Projects
![]()
No projects found
Investors
New
DAO Treasuries
DATs
Lend / Borrow APRs
Chains
Categories
Spaces
TWITTER vs THREADS? | Twitter Data Scraping #AITownHall
Recorded: July 6, 2023 Duration: 1:25:01Space Recording
Short Summary
The discussion covers trends in platform adoption and user migration, innovations in AI and blockchain integration, and technological advancements in mobile computing. The importance of data quality and reasoning capabilities in AI development is emphasized, alongside the evolving landscape of social media platforms.
Full Transcription
Eugene, kan du høre meg?
Jeg kan høre deg bra.
Hei, mann, hvordan går det?
Jeg går bra, GM.
Kan jeg tala om hvor du var i dag?
Jeg tror du er åpen om det, men jeg vil bare spørre.
Jeg er all good for det, ja.
Ja, så du har jobbet på Metaman, så du er perfekt for denne diskussionen.
Jeg har noen tanker, sikkert.
Ja, det kommer til å være en god.
Det kommer til å være fun, mann.
Jeg trodde aldrig at det skulle bli så mye traksjon, men hvordan fungerar traksjon?
Den initiale pumpen, er det traksjon eller er det vanvittig når noe er hypet uten traksjon?
Traksjon er denne turen og den avgjørende minuten som er spennet per dag.
Er det det du bruker å forstå hva traksjon er?
Yeah, definitely. I mean, engagement comes up with a lot of things,
and I think we'll have to see what the churn is like, right?
So, you know, like what the MAUs, the monthly active users,
what that eventually is going to be.
I mean, right now there's a lot of hype,
and the on-ramps are just so easy to get onto here, right?
Like, if you have an Instagram account, suddenly you have a lot of followers,
so the network effects are just built in.
But definitely in the, you know, inside Facebook, even for VR apps,
but obviously for Instagram and others, you know,
you think about, you know, time in app,
you think of how many people are there,
All of those things make a giant difference.
So there's like revenue, obviously, is part of it too,
but that's usually last in the funnel
when we do our KPIs internally.
Anyone calling this a success,
it's way too early to even utter those words, correct?
Ja, jeg tror det er definitivt tidig.
det har hatt noen ...
De har prøvd å kopiere noen ting, og det har falt.
Men jeg vil si, aldri understå Facebook-Meta, ikke sant?
Jeg mener, det har dømt firma som Zynga,
det har dømt dem, og det har dømt firma som Snapchat.
Jeg mener, se hvor snaptchat-markedkappen er nå.
Så jeg ...
Jeg mener, jeg tror at noen firma bør være bekymret.
Jeg mener, en interessant ting er, i VR ...
Jeg var en del av Oculus, det var sånn jeg ble en del av Facebook.
you know, Zuck was there at the fridge
in our little
startup and saying, hey, you know, I bought Oculus
for 2.3 billion
because I want to make a
VR accessible to a billion people.
Ten years later, hasn't happened yet, but I know
he's still on that warpath, but social media
apps, man, I mean, that's bread
and butter for Meta, so this is right in their wheelhouse.
Definitely a threat.
I've got another question for you, Gene, by the way, surely I've sent you a
call, so you've got to accept it. Another question
I have for you, man, is there examples where
Meta launched an app, a social media app,
or Meta or others?
I know there's many examples outside of Meta,
but is there examples where Meta launched a social media app,
got a lot of traction, a lot of users,
adoption was quick,
and then it died off relatively quickly, no one ended up using it.
Has that happened with Meta? Because that's relatively common among social media platforms.
Even BitClout, the one I used, the decentralized version of Twitter,
it blew up when it first launched and then just died down to pretty much nothing now.
Yeah, I mean, look at Clubhouse, right? I mean, look at what Twitter did to Clubhouse, right?
I'm curious to see what the exact engagement numbers are.
That has happened several times with Meta, and actually specifically in the VR apps, right?
So right now, Meta Horizon Worlds, they put...
tons of money into it, gigantic team,
and they can't compete with startups like VRChat and Rec Room, right?
Meta Horizon Worlds is literally the fourth iteration
of Zuck trying to create a social platform in VR.
