Projects
![]()
No projects found
Investors
New
DAO Treasuries
DATs
Lend / Borrow APRs
Chains
Categories
Spaces
#FinanceDaily: Twitter Fights MSFT | Debt Ceiling Deal W/ Robert Wolf
Recorded: May 19, 2023 Duration: 1:41:01Space Recording
Short Summary
The transcript discusses various aspects of AI, including strategic partnerships, innovations in AI technology, and emerging trends in the industry. Key points include the integration of OpenAI's tools into Microsoft's Bing search, the potential for AI to transform industries by making production faster and cheaper, and the expectation of new jobs emerging from AI advancements. Additionally, the transcript highlights concerns about data usage and the potential for AI to disrupt traditional job markets, as well as the democratization of technology enabling global talent to compete in innovation.
Full Transcription
All right, while everybody's waiting, I'll tell you what I've been up to this morning.
I downloaded a chat GPT iOS app, not a sponsor or anything.
Just interesting.
And holy cow, it's so much better.
Open AI has its own app now, officially, and it's really good.
Yeah, you just go to the App Store and download it.
Again, we're not affiliated with them at all.
In fact, we're going to talk shit about them in a few minutes.
Okay.
Because Twitter is talking shit about them, so we'll repeat what Twitter is saying.
But the iOS app is actually legit.
You can even use it like Siri.
They have voice.
It's crazy.
I am a little worried that they know everything I'm going to be doing on my phone.
That's kind of scary.
But yeah, outside of that, not bad.
Yeah, we're probably going to wait like maybe three or four more minutes before people jump on.
Lots of news to go through.
Is everybody downloading the chat GPT app now?
Is that what's happening for people that are listening?
Yeah.
It's crazy.
How is it any different from the BingChat app?
It's like way better because the Bing chat app is like,
it provides too much search.
Like I don't need the search.
I actually feel like the search is stupid.
That's like my primary use for it.
That's all I do with it.
Yeah, so this is like full conversation.
Like think of chat GPT, like the website,
but it's connected to iOS.
So now you can start using...
It like Siri, like I was saying.
It's kind of crazy.
Series sucks in Europe.
Because I can't see a map still.
I don't know.
The Bing Chat app does have voice, too.
You can use it like Siri too.
Yeah, but again, it provides search results instead of just chat.
I don't want.
Well, it does do chat, though, by the way.
It does do that.
No, no, I know it provides chat, but it supplements it with search.
And I don't really want all of that.
It's too cluttered for me.
It feels like...
Oh, God.
You won't...
That's interesting.
Yeah.
You just want to...
If I wanted to use Yahoo.com, I would use Yahoo.
You know what I mean?
Like, it feels like...
Search feels like Yahoo.com now.
Or MSN, like, back in the day.
I just want my question answered.
I want some sources that I can double check.
And it does all of that.
I don't need all of the other stuff.
I feel like if you're still searching for search...
you know,
you,
so I guess this might be a good thing
to talk about later,
but what you're saying,
I don't want all that.
I want an app that does this.
You're describing in my head
to the Bing chat app
because it does like,
I turn it on precise mode.
I just ask a question
that answers that very specific question
and give sources underneath.
Yeah, but I've used it.
I have the Bing app as well.
Just try using the opening.
I want you'll see like the user interface
is like incredibly better.
Bing is still trying to make money from ads and shit.
And it's like, guys, like, I don't want any of that.
And the precise versus not precise and all of that,
I want to be able to talk to it.
And I want to be able to, like, prompt it the way I like to prompt things.
So, like, usually I'll use prompts, like, assume I'm a scientist.
Right.
Assume I'm an expert, how would you explain to me your decision making behind X, Y, or Z, and it's able to talk to me in that way?
Whereas what Bing does is it actually takes the prompt, and, you know, it actually hilariously does TF, IDF, which is a type of, you know, confirmation or transformation of the prompt itself.
So if you actually look in the Bing app, it'll tell you the words it used to search.
And it'll be like some of the words that you had in your.
So they're like changing your prompt to make it quote unquote more precise.
And I'm like, no, my prompts are always precise.
And even if you do that, again, like when you're doing your search on Bing, you'll see the actual words they took.
And then they use that.
And so to me, I wanted to use the real prompts that I'm talking about and the,
affect specifically that I want.
I want it to talk to me.
Like,
that I'm actually someone who knows what he's talking about.
And I think that or like...
I'll have to compare him because, like,
my brother's a dentist.
And so I was like,
asking me a question that only a dentist would know that's...
And it was doing it.
It was like, he's like, that's crazy.
Like, that's exactly what I would have told someone.
So I'll have to compare them.
Yeah, you'll see the difference really in the in the prompt side of things.
And I wanted to tell me how I want it.
Not just like what I want.
That's that's actually the main point.
Alan, I can hear you chewing, my friend.
I don't know what you're doing, but.
No, I need to know what app you're using.
I'm chewing because I don't understand which chat, G.B.
Which one are you using?
Oh, there's like a bunch of.
Weird ones. There's a thanks for for I mean, I was like, oh, is he like on on there just chewing gum?
OpenAI has the official open AI chat GPT app now and the official chat GPT app is is on the app store. There's been some versions of it before, but this is the most updated version and
Like Bing and some of the other people, they've included Whisper, which is the audio version, so you can just talk to it.
But, you know, it's really, to me, I don't need some, but what Bing is trying to do.
So, Alan, I hope to answer your question.
Open AI, chat, G.
No, you gave me two different things, Ultimate and something else.
Open AI.
Open AI.
Chat GPT.
Okay.
Okay.
thank you,
thank you very much.
You got it.
And I would go for the,
it says on the app,
the official,
JPT app.
And then when you look at it,
it should have open AI as its developer.
Both of those things are critical.
There's a ton of copy cats.
and,
So this is, you know, the app that everybody's kind of, or, you know, the web version has gone obviously viral.
This iOS version is much easier to use.
So, Justin, what I was trying to say earlier is Bing does a really good job of telling me what I need to know, like a really good job.
But it's not telling it to me how I want it to tell it to me.
So that's been the difference between search versus what Open AI does.
It's like if I want to tell me...
you know, how fucking Shakespeare would have said it, then I, that's my fucking decision.
You know what I mean? And I want, sometimes you want something, something weird like that.
Like, or, you know, and so Bing is still doing search. Open AI is truly, in my opinion, generative where if I'm in a fight with my wife and I need a good text message to send her, I can just ask this app now to generate that for me.
Like, you can't do that with Bing search, in my opinion.
Hopefully that was helpful.
And I do not use, just in case my wife is listening, I do not use chat GPT for our fights.
But, you know, the idea really here is like that's what's more exciting to me, more than anything else.
You know, it's really around, you know, can we, to me, search is no longer a problem.
I can search in like a million different ways.
But to generate something is much more interesting.
So I don't know. I mean, I've never used Bing to specifically generate messages.
Like, what would I write in this situation?
But, so think of it as more like fill in the blanks.
But, yeah, just for everybody that's joining, I know I went on a little bit of a tangent, but super interesting, opening eyes.
new app is
pretty legit.
I actually really enjoy it.
One of the things I used when I first downloaded Bing
just to mess with it, because
it uses some of the features or whatever
from OpenAI. I don't understand the details, but
is I live in Iowa.
I live in Iowa.
I drive 20 miles to work every day.
My car gets 30 miles per gallon average.
How much will I pay annually in gas tax?
So it, like, literally pull all the numbers from the correct sources, calculated the gas tax, and gave me an answer.
Like, that's, that was one of the first tests I gave it, which is...
Oh, man, that's still a search function, Justin.
Oh, man.
I'm going to...
You know what?
I'm going to use Justin as, like, a good example.
Yeah, that's what I'm saying is, like, I don't understand the...
I'll need to...
We're going to talk about this.
I did use Chat GPT, and I was like, eh.
Like, you know, I didn't pour a ton of time into it, but I was like, I don't really see how,
at least for my uses, what I was using it for...
Aren't you in PR?
Yeah, yep.
Okay, all right.
So, let's say that you have a customer that you want to generate in their voice.
What if it could look up all the different articles that have been written in that voice,
and based on that, generate...
based on the bullet points that you put in there,
their voice of what it looks like as a first draft
that you can then edit.
You could use this chat,
GPT app to do that in like 30 seconds.
Yeah, okay, I see what you're saying.
So, yeah, because Bing caps out at about,
I don't know what is it, like a couple,
like 100 characters or something.
Yeah, but Bing also doesn't generate,
it doesn't generate that entire press release.
Yeah, correct. And that's really good point because I would never use chat GPT to do that stuff. I still don't believe in it's in I don't believe it's good enough yet. Oh, oh no. Not even a first draft, Justin? No.
I don't know. I would try chat TP4. It's very good.
Exactly. I was like, dude.
Yeah, you're missing out there.
I'm all about it, right?
Like, I would love, I think writing press releases are absolutely awful, and I think they're,
you know, I hate them.
So I actually tell my clients don't do them.
So I would love to use it for that.
I'm not opposed to it at all, but I just, when I did, like the functions I'm trying so
far, I find that it uses a lot of words that are kind of common because it's basing its patterns
off other breath.
Yeah, but you're using the wrong prompts.
I'll push back.
Are you using the lower-
No, that's fine.
That very well could be.
Yeah, so like the, and I'll explain to you what I mean by that.
