeah,
like an angry person.

Speaker 2 (01:05):
Yeah I had that. I could imagine you. When I
was recording that video, I was like, I bet ed
is well now the great time exactly? Well, I mean,
let me just give the context, right, So I'm in
an interesting situation for observing this because you know, I
am a computer scientist, so I'm not afraid of the technologies.
I'm happy to talk about transformers and feed forward networks

(01:27):
and diffusion models and like, that's not that scary to me.
But I also write about technology. My main journalistic home
is to New Yorker, where I write a lot about
you know, I do AI journalism there, so I'm up
on that as well. And so I'm often noticing things
in journalism. When I see AI coverage that is faulty,
it really catches my attention because I have a foot

(01:49):
in both of these worlds. So there's a lot of
good AI report out there. There's also a lot of trash,
and I really wanted to help people figure out how
do you sort it? Like how do you figure out
if you're reading something? Should I pull the ripcord on
this article? Like this is not helping me, Like how
do you know if it's good or not? So I
was like, Okay, here's what I'll do is I'll come
up with the three most common traps I see in
AI reporting that makes me want to, you know, throw

(02:10):
my iPad at the wall. And I came up with
three and I'll just give you the three names. I
made up. All these names don't. I don't think they're great.
My producer thinks he loves them. But here we go
vibe reporting that's number one. So that is where you
will omit certain facts and put loosely related quotes next
to each other in a way that creates a general

(02:33):
vibe that you want to be true but it's not
quite true. So you don't actually make a concrete claim
that's not true, but you imply that claim by what
you omit or what you put in your story, or
what quotes you put next to it. So you'll put
a quote, for example, about layoffs at the gaming division
at Microsoft, next to an unrelated quote about concerns about

(02:56):
AI and it's impact on jobs. Now you have the vibe, oh, man,
all these people just got laid off because of AI.
AI is taking jobs, where in reality the layouts had
nothing to do with AI. But you put these things
next to each other, you give that sense. Then I
had a mining digital ick. So to me, that's any
AI story where you take an example from the edges
of AI, like something that wireheads and San Francisco are
up to, and you just tell a story that's unsettling

(03:19):
without talking about any of the technical details like well,
what's different here? Was there a technical innovation we need
to know about, and not discussing any concrete implications. Oh,
this means this is going to change in the future,
or it's going to have an impact on this sector.
You're just telling a story to unsettle, and I think
a lot of the coverage of molt Book and open
Claw fell into that. And then finally this astonishment, which

(03:40):
is more of a YouTube phenomenon than a print journalism phenomenon.
But that's where every single thing that happens in AI
is insane, amazing, terrifying. Everything is going to change, and
so you constantly create this atmosphere of something seismic just happened,
so that the consumer of the infra ends up in
a bit of a panic, like I can put my

(04:02):
finger on exactly what's terrible, but like everything terrible is happening.
Those three traps, to me, should be automatic rip chords
from what you're reading or watching.

Speaker 1 (04:11):
The only disagreement I'll have is that you would say
that this isn't the majority of AI journalism, because I
actually argue it would be, especially that kind of the
beginning one. The vibe reporting is very common because The
big one right now I'm seeing is this AI software
AI is replacing software thing. If you are a reporter
and cal is not making this statement I am. If

(04:32):
you're a reporter and you bring up anthropic cowork around anything,
you are wrong. I was about to call you a name,
but I'm being nice today for some reason. All Right,
you're a dip shit, because you are a dip shit.
If you look at claude cowork, which is a thing
for fucking around on your desktop, and you say this
is going to compete with Salesforce, you just don't know

(04:53):
what you're talking about. You are wrong. But then one
abstraction higher is this idea that claud code is going
to destroy SaaS software as a service, and the idea
being that software as a service is this thing that
people are just going to stop building their own CRMs.
They're going to stop building their own per seat software things,
but they're going to build it internally. This just I

(05:13):
don't know if you've seen this, cow It just reflects
a complete lack of understanding of how software works. Yeah,
because you don't pay Microsoft for Microsoft three sixty five
teams because you can't build your own word or what
have you. I don't think you could. But nevertheless, it's
also because they maintain it, because they make sure it
doesn't break, or if it breaks, they fix the bit.
So they make sure that it stays up all the time.

(05:37):
They make sure it's accessible that has secure log in.

Speaker 2 (05:40):
Well, look, there's a few things going on here. One
of these things I reported on last month. I did
a big New Yorker piece on agents. Right, so this
is relevant to the cloud code and how people are
thinking about the current future, and there was this basically,
here's what seems to have happened. Cloud code and these
other command line interface agents can do really cool things,

(06:00):
and I'm using the work cool here in a very
careful fine, yeah cool.

Speaker 1 (06:05):
I would like you to be like completely like give
people the actual explanation here.

Speaker 2 (06:09):
Yeah, so cool, meaning like Oculus. Right, you put on
the the Oculus visors for the first time. Everyone had
the same reaction. This is really cool. Now that's separate
from that's a trillion dollar business, Like, let's put a side,
but this is really cool. I'm seeing three D in
a world where it tracks my head. Cloud code and
other command line interface coding tools became like that. For programmers,

(06:32):
it was like really fun the watch it doing multi
step execution of the construction of demos or this or that,
and the reason why it could do those cool demos
it was sort of well suited for that world because
that's a world that exists only in text. So cloud
code works on a command line interface. It's all text based,
and it works with a filesystem. You can write files,
edit files, sind files to compilers. It's all exists in

(06:54):
a small number of commands on a command line interface.
It's all in a world of text, because that's where
computer programs are built. That's a perfect case for lms,
which love dealing with text, and they love dealing with
structure text like computer code. There was an extrapolation that
then happened, and really this caught on January of twenty five,
which is where you first began to get this sentiment

(07:15):
of oh, it's doing such cool things over in the
world of command line interfaces in code. Certainly these agents
can now soon do similar cool things and like all
different things we do on a computer, and that is
what laid this foundation of people were so impressed, programmers
were so impressed by the coolness of what was happening
with cloud code and the other command line interface tools

(07:38):
that they extrapolated that vibe over to other computer usage.
Is the point of that article I reported was, Oh,
it turns out like everything else we do on computer
is much harder. It's not six text commands on a
command line interface creating structure code that you can compile
and test to see if it compiled or not. It's
much messier. The interfaces are visual. I don't realize what

(08:00):
complexity goes into the things we click and select doing
something as simple as even trying to just book like
a hotel in a new city or something like this.
And if you use a language model as to underline
logic and decision engine of making actual actions in the
real world, well, the language models make things up and
get things wrong or a little bit wrong twenty percent
of the time.

Speaker 1 (08:19):
And a little bit wrong in computer science means breaking everything.

Speaker 2 (08:23):
Well, it means a lot like it's okay in code
because they say, oh, that didn't compile, let's try again.
But when you're booking a hotel room, as I sort
of detailed that article, it means like you ended up
in the wrong city two years from now, and the
room costs six thousand dollars, right, and so it just
didn't work. So twenty twenty five was supposed to be
the year of the Agents and it just didn't work
and they don't really know how to fix it. But

(08:44):
we're still vibes. This is Vibe reporting. Yeah, but cloud
code is like really exciting, and so why can't we
do that with everything else in the computer. It's actually
a much harder problem than they were letting on.