So there's Facebook Spaces, there's Facebook, etc.
And I think the problem with VR in general is it's a disruptive innovation,
whereas you could argue Clubhouse is an incremental innovation,
meaning it's easier for existing players to copy that technology.
Facebook has had problems in...
and issues in the social spaces, right?
So they had various apps, I think one was called Paper,
and other things in the past that tried to copy it,
got traction and didn't work.
So there is a history of that, but I mean,
you know, once Zuck is laser-focused on something,
he'll throw so many resources at it.
And the other thing to keep in mind is meta-engineers,
I mean, this is a stereotype which I hesitate to propagate,
but I think folks in the Bay Area have a lesser opinion of
av twitter-ingeniører, jeg tror det er ufært at det er veldig talentøre men vi
kjenner at twitter-stafet har blitt utsatt, det er ikke så mange
og meta bare har en armad av gode ingeniører, så de har også
gode tekniske talent som kan ta over problemet som ikke bør bli
overvåkt særlig med alle glitchene i twitter-spacene
og ting som det er så ja, direkte ansvar, dette har skett før
det har skett forløp men jeg tror meta er på vei for en virkelig krig her
og de har allerede en god
før vi åpne panelet, Sully du er veldig
bullish on threads surprisingly, or I misunderstood?
Just checking your messages in the group, are you bullish on threads?
Yeah, yeah, I think it's going to be big,
because when you look at the level of people who are engaging in and out,
and you look at the situation where twitter is already struggling
yeah i do think it's going to do really well you already seen the level of like
take up what is it like 30 million in 16 hours how many people are using
twitter as an example 337 million i think it was
so already 10 they're expected it's not 10 yet it's expected to be so
so twitter um
30 million
37 million is 10%
yes that's expected by tomorrow
but I have a statement
let me know
maybe this is
no no
there's messages going out
that they've already reached
30 million in 16 hours
I was reading it
that's fucking mental
I've got to make a statement
and I want to get some thoughts on this
and I'll give you guys the mic
to kind of take over
and get some expert thoughts here
but the statement
the reason I'm speaking in this space today
generally I'm not
I'm pretty quiet here
but it is the following
Threads can succeed alongside Twitter, because despite the user experience being the same,
like very similar design, very similar app, there's a fundamental difference that cannot be copied,
because then threads would not be threads, and that's the way they deal with censorship,
and that changes the entire experience.
So Twitter is obviously anti-censorship and really focused on free speech,
Threads er din typiske sosiala medieplattform,
likvældig til andre plattformar under Meta,
likvældig til Instagram, Facebook og TikTok.
Så denne fundamentale forskjelligheten
ergjør at de tog sammen som to plattformar
med to eksperimenter.
Kanskje Eugene kan kikke det ut,
og da tar vi det til panelet,
men er det, du vet, Meta Prime lurer på meg,
er det en god bestemming,
eller er det ikke sann?
Ja, jeg tror det. Jeg mener, klart, ethosen av founderne eller folk på toppen, jeg mener, jeg vet Twitter har en ny CEO, men Elon Musk er effektivt i mange måter i kontroll, og har en viss ethos av fri språk og andre ting.
Jeg mener, Zuck har en veldig annen ethos, og det har vært mye diskussion om dette.
You know, I think it's kind of like a country, right?
The way that an org is structured really affects what, you know, what the product actually ends up being.
So I think that you make a good argument there.
I mean, you might argue, and I know like in the decentralized crypto, for example, folks are not the biggest fans of meta generally, like as a company.
But I think there might be something there.
I mean, it doesn't...
Det betyr også at TikTok er veldig succesfull, og YouTube er det samme, det er veldig andre format, men de er både å gjøre video, så det er sikkert en verden der det kan være trull.
Men ja, ethos betyr, jeg tror, i slutänden.
Men ja, kanskje vi borde gå til Metaprime og noen andre, det ser ut som om det er sparking ganske debatt, så Meta, vil du gå inn på Meta?
Ja absolutt, og jeg er stolt over at du bringer denne temaet, for jeg tror det er en bra tema.
Mario, mye kjærlighet til deg, for jeg tror at du alltid er bekymret, åpen og interessert,
og at det alltid er gode samtaler.