So you have to be incredibly specific and there's a lot of tone related things that are, they are,
so think about what you're thinking in your head as you're writing in terms of like,
oh, I want to write this in this tone, in this voice.
now you can actually tell it to do that and you can actually say hey i want to use your
the vocabulary at this reading level and it will actually take all of that into account so
uh and and by the way the more you play with it now it has memory so the more you edit it on there
it'll learn your voice and the way you like to speak so it's as if you have 10 justins which
sounds scary but you have 10 justins running around um
writing and I know press release that is probably useless especially in the age of AI but you know but putting together
more tailored more personalized versions of what people are doing when they're doing outreach when they're
outbound what is happening right now is and I don't mean to say this in a in a in a you know in a overly doom and gloom way
creative energy when it comes to actually generate like no one's very few people think of
PR as a place where chat GPT will not completely disrupt it like like if you're sitting in
PR and you think that it's not going to affect PR
I think, you know, I would say, like, you haven't really tested the essence of it yet.
No, I absolutely think it's going to change the entire industry, right?
Like, you know, and I use it heavily.
Like, for example, I'll say, instead of a tool like muckrake, which a lot of PRs are familiar with,
I'll just tell it, hey, who are the top 10 journalists writing on this topic?
You're still using it for search, Justin.
I'm sorry.
But that's the bulk of the time that, that's like the bulk of...
So understanding who I'm pitching and understanding them at a very deep human level is the bulk of my time.
And now I should clarify this with I...
I do PR very, very differently, what probably most people think sounds like a time waster.
But most of my time is spent really understanding a person's writing style,
why they write the way they do, the certain words they use.
It's a lot of nuanced stuff.
And I'll use their phrasing back at them so that they...
Exactly.
There's a lot of things that are very nuanced.
Yeah, so just to be clear, and we'll move on now.
But I was going to say that...
and maybe Alex or somebody else can also jump in.
But what I was going to say was that the beauty of chat GPT is not that is making search better.
I just want to be very, very clear.
If that's what you think it's doing or that's what anybody that's –
It's actually in generating new content, actually, is the generative side of things.
generate new information.
It's not even searching the internet
and providing it in a concise, summarized way.
Some people really like that function,
but actually there's better ways to do that,
much, much better ways than using this
because there are ways that you can do that
without having as many hallucinations.
which for people that are listening,
hallucinations are,
it just makes up the answer
and thinks it's the right answer.
You know, there's much, much better models.
And, you know, like the entire search paradigm
has been built on an algorithm called TFIDF.
And you can use that with some things that look like
Open AI, but actually aren't.
and you can have reinforcement learning to improve that.
That's not the issue.
The thing about generative AI that's super exciting
is the transformer technology,
which is actually allowing us to generate using inputs
and coming up with unique outputs.
Those outputs are what's exciting.
And that's what I think is very,
and it's very, very relevant to our conversation
around Twitter and Microsoft.
Wanting to let anybody else weigh in.
If I'm completely off base outside of Justin,
does anybody else, does anybody disagree with me about,
about sort of my underlying supposition.
If not, we're going to move to this crazy news around, you know, Twitter and Microsoft
continuing to tussle.
There's a lot of background, a lot of baggage around this one.
Yeah, I mean, this would be maybe good for another room sometime, but I would love to dig
into the nuance of language and how I use it and why it works, et cetera, et cetera.
But, you know, I don't want to embarrass you in your own room today, of course.
So we'll just let it go.
Sure, man.
That's great.
I also don't want you shilling, so we're going to move on.
Just kidding.
All right, so let's get this done.
So, you know, just to be very, very clear, you know,
we are not going to take sides in this conversation in terms of, you know, who's right, who's wrong.
But I do want people to know the information around this, what I would call big issue.
I think Ellen got dropped down, which is...
You know, which is around the background of Elon Musk and Open AI and now Elon Musk and Microsoft.
And again, if I misspeak, I did a little bit of research on this, but if I misspeak, please do correct me.
So Elon Musk was originally involved with Open AI as one of the founders of OpenAI.
He invested about $100 million and was very instrumental in the early days.
Open AI at that time was a nonprofit.
They spun off a for-profit arm, which was then backed by Microsoft.
Microsoft and Satya Nadella have now used Open AI integrated into their Bing search, which
is now an Internet Explorer or Edge, whatever they call it now.
And it has become, a lot of people have switched over to that.
because it integrates all of these tools from OpenAI.
Open AI is also building plugins,
which I think are probably the most exciting part of it,
where now you can use the rest of the internet and build agents.
And so clearly Microsoft has gotten significant benefit from OpenAI.
So the question becomes, what's going on between Twitter and Twitter,
and open AI.
Well, one, I want, again, I'm trying to get full context.
This is not, I'm not going to just talk about, you know, the, the letter, because I think
it's, it's completely unreasonable to just talk about the letter.
You know, Satya Nadala was actually reached out to by, by Musk, you know, when he was
fundraising.
And.
Let's just say that it was not the best conversation, but, oh, hold on, must just tweet it.
Oh, I think, no, this is an old tweet.
I was looking for this information.
So, you know, it did not go well.
They didn't end up investing.
And he said, oh, I'll get back to your Microsoft Teams or something.
I can't remember.
Something of that sort.
But I was looking for the original information.
But...
Yesterday, a lawyer for Twitter owner, Elon Musk, accused Microsoft of misusing the services data and demand and audit from the software giant.
The letter was addressing a pretty narrow set of alleged infractions by Microsoft, specifically around their utilization of the API.
And just so people know, you can go and get analytics off of the API today, right?
Sprout social and others use it all the time.
And so, you know, I think the key here is to understand that Twitter believes that their API was used inappropriately to train open AI.
That is the crux of it.
There's a lot more details around that.
You know, it overused the services data,
such as it's exceeding reasonable request volume or excessive abuse, excessive usage.
So they're coming up with funny little terms because nobody saw this coming, right?
And so they were using Twitter outside of the terms of agreement of the API.
That is the crux of the argument.
But this is not all that it's about.
There's obviously a much broader conversation here on X.AI.
But I wanted to start with just the basis of this.
You know, one, you know, the legal basis is obviously outside of the scope of this.
But from a business perspective, is this smart of Twitter to do?
Does it make sense for Twitter to sue Microsoft right now?
Depends on what they have, right?
Do they actually have good evidence
outlining they did this?
I'm unclear on if they actually have
sufficient evidence to show that they broke some laws.
So according to the articles that I've read,
they actually, you know, the Microsoft,
so to give you some, yeah, according to the evidence
that I've seen, Microsoft retrieved more than 26 billion tweets
in 2022 alone.
So through the API.
And that volume...
according to Twitter is outside of the bounds of what their services agreement allows.
So that is the point that they're making.
It's the volume of what they've done.
Say again, Rob, you cut in and out a little.
Sorry, I'm rugging.
I'm just in the car.
That does seem pretty excessive.
Is it coming up?
Yeah, Rob, Rob, I might have to.
Yeah.
So that was across the entire year.
Rob, you're cutting in and out a lot.
So I'm going to, yeah, that's across the entire year.
Yeah.
But the, so, you know, the only challenge is how vague this letter was.
And I am confused about what the strategy here is.
Is it just to try to, you know,
go after Microsoft?
Is it a personal vendetta?
I actually think it's a personal vendetta,
which is probably what's becoming very obvious
in the way I'm saying it.
But Alex,
wanted your thoughts on this.
You know,
what do you think about the letter that's gone out?
Are you concerned around
Twitter,
you know,
wasting their time chasing Microsoft?
Or do you think this is a viable strategy?
Alex.
Hello?
All right.
Alan.
I know that, oh, there you go, Alex.
Oh, Alex got kicked off.
What just happened?
Hold on.
Let me bring him back up.
It's definitely Friday.
100%.
Yeah, so, yeah, go ahead, Alex.
Sorry, what was, I didn't get the prompt.
Yeah, yeah. The question was, you know, what do you think about this, this letter that was sent out? Do you think that there's a viable strategy here? What is Twitter doing here? Clearly, we keep hearing how Twitter doesn't have time or money to spend on things, but yet they're going after Microsoft. That doesn't seem like a very smart idea.
Yeah, I don't know. I've not looked at the issue very closely, but I heard the earlier discussion about whether they're breaking some kind of law or something. I can't think of anything like that. It would have to be something in their terms of service where, you know, we've given you API access to get certain data and you're allowed to use it on certain terms and you've used it on terms that you're not allowed to use it on.
This goes back to a bigger point, though, which is Elon, obviously, he's been consistent, right?
He's been saying that warning about AI, warning about AI, but not until Open AI, you know, chat GPT came out and got all those users.
Not until then did he start going to the White House and going to, you know, lawmakers to talk about putting a hold on this thing.
So I am kind of, I mean, I'm a huge Elon fan.
I'm a big fan of what he's known with Twitter.
I'm a Tesla shareholder, but...
I am a little concerned that maybe he has this temptation to use the law to slow down some competition,
because sometimes that's what regulations do sometimes.
Some bigger companies like to use regulations to slow down disruptors.
So I kind of wonder if maybe that's in the back of his mind.
Certainly can't prove that.
I don't have any evidence of that.
He's never said that.
But I do have that concern.
And this is maybe a continuation of that.
Yeah, I mean, again, I'm just not familiar with
With why, so again, just to kind of remind everybody,
there's been a broad push across the industry
to try to slow down the development of AI
because there are concerns around safety
and I actually think partially around copyright,
but in reality, incumbents do this all the time, right?