Speaker 1 (08:55):
Well. The thing is I've seen, especially like in twenty
twenty six, I've seen a lot more cloud code stuff.
Was there's been a very big consent manufacturing operation going
on right now. Wall Street Journal, Atlantic, CNBC Didrew Bosa.
This is a statement from me, not cow did you
both from CNBC should be fucking ashamed of herself going
on CNBC every day just going cord Code's got a
destryal software. She's on Twitter because she was able to

(09:18):
vibe code a some sort of Monday clone, which is
just like a it's a project management tool. It's like
I made some software that worked. This is everything now.
But that's kind of what you've been talking about with
this Vibe reporting, where it's like I did a small thing.
Now all things will be done in this manner, whether

(09:38):
it's possible, God no, God no. But she she, like
many reporters, are able to find a lot of people
who are invested in AI, who will absolutely go on
TV and say, Yep, it's completely true, that's gonna happen
one hundred percent. It's just so strange because it makes
me feel paranoid and kind of conspiratle when you look

(09:59):
at so the majority of news about AI, and it
is this viper POI. It's these vast extrapolations from sixteen
thousand job losses Amazon. They mentioned AI. This plus this
equals the AI is replacing people. It's just so it
makes me feel like uncomfortable with the world.

Speaker 2 (10:19):
Yeah, well, can we sit for a moment on that
Amazon example, because I think it's a great please please,
it frustrates me that one frustling. All right, So Amazon
lays off sixteen thousand people, right, all right, it's covered
in two different ways, so the vibe reporting way it's
covered I in my newsletter, in my podcast, I looked
an example from Quartz and it was covered as clearly

(10:41):
intended to imply Amazon laid off sixteen thousand people because
of AI. They're being replaced by AI. They put the
subhead of the article was the CEO of Amazon talking
about how AI is going to increasingly disrupt the job force,
and then the article itself, no alternative explanations are given

(11:02):
for these layoffs. They kind of just give the details
of like here's how many people are laid off and
here's where they figured it out, and then they put
a couple of quotes in there about AI being very
disruptive and being able to automate jobs. It turns out
those layoffs had nothing to do with AI, and you
can find other reporting that focused because it was reporting
that was for the financial market, so it was trying
to focus on what the hell is actually happening, and
the deeper reporting was like, yeah, they laid off a

(11:22):
bunch of managers because like a lot of tech companies,
they over hired during the pandemic because cloud computing became
much in demand during the pandemics, So a lot of
tech companies over hired during the pandemic and they're all
shedding those jobs again. And Amazon's pretty ruthless about this, right,
They're always looking for excuses to fire people, and they said,
we have too much bureaucratic bloat, there's too many managers.
We're going to fire a bunch of these managers we

(11:44):
hired so that we can be more lean. Again, that
has nothing to do with AI. In fact, this is
the second or third round of these firings that they've
planned to do, the first round occurring before chat GPT
was even released. Like, this has nothing to do with
jobs being replaced by AI. But then you can vibe
report it because like, well, technically speaking, Amazon also is

(12:04):
investing money in AI products, So technically speaking, money saved
by firing these managers could be respent on AI. So
you can say with a like a semi straight face,
they fired people because of AI. But clearly, you know,
the impression that you're giving to the reader is that
they were replaced by AI and it had nothing to
do with it. And I heard, by the way, so
I wrote about that. I put in that video. I've

(12:25):
heard on background from multiple Amazon executives since they were like,
we were completely baffled by this coverage that was implying
that people were being fired here to do with AI.
This is just what we do at Amazon. We're ruthless,
Like if you're not earning your keep, we fire you.
They were completely baffled by that coverage and like things
for pointing it out, like what.

Speaker 1 (12:43):
I go to be honest, cal if I got an
email like that from an Amazon executive, A tont go
fuck themselves? And I mean this nicely because Andy Jesse
last year June seventeen, twenty twenty five, the whole thing
about today and virtually every corner of the company we
use in generate if I had to make customers lives better.
I believe Amazon benefits from that obfuscation, and I think
they deliberately fuel it. Now that may be executives who

(13:06):
disagree with.

Speaker 2 (13:06):
It, well, these were lower level managers, right, so not executs.
Shouldn't say those like people who work there that were
like oh yeah, no, no, no, they're not firing people
because like AI, they're just being brutal.

Speaker 1 (13:16):
Right, right, they're the people who tell the truth.

Speaker 2 (13:18):
Yeah, exactly exactly, But no, you are right. I think
for the there has been a lot of vibe reporting.
Basically the end I've been covering this for the last
two years, the entire sequence of job reductions that were
post pandemic corrections, so the entire tech industry over hired
during the pandemic. The entire tech industry cuts ups in
the last few years because they hire too many people
and now they have to correct back to where they

(13:41):
were pre pandemic. Consistently, across the board, these cuts are
vibe reported is due to AI consistently. You see exactly
that story. We see it, and I'm a computer scientist.
We see this in coverage of computer science majors as well.
Same idea. Computer science majors historically directly tied to the
tech industry. If they're cutting in a downside, majors go down.
If they're hiring. I mean, it's just economic. It's not

(14:03):
surprising when the technistry is booming, we get a lot
more majors because they're good jobs, right, And so computer
science majors went down in the last year or two
as the tech company started cutting. That was reported in
the Atlantic as kids are not majoring in AI because
AI makes the the.

Speaker 1 (14:21):
Surprise majoring and kompsa.

Speaker 2 (14:22):
You mean yeah, because AI is going to do all
these jobs. Because we've had these cycles every five years,
we have this cycle there, this is nothing anyways.

Speaker 1 (14:31):
No, no, I'm and the Atlantic has just done a
piss poor job with this stuff. I'm I'm heighten and
I'm heighten on them because they had an awful claud
code thing. They just you know, I'm gonna bring it
up and read the title. I'm not gonna say the
reporter's name because I think that that's that's main. But
let's look at this move over chat GPT. You're about
to hear a lot more about Claude coding. Know why

(14:53):
you're gonna hear about that? The fucking Atlantic is writing
about it, and it's just over the whole days. Alex
lieberm And had an idea what if he could create
Spotify rapped for his text messages without writing a single
line of code. Liberman, co founder of the media outlet
morning Brew, created I message wrapped. I just want to
start here and say that guy doesn't do any fucking work,
and I know people who work there like I just

(15:14):
want to start by saying, if that's the best you've got,
and that probably involves throwing a giraffe and the entire
zoo into a vat of acid to make the GPUs move.
You're meant to read this, and the deliberate effect you're
meant to have is you're meant to be scared and
excited like it is a kind of a mishmash between
the vibe reporting and the I guess it's the folksting.

(15:38):
In fact, this might be a triple score because this
is meant to feel you, make you feel with discomfort,
but also make you freak up but also be excited,
a rare triple score. It's just stories like these pissed
me off because I'm fine with people going this could
do this. I don't mind it happens, But when it's
just like, hmm, feed me the slop right into my mouth, asshole,

(16:00):
make me feel, make me feel all the marketing things
all at once. First of all, I'm not impressed by
the idea of doing I message s grapped. That's not
That's not a thing a human being with that with
like friends and hobbies does. I've been I've been watching
the first season The True Detective. I've got many more

(16:21):
shows I could watch. Never once would I think what
if I could get it wrapped of all my messages?
What a fucking psychotic thing to do. But when you
read this, it's like yours. You spent your holidays with
your family, wrote one tech Polticy expert. That's nice. I
spent my holidays with Claude code.

Speaker 2 (16:38):
Well, it's fun.

Speaker 1 (16:40):
That's cool.

Speaker 2 (16:41):
It's fun to create, uh these sort of demo apps
if you're like engineering minded, So in that sense, they're
like most Yeah, the most cynical analysis here is that
cloud code is like model trains for engineering minded people.
It's a fun hobby I can put. Look at this
like I made a thing that can Like I was

(17:02):
talking to a friend of mine yesterday, computer scientist, and
you're like, oh, I built a thing where I can,
you know whatever, email an appointment to a thing that
gets parsed by an LM and then goes to my calendar.
That's just fun for him, in the same way that
someone else might be like, hey, I built a cabinet
and like it's really nice, Like look, I got the
wood to go together, and like I'm proud. I could

(17:22):
just on a cabinet or whatever.

Speaker 1 (17:23):
Except you build, Except you built something with your hands,
and a cabinet can have things put in it. It's
I just it feels like Lego almost, Like it's more
like it's toy software that has some functionality. Because the
big thing I'm waiting for with the vibe coding stuff
is an actual product, you know what I mean.