Mena, mena, ikke gjør dette igjen, neste gang du vil ta på mitt vis, gjør det,
ingen behov av å smythe det ut.
Ber mig, hvordan er jeg feil? Hvordan er jeg feil?
Jeg er åpnering deg, fordi jeg definitivt er i forhold til å oppenar og i forhold til å oppenar.
Og det er noen grunner til dette.
er i minstens den enge grunnen...
Jeg retweetet ingenting,
det er en fantastisk account, Fireship,
han gjør all den gode dev-nyten,
det er veldig kul og mimi, men...
He put out a tweet today that was like,
oh man, another Twitter clone,
and he just listed like 10, 12 major Twitter clones, right?
And so it's not like there is a lack of alternatives to Twitter,
and that's why Twitter is dominant, right?
It's very much the opposite.
There's a ton of Twitter alternatives,
and Twitter is still dominant.
And so in my mind,
If you're going to beat out Twitter, you need to do something that's unique.
You got to do something that's going to pull people.
And, you know, Solomon was talking at the beginning.
It's like, oh, they're witnessing all these really great numbers and all that kind of stuff.
You're going to see that a lot of time.
You see this in crypto all the time.
There's a lot of people that like pile in right at the beginning.
Right.
You get those big spikes and then you have a giant drop off.
Right.
As the kind of newness aspect of it fades.
And then.
You're gonna actually see viability you're gonna see viability if like if it starts to grow from that point on and so I think
It's really really early to call whether threads will be good or not
I also think that the demographic
Please carry on cuz I'm not trying to interrupt you and so at what point
What number of users in what period of time would you think?
damn like
This is now concerning, and what makes this different is, this isn't some like Rando, this isn't Blue Sky, this is Zuckerberg who's doing it.
It's like basically his main rival. So both questions please, Meta, that'd be brilliant, Pons.
Ja, og jeg tror det er fairt.
Jeg tror det er for tidlig å kalle,
og det er vanskelig for meg å si,
in en år og 32 dager,
det er da vi kommer til å kunne...
Det er noe du vet,
men jeg tror det er ikke noe
som kommer til å hende
i de første par måneder.
Jeg tror du kommer til å må vente.
a year to see this play out, and you'll start
to see the trends before the beginning of the year,
like that year closes, but I think it's going to
take some time, because again, I do think it's going to
follow that pattern
of very large mass adoption
at the beginning, because your point, it is
meta, it is Zuckerberg creating this and all this
kind of stuff, and you have a lot of people who are
hungry for an alternative, or at least
you have a lot of people who are vocal on Twitter
like, oh, I'm going to leave, but then they never leave, right?
So I think that there's
a desire of a lot of people
to make this work, because there's a lot of people
that don't like what Elon's doing with the platform.
They don't like the fact that he's
kind of made more free speech
more ability to kind of, you know, challenge the narrative.
And so I think that there is kind of a desire to migrate to something.
But the thing is, what is Threads going to offer long term
that's going to make it better than Twitter?
Right?
Right? And that's a very important question
that I think that if people want to
prove to me that there's viability in threads,
they'd really have to answer that question.
And I would say that even if they
think that there's things that are more viable,
I would also say that Elon has really
big plans for Twitter. And, you know,
it's yet to see if they'll play out, right? He wants
to morph it into X. He wants to
make it into something that's
Really what he bought when he bought Twitter was not just a social media platform.
It was a network.
You have amazing minds here on Twitter, and that's a huge part of the value of it.
A lot of smart people who are here, they want to discuss things.
And even if you create something that...
might be better on paper, if you don't get those smart minds to migrate,
it's not going to be as good.
If you just get a bunch of normal people to use it, it's like, eh, it's fine.
But the reason people come to Twitter is they have really interesting conversations
with really smart people who are engaged.
And those smart people have built networks here over time.
They have followers. They have invested into the platform.
Så, inte bara må du skapa noe som er bedre,
du må virkelig offer dem noe som får dem til å forlora,
eller i minst splittra deres tid som er viktig utenom det de allerede har her.
Og jeg tror at det er en veldig høyr ordre.
Meta, don't we see that happen in other social spaces, right?