Like, they want things to slow down.
So, Rob, you know, I don't know if your audio is better now,
but, you know,
Is this just another effort to slow down AI?
Or do you think that there's a part of me that thinks that maybe Musk is trying to get regulations out,
that he wants court cases that stop companies from just taking information from your website
and using it to train their AI.
I think that might be a large part of it.
Yeah, I don't know if my old days better now.
Way better, way better.
Perfect, yeah.
So I think there's two prongs to this.
The first is that 100% the letter that was originally sent by all those CEOs and Elon and others
you know, it just smacks of competition slowing.
There was no way anybody was going to turn anything off or slow things down to accommodate
competition catching up.
Broad a broader point here is that this has got less to do with AI's, in my opinion,
this has got less to do with AI's production and more to do with stopping open AI's production.
and maybe Google as well in terms of sliding them down.
There's two reasons I would think that Elon is doing this.
The first is to try and paint Microsoft and Open Hour as the bad guys.
That would be a good PR kind of move.
The second is to probably try and get disclosure.
Because that would be very useful to know, like, exactly what they used, why they used it.
Because they'd have to disclose not only the information they had, but what they did with it, which would be very helpful to him in developing a competitor.
So that seems like a very good use of money to me.
But I don't think there's probably much he can do in terms of grounds.
If there's been some policy violation, then that's going to resolve itself with essentially a settlement for money because of the data drawdown, additional data drawdown.
They were outside of that usage.
So that's going to be a money settlement, which Microsoft is probably going to be more than happy to pay to try and settle this before it gets to disclosure.
And if Elon then says, I don't want...
you to settle, I want to go to court, then we know he's going for disclosure of information
at that point. That would be my guess. I don't know, but that's my, that would be my sort of
theory on that. Yeah, Adam, I've seen that you've been tweeting about this. Did you have any
insight for us? I do, actually. So, I think it was
approximately 100 billion tweets that were analyzed.
uh pre-elon uh purchase and uh so the amount of data that exists on twitter um and to give that
you know freely or yeah without any cost yeah yeah exactly so you have to stop that i mean
he paid 40 i had him i push right he has it they already have it exactly
Okay, so...
This is why Twitter was losing money before.
We've established that, but he already has...
They already have the data, so they can't, like, claw that back.
That's what brings the lawsuit, so...
And then now he's going to use the AI he's developing to, in order to train it, basically, further, I would say, if that makes sense.
No, that's super helpful.
And again, you know, for people that are not aware, you know, Twitter is a very exponential...
platform. So every day, they're generating more and more and more information. And really, I think, Twitter, comparing Twitter to the rest of the internet, Twitter is as quick, you know, is a very quick way for...
uh people to find out what's going on in the news so it's it's to the minute to the second you can get
this information and you know we were talking a little bit about x.a i when when the announcement
first came in about truth gpt which by the way i still think is an awful name but uh but you know
i think uh it might also be affecting my liberal sensibilities but i was going to say that you know
it's provocative it's like uh uh uh
from Zoolander when he's like, it's provocative.
But I was going to say that, you know, ultimately, while that sounds great, you know,
the thing that makes it really interesting, in my opinion, is the fact that it would be probably the most up-to-date AI.
Because if it's being trained real-time on Twitter data, and let's say that they can stop...
open AI from accessing Twitter data in real time, then perhaps it can, it can serve as an
AI-generated news source and really affect corporate media. But this fight with Microsoft
might just be around, hey, you've done this already to us, we're going to get some sort of
cash settlement, we need the cash, and do we stop them from continuing to do this so that we have
some edge of...
with our new platform that we're going to be building,
that essentially will be what I would consider a clone.
I think it's just a little bit more than that.
I think it's got, you know, at the moment,
we're having huge conversations on these spaces and others
about trust of AI platforms and where they're going and stuff.
This seems to me like if you're opening up a competitor
and you have the data that was trained on,
and you think that you've got at least a smither or something
that could slither or something that could get out in the public
that discredits...
your main competitor that's going to make people or make people feel like they distrust that more,
then that's a huge benefit for you if you're launching competitive products.
That would just be my 10 cents on that.
I think that's kind of, it's got to be one of the,
I definitely think it's disclosure.
I definitely think it's just to stick it to them and try and make it so that they're less trustworthy in the eyes of the pub.
I also think there's something really interesting going on,
which is with more lawsuits around,
you know,
information on your website
being used to train AI,
I do wonder if it's going to lead
to a little bit more clarity
around where,
where's the line,
right?
Right. Like if if people are spending a ton of time putting together information and content and then that content gets scraped into the into some generalized source, right?
Like I mean, I just think about people that have like newsletters and even articles and they're putting content for SEO on their websites and they're spending all this time and money and effort.
And then chat GPT just takes it and puts it in.
And even though Justin is right that it's not equivalent to a content generating person that's in like a human that's really, really highly skilled, it's going to get there.
I would say within 12 to 18 months, it'll be better than most human people.
humans at generating content when it comes to like written content that's non-niche so like you know
when it comes to generalized content you know a simple example of this is if i wanted open
AI to generate uh some information for patients around allergic rhinitis it could potentially
generate a really good version around the diagnosis and treatment of allergic rhinitis i'm giving
this example because we actually tried it and
you know, it generates the first version,
and then you can edit it within like 30 seconds
instead of generating the whole thing.
And so, you know, it does a pretty good job of that,
but sometimes it hallucinates.
But now with GPT4, it doesn't hallucinate
in medical stuff that much at all anymore.
And so, you know, you can generate the first version of the content
and then make sure that a licensed professional reviews it
and make sure to update it.
So, you know, now suddenly the work,
you still have to edit the content,
but the generative part of it can become significantly reduced.
Well, that's all great until you start asking yourself,
well, where did the first version of that come from?
right it didn't just come out of nothing right and so i think for a lot of people that create content
i'm not one of those but you know the people that actually do the hard work of creating written content
these large language models there's no regulation stopping it from scraping their information
and putting it into a general one and then saying hey this is generative it's not just
just copying.
That's the key right there is we're kind of back at that era.
I worked for IHeart for a short period.
It was very, very part-time.
I had a radio show there.
And I remember when I came in, I was like, wait, why are some of our website, like, I can't
rank for SEO purposes.
I can't rank for certain terms.
And one of the guys who knew about web, he said, oh, yeah, years ago when Google was
out, they literally blocked search engine traffic.
Because they didn't want Google stealing their data.
So the result was we could not show up in search.
Now, they've since fixed that.
But I think we're probably going to go through that same phase again
where people start to block AIs with like the Robots.TxT file.
but like I'm just telling my clients, hey, you know, the AI is the giant robot in the box,
and it knows what's in the box really well, better than you.
But outside of the box, it doesn't know anything.
And so if we trust companies like Microsoft, I mean, if Microsoft works with content creators,
they're going to do great, right?
They're going to do really, really great.
If they don't, I imagine everyone will start to block them out and it'll be a hard road.
So the working strategy for my clients right now is,
you need to be the one to put fresh inputs into the box to teach the robot.
Here's some new information you didn't know yesterday.
And if you were the one who is constantly feeding new information into the box,
into the robot's head, you are the one that gets sourced all the time.
So that's basically the...
But if there's any information in the robot's head that is already known,
don't waste your time on that.
So for example, like in SEO, people call it like the skyscraper technique, which is take a piece that's ranking really well, make it deeper, make it better, you know, get more resources in there.
And basically people just Google stuff and make a deeper piece.
AI search has effectively eliminated that strategy, which is good because that's just junk crap content.
No one wants to, when their car is broken down on the side of the road, no one wants to do a quick Google search for how to jumpstart their car and have a three-year history of the automobile, right?
So it'll be better for everyone.
Well, one question I have for Brian and Neely is Brian, you know, or Neely actually first,
are you starting to think about all the content?
Because you generate really high quality content, you know, and your team is working on reports.
You guys are thinking about things very, very differently.
But like if these AI systems can just scrape your information and include it into a larger content,
the larger ether, as I'd call it, like in the back, right?
Put it in their code base, you know, does that worry you?
Do you start thinking about putting things behind the paywall?
How do you stop all of this?
I mean, this is like a more generalized question because Twitter is making us ask this question right now, aren't they?
So for me, you know, it's not uncommon for me to be in a boardroom meeting and I'm sharing, you know, some sort of economic insight that maybe is not out there in the public domain and, you know,
I get this, there's this awkward moment that occurs where someone in the board will say,
okay, how come we've not heard this before yet?
Right?
Like, they're like, how did two chicks in a calculator figure this out?
We haven't heard this or read this elsewhere.
Right.
And my response to them is the same as it is to AI.
Ina Garten has had the same access to flour, sugar, milk, and butter that I have.
Her pancakes will always be better.
Okay.
Her pancakes will absolutely always be better.
I happen to be able to paint pretty decent synthesis around consumer insights with economic data.
Been doing it for 26 and a half years.
I know where the bodies are buried.
Now, can AI figure that out?
Yep, they probably could.
Can they deliver it with trustworthiness, um,
kind of insights as to what to do because we speak sell side right we speak boardroom can maybe
eventually it could be there well it could just read all of your old work and speak like you yeah
I mean we don't we all of our works pay we don't have it in the ether we actually there you go that's
yeah we pay wall at all um but in just for the clients who retain us you know not even clients who project us
so like it's so my question like my audience is like you know 10
But what if it got a subscription, one subscription for all of the different things and maybe somehow use all of your Microsoft Word information if you use Microsoft Word?