Speaker 2 (17:46):
But you know that doesn't go well. I mean, look,
this is the story of This is the story of maltbook,
right because.

Speaker 1 (17:52):
Oh yeah, tell me more about moltbook. We were just
talked about this.

Speaker 2 (17:55):
The whole can of worms to open there, I'm just
going to open like the top of the nearest can,
which is before even getting the what molt book is
and is not. You know, it was in the news
all the time. It was vibe coded and immediately was
just full of terrible security holes because it was vibe coded.
And it turned out that you could get the API key,
so the key you use when you access the paid

(18:16):
service to get used in LLLM, you have an API key,
so they know who to charge. You could just steal
everyone's API key who was using it because the guy
just vibe coded it and so no one was actually
there looking at the code. But what but what I
think is going on? Okay, So here's what I would
like to see. I would like to see more reporting.
That would be how are people using X Like to me,
that's very interesting, yeah, right, how are.

Speaker 1 (18:36):
People using agree?

Speaker 2 (18:37):
Yeah? And the problem is those answers right now. And
this is confounding. I think that people who are very
excited by the potential and coolness of these things in isolation,
the answer to how are people using X is often
not nearly as exciting as you would guess. And I
think the reality with I mean, I go quite have
my arms around cloud code. I do know there's a
lot of people who are building kind of like internal

(18:58):
tools or personal tools with it, which I think is cool.
Most people aren't interested in that, but for some people
they are. And you know, I think it's fun. Computer
programming is fun, right, so like the ability to make
a program work is and some of those tools are useful.
But this is that's not a major industry on its own.
You know, I'm designing a tool for my small company
that makes it easier for us to do whatever. That's cool,

(19:19):
but that's not like a trillion dollar industry. I can't
get my arms around yet exactly how professional computer programmers
are using it. They're all talking about it, really yeah,
but I can't get I can't get my arms around
it yet. Yeah, what do you what's what's your sense of.

Speaker 1 (19:34):
I'm gonna be honest, I was going to ask you.
I was literally going to ask have you heard people
using it? Because if you go on X, the Everything
app if you put on pulhasmat suit, you go on
X and you go and look. And the way that
people talk about this is like they have connected into
the matrix that they are now a thousand X engineer.

(19:57):
But when you go and look at what they do,
no one actually says. There's always these these kind of
vibe store, the vibe tales, the mythology where it's like
I had a problem vaguely that would have taken me
X number of hours, but when I used claud code,
it solved it immediately and caught two bugs that I
didn't know about. It's very Marine Todd's style. Then everybody

(20:18):
clapped yeah, and it's you were a computer science teacher
and you you don't know either, which makes me think
it's not as big a deal as people are saying.
My other bit of evidence is that at the end
of last year, Anthropics claud Code revenue was one point
one billion annualized, so about ninety to one hundred million

(20:40):
a month for a revolution. That feels low somehow.

Speaker 2 (20:57):
Yeah, I mean what I know as a computer science
you can't rate performance oriented code. You can't write safe code.
You can't write code it has to sort of juggle
sort of complex out of scenarios. I mean, you just
need good programmers' eyes on it building this code.

Speaker 1 (21:13):
Why is that? Is there a way of explaining to
a known code of why.

Speaker 2 (21:16):
That is code? There's like a poetic element to it,
you know, it's Writing good computer code is difficult. You're
often you're dealing with, you know, what is my problem here?
And I want an elegant way of sort of satisfying
this problem. You're often drawing from pretty nuanced algorithms and
data structures to try to figure out how am I

(21:37):
going to organize information and efficiently access it? When performance
comes into play, there's a lot of really subtle decisions
to make about, you know, how am I going to
store use things in such a way that we don't
get bogged down when we're trying to execute things. It
just becomes I'm I don't code as much anymore. Exo'm
a theoretician, but I did my whole life since I
was you know, seven, And it's a there's a it's
an art form. Right When you're using cloud code, you're

(22:00):
not really supposed to look at the code, and so
I think that takes a lot of uses probably off
the table. It's like the use cases. You're supposed to
have these different instances of cloud code running, and this
one's going to write code, and then this one's going
to write test for that code, and then the cloud
code is going to run the test and then try
to fix the code if it doesn't matches the test.
But your eyes aren't on the code at all. And
there's I mean, obviously for a lot of programming, that's

(22:21):
an issue. But then the other thing I've heard about
the computer programming industry is it's very stratified. There's a
smaller number of like really good, serious programmers that produce
like ninety percent of the really important, valuable code on
which everything runs. I don't think they would touch cloud
code with a ten foot pole, like they're good at
what they do. And then you have these like huge
strata of people writing like JavaScript and sort of hacking

(22:44):
together Python, and you know it's like not very good code.
And then it's I guess you could replace some of
the functional it's functional enough, but I can Yeah, I
can't get my arms around it yet. But I do
get you know, I have a lot of sources, so
I hear from you know, I do hear from people
that are talking about how cool cloud code is. Is cool.
I hear from a lot of professional programmers that are like, yeah,
we don't use we can't use this. We're trying to
write serious programs, like we have to sell this software,

(23:07):
like this is this is not solving a problem we
have or you know we so you know, I don't know,
But the problem is that that's what the reporting should be. Hey,
here we are at this company looking over the shoulder
of people. What are they doing. Let's talk to the
engineering teams on background of this tech company exactly what
role is going on here? And I think what a
lot of reporting has become on AI is your hype laundering.

(23:30):
So you look at the discussion about the technology happening
from more engineering minded people, you convince yourself as a reporter,
I can't understand the engineering, but I will trust the
people who do, and then you launder what you're sensing
from that hype into your articles, not realizing that like
nerds like me, we get hyped up about stuff, and
we get super excited about stuff and we go create like, yeah,

(23:51):
you can't just launder our hype into this is what's happening.
And so it's like reporting on a war where you
have no one embedded, there's no one actually on the
ground where the battles are happening. You're just responding to
the press conferences that the generals are holding, you know,
back in the Pentagon. Like it's not a way to
report on what's actually going on.

Speaker 1 (24:12):
And I think the other thing as well is there
is a if you don't do this hype laundering. I
don't know how these If you're a report listening to
this and you have a thought about this, send it
to me anonymously. Is it trod Thatt seventy six on signal.
But my thought is as well, is there's probably a
problem with rationalization as well, because if you look at
this and you say, Okay, well it's kind of cool,

(24:34):
it's fun in whatever indeterminate way, it doesn't seem like
serious software, like actual real deal software is being made
with it. But then the CEO of Google says thirty
percent of code is written in AI, which is bullshit.
And I've heard from so many software engineers, well, it
couldn't be that everyone's just wrong, right, It couldn't be

(24:55):
a case of that everyone is making the most egregious
capital expenditure fuck up of all time. This will be historic,
I believe, worse than railways, digital beanie babies, but done
at the scale of laser tag arenas. Now, that can't
possibly be that, because everyone else is saying this is
exciting and good. And at that point they choose instead

(25:16):
of being like worried, instead of being a bit anxious
about this, they say, well, Amazon Web Services spent a
lot of money, so this spends a lot of money
money too, so this is actually good. It's actually good.
And indeed these people seem emphatic and excited, and I,
as a non coder, can build a fudged CRM that

(25:38):
probably would not withstand even the laziest haacker. I can
do this, and thus it will extrapolate further from there.
And what sucks is what the people that I believe
actually will be hurt by this are retail investors. I think,
regular people buying stocks in these companies or selling software
stocks because they believe that color code will replace them.

(25:58):
And altomately, I think it's just gonna be a blood bath.
For people's furrow. One case that could have been avoided,
except it would require reports to do something uncomfortable, and
I don't think they want to do that. Ema.