And also, how many people who are using Twitter use Facebook products?
So Instagram, WhatsApp, right, etc.
Even the never Facebookers, like a lot of them use WhatsApp, for example,
which is very much a meta product that they acquired for a ton of money.
So I think it's an open question, but you bring up great points.
So I love that you kicked it off.
Maybe we'll go to Kareem and then some others.
Yeah, so Kareem, go for it.
And then we'll go to Kristen and GP.
Thank you guys.
I just want to go to Karim.
Meta, have you signed up for Threads?
I'm going to ask the same question to everybody.
Okay, go ahead Karim, sorry.
I did actually signed up for threads, because basically I need to see what is happening there.
And regarding Meta's point about migration from Twitter to threads,
I don't think this is the pivotal point here in this equation.
I think what Meta...
threads, the make it or break it point for threads,
is the adoption and the acceptance of Generation Z
and Instagram's demographics in general for that product.
I think it is more of a shocking thing
for Generation Z.
Generation Z is not used for
the text form of content
or consuming or creating
the text form of content that much.
So when you suddenly introduce
a platform that is mainly about
counting on text as the primary content form,
it would be extremely shocking.
Would they accept it? Would they adopt it or not?
That would be the make it or break it point.
Whether it would reach one billion users or not,
that acceptance and the adoption point would be what makes it.
or Brexit as well.
So here, all what we need to see
is would Generation Z accept it or not.
From my followings
and the people I follow
and my experience today with threads
I didn't see so far
a single Generation Z person
On my timeline.
Everyone I saw was those who are
whether just pulled from Twitter
and want to go there,
or those who are just trying out
what is happening there, and so on.
That's an interesting point about text versus not text.
That's what is going to determine
if it will succeed or not.
I think it's an interesting point, and one to unlock about Gen Z versus text.
That's such an interesting point.
Kristen, let's go to you. Welcome to the space.
I think you haven't been in the AI Town Hall before, so welcome.
We'd love to hear your perspective.
Thanks for having me.
So I like a lot of what I've heard.
I agree with a lot of the commentary that I've heard here so far.
I shared some a bit last night.
and on air this morning, which is that I think one of the challenges that I see is that everyone's just trying to be a Twitter killer,
and yet we haven't actually seen anything kill Twitter.
So we've seen this narrative over and over and over again, that it's the next hot thing, it's going to replace Twitter.
There's this mass migration of early adopters who are typically the same people that adopted new technology,
but then in a few months...
It's a ghost town.
So I don't think that we can really, you know,
hedge any sort of bets based on what we're seeing
within the first week of the new user activity on there.
I think we need to give it time.
I also think it's pretty bare bones.
And for the people who always like to say,
oh, it's beta, although I don't even know
if that one is beta, but, or,
oh, it's give it time, give it time, right?
I think that's...
just an excuse like they've had time this is this is zuckerberg right when they launch something
people expect it to be polished and to be a certain way um there's a lot of things that if you really
want to be a twitter a viable competitor to twitter right twitter right now is where the action is at
like there's no spaces integration uh on this quote-unquote twitter killer called threads there
There's a lot of things that are just missing.
I can't edit my posts on there.
There's no edit feature.
There's no highlight feature.
There's no pin feature so that people can see that something's important on a profile.
I can't direct message anyone right now.
Listen, maybe all of these things are coming.
But I think that if you want people to truly think something is going to work,
to kill off Twitter,
the media or politicians
are used to a certain experience.
So you have to create
that level of experience or better.
And what you can't do is say
this is going to kill off Twitter
and give them something that is worse.
and something that has a limited functionality and limited features.
Now, I will say, I personally like the threads,
I guess the ability to write on there versus just Instagram,
but what I don't like is that I feel like I spend a lot of time decoupling
and removing myself from...
being a daily active user of Instagram when the pandemic happened.
I think that in general, at least for women,
Instagram can encourage a lot of things that I think just aren't super positive for mental health,
and so I was pretty proud of that.