I don't know what you used to put these reports together.
I have no idea.
Tablets, Sanskrit.
Yeah.
So, you know, like, what if it could actually get access to the source data and was able?
He just completely missed that, by the way.
Yeah.
Thanks, Justin.
You know, I'm just moving on.
My point is, though.
Okay, it's fine.
I guess what I'm trying to say is I'm not fatalistic about it
because I think where AI is going to help me as an individual.
Right.
is it gives me less time of trying to find what I need to know and more time for me to wonder.
And that's the difference.
I find that so interesting, Neely.
I happen to be a reasonably good wanderer.
But Neely, that's still search.
It's not the generative side of all of the stuff that's happening.
I somewhat disagree with you on that.
You don't consider that search still?
It is interesting because Donish's priorities are like creation,
and Neely, you and my priorities are getting the information we need.
This is actually a really fascinating discussion.
So that's the delineation, right?
I think that people are potentially, in my opinion,
underestimating where we're going to be in five years
in terms of what these AI systems can generate and create.
And I think that's what I'm pushing back at.
I want to be really clear, Donish, I am not underestimating AI.
I'm telling you AI is going to find me more minutes in my back pocket to do what it is that I do well.
So my question is, do you think that in five years you're going to use, you're not actually going to generate first graphs of things at all?
And that AI is going to generate the first draft.
And then you will edit it.
Or do you think that you will still be using it for search functions where you're getting all of the information?
It does a really good job of getting you the specific information that you're interested in, getting you all the ingredients, and then you're still going to be cooking.
Like that's what I'm trying to understand because I think the way it's sounding from the other side is that you think, like, look, it's going to help me get better ingredients.
It's going to help me get better insights.
And then I'm still going to be the one cook in.
And what I'm telling you is that someday, you know, you may not need to be the one cooking.
You might be the one that supervising 10 things cooking at the same time.
So there's a difference in mentality there.
And so I'm trying to show that because that's the crux of the argument of what's going on,
why people are saying, hold on, this is bigger than we think it is.
Because just a better search or better analytics or better...
information gathering is not what people are worried about.
That is not super scary, like, at all, right?
And so if that's what you think is going to happen in the world,
then I can understand why you're not worried at all.
But if it's, hold on, this can ultimately generate music better than Drake.
This can generate...
writings by using Neely's voice and combining it with Justin's voice and making a new voice, right?
Like those are the things I think people are really interested in.
And, you know, I can understand why people who have built their entire careers on, you know,
writing really, really high quality content can't even imagine that that's going to happen.
But like, it's, in my opinion, sort of like,
the people that worked at IBM and were computers there, which was like actual mathematicians,
essentially, thinking that the computer will never get there. It's like sort of, that's the issue,
I think. Let me put one final point on it. And let's go back to Aristotle. To me, what gets to yes
in decision making is trustworthiness, which is around authenticity, rigor and logic, right, as well as empathy.
I don't think AI is going to fully unlock authenticity or empathy.
I think it will do better, but I think there's still a human component that we require with trustworthiness.
The authenticity and empathy can only be bridged with human.
I do believe AI will be able to significantly amp up the logic and rigor component of the Aristotelian 3.
And I do believe that AI will ultimately lead to more trustworthiness, but in the hands and integrated with humans.
So that's, like, to me, it's around trustworthiness.
And humans still require a part of that.
I think that's a fair, fair commentary.
And so, Brian and Ed, you know, I wanted to get the other side.
I don't know if you agree with me here.
around the generative side.
But Brian, I saw you do 100.
So I wanted to get the other side of it.
You know, I understand Neely's point around that, but I still want to return.
Well, I want to add on to that real quickly and then go to, then we can talk to Brian and Ed.
But just briefly, there's a really amazing book, Dan Ariely.
He's an incredible researcher.
Yeah.
And basically his whole premise in several of his books is how humans are irrational.
We think we're rational, but we're extremely irrational.
And part of that irrational behavior is the ability to make connections across industries that no one else actually sees.
And I would dare to say a robot would never see.
So because it's irrational.
So if you could talk to a sentient AI that has feelings and emotions and they'd say, the robot would say to you, how did you make a connection between these two totally unrelated things?
Your response would be, I don't know.
It just occurred to me.
And I think that I don't know part, that subconscious part of the brain is what, because we can't describe it and we don't understand it is what we're discounting.
in this conversation. I've reached out to Dr. Scott Berry Kaufman. He's a renowned psychologist guy, brain guy.
And I was like, dude, we need your voice in this conversation because you know all the stuff that is not being talked about in these spaces.
So I'm trying, he's like, I don't know, man. I'll look into this.
So I'm trying to get him on one of these stages. But I think that could provide a lot of perspective.
So Brian, you know, wanted to get other voices in here too.
Brian or Ed, you know, how do you guys think about the role of AI in terms of actual?
And, you know, especially like with your own content specifically.
So, you know, Neely mentioned that most of our content is either private with customers or behind a paywall.
You know, are you guys worried about putting your content out there for free?
I know you guys have some free content.
Like, what if it's being used to train AI?
Are you concerned about that?
So, yeah, I mean, it's interesting. I don't know if I'd say I'm really concerned. I don't tweet for a career, so I think I'm kind of separated from the issue. I'm not relying on getting paid from my tweets. I used to run a news site, a media site, Hill Reporter.com back in the day. And...
If I was running that today, I would definitely be afraid.
I'd be, like, it's crazy how quickly people can just take a couple snippets from a couple
news articles and rewrite an article.
And there's, like, no way at all to tell that.
this was taken from one side or the other.
So it's definitely going to,
it's definitely going to be a problem sooner rather than later.
An AI can tell, though.
Justin, let Brian for sure, but go ahead.
So, so you're saying AI knows that it's taken from somewhere else.
Is that what you're, or you're just?
Yeah, I guess that that was a picture.
Like, like, I, I think that AI,
AI can tell to some extent right now, but if you,
Say you wrote a program, you said take take like a sentence from New York Times related to this article, a sentence from Washington Post, sentence from New York Post, and combine this for a paragraph and use different vocabulary. It's pretty hard.
for the AI to distinguish that. I mean, there's going to be so much that comes about. I think
where it gets interesting is when AI gets more personalized. So like everybody's going to have
their own AI bot that helps them out. And what that will happen is that AI will learn how you write,
how you tweet, the things you say, the tone that you say them in. And that's...
That's going to be a useful tool.
I've tried asking chat GPT to create a tweet around a topic,
and you can tell it's nothing like I write.
It's just boring and bland.
And even if you change the prompt to be more like, you know,
like use different words to prompt it to write in a different style,
it's still...
I'm not going to say it's easy to tell, but if you know me, you know that I didn't tweet that way.
And I can tell from replies in my comments to my tweets that several people are using chat GPT.
It's they stands out.
They're longer tweets.
They're very boring and, you know, like to the point.
But yeah, I think it's going to change things.
It depends where you're standing.
Like for me, I don't, I'm not worried about it, but I'm also not relying on Twitter revenue for my career.
So if I was, I think I would be a little bit more concerned.
So one really simple thing to remember is that we're trying to compare what computers, it's analogous to comparing what computers would end up becoming when all we have is Pong.
Like, we don't know what the hell is coming.
And I think that...
People drastically...
I'm sorry, I didn't mean to cut you off.
People are drastically underestimating how fast this is going to progress.
I mean, when you have all these companies competing against each other and all these
companies jumping in, I mean, Apple's jumping into it, Twitter and Elon are jumping
into it.
You have Google.
You have Microsoft.
You have OpenAI.
It's just going to explode.
And...
like competition leads to innovation at a more rapid rate and AI is already rapidly increasing
its skill base if you want to call it that I can't imagine what the next three or four years
are going to bring yeah and I'm just speaking from one side of this stuff sorry yeah so I can walk
yeah working theories are helpful but like people trying to understand theory sorry working theory
would be nice for a change what do you mean
Siri on your phone.
Yeah.
But what thing Siri?
I thought you said theory.
Sorry.
Well, Apple's, I just read an article that Apple's really investing now in AI.
They obviously, they missed the boat with Syria.
I mean, that was a perfect opportunity.
Also, Amazon with their echo devices, I mean, you would think they would have been
more ahead of it.
That Apple's underlying framework,
did not, like, around security and around privacy,
it kind of slowed down their AI development.
And I think that it kind of speaks to the fact
that the only way to speed up development of AI is by,
intruding on privacy and security.
And I think that's the underlying challenge is Apple is trying to stay true to its values
while also not falling completely far behind.
You know, one thing I will say, just so people are aware, you know, the, they're,
So my area of expertise actually academically was active learning,
which is part of semi-supervised reinforcement learning.
And so I actually worked very strong.
That was my academic work in medicine.
And so, you know, this is an area of expertise.
So that's why I'm speaking from a...
You guys need to understand.
I'll give you a very, very simple example.
Audiograms, which is, again, my area of interest.
You know, we have been working on trying to improve the audiogram at a crazy level.
Previously in my research when I was at Washu.
And people kept on trying many, many different techniques where humans were being augmented.
And ultimately, the technique that has come out is using...
you know, reinforcement learning through human feedback, which is the Oracle strategy,
which is what actually is underlying the real improvements in AI with Open AI.