Speaker 2 (26:12):
I think there's two things going on. This is what
I've decided with reporting. First of all, it's an asymmetric risk.
So a lot of reporters are like, look, there's not
a major risk if I'm excited about this and it
doesn't pan out, because we could be like, yeah, surprisingly
this didn't pan out and with some factors we couldn't see.
But they do feel like there's a huge amount of
risk of saying this is not a big deal and

(26:33):
then it is, and so it's definitely an asymmetric risk.
We saw a lot of this during like COVID as well.
Right is like you wanted it was less There would
be less harm if you were too alarmist about something,
but there could be a lot of harm. They felt
like reputationally, I feel like this is not a big deal,
and it was, and so there's definitely an asymmetric risk profile.
There's also like a meaning defining profile. It's just really

(26:56):
exciting to think everything is going to be disrupted in
chain unrecognizably. It sort of gives a focus and meaning
to like an otherwise somewhat chaotic and disrupted world that
we're in right now. And so there's that aspect too,
is that people want to believe there is something massive

(27:17):
about to happen because in some sense it wipes away
all the bad stuff that's happening. Who cares, none of
this is relevant because this much bigger thing is coming.
So I think that gets wrapped into it as well.
I think the economic reporters are more on this because
like their whole job is to try to I mean,
they're not on it. I think after your work they're
on it more.

Speaker 1 (27:34):
But like, oh no, if I agree, I am reading
Big Series. There's an article in Bloomberg I'm trying to
shove through archive dot is because they were so main
and unfit to my friend Steve Burke at Games Next
as that won't pay them. But it's more shit about
like the software narrative, the fact that software is being
disrupted by Claude code. You get the same pallid reporting

(27:57):
from Bloomberg and even the Financial Time about Anthropic and
the fd is generally pretty good. Well, they're like, yeah,
they're just gonna make thirty billion dollars next year. It's
just fanciful. It's there's the skepticism, the cynicism doesn't exist.
And I get I agree on that there is a
bubble reporting.

Speaker 2 (28:15):
The last fall, there was a period where everyone did
the bubble reporting. After GPG five, you had a two
month period where every major publication did bubble reportings. But you're, yeah,
I guess that did kind of die off.

Speaker 1 (28:28):
It died off because they hear one nice thing from
Jensen Wong and they're like, well, I'm sold.

Speaker 2 (28:35):
Like that jacket that jackets looking pretty sharp. I mean,
someone who wears that jacket. How could they be wrong?

Speaker 1 (28:41):
Would a man that has a shiny jacket like the
Gleat singer of Corn wears at concerts? Would he lie?
And it's just what really bothers me. As far as
the economic reporting though, is the Oracle, because it's like
Oracle needs to prepay three hundred billion dollars over five
years by op AI, a company that, if we're to

(29:02):
believe reporting, which I do not, they made thirteen billion
dollars last year in revenue and lost. Sorry they've made
thirteen billion and lost. They claim nine. I think it's
probably higher. How are they meant to pay thirty to
sixty billion dollars a year in a year, And everyone's like, well,
they'll work it out. How's Oracle meant to build those

(29:23):
data centers? Those data centers are gonna cost one hundred
and eighty nine billion dollars. They've raised one hundred billion, Okay,
how they meant to do that? And everyone's just like, oh,
they'll work it out. If I wish I could do
this with the fucking bank, I need a five hundred
million dollar house, I'll pay for it at some point,
all right, it's fine, and the news would run articles

(29:44):
about my genius housing purchases. It's just it's one of
those things where and you said, well the asymmetric risk
doesn't exist, it does for some of these reporters because
I've been saving their their bylines for years. Because I
actually think that there needs to be some sort of
reckoning with this, because if you look back, there are
major financial outlets that did the same, that were literally

(30:06):
propping up Sam Bankman freed two weeks before FTX collapsed,
who then went on immediately to cover AI fucking and
now they're peddling bullshit for anthropic and sorry, I'm kind
of hating. Well, I guess it's hated season, so I can.
It just frustrates me because regular people are being scared.
They're being scared by the kind of astonishment which I

(30:29):
actually loved the astonishment reporting of like, oh well, open
clawd has proven open claw has proven that ai agi
is here, or they've built their own social networks that
we should be scared, scared insane did Yeah, And it's like,
I assume you saw that for the listeners. By the way,

(30:51):
open Claw had this this fucking claude bot whatever it is.
They had their own social network where the quote unquote
lms would post, but it turns out that those posts
would just make by their owners.

Speaker 2 (31:02):
Yeah, and this is just I mean, this is a
good case study.

Speaker 1 (31:04):
Right.

Speaker 2 (31:05):
This one bothered me because like writer friends I know
who are not technology related, we're texting this to me.
They were worried, right, They're texting me these articles like
this seems bad, Like this seems like something really bad.
They were really getting the digital ic really strongly off
of these articles. And I would start reading these articles
and I say, well, there's no discussion of what is
the technical breakthrough here? And what are the concrete implications,

(31:28):
Because it turns out there was zero technical breakthroughs. There
is no new AI technology connected to open claw, which
used to be called molt whatever mold or open clawed
or whatever open claw is, I think where they ended
up right, which is an open source library or framework
for building AI agents powered by elims. There's no new

(31:50):
AI technology involved in this at all. The agents you
build are just accessing off the shelf elms that we're
all using for chatbots anyways. You can aim it at
whatever commercial chatpot you want. There's no new framework for
how the agents work. It's the same sort of react
loop that we've been trying for the last two or
three years, where you just you have a program. It's
like a bit of Python code that asks an LM, hey,

(32:11):
here's what I want to do, here's the tools you
have available. Come up with a plan, and then it
sends it back a plan, and then the Python code
takes the first step out of that text and says, okay,
here's the tools available. What should I do to execute
this step? And then like the LM will give you
some steps and then the Python code runs those steps
and then you just go back and forth right. That's

(32:33):
like this basic react, which.

Speaker 1 (32:34):
Is exactly how Manus worked as well.

Speaker 2 (32:37):
Everything there's nothing new about this.

Speaker 1 (32:38):
Manus was this AI agent that Facebook might be Meta
might be acquiring. They literally when you use it, it
just writes prithon code for every step.

Speaker 2 (32:47):
Well, so it's yeah, so this, I mean, there's nothing
new here technologically. The only thing that was new about
openclaw is it was open source, so it made it
easy for anyone to write bad agents. And the other
thing that made it interesting to wireheads in San Francisco
is that the commercial products where people are trying to
build these companies, they have common sense security, like, well,

(33:07):
probably if it's just a Python program blindly doing what
an LLM tells it to do, like it probably shouldn't
have access to like your credit cards or to your
hard drive or whatever. At open cloud like you can
just you can give it access to anything you want
on your computer, and so you could build really cool
demos that are also like incredibly insecure and unsafe. And
so it was fun for hobbyists, but there was no
new technology there. Nothing was new, So it was just

(33:30):
a way for other people like hobbyists to build their
own agents. That were like less safe than the more
carefully built Like what people don't here's a story people
don't know is like in the immediate aftermath of chat GPT.
So we're in early twenty twenty three, right, so chatchepts
come out. Open AI goes on a road tour of
major publications, right, and they're to try to Hey, here's

(33:51):
what's going on. You need to write about this. In
early twenty twenty three, they're like, the next thing we're
going to offer is something like these agents. They call
them plugins, and you can install plugins that basically can
do actions on behalf of your LLLM queries. You could
have like a book an airline flight plug in and say, hey,
chatchipt book me a flight, and a language model can't

(34:11):
do anything but produce text, but then the plugin could
take the text and go and book you a flight.
And they're like, this is the future, obviously, and that
project disappeared because oh, it's incredibly unsafe and unreliable to
have code that could interact with the real world that's
following commands from an unreliable who's stayed an LLM like
that went away not because the technology was hard, but

(34:32):
because it's not safe. So there's nothing new. My whole
article on agents, they tried this all at twenty twenty five,
and we're failing to get this type of agent to
work for anything outside of basically like producing computer code.
So open clause is nothing new. It was just a
way for other people to build these things. The only
interesting stories were security whole stories. But it was reported. Man,
I was listening to the All In podcast and they

(34:53):
were tasing why this is the future of AI, Like
this is it. Everyone is going to have of an
open claw agent as like a personal assistant, and this
is like the biggest thing in open AI. And I
don't know who it was. One of those guys was like, yeah,
we replaced our podcast producer, you know, with this agent
who can like email guests on our behalf and book

(35:15):
it on our calendar. And well it's costing about a
thousand dollars a day right now to run it. But like,
I don't know why we're going to need employees in
the future.