Now that, you know,
like not super excited to
be back in this meta
ecosystem and I also feel like
for all of us who've talked about, you know, myself
included, railing on
censorship and big tech for the past few months
now suddenly like those same people
have all
just
jump ship to try and test this
new tool. So it also makes me
wonder, right, for the amount of time we
spend on here talking about
big tech censorship, and this is terrible,
you know, then why are we all
like jumping into this pool
where we know that, you know, the EU
for example, like, it's not launching
there right now. Why? Because of privacy
concerns. So I think, like,
for this conversation to have meaning
about privacy
and censorship, right, we have to
really
like put our money where our mouth is and say okay i'm not i'm committed to certain um privacy
practices and i'm committed to to not putting myself back in the thing that i claim that i stand
against and so that's something i was thinking about with that i'll give it time you know it's
only been a day it's hard to have i mean these are initial hot takes based off of being there for a day
um but so far one day
And it's sparked so much conversation already. It's amazing.
Thanks, Kristen, for that.
GP, we've discussed large tech companies in the past here on these spaces.
I'm excited to get your take on this development.
Ja, takk, Eugene. Først og fremst, jeg bruker ingen meta-produkter.
Først å tolke meta-produkter og threader,
jeg tror det kan være interessant å forstå rollet av rakelimitene,
uten at det er en skriping,
om folk forstår de unnøyde billene til Amazon,
av hvilket Twitter er i,
arrearer av 70 millioner,
og de unnøyde billene til Google.
as a reason for the reduction of the speed and effectiveness of the platform.
As he looks to migrate off the dependency on both of those two platforms,
people are not aware.
Elon Musk has had a demand for payment from both of those firms,
and both of those firms have...
threatened on July 1st, threatened to withdraw services,
in particular AWS, it would not pay for advertising,
and it also said that the contract that Twitter is currently subject to
is worth $510 million over five years,
and was signed before Elon Musk took over the company.
He does not want to pay that bill,
and with AWS, similarly, the contract says
for cloud computing services is also in arrears,
and a key part of what keeps Twitter ticking over.
Now, Google Cloud have a $1 billion contract,
or Twitter has a $1 billion contract with Google Cloud.
Now, people are focused on Twitter data scraping
as the reason for the rate limits.
I'd posit that it has a lot more to do with Google contracts
and the contract that Elon Musk does not want to comply with,
and similar with AWS.
Also, with respect to censorship,
let's not pretend that Twitter is a free speech platform.
Linda has specifically spoken to the EU.
The fact that Threads is not available in the EU
has nothing to do with a moral compass within the EU
regarding the censorship industrial complex.
It has rolled it out at scale here, and it continues to do so.
But, GP, what are the alternatives, though, to free speech?
I mean, I kind of agree with that sentiment, but where else do we go, right?
I mean, is Threads going to be a better place? Is Clubhouse better?
Like, you know, we all seek free speech, and there are limitations to it,
but can we, I mean, I'd love to know if there's an alternative, right?
Yeah, absolutely, there is an alternative.
I mean, there are, of course, Macedon and all these others, but they don't have the scale.
Of course, and this is the issue.
I mean, anybody who thinks that it's an overnight migration,
just like with DAPs and Web 2 to Web 3 migration
for enterprise business processes,
it's a fallacy.
This thing will take years.
Why am I here speaking on a platform
that I don't believe represents the values
of the pillars of democracy
and the pillars of free speech
and the pillars of non-surveillance?
I'm here because it's the only platform available
where bright minds collect
and speak to each other.
Therefore, I am forced to make a Hobson's choice.
I can either not participate,
and therefore deplatform myself,
or I can participate,
and in participating I can answer the question
which you've just asked,
which is, the future of secure communications
and censorship-free communications
is not with the migration of Web 2.0 behemoths to Web 3.0.
It is with the open source creation, which is very active,
of alternatives using the intersection of AI blockchain on Web 3.0,
using a tabula rasa of clean data,
free from the impositions of the tech behemoths
and the unfair playing field that entrepreneurs face in Web 3.0,
when faced with Web 2.0 behemoths who will buy them out,
or they either have to sell out to them or US VCs
and fall right back into the same tyranny.
So, JP, just like on what you said,
I mean, you are right in the sense of Twitter isn't a complete free speech platform, and we know that Twitter is ideologically driven.
So, for example, we know that, let's just take some examples.