When they're using these tools, it's getting incredibly better.
We can get down to like a few hertz worth of information for some context with audiograms
today with humans.
It's,
thousands and thousands of hertz difference between one tone and another,
and you can't tell what you're hearing is between a certain tone and another one.
With AI, we can do it down to single heart.
So we can know if specific hair cells have died.
I mean, it's like crazy.
It's like the level of information that you're getting from AI is not just better than humans.
It is impossible for humans.
When you look at the work that Google did with computer vision on retinas,
they could look at someone's retina and...
tell whether that retina was a male or a female.
Just by looking at the retina, do you know what percentage of doctors can do that?
Zero.
It's not something that can actually be done.
So, you know, we have no idea what's coming.
We have never been in this situation before.
And I think that, you know, thinking about it from a, hey, look at it today.
So I agree with Brian that today it kind of is not very exciting.
But with enough information and Brian kept on feeding it its own information, I think we would see it developed significantly.
And I think you just have to go back to people that were there for the early days of computers.
and you see where it was then and where it is now, this is going to be even faster and more exponential.
So, you know, I'm kind of excited about that.
And I want to go to Jeff in a little bit, but Jeff, you know, obviously you're an expert in this space.
But Trevor, I wanted to make sure that, you know, you wait in as well because I think a lot of people are underestimating AI.
Do you agree with me?
Yeah, I do. I think one of the things that's going to be really, really fun to watch over the next couple of years is the delineation between AI and automation has been up for debate for several years, right?
Like RPA process automation, implementation into companies that,
You know, we're kind of laxedaisal in terms of the adoption of technology in a certain sphere,
where that margin of time or productivity has been completely missed in a vast number of industries.
And I think there's kind of a crux here where eventually you are going to start to see
Process automation start becoming a more common place in small to mid-sized companies that find, you know, really obscure mundane tasks that are now taken over by, you know, some sort of automation process.
But in order to do so, you have to have AI that can help generate the process automation. I think that's something that,
is even more interesting to me because if you can have industry specialized AIs that
understand context a little bit more relevant to a specific problem, now the AIs can actually
generate some sort of reduction of mundane and allow human interaction into things that are a little
bit more complicated and abstract. So I think there's something to be said for the fact that
We all kind of, you know, think of the big, scary AI and the generative.
But I think the generative aspect that becomes most interesting is when you have a continuous cycle of a work process in a day, that you don't necessarily have the time or effort to be able to have a side project to automate.
That's where AI can have a huge influence on a reduction of layman skills or redundancies that I don't think we're necessarily highlighting.
Yeah, I mean, I do think that, so Jeff, do you agree about the generative comments on generative AI with Trevor just mentioned?
I do. I do. I think it's going to continue to accelerate and it's going to be huge.
And just I think what, you know, what you see happening right now just in terms of the solutions.
I mean, I think chat GPT may have mentioned this earlier, released a ton of plug-in yesterday.
If you go to bar.google.com and use their solution, it's already connected to the internet.
So you're getting real-time feedback on things.
I think it's going to be huge.
The area that I employed it a lot is in the manufacturing environment in terms of, you know, started in quality control.
And it's really kind of moved now into the robotics and automation element.
Yeah, you know, I wanted to, go ahead, Trevor.
Why don't you respond?
And then I wanted to go to Robert on his thoughts.
Go ahead.
Yeah, no, I was just going to ask Jeff, you know, I think, you know, the quality control space,
one of the things that I think you were bringing up, doctor, is that the, you know,
we don't even know what we don't know in terms of spectrum, right?
I think it's interesting to think about, like, the way that humans, you know, see light,
I think is a great example of what AI is going to begin to start to normalize as something that's easy.
So I'm curious for Jeff, like, where do you think that we go from here in terms of kind of the obscurities that perhaps we may not have ever recognized that have been right in front of our face this entire time that now AI can kind of.
perhaps tie some of these unrecognizable variables and bring them to the forefront.
Yeah, one of the big things was in, you know, like in manufacturing these, like the physical,
you know, so, you know, basically doing unit by unit optical characterization, like, before the
assembly was complete. And then actually comparing, like, for example, the position of a screw to, you
you know, the RF performance of that device and actually being able, you know,
intra-manufacturing process to be able to say like, hey, look, you know, when these are,
when these screws are not properly torqued, you know, your, your output power is drifting.
And, and, you know, before that would have been a unit that was, you know, closed up and moved
to the bottom of the line and then,
you know then it tests and it failed and then have to go to the repair bin and now you you can do
active alignment and you could you can correct these things before they ever hit
a backlog or repair pile.
So, I mean, in the manufacturing environment, it's been actually way before generative.
It's been in use, but it's really, it's really kind of accelerated, I was saying, in the last
three to four years.
So I think you're going to be able to, you know, really take just a whole other step.
in terms of defect prevention and just accelerating the product development process.
Yeah. And so for, we have a lot of people listening now. So I just want to make sure to plug this. If you are working on an AI company, reach out to Mario and his team. They're actually working with some of these companies and the VC funds that support them. They're also doing some Shark Tank style pitches, not just for AI companies, for other companies as well. So feel free to reach out to Mario, you DM him. And he's working with them.
closely to help them build community and try to get access to more investors and information.
So, you know, I wanted to switch the conversation over to something that is incredibly poignant right now, which I don't know if this is a victory lap yet.
I'm not sure where we go, but we have Robert Wolf up here.
So I wanted to ask him about the debt ceiling.
I know it's a pretty hard turn, but it's an important one.
You know, it seems like both sides are at the table, Robert, and they're negotiating actively.
You know, I was interested in your thoughts around Biden cutting his trip short and skipping the quad meeting.
But overall, what are your thoughts on this debt ceiling conversation?
Are we going to see resolution next week, in your opinion?
Yes.
To answer in one word, as I've said all along, yes.
We're not going to have a failure of our nation,
even though we always like to go close to the edge.
I've made a few comments about this on the past.
I don't view this that similar to 2011 or to, you know,
when the last time we went through this, as I mentioned, at that point,
Bainer had, you know, a 24-seat majority in the house.
This is a four-five seat.
I don't think we're going to see, sorry, guys, a trillion dollar digital coin.
I don't think we're going to see the 14th Amendment.
I don't think we're going to have all these different bells and whistles to try to do something in a nonpartisan way.
It's going to be...
Robert, no bricks?
No, it's going to be, it's going to be, you know, they're going to pass the debt ceiling, and you're going to have...
you know, spins from both sides. Biden's going to be clear that they passed the debt ceiling
cleanly and he didn't negotiate on it. And then McCarthy's going to say that they got some changes
for the budget, whether it's giving back some COVID money, whether they look at 2022 baseline,
whether they, I don't think they're going to change the work week stuff. And I think the vote's going to
end up being a combination of,
Democrats without some of the progressives, Republicans without the Freedom Caucus.
um you know it will feel ugly for the next few days you know bitching and moaning you'll have
you know everyone on the bully pulpit saying you know the freedom caucus members or the
progressive members yelling and screaming that this doesn't work but at the end of the day we've
passed it 78 times in the last 60 years 49 times with a republican president 29 times with
a democratic president this will pass
And then they'll go into the budget and there'll be some changes.
My guess is military will be upped a little, which the progressives won't like.
Discretionary, non-military will probably get hit a little.
And so, you know, there'll be a little give and take, but it will pass, but, you know, won't feel good.
Do you think, Robert, there's going to be any impact on growth with what the, with the, you know, if it was a completely clean bill, then, you know, none of the current entitlements would be removed.
Are there going to be, is there going to be anything that would actually affect the economy in any way?
I mean, there's been some talk with Deers.
Well, uh, or.
earnings this morning that there might be something related to to farming and farming subsidies
anything that comes to your mind that you're worried about let me just be clear from my
perspective it will be a clean debt ceiling vote
And you won't, you know, that's just my view.
This is going to be, you know, that's what the vote's going to be on.
How do you define that, Robert, for our listeners?
You know, everybody here, while being the best-looking listeners on Twitter,
they don't always know what we're talking about.
Meaning when there's a vote, the budget will not be attached to it.
So this will be a clean debt ceiling vote.
That being said...
the republicans will say that you know with respect to the upcoming budget you know there'll be
covid money coming back there'll be some things but it will be tangential to the debt ceiling
that's just my opinion but either way the debt ceiling will be uh passed with respect to
your view on gdp actually you know you know me i think inflation's going to have a much longer tail
I think the things we should be looking at is whether it's the infrastructure bill, the Chips Act,
you know, even what Musk was saying the other day about the Inflation Reduction Act and tax credits with respect to,
now that Tesla is even part of the tax credits for EV and clean and electric cars.
All of these things is going to add to growth.
Infrastra, I actually interviewed Anita Dunn yesterday at the Sault Conference.
She's the senior advisor for Biden.
And, you know, one thing I mentioned is, you know, with all the infrastructure spend,
and I'm just reiterating with many of you know, you know, during COVID, we had durable goods inflation, right?
Peloton, computers, furniture.
Now we have service sector inflation, mainly wages, labor.
Labor.
a tight labor market with that type of thing if you add infrastructure which is the fastest multiplier
of GDP growth for every dollar spent it's a 1.6 times multiplier you're going to in my opinion
even see a little more um you know inflation with respect to wages and so yeah i i don't think it's
gonna i don't think what the the sideshow you know
comedy going on in Capitol Hill is going to really impact GDP at all.