Speaker 1 (35:23):
Oh yeah brother, Yeah.

Speaker 2 (35:25):
So anyways, like it's a non if you said what's
the technical implications, not if you say, what are the
concrete implications? The best story they can have is like
maybe when people I don't know what it is. All
the companies will try to build these things for years,
so I don't know, but it was reported like just
something icky is happening, and that molt book was an
application that was built for these agents to communicate on

(35:48):
like a Reddit style social network. They vibe coded that
framework and it was full of security holes as we
mentioned before, and there was like a small number of
users that created a huge amount of agents, and they
were just kind of like prompting and prodding their agents
to produce, like let's talk about creating around church or
killing humanity. It's like, guys, this you have Python code
asking lllms to like write text in the style of

(36:10):
like the matrix, and you're posting it on a fake
social network. The real story here should be where they
don't these people have things to do? Don't they have jobs?

Speaker 1 (36:20):
Like what is y? Yeah?

Speaker 2 (36:20):
Yeah, yeah, really brilliant my model train, I added a
tunnel so that one really got no.

Speaker 1 (36:26):
I actually will push back on that. Mudile train listeners
I think make up twenty percent of the listenership of
this podcast course. Also, modile trains are cool. Signaling to
my fellow autists out there. But you know, mudtile trains
are a physical thing which you build it, you build
a little a little city. I think that delightful. Respect
to those with this. It's like if they and you

(36:47):
know what, if they were just saying I have been
fucking around with some software and it did something cool,
I'd respect the shit out. Then I'd be like, yeah,
enjoy yourself. It's going away. This is a little subsidized rites,
but have fun with that. I don't know why you
need a max studio, but good luck. No, they are like,
this is the future. But I'm gonna be honest, cal
My real question is what does openclaw actually do? Because

(37:08):
when I went and read what it actually I read
so many posts you said it books callen does and
does this. I could find no proof that anyone successfully
did that.

Speaker 2 (37:18):
And it doesn't. It's it's just a series of library
calls that you can use to build your own program
in theory that would do that. So open claw itself
is just like a series of like interfaces and hooks
that makes it easy to write a program that talks
to other services basically and makes calls to LLLM. So
it's just like a wrapper in which you can write
your own code, and some people are trying, Yeah, like

(37:40):
you can you can write, you can have it talk
to your calendar, you can have it talk to your
email and theory and.

Speaker 1 (37:46):
Look, but does it doesn't work?

Speaker 2 (37:48):
Well, it's it's just asking an lell you can simulate.
I mean, I did this for my New Yorker piece.
I was like, look, I can just simulate being an agent.
Just ask any LLM, here's what I'm trying to do.
Give me like the steps for doing this or whatever,
and like anything else you ask of an LOLLM, it
will give you answers that sound very reasonable in general
and then have like a lot of issues in the details.
Which is why agents based on llms have struggled because

(38:10):
if you it's fine if you as a human are
talking to an M, because you don't realize how much
filtering and tweaking. You're like, well, that's kind of ignore
that and this is good, or let me ask you
to redo it. It's really an issue if you write
a program that just says I will do whatever. I'll
ask the LLM for a plan and then just do
whatever it says because it doesn't know that doesn't really
make sen or like this sounds generally generally reasonable, but

(38:34):
with like issues in the details. Does it work out
when you execute it? I mean, in a practical sense,
can it? They said only all in podcasts, three dumb
bitches saying exactly, that's the main reference. Don't use that
word usually those people sitting around going blah blah blah.

Speaker 1 (38:51):
Oh we've made it and replaced our podcast producer with
an LM that can make appointments and send emails in practice.
Is that true? I severely doubt that. I just I Also,
you're spending on thousand dollars a day, so you're paying
three hundred and sixty five thousand dollars a year for this, right?
Are you really or are you just? Have you? If

(39:14):
it is it completely it probably isn't. And that's the
thing I don't I get why they all in guys
do it because they probably have investments and they're boosters TVPN,
same fucking deal. It really rules that the two largest
like tech things in the valley are just state media
book for Silicon Valley. What gets me is when you

(39:36):
get like the Atlantic CNBC Business Insider in places like
that doing stuff like this, and it bothers me because
they don't even need the hate on it. They could
just say, yeah, I could do this, it's pretty cool,
right yeah, But I guess that that doesn't get the clear.

Speaker 2 (39:51):
It's not an interesting story. That's not the real story
is not always that interesting. It's like like hobbyists are
building these tools that kind of do cool things but
make mistakes, and it's a little expensive. Like the most
interesting story out of open claw is the only thing
I think is going to be impactful out of it
is because it was so expensive, right, like that thousand
dollars a day. What it's forcing people, these hobbyists to

(40:11):
do is to turn to like cheaper alternatives for models,
and that I do think is significant, this idea that
there's open source models out there, they're significantly cheaper than
trying to use like open ai or claud. More and
more people are running their own local models because it
turns out, for you know, most specific uses you might
use in LLM, you don't need like a super fine

(40:33):
tune trillion parameter beast of a model running in some
data center somewhere. It's like, you know what, I'm parsing
my email that try to extract like suggested times. I'm
fine with like a twenty billion parameter model that can
easily fit in a single GPU on a thing I have,
you know, we share in our office or something like that.
So to me, that's the most interesting story out of
open claw is the bad news it could be for

(40:55):
the big companies as people get more comfortable with we
don't needed these formula one car versions of language models
for the stuff we're actually doing. We're fine with the
forward focus, right, And I think that's a transition that's
going to That's a transition that to me is more interest.
That's what we should be writing about. Like, that's an
interesting idea to me is that you have these huge,

(41:16):
high valuation companies, but you also have all these open
source models, like the weights are just out there in
the public domain that can do ninety eight percent of
what people care about, and you're beginning to get low
cost competitive services where people can just spin those up
in cheaper data centers. That's interesting to me. That's an
economic story. AI creating its own church is not as
not as relevant to me.

Speaker 1 (41:50):
I'm already working on a story full this Friday in
fact oh my premium newsletter about the fact that actually
the margins of serving GPU compute are dog shit like
the best of the best of like a co location
place like Applied Digital, it's like twenty seven percent gross
margins and that's at one hundred percent utilization. Anything below
point seven they're burning cash. I hear this date center

(42:12):
out in North Dakota losing a million dollars a day.
That's the thing. This is even with these low these
lower cost models on device could be interesting.

Speaker 2 (42:20):
That's what's going to happen. I think, I think undevice
is what's going to happen.

Speaker 1 (42:24):
I just remembered something. This is a classic bullshit story
that I see every so often. So have you read
any of the stories that are like, yeah, Claude can
now work for hours uninterrupted? Have you read about these?
You heard this? You've seen this.

Speaker 2 (42:37):
Work for our Oh yeah, yeah, we were talking about
the Yeah, the multi step agenda execution.

Speaker 1 (42:42):
Well, no way, it keeps coming back.

Speaker 2 (42:44):
Yeah, yeah.

Speaker 1 (42:45):
AI on the CNBC AI on the verge of eight
hour job shift without burnout or break is it is
twenty four hour our AI workday. Next going to just
censor myself what I was going to say there, because
it's just like yeah, it can work for hours. Is
the output good?