It's just a bunch of, it's just a bunch of populist rhetoric on both sides, yelling and screaming.
Talking about politically though, oh Trevor, go ahead.
I wanted to go to Ed real quick before you go, Trevor, just really quick.
Ed, you know, from a political perspective, look, with McCarthy bargaining a deal,
yeah, it could hurt Republicans, but it won't hurt any of the candidates directly.
with Biden on the other side orchestrating this deal despite and by the way I will show my political
leanings I think people have been underestimating his ability to get deals done and I think this is another
example of how he's getting deals done and negotiating deals but ed you know is he going to lose
even more face with the primary voters you know because there is a challenger out there I don't
completely agree with the challenger I don't even know if they're going to run a full primary process but
Is, you know, we saw Bernie come out.
We saw other people come out.
Are you concerned that the Democratic Party is going to turn on Biden for making a deal?
I don't think so.
And I mean, the primary, it's really worthless, right?
I mean, there's no chance that Biden is not reelected in the primaries.
I think, like, I agree with you.
I think Biden is good at making deals.
I also think that him and McCarthy are probably going to kind of agree to both take the win here.
I think I agree with Robert.
They're both going to kind of claim to have won.
But I don't think either one of them are really going to be opposed to the other one taking some of the credit.
I think that's probably going to be part of the deal itself.
Yeah, you know, Trevor, do you agree?
And just wanted to get your thoughts also broadly.
Yeah, I think so.
I mean, I was going to actually ask because we had some of this discussion yesterday,
but to Robert, you know, you bring up an interesting point about the, you know,
the wage inflation, obviously, and, you know, the specific correlation to, you know,
direct construction.
As you know, material prices have been fairly parabolic the last year.
three to four years and now we're starting to see you know material prices come down and we're
also seeing the inverse with with labor where labor has seen you know quite an increase so
i i'm curious you know as as we start pushing the narrative of more infrastructure spending is that going to
tighten the cost for builders even more, like generic home builder and commercial builders
that are already fighting wage inflation from, you know, a labor perspective or a trade perspective.
And now you're going to even have a more immediate price war with general contractors
and, you know, generic infrastructure milestones that the government's wanting to push.
Just curious of your thoughts on that.
Yeah.
You know, it's an interesting point.
I haven't put too much time into thinking,
but my probably somewhat uninformed answer would be yes.
I think you'll have some supply chain concerns,
and some people start looking at pre-purchasing.
The answer would be yes, but for,
it's hard for me to assess what's going to happen in commercial real estate,
new projects, getting funding.
But certainly, you know, the overwhelming,
amount of infrastructure spend and then look at the chips act i think there was there's been over like
400 billion of private sector investment into manufacturing in the chips act too if you look at the
chips act and then you look at the stuff that's going to happen with the inflation reduction act
predominantly on the climate action side with all the tax credits going to you know build
built clean energy stuff and renewables
All of those type of spends, yes, different materials and somewhat, but all those different type of spends happening the same time.
is one of my reasons I think inflation has a long tail
and why I don't think we're gonna go into a recession.
I think the jobs market's buoyant.
You know, we've seen, if we're being blunt,
retail sales came in stronger,
consumer credit confidence came in stronger.
So, I mean, consumer sales.
So we're starting, you know, we haven't seen,
you know, we've seen inflation come off the peak,
but no one's seen the slowdown.
Robert, you're not seeing a slowdown in consumer discretionary?
Well, look what happened in retail sales the other day.
Retail sales decelerated, Robert.
Only with now.
Your-over-year, unadjusted basis.
Right, but wasn't that?
From a high single digit in January down to flat in April.
It's empirically.
It's 25% of GDP and it's decelerated.
Yeah.
Just a question.
What I thought that it came in, yes, decelerated, just like inflation has decelerated.
But wasn't retail sales one higher than expected?
Number two, wasn't the spending more than expected on non-large purchases?
Yes.
if you're referring to PCE, yeah, before it gets revised, and that's always the problem with
government data when you're in the middle of a cycle shift is, you know, the data gets revised.
We're going to see it with jobs. Jobs has definitely been revised, and we're going to also see it with retail sales. In fact, the
The Census Bureau not only lowered the March numbers, they lowered the last 10 years of data, particularly the post-pandemic numbers with their most recent revision.
So only because I swim in this data, do I know this data?
And I understand how the media will sometimes take the seasonally adjusted month-over-month numbers.
But the way that retailers report it and what they're actually reporting today,
and Foot Locker is a great example.
I mean, Full Locker's down 25% right now pre-market because of sales slowed down since March.
And other than the grocery dominant, you know, retailer that is Walmart,
pretty much every retailer, Home Depot included, has talked about softness this spring.
So empirically retail sales is, in fact, in a deceleration mode.
So to Robert's point, though, I guess the assumption is that as the government gets closer to really deploying large amounts of infrastructure dollars, do we think that the lag between, you know, this deployment of funds and infrastructure is going to be enough to, let's say, Art Laffer's notion of trickle down, you know, is that going to, you know, brace enough confidence?
confidence where there's support for, oh, okay, now we have infrastructure spending, you know,
people are starting to get, you know, blue-collar jobs that are fairly well-paying.
And you also have price competition on the wage impact of that, right?
So I think there's there's some element of as the government deploys this capital to reroute
infrastructure and improve it, that there are going to be some sort of tailwinds that help
You know, I think some concerns or fears about financing, about stability within, you know, builder sediment, et cetera, that maybe we haven't taken into account, you know, over the last, you know, six months when we've been having discussions about commercial real estate.
Yeah.
I mean, I just wanted to go back to retail, though.
Robert, does that change your perspective on recessionary pressures?
If you're seeing retail slow down this significantly?
It doesn't. I'm not questioning what she was saying. She's an economist and this is what she lives and does all day. And I'm not questioning that there's a slowdown. My only point is it's not slow enough. The labor market is still tight. Yes, there's going to be revisions, but
in my opinion the one underlying thing that will continue to be in front of us is inflation is
going to be at four percent to north for the foreseeable future and i think and and you know i could
easily be wrong i think actually wages is going to outperform um inflation and so i think it stays
there a bit longer um
And so I think as long as wages kind of stay somewhat buoyant and the labor market stay somewhat buoyant,
You know, I think we're going to see spending, you know, stay somewhat buoyant.
I'm not saying it's going to go where we were.
I just think it's, I don't think we're going to be in a recessionary mode, but that's, you know, that's the flip of a coin we're in, right?
You ask the one person, they say something else.
Rob, I got a question for you.
It's really funny.
I mean, we went from the, hey, the other thing I just want to add on consumer confidence and just the consumer in general.
You are going to say, I mean, someone brought up the Biden thing.
Biden's not even going to really be campaigning until early 24.
There is no primary for him.
Who's ever running has zero chance right now.
I was named after Robert F. Kennedy.
Okay, just so everyone's clear.
I'm Robert born in 62.
My brother's John, born in 1960.
I was vice chairman of the Robert F. Kennedy Center.
I'm a huge fan of the Kennedys.
He's not going anywhere in this election.
Marion Williams is not going anywhere.
Marion Williams not going anywhere in this election.
Biden, as someone said earlier, he is going to be the Democratic candidate and won't even be campaigning until 24.
And so he's going to just be out on the road touting infrastructure, chips act, manufacturing, all these things that will give some wind in the back.
on what I would say is the naysayers on the economy.
And I just think that will be a larger bully pulpit at this time.
I just think that the average Democratic voter is struggling with the fact that, you know,
that the president that is, that we're supposed to be supporting is going to be 86 at the end of it.
Yeah, it doesn't make a difference.
What does is a Democratic voter.
Robert, you're saying there's no campaign.
strategy difference from his last campaign to this one?
No, I didn't say that at all.
I said, no, there's a huge difference.
Well, you have a lot more to like that.
No, no, time out.
Time out.
I did not say that.
It was a joke.
No, I thought that was clear.
No, I know it was.
And you gave me a softball, so I'm going to hit it out of the park.
He's the president of the United States.
He's not candidate.
He's not, he's the incumbent.
Makes a huge difference when you're the president.
And the Democratic Party and the president's going to take advantage of him being the incumbent.
That's all I'm saying.
And so it's no different than President Obama.
He didn't campaign until about a year later after when he was the incumbent.
My point is not about political politics here.
My point is that when you're the incumbent,
You're actually just going to keep touting.
And I think to someone's point earlier,
he's just going to keep touting his legislative victories that were bipartisan.
Chips Act bipartisan.
Infrastructure Act bipartisan.
I think that's what he's just going to keep telling.
It's a shame that one of the best jobs.
What I did learn is that I need better.
Hold on.
Hold on, hold on, hold on.
Everybody can't speak at the same time.
So we're going to go to Eugene next and then back to Robert.
Go ahead.
Yeah, hi, thanks.
You know, one of the best jobs is being former president of the United States.
And it's a shame that given, you know, given the age, that that is going to become less
and less something that buying gets to enjoy.
Robert, a question, you know, I'm actually more in your camp around where inflation is
headed, generally around where your numbers are, maybe even higher.
But, you know, we've been talking a bit about AI.
And AI is, you know, how.
has the potential to be extremely deflationary, right?
And, you know, people are worrying not about the destruction of blue-collar jobs this time,
where the, you know, the early stage of the internet, Web 1, Web 2,
was a destructor of jobs in the blue-collar sector.
Now with AI, where it is, it's going to be destructive to the white-collar sector.
How do you see AI and its forces affect the inflation over the long run?
Yeah.
So it's a great question, and I've been listening to a,
incredible group of smart people on this call before you guys let me chime in i i probably have a
different take um and i'm going to just i i think the question is how long is long before it impacts
the job the jobs market and inflation and growth
I'm just going to bring you back a little that, you know, I was one of the first individuals that started a drone service company.
And we were, you know, all hyped. Drones was going to change the future on how, you know, and change jobs, whether it's looking at utility lines and pipelines and disaster recovery and all these things that was going to change. And
And for five years, you know, we raised a ton of money and it was, you know, amazing things we could do with, you know, whether it's looking at anti-poaching in Africa.
I can't tell you the use cases for drones. Regulation slowed it down to a snail's ass pace.
Couldn't fly beyond a site, couldn't fly near airports, couldn't fly near urban areas because of, you know, privacy.
You couldn't believe how regulation stalled so many ways to use it.
We saw the opposite of that with crypto, where, you know, regulation had a lot, the lack of regulation in some ways you could argue, let it be come the Wild West.
You know, we've seen it with social media where lack of regulation in some ways has allowed it to be, you know, these open forums, whereas news companies, you know, have to live by a different regulation.
You know, so regulation has been slow to catch up. Some ways that's prohibitive in some ways that allows for this open architecture.
It's hard for me to assess what that means for AI,
how fast it's going to move,
and where regulation comes in or doesn't come in.
And so the only thing I would say is I don't think it's going to impact the jobs market
as fast as we think.
And let me just segue kind of off topic.
But I remember when I was running UBS and we kept acquiring
you know, different financial institutions. And I became the COO and I was technologically illiterate.
So I had to have a big learning curve. And the one thing I learned is no one likes to decommission anything.
So you keep on having new things, but you don't decommission.
what was old because you're scared to unplug something or you're scared to change something.
So yes, AI is going to move incredibly quickly, but that doesn't mean all of a sudden,
you know, the guy who climbs the utility pole, okay, isn't the one that's turning the screw
as opposed to the drone that's inspecting it.
So my only point is things sometimes go slower than we think.
change goes slower.
And so the question
long-winded, I apologize,
is when you say how long,
I don't know how long, but I think it will be
longer than what we think.
Yeah, and Jay, I know that
you had a separate question or comment
around what Robert was talking about, but I wanted to
let you weigh in. Yeah, so on AI,
you kind of have to look forward.
And, you know, I think
the drone comments are very astute.
I also think that
You know, with crypto, you saw the government kind of let people do whatever they wanted, and now, you know, you see what happened to Silvergate and SBNY and how they're coming after everyone.
The interesting thing is that with AI, unless it's regulated, I think it does pose a very, actually a near-term risk of jobs.
You saw the IBM announced that 8,500...
employees were going to be let go and their jobs were going to be automated. And I actually
think that, you know, Wendy's with their strategy, frankly, I think every data entry job in any
every call center job in the country eventually could be automated if it's not for regulations. And
that's just the very tip of the iceberg. So, yeah.
you know, it may take several years.
It, you know, it could actually even be regulated.
But if it's not regulated and you let the free markets do what they need to do,
I think it poses a bigger risk than a lot of people realize.
And I'm saying this as someone that's not the best technological, you know, I'm not a coder.
I'm not a technologist.
But I think it is a deflationary force.
And it kind of ties into the fact that, you know, while inflation is relatively stable now, we could see deflationary forces over the next couple years.
And I think the consumer data is showing you that things are slowing down, whether it's airfare data.
prices are coming or slowing down whether it's even Walmart yesterday the beat was on grocery it wasn't on it wasn't on
you know general merchandise general merchandise margins were done very materially and they actually talked about a weaker consumer and let's not forget that student loan payments on average four hundred dollars a month
for the average US consumer are starting probably at the end of the summer.
So, you know, I think going into the second half the year, you know, and then let's look at
LEI.
LEI is down 8% year every year, down 13 months in a row.
It is the fastest decline and longest decline since April of 2007.
And I know we're coming off of high growth numbers.
But the interesting thing to note the LEI is it would have even been lower.
Do you know what held up the LEI?
Stop prices.
Stop prices were the biggest positive contributor to LEI.
If you didn't have positive stock prices, we'd be even lower.
It's still a lot.
Yeah, sorry, Robert.
Did you have a response to that?
We just think that we're over.
We're perhaps, I mean, here's the thing that's really interesting is that people that study this every day believe even they fall into two camps.
And that's why I wanted to bring Shakar up.
So Shakar actually is a technologist.
He works at Copy AI, which is interesting because we were just talking about PR.
and content generation, and they've been doing this, and they're growing at a crazy rapid pace.
You know, I just know Shakar separately, but, you know, Shakar, do you agree with J's supposition
that we're actually going to see deflationary forces from this?
Or do you agree with Robert that's going to take longer than we think?
Absolutely.
I agree with Jay. I think we're going to see massive depletionary forces.
And honestly, I think the main thing that's preventing that from happening much quicker is one, obviously, you know, the regulatory question because I think a lot of companies are waiting to see, you know, really what's going to happen there.
But really, it's just fundamentally the adoption curve.
I mean, we've seen this impact already happen within tech.
What we know about tech is they tend to adopt tech faster.
What I'm seeing in my day-to-day, and literally, you know,
I spent a lot of my time working with customers, helping them actually, you know,
deploy these kinds of use cases live in production.
is, you know, even typical laggards are recognizing this moment.
And I think a big part of that is really the, and massive, you know, growth of Chatsybt.
Now that the genie is out of the bottle and people recognize what's possible, I think, you know, we're going to see a quicker adoption realistically because the massive transformative impact this can have on industries.
Now, there's kind of two sides of this deflationary argument because I think on one hand, yeah, you know, things being faster or cheaper to be able to produce would have, you know, top down price pressure that, you know, theoretically should make goods cheaper and, you know, better to be able to afford for a typical consumer.
I think the other thing, though, to realize is that when you think about these macro trends and how they're colliding,
there could be massive other impacts on the economy.
And Robert, I'm super curious on your take on this.
So take like the VC market as an example.
What we're seeing, right, is typically these companies are investing in software
because there's massive ROI, typically easy to build a moat.
You know, there's gains in the long run because it's hard to replicate.
If the cost of producing code and the cost of producing software goes to zero, and at the same time, we're seeing trends like, you know, plugins for GPT actually driving, you know, a different internet experience than what we're typically accustomed to, then long run, right, what does that mean about investing in software?
Taking a step further, just very, very native to copy AI, we are 36 people with 8 million users, and we've done that in 28 months.
And a big part of our growth strategy is really just leveraging AI to be able to scale.
So what does that mean for the late stage VC?
Is there really even a point to, you know, raising these giant series D, series E rounds?
Not saying that, you know, that's going to be the massive impact on the economy.
The point here is that I think that there's unintended consequences and, you know, hard to anticipate impacts because of the transformative nature of this, right?
This is not just, you know, infrastructure moving to the cloud.
Anyways, that's my rant.
Yeah, Robert, your thoughts?
I think it's very thoughtful.
I need to think about it.
Yeah, I mean, unintended consequences always comes with change,
some for the good, some for the bad.
I think my point is not at all that,
whether it's chat bots or AI or however we think of machine learning
and robotics isn't going to change the labor force.
My point was more of the timetable.
And so I just want to be clear on that.
I also want to equally be clear that, you know, productivity gains.
We've always kept talking about productivity gains over the last nth years,
especially when we went to a shared, you know, we had a decade of the shared economy,
Uber and Airbnb and, you know, all these different things.
Oh, it's going to impact the labor market like we've never seen the shared economy.
Because of all the productivity gains.
Actually, the labor market went up.
My only point is there's always new things coming that adds more jobs.
We, you know, startups, entrepreneurialism, it still drives.
so much of, you know, the future and jobs.
So my only point is that as skills change, new jobs come,
I'm a little less pessimistic on the impact of AI on the overall labor force.
I think it's great.
We have productivity gains, but I think new things come up from it.
And so I'm not as...
I'm probably not as concerned as I should be,
which is why one of the reasons I like coming on these calls
to listen to smart people who live this 24-7, and I don't.
No, it's fascinating.
And honestly, I wanted to get another voice in here.
Cal, thank you for joining us.
I just want to let you know, guys, I have to hop in the next two minutes.
I apologize.
Yeah, no, absolutely.
Cal, you know, you were at Best Buy, you were CEO Best Buy Asia.
You've also invested, I'm sure, in AI companies advantage and then now what you're doing.
You know, um,
You know, what are your thoughts, one, on the retail data on what, you know, Target and Walmart and, you know, consumer discretionary?
Are you concerned about early recession fears?
And then what Shakar and Robert were talking about around AI and, you know, our earlier conversation around AI, just putting it all in context around where do you think inflation is going to go and, you know, do you think that AI is going to have an impact?
Yes, so here's a thing. What a wonderful group of people, different points of view.
Long term, I'm with Robert, right? I think, I think, especially the U.S., you know, it's going to be different in different countries, but I think we'll...
We'll get through this and it will create new jobs, different kind of jobs, and a kind of total restructuring of how we think about blue collar, white collar, physical work, empathy work, all that kind of stuff.
That's going to happen, right?