Speaker 2 (43:03):
Does it a problem? Yeah, just a lose?

Speaker 1 (43:05):
Does it? Will?

Speaker 2 (43:05):
Just keep reading? Call it? Recall it? Recall it like
I can. I can write a Python program that calls
an hour for twenty four hours.

Speaker 1 (43:12):
If you need me to burn something for hours on end,
I just give me some gasoline. I got your baby.
I've sort this right out. Be much cheaper, but it's
like yeah. By September twenty twenty five, Claudes on It
four point five is reported to run autonomously for up
to thirty hours, reported with com I mean, but that's
more vibe reporting. It's you're meant to read that and go, wow,

(43:34):
this thing is performing tasks that are useful, that execute code,
that do something, and in that thirty hours that is
equivalent to thirty human working hours versus they sat there
and pissed their pants for like twenty nine point five
of those at least well.

Speaker 2 (43:50):
And also that those are like specialized TSK typically that
have clear milestones and testing, so it's like it can
do something. The loop can try and try until it
passes the test, and then you know that's done, and
then you can move on to the next step. And
then you can keep calling to l M and executing
until it passes the next test and you can move on.
And in theory like it's yeah, they're avoiding error cascade

(44:12):
because it's testable and they can keep retrying or something
like that. I mean, this was like the meter had
this problem with their graph of how long of tasks
AI can now do, and it was like the sort
of like super arbitrary decision of like, oh, here's a
test that takes a human five hours, Now AI can
do that. Here's a task, whereas really more about like
how many things in a row can you do without

(44:33):
the errors cascading out of control or something like that.
It's very different.

Speaker 1 (44:38):
You're being way nicer than you should be, in my opinion,
just because they don't even explain. They don't even say like, yeah,
and we managed to make it through it and let's
go bus right hours c NBC just fucking woo just
for tearing their shirts off and screaming.

Speaker 2 (44:54):
But it goes back to the two things. We need
this report and you need technical the details of the
relevant technical innovation, and a discussion of concrete implications in
your future implications of that breakthrough. That's what I'm always
looking for, So, like, if you want to cover a
story like that, like, well, what does that mean? What
is the technical breakthrough? Right? What does this mean technically
you can do thirty hours of work or eight hours?

(45:14):
What is the work? What was changed? What did they figure?
What was happening before? What technical breakthrough made this possible now?
And then what are the concrete implications? What specific things now?
Can we directly this will now allow us to do? What?
Tell us what jobs are going away? Tell us what
tool we're going to see, like you have to, but
we avoid that, right because then you're putting your chips
down and then those things don't come along, And so

(45:37):
I'm always looking for that. What's the technical innovation and
what are the concrete implications? If you don't have that,
you're mining emotions.

Speaker 1 (45:43):
You're mining emotions or just helping boost stook. That's what
really bothers me because I hate that they're scaring people.
I really really hate that they're doing marketing like they're
just doing like I run a PIAFA. I've dealt with
ether these states startups for like fifteen years. There isn't

(46:04):
a single early stage startup that would get a percentage
point of this bullshit, Like you email the reporter about
like a series a start. They're like, all right, all right, motherfucker,
how much revenue are they profitable? Yet? Why not? Why
not explain to me right now? Anthropics like we're gonna
gonna burn a hundred billion on training. I guess what

(46:25):
do you think? And they're like, yeah, I love it.
I actually think that's huge, that's great. I'm not And
to be clear, I think that this scrutiny should be
from the beginning to the end. I think that everyone
should face this scrutiny. Yeah, I'm not saying that I
should get an easier I'm saying, actually everyone should. Anthropics
should face the most brutal scrutiny, and I guess that

(46:47):
they don't because they want access. It's just very dull
and annoying, and it's just helping already rich guys like
Wario Ama Day, who should not have more money. Listen
to him speak, he needs he needs to face some stress.
I think some stress would help him grow. But caw,
as we wrap up, I did actually have like a

(47:07):
technical thing that wants an idea of being perclaying. So
I think that this term training with models is being
misused and used in a way that is kind of
vibe reporting style, which is they use training as a
word that suggests that it will stop, that they will
stop training these models. But from correct me if I'm wrong.
Training is everything from building a new model to updating

(47:30):
and models current parameters. Correct.

Speaker 2 (47:32):
Yeah, there's pre training and post training is the right
way to think about it. So the pre training is unsupervised.
That's where you take all the text that's ever been written,
and you will take a real piece of text written
by a human, and you'll cut it off at an
arbitrary point, and you'll tell the language model your guess
the word that came next. There's a real word that
came next. This is a real writing Guess the word
that came next, and then it guesses, and then you

(47:54):
adjust the weight so it gets closer to the right answer.
That's prehen you adjust the waits. What are you doing,
You're just so you're a training algorithm called back propagation.
This is a Jeff Hinton, you know innovation where you're
you're going through and you're adjusting the weights all the
way through the layers in such a way that the
answer it gives for this particular test gets a little
bit closer to the right answer, because you know the
right answer that's pre training, that's pre trains unsupervised. So

(48:17):
you take Hamlet or you take Dickens like the best
of times it was the worst of and then you
give that to the thing what were to come next?
And you know it says Bacon like, we're going to
adjust these weights now in a way that like your
answer gets a little bit closer towards times. Right. Okay,
that's pre training. Then you get post training, where you
already have trained, so you have this network. All the
weights have been set through this massive multi month you know,

(48:39):
billion dollar pre training, and now you want to go
through and you want to tweak this to avoid certain
types of behaviors or to influence towards certain types of behaviors.
So for post training, it's almost always based off of
you have inputs like prompts and correct answers. This is
the right way to respond to this question, right, So
you've you've you have of questions and answers. You give

(49:02):
the prompt to the LLM, it spits out some answer,
and now what you're doing you're using is called reinforcement learning.
It's a general technology, but you're using techniques from reinforcement
learning the sort of like zap it like you would
zap a dog when you're dog training it to if
it's a bad answer, you zap it, so you get
those weights away from that answer, and if it's a
good answer, you give it a treat like Okay, this
is post this is post training. And so that's where

(49:26):
you've moved past the word guessing game, which is where
all of the sort of general smarts comes from these models,
And now you're you're you're doing the sort of zapping
and treat training around very specific things, right, so this
is where you keep going. Yeah, So like so you'll
go through and like ask it questions where the answers
might be like about building bombs or whatever, and every

(49:46):
time it spits out an answer about a bomb, you
give it like a really bad negative shock, and you
like definitely turn off those circuits, like we don't want
you to spit out answers by bombs, Like that's where
all the guardrails come from. Or if you want it
to get better at doing like a particular math exam,
you can give it like lots of questions from that
math exam and then you have the right answer and
you can kind of zap it to move it towards

(50:08):
what the answers look like on this math exam. So
so post training is more focused. You have particular types
of behavior you're trying to sort of instill in this
already pre trained massive network. That's mainly where the focus
turned after GPT four. So GPT four was like the
extent of pre training making it smarter. After GPT four,

(50:29):
trying to make those models even bigger and pre train
them longer didn't lead to much performance increases. So everything
we got between four in the lead up to five
was post training, and that's when they began focusing on
metrics because if I have a particular metric, I can
post train a model to do well on that test.
And so everything became about metrics and post now we
can do this thing better or look at this thing

(50:51):
we do better, And so that's kind of the game
that's played now is we do lots of post training
that requires like much more specific data because you need
like right answers, pairs of props, and correct answers. So
only certain things we can do this with. But that's
the game we've been playing since like twenty that's one.

Speaker 1 (51:08):
Do they do that with models that exist now? So
this is basically updates, right?

Speaker 2 (51:12):
Yeah? They do it on a semi regular basis. Yeah,
but they usually there'll be a name change, like GPT
five to two is different than five to one, different
than five.

Speaker 1 (51:21):
But don't they update the current models?

Speaker 2 (51:24):
They don't repretrain them. That's too expensive.

Speaker 1 (51:27):
But they no, no, I'm not saying that. And say
do they post train the current models to make them
better at stuff?

Speaker 2 (51:32):
Yeah?

Speaker 1 (51:32):
To tweak things.

Speaker 2 (51:33):
Yeah.

Speaker 1 (51:33):
So that's This is a very long way to get
to a point I'm kind of making. Which is one
of the problems with VIBE reporting on this is training
is framed as this thing like inference is framed as
op X that is permanent that you cannot avoid influence
being creating the outputs. Training is framed as this R
and D mysticism, which is just out there and you're

(51:56):
not like training. They never say this, but you hear training,
you think, oh, you train and then you perform, and
so training would end. But from what I understand, training
is as common and necessary an expense as influence at
this point.

Speaker 2 (52:10):
Yeah, it's the only way you improve or update things, right, Like, so,
if you had a Microsoft through sixty five software, you're
constantly sending updates and patches and whatever is you like
to add new features or whatever. In the AI model world,
it's it's yeah, it's posting more rounds of post training.
It's the only way to make any sort of improvement
or fixing bugs. Like you're like, oh, here's something that's

(52:30):
saying that we're really upset about. Okay, let's go in
and do some zapping. Let's get out to zapper, give
it a bunch of examples of saying that bad thing,
and let zap and say don't do that. And now
it's like very unlikely to do that. So yeah, they're
constantly they're constantly doing that. Otherwise, your a stasis, which
is different than the way most people actually think of it. Differently,
they think that the model is somehow like learning online

(52:51):
that like, which as it's absolutely not true that as
you talk to it, it's learning and it's getting uh,
it's getting better. And they're like, but wait a second,
something I talked about earlier, it must be getting smarter.
The model doesn't care about you, just your local software.
You don't realize this. It's including like huge bits of
stuff you've talked to it before. In the prompt that's

(53:12):
going to the model. It's like, here's a bunch of
stuff that Ed has submitted to you in the past. Okay,
now here's his current question. So it's the model hasn't changed.
The model doesn't have memory a canunt of justice. It
doesn't adjust its memory in real time. It's all static.
There is no dynamic memory involved in these models. They're
fixed until you go out and post train it and
then you bring the new weights back in the data center.

(53:33):
And that's the thing. Do an inference now.

Speaker 1 (53:35):
But that's the thing like this. The reason I bring
it up as the kind of closing vibe story is
because training is very clearly getting more expensive what they
say profitable in inference, which again doesn't really make sense
to me. But putting that aside, they're not. But even
if they were, if training never stops, then who cares

(53:56):
about profitable in inference? Like it just means that you
will get more or expensive forever. Inference is just yeah, yeah,
differences pooh, training is pooh. That's just that's I'm not
putting that one in.

Speaker 2 (54:09):
A yeah, okay, it's an interesting question, right because it's
also getting like the what's an altman who is making
those comments about well, we just didn't have to pay
for the training. This would be part I.

Speaker 1 (54:18):
Didn't have all these expenses, I would be so profitable.

Speaker 2 (54:22):
The one distinction that's maybe relevant there not to be
an apologist, but the one one distinction that is relevant
is pre training versus posts. So pre training is insanely expensive, right,
because you're you're training something on all the all the
words in the world, and it takes months. So it's
it's just like you're running a data center that's going

(54:42):
to have nowadays up to six digit GPUs uh running
full time for months just to get that pre training done,
and you have to pay for all of that, right,
So that's all right time, you're not getting money, you're
just paying for training. Post training is also expensive. It's
not that expensive though, because as expensive as pre training,
because each post training session it's a way, way, way

(55:02):
way smaller data set that you're post training on. It's like,
all right, we generated like ten thousand examples of people,
you know, responding to questions in a racist way, and
those ten thousand examples will use to reinforcement learn and
try to move it away from answering those type of
questions in a racist way. That's like not that big
of a data set compared to we're going to train
this on every word written that we have access to.

(55:25):
So the post training is not as expensive as pre training.

Speaker 1 (55:29):
Unless you're doing it all the time.

Speaker 2 (55:30):
Yeah, that's true.

Speaker 1 (55:31):
That's the thing if you're doing it all the time
and it takes months of pre training, but you're constantly
doing something like post training for months. Yeah, I mean
post training not functionally the same thing.

Speaker 2 (55:43):
It's the same thing. I think this is a fair point.
When you when you have a particular example that you're
post training on like one prompt with an answer, it's
kind of like you're doing inference in reverse, right, So
you're going from your back propagating from one side of
the network to the beginning, as a post are going
from the beginning to the end. Now, it's more expensive
than that because when you're just doing inference, the fundamental
operation is basic multiplication. You're just multiplying numbers in the

(56:06):
big table backpropagation which you use to training. You're it
is multiplication. When you do a bunch of it's a
bunch of derivatives because you're constantly you need to calculate
like the derivative of these. I mean like it doesn't
really matter, but you kind of need that. You want
the derivative because you want the gradient to scent to
be towards like better and away from worse and derivatives.

(56:27):
My understanding is this is like more expensive operations per
weight that you're trying to change because you're not just
multiplying a number, You're having to calculate derivatives and becomes
a little bit more complicated. So yeah, it's like inference
in reverse, but also like a little bit more expensive.
So if you if you have ten thousand sample question responses,
you're going to use the post train, it's kind of
like ten thousand users sent prompts and there are particularly

(56:51):
expensive prompts, and you had to pay for that instead
of like them paying for it. So yeah, it's good
to think of it as like inference and reverse.

Speaker 1 (56:59):
It's also an ongoing cost, Like it's everyone is leaning
on this idea that this stuff will magically become profitable.
I don't know if you've seen the cash flow diagrams
of anthropic and open ai. But there is a mysterious
math going on where year twenty twenty eight, twenty twenty
nine they just become profitable.

Speaker 2 (57:16):
Well, I was going to ask you about this. Last month,
there's this announcement that open ai had some massive increase
in revenue.

Speaker 1 (57:26):
Yeah, well this is actually a great vibe reporting thing,
and that's annualized revenue. They said they hit twenty billion
in annualized revenue, which would mean one point sixty seven
billion in a month. Now, important details. We don't know
how they're defining a month. We don't know if they
mean twenty eight days, twenty nine days, thirty days. We
don't know if they mean a calendar month, so the

(57:47):
month of November or December, or do they just bean
any thirty days. We don't know if they're doing insane math,
which happens very rarely, that this company feels like one
that might do it. Just my gut instinct is they
may be doing his seven day period and we're going
to turn it into a month like they We don't
know how they're doing this. And they also coupled this
by saying that as compute grows, revenue grows. I don't

(58:11):
know if you seen this.

Speaker 2 (58:13):
No, it's like, say, their formula, yeah, okay.

Speaker 1 (58:15):
Yeah, it's an insane formula that does not map to
any economics. Like it's just it's the kind of thing
that if we had a functioning business and tech press,
that would just be scrutinized to the bone, that would
just be ripped apart and say, what the fuck does
that mean? Because if this were true, if you simply
add more gigawattch at more revenue, then you would simply

(58:38):
print more money, like it would be a money print
of us.

Speaker 2 (58:41):
Like my movie theater did well. Therefore, yeah, if I
build one hundred thousand movie theaters, we're gonna make one
hundred thousand more you know, times the amount of money.
And it's like, well, wait that theater was in Manhattan
and it was like really well run, and they're yeah,
well okay, I assume I assumed as much trying to
understand that story. Okay, yeah, the annualized revenue stuff, I

(59:01):
think I even use that term talking to someone I
credit you as like Trum would say, look out for
annualized revenue.