So value will eventually, but we're going to have this, like, I don't know how long this disruptive period is going to be and it's going to be very disruptive.
And the reason why I said that, oh, let me just go back to the retail thing.
I nearly spot on what she said about retail.
I think discretionary spending is really on really tough right now.
You've actually seen public signals from Corey Berry as well from Best Buy recently.
You know, those are discretionary goods and that's where people go first and downgrade or, you know, don't upgrade.
But then upgrade cycle will come next year and the year after she signaled that to on consumer electronics.
And.
which is a good signal for discretionary items generally.
And home improvement slightly different,
but, you know, because you kind of do home improvements
when you're feeling like you might not have a job next year
or the year after, maybe.
So it's got slightly different dynamics,
but how much you spend is definitely, you know,
it's an indicator.
And I think, so Neely's dead on on that.
But what I'd say is, you know,
You know, the blue-collar jobs is what I'm watching, right?
I used to be a consultant.
I used to be a partner in a couple of consulting firms.
We created our own consulting firm, et cetera.
I used to be at KPMG.
You know, these firms are going to suffer big time.
And right now, in the short term, next three or four years,
there's a little bit of complacency because people are confused,
so they'll go to their nearest consultant.
who are hanging out in the offices, the account managers and all those people, say,
what do I do about AI?
And they'll be churning out all kinds of, you know,
the chat GPT documents to tell them about AI is going to do, right?
And maybe listening to this call and coming back to them with some insights.
but they'll be like running really fast.
And that's what they're all doing.
All the consulting companies, that's what they're doing.
And you see some of McKinsey's reports,
really drab reports that are coming out on AI.
But it's classic.
They're using their old formula to tell people about
what is actually gonna disrupt them, right?
Which is very interesting.
But you know, so it'll take a little bit of time,
but the signals are there.
PWC almost like signal that it's over
by doing the deal with AI and Microsoft.
Who's gonna create most of the value there?
It's, you know, it's not the leverage consultants that they've hired from MBA schools for, you know, for 60,000 a year and they're going to leverage them at 100,000.
That's just not going to happen, right?
Something different is going to happen.
And I think you said it or somebody said it earlier where, you know, maybe what you have is these, these, you know,
there was a cooking analogy, I think Neely had it earlier.
But you had all this, you know, you could manage a bunch of cooks doing interesting things.
And that's a different job.
That's orchestration.
That's, you know, and those jobs will emerge.
But I really do think that the blue collar and then that my view on the economy is what happens is the blue collar jobs are the ones that,
Sorry, the white-collar jobs are the ones that, I got that confused, sorry, the white-color
jobs are the ones that we all, like, kind of hold on because they spend money, you take the
jobs, you know, they're spending money in, you know, at Home Depot, buying consumer
electronics.
But guess what happens?
The highest earning person, the consultant loses their job, they're earning like 200,000
a year from IBM or something.
Now you've got this kind of, oh, boy.
our family is not going to earn that much money.
We were all set, right?
But it's not going to happen.
And then you start reducing it.
So I think it's going to be deflationary in the sense that people will spend less.
And they will start worrying about whether they can pay their mortgages, etc.
Because these jobs, there's a lot of them and they pay a lot of money for very mediocre
amounts of value that people create. It's just the fact. We all know it. There's middle management layers
and that's we're starting to see that and that is going to be so disruptive. Anyway, I hope that helps. I
I'd love to jump in on this.
Hold on Shigar.
I just want to let Robert win before he's got to jump off.
I just want to let all you brilliant guys know.
I used for Mother's Day cards, chat GPT,
and it got very mixed reviews.
I just want to let you all know that.
That hallmark may be in trouble.
Holy joking.
Listen, I appreciate you guys.
Let me join in and chime.
No, I appreciate you coming on, Robert.
Always the pleasure.
And I will say that we were all.
I look forward to next week when we passed the debt ceiling.
And you can say Robert just said yes.
Oh, God.
Here we go.
Another victory lap.
Robert just comes up.
I hope we pass it, but you guys saw the Bernie Sanders and then the right wing backlash
yesterday.
Yeah.
Yeah.
We're eventually going to do it.
We know that the progressives, this is what I said earlier.
You're going to have progressives vote now and you're going to have the Freedom Caucus vote now.
And it's going to be bipartisan because you're going to get the left and right, which is still the majority of...
of the house, center left and center right. But you know, you'll probably lose what, the Freedom
Caucus is 75 boats. So there's no question that Hakeem Jeffries is going to have to come up with,
you know, probably everyone outside the progressive group to make sure this passes.
Hey, Robert, before you head out, can I add something to your earlier point?
You said something that I want everyone to hear.
You mentioned that this will create new jobs we can't perceive, we can't imagine.
And to that point, in the Industrial Revolution, it brought on the advent of free time.
I tweeted this at you.
And we could have never seen...
the growth of the entertainment industry and professional sports that right now make up a few trillion dollars,
a few trillion dollars collectively.
So to that point, who knows what we're going to see.
And it's still the only thing you watch live.
For now.
For now.
Although I saw ESPN may go to streaming too and even come off cable.
But there we go.
I appreciate it.
So thanks, everyone.
I love listening to what you guys have to say.
It's great learning from me.
Appreciate you coming on.
Thanks, Robert.
So, you know, on that note, I know that there are a few comments to close.
Shakar, I'll let you go next and then Eugene you get to close today.
So, Shakar, go ahead.
So I love that segue that we just landed on because I think, you know, in my bearishness on what I think is going to AI is going to do in the short run, what's lost is that I'm actually a complete AI optimist.
I think, you know, the bigger picture here is actually the opportunities that AI is going to create in the long run.
And when you think about, you know, what it takes to be able to start and build a company today, you know,
A lot of the people, a lot of the smart McKinsey consultants that Cal referenced, right, are absolutely great folks that would love to be able to, you know, build their own companies.
But the scarcity of VC capital or, you know, their limitations of their own budgets or, you know, time investments or really what's limiting them from doing that.
So to that end, I think, you know, we are going to see massive transformation.
We're going to see more companies be born, more opportunities be created.
And that's something that everyone should be excited about.
That's going to create a massive excess in the world.
I think fundamentally, if I'm, you know, seeing the world of software correctly,
you know,
What that could mean, right, is a future in which software is just in time.
People don't buy enterprise apps in the same way that they think about enterprise apps today, where you have to wait for feature releases.
Instead, you know, you can literally just describe what you want and software is assembled on the fly to be able to, you know, solve your problem.
That micro trend, you know, maybe sounds a little far-fetched today, but living and breathing this space, I'm super bullish on the fact that it's going to fundamentally change a lot about the way that we interact with, you know, most systems.
So that sounds like an incredible future.
Eugene, what do you think?
Yeah, definitely.
I am also similar in Shikar's camp in the sense that I'm a short term.
I believe that there's going to be a lot of creative destruction a la Shumpeter in the short term
and quite a lot of opportunities in the long run.
I think there's going to be a,
AI is going to be a great thing for just human, just, I guess, human progress overall.
But I think the times that are dark ahead are going to be tough,
and it's hard to see what the timing is going to be like.
You know, specifically on the VC market,
I mean, for anyone who's used chat GPT to code,
I mean,
it codes pretty darn well.
I mean,
it has a lot of specific mistakes,
but if you have somebody who just knows even the bearish thing about software,
you can actually get it to do an amazing amount of stuff, right?
Like one person,
one developer created angry birds,
recreated it in two hours using unity,
right?
So the thing can code in C sharp,
right?
And they can code in a lot of other things.
you know,
like the whole question,
I think Shakar and some others brought this up earlier,
but it's like you have,
you have,
VCs who raise all these large funds, certainly in the COVID inflation era.
And basically, instead of needing 50 people to get into an MVP, you need like an order of magnitude less than that, right?
Anywhere between one to five people can produce the kind of MVPs that 50 people could earlier.
And you know, you think the VC market and the tech market is actually sort of the early...
indicator of what's going to happen in the economy, right?
Like our research shows that VC-backed companies, even though they're small part of GDP in the
beginning, eventually grow to become greater than 20% of US GDP in the long run, right?
So that's kind of an incredible thing to think about.
And if you can say that you could start the company at less, what are going to be the impacts
to, you know, I mean, a company has five people instead of 50.
That's 45 people, let's say, just in the early stage, just in the early stage of
of people who are no longer needed, what's going to happen?
And then another thing I worry about is, you know, where is that going to happen, right?
So, you know, the U.S. and particularly Silicon Valley, center of innovation, right?
I mean, I've been on here, you know, saying Silicon Valley's praises relative to even other parts of the U.S. geography,
even places like New York as being just the center of innovation, particularly with AI.
But now the question is going to be like, well, what about...
you know, when you can have these few people, like, why do those people have to come from those places?
Like, what about Africa, right? Which has the youngest demographics.
And why aren't these, some kids in Uganda going to just, you know, start to eat all of our lunches, at least on Silicon Valley, right?
So I guess that's an open question.
Actually, I don't have a strong view on that, but I think it's an open question maybe for another time.
No, I appreciate that.
On that positive, optimistic, slash negative note, we will...
Head on out. Thank you so much, everybody, for joining us. We'll be back on Monday, 8 a.m. Eastern.
And I think Mario has a space tonight that might be a finance-related space. So please feel free to
follow everybody that's on stage. Thank you, everybody for coming up. Appreciate you all. See you
Monday morning, 8 a.m. Eastern. Thanks.