Speaker 1 (59:07):
It is. The funny thing is with that as well,
it's more viabe reporting because arr standard. It's very standard
in SaaS companies that sell on a per seat basis,
So a sales force would do a well even they
are doing annualized now with the AI revenue, but it
would be a software company has one hundred seats they

(59:27):
sell to a company and they charge fifteen bucks a
bit a seat per month, and they charge that annually,
and the actual cost of each user is fairly measurable
because they're doing CPU based stuff. Like it gets expensive
at scale because there's a lot of people, but it
doesn't get multiply multi I can't say the word. It
doesn't get much more expensive as you grow with AI.

(59:50):
It's actually because of the way large language models work,
your most excited customers are your most expensive.

Speaker 2 (59:56):
Yeah, because they blow whatever whatever their monthly revenue is,
whatever their cost is, they blow past it, so.

Speaker 1 (01:00:02):
You can't even do a percy revenue. It's just every
all of these things. When I say them out loud,
I'm like, I feel like this should be more obvious.
It's why I loved your video because it was like
thank you someone else, Well, here's the.

Speaker 2 (01:00:15):
Question, here's my here's my question, here's like the dangerous
question I actually put out a video last week. It
was like dangerous question. We're now on year three or
four of like the I'm counting new year starting like
New Year twenty twenty three or whatever of people saying,
oh my god, this is so cool. These massive disruptions
they're going to change everything is imminent, and year after

(01:00:38):
year we've said that I'm not yet seeing the massive disruptions,
like not the not the stories of what might be
disrupted or the stories of what's different, but like, how
many years do we have to go without industries crumbling
or major new economic players that didn't exist before, or

(01:00:58):
complete restructuring of huge companies around this technology? How many
years we have to go? So I did a video,
right I went, I found the Reddit thread where someone
just asked this question. This was from like earlier in
the month. They were like, outside of the vibe coding stuff,
what what are the like, what are people? What are
the big tools? Like, what are the big things that
have come out of this technology? They're changing things? And

(01:01:20):
I read through this whole thread and it was interesting,
there's not people don't have much like well, you know,
like it could these are like really small case studies.
I used it to help, you know, gather clean up
data that I got from whatever. I was like, this
is like such a nerdy specific use case, and so
I went through that thread in a video and this
has kind of been my question. It's like, it is
very cool technology, but how do we know where this

(01:01:41):
is going to fall? Like to me, the scale is
it would go like this, like blockchain software, then Oculus VR,
then maybe internet to electricity. Right, so we're gonna have
a scale of disruption.

Speaker 1 (01:01:56):
Right.

Speaker 2 (01:01:56):
Blockchain software is something where the premise it's made no
sense and it was never going to get off the
ground and there was going to be no impact on
the world. And you know, because I'm a my training
in CS is in distributed system theory. I was there
in twenty twenty saying guys, let me just tell you
this is nonsense. No Web three is not about to
take off. None of this makes sense. And that was true.
That did nothing. Then you have like Oculus VR. It

(01:02:18):
really is cool, right you put on these things like
that is awesome, Like I love this technology, but it's
having a hard time having any real major impact because
people aren't sure what.

Speaker 1 (01:02:27):
Times use of it. Also, most people don't necessarily have
a great experience initially because it's extremely dependent on where
you are, who you are, the size of your skull
not kind of yeah so.

Speaker 2 (01:02:38):
Big right for a limited group of people, it's really cool,
but it fails to come out. Then the Internet is
like really disruptive, changed a lot of things more. It's
not so much as like whole industries disappeared or whatever,
but it like changed the way a lot of industries
actually function. And then like electricity, you could say, like
it just completely changed what day to day existence of.

Speaker 1 (01:02:56):
Business was, like how existence of exactly?

Speaker 2 (01:02:59):
Yeah, And so that the question like that everyone should
be asking is where is the ENERVEI going to fall
on this? And you know, I would say, right now,
this is what gives me yelled at. I'm not saying
this is the prediction necessarily going forward, but right now
I don't think it's got passed much farther past the
oculus part of that scale where it's really cool. There's
very cool things. Chat Gipt is very cool. It's very

(01:03:19):
cool that it can like have that comprehension and no
one thought it could do that, But we haven't yet
figured out what.

Speaker 1 (01:03:27):
Comprehension. It's no, it's a text I mean.

Speaker 2 (01:03:31):
We take it for granted, but for CS people, the
ability that like I can add, hey, give me text
that like whatever in the style of a poem that
does whatever, and that includes a character from Star Wars,
and then it can give you text it does that.
That comprehension, like for computer scientist was like, oh, we
didn't really know how to consistently one yeah. Yeah, that's
like very cool. Yeah, but we haven't got past this

(01:03:53):
is like the the surprising thing of this field is
we're not really past the oculus stage yet, like where
they're like, for certain this is vibe coding is really cool.
The comprehension is cool, so like sora is weird, but
like it's cool, you can do that, But the markets
are not None of these have markets yet, right, Like that,
there's not big markets in any of these yet. And

(01:04:13):
will how far will it go from oculus to the Internet.
To me, that is like the number one question, the
number two question, the number three question of all reporting
on this, and almost no one's talking about that. It's
just hype laundering. We'll take this hype, will extrapolate, it
will react to that extrapolation. That's kind of what reporting is, right,
now in AI, Whereas to me, this is the hugest question.

(01:04:33):
If this ends up oculus, retail investors are going to
get screwed. If it ends up Internet, all right, that's like,
that's like a really interesting significant story. If it ends
up electricity, obviously that really matters. But like, I don't
know anyone who actually thinks it's going to be that disruptive.
Not the current technology. That's the story to me. Not
let's like extrapolate us. You know, hey, what are these

(01:04:54):
things are creating a church? Or let's let's let's hype
launder extrapolate that and react to our extrapol That's not
really reporting so much as speculative fiction writing. I guess
I don't know quite what to call it. But this
is the real question, where exactly are we now and
what are the possibilities of like where this is going
to go positive and negative? I don't we have enough
talk on that.

Speaker 1 (01:05:16):
I fully agree, Cal, it's been such a pleasure in
having you. Thank you for joining me.

Speaker 2 (01:05:20):
Always happy to talk shop, always happy to hate with you.
I guess in your season, I like.

Speaker 1 (01:05:24):
It, Hey, season is the best. We will be back
this week with either a monologu or an interview. I
have not decided, because I've got wonderful Corey Quinn interview
I just did, so I'm considering putting that in a monologue.
You'll find out on Friday. Anyway, this has been Better
off Line. I'm at zychron Subscribe to the premium, download
a T shirt, whatever you desire. Thank you for listening

(01:05:53):
to Better Offline. The editor and composer of the Better
Offline theme song is Matosowski. You can check out more
of his music and audio projects at Mattasowski dot com,
m A T T O S O W s ki
dot com. You can email me at easy at Better
Offline dot com or visit Better Offline dot com to
find more podcast links and of course, my newsletter. I

(01:06:15):
also really recommend you go to chat dot Where's youreaed
dot at to visit the discord, and go to our
slash Better Offline to check out our reddit. Thank you
so much for listening.

Speaker 2 (01:06:25):
Better Offline is a production of cool Zone Media.

Speaker 1 (01:06:28):
For more from cool Zone Media, visit our website cool
Zonemedia dot com, or check us out on the

Speaker 2 (01:06:33):
iHeartRadio app, Apple Podcasts, or wherever you get your podcasts.

All Episodes

Episode Transcript

Host

Ed Zitron

Popular Podcasts

Las Culturistas with Matt Rogers and Bowen Yang

Stuff You Should Know

Dateline NBC

.css-15opob5{left:0;position:absolute;top:0.8rem;} All Episodes

.css-14f5ked{margin:0;word-break:break-word;display:-webkit-box;-webkit-box-orient:vertical;box-orient:vertical;-webkit-line-clamp:2;overflow:hidden;}Hater Season: Cal Newport on AI Reporting