We know, for example, Ye and Nick Fuentes were banned, but at the same time, Musk boosts,
extreme islamophobic posts.
So we know that there is some kind of ideologically
driven aspects of Twitter
and it's not a free speech platform.
But, on the other hand, you have a scenario
where, despite that being
the case, which it is,
When you look at it from a holistic point of view, a lot more is allowed on Twitter.
You're allowed to say a lot more things on Twitter that you're definitely not allowed to say on Instagram.
You're not allowed to, like there's many a people banned on Instagram.
And even Zuckerberg wrote on there that he, you know, wants it to be a friendly app.
So what does that mean? Friendly app means that, you know, there's going to be very much information.
liberal orientated
there's gonna be certain verbiages that's not allowed
so there is that aspect of it
so I think there's like more nuance to it
I mean in terms of Masterden, it was just unworkable
that app was horrendous
what gives, in my view
Instagram, sorry
threads a chance is the fact that he's got the entire
you know, behemoth of Facebook
Instagram, it's Zuckerberg
it's not just some like rando
but Brian, I'd like to hear your thoughts on this
Ja, bare en siste ord, og da tar jeg av mikul. Det er alt veldig bra i den globale konteksten.
Stats- og nationalskene, Sulaimh, som de som er i Republikan i Irland under kommisjon,
NAMAN, som bare har blitt laget, mediekommisjonen, som har utfølget institusjoner i dette
stat som er uavhengig av organisasjoner som en utstensi av armat i staten for politisk
sensurskap av medier, vil utvikle folk som meg fra å partisipere på denne plattformen
by the word of the statute in my nation.
And those same laws are hitting all 27 European Union nations at the same time.
And they include the hate speech acts which say and poorly define any form of hate,
but allow anonymous deplatforming of people based on hurt feelings
and hurt emotions, and people who simply disagree with you,
and because they are put into protected categories,
my free speech will be removed from me somewhere between the next six weeks and three months.
The only reason it hasn't been already is that I and a number of others
are standing outside the presidential palace on a daily basis here,
and that there's been a huge back...
backlash with respect to all...
Hey, GP, I think you're...
Yeah, so I'm just saying...
You're breaking up, anyway, let me just get Brian...
You're breaking up, GP,
but let me just get Brian in on this.
Brian, what's your thoughts?
Først, jeg vet ikke om dere har lagt ned de nummerene, så jeg vil ta ut de nummerene her, fordi det jeg vil si er relativt til alt dere alle sier.
Threads, det siste jeg har kjøpt, har 30 millione brukere over natten.
Nå, for å forberede det, Twitter har peakt på rundt 440 millione brukere, og det er under 400 millione i dag.
Substack has 35 million users
Mastodon has 4.5 million
Getter has 3 million
True Social has 2 million
Noster has 720,000
and Blue Sky is still under 200,000 people
Så, når vi snakker om å bli deplattformat, er det virkelig det større her, fordi vi alle har ulike ideer om hva frihet eller språk er, og hva sensori er.
GP var på vei til å komme inn på mye av det nå, i stedet for at det hender at de to grupper begynner å fjøde, og det er ikke virkelig som om regjeringen censurerar deg, du ender på å bli...
These huge censorship waves from people who are just emotional and just don't like what you say.
Soleiman and Mario, you've gone through this recently yourselves,
because the head of this platform, Elon, had retweeted a space that you had done on Russia,
and there was this huge wave to try to get you deplatformed from this place.
And if you were to get deplatformed, right now, Threads is the only viable alternative, not for social audio. There's no social audio on there now. If you were to get deplatformed and have to try to run this show somewhere else, you'd have to go somewhere else. But for an alternative to Twitter, if you find yourself...
Under the knife of the Twitter mob, the cancel culture,
Twitter is where cancel culture came from, by the way.
The entire idea of cancel culture was born on Twitter.
What they really mean when they say cancel culture is outrage on Twitter
and people trying to do whatever sneaky things they can do
to get you de-boosted or shadow banned or even removed from the platform
by having all of their alt accounts report you until you end up getting removed.
And I've had this happen to me on every platform I've ever gone on. I have not been deplatformed. Some of them I have. I've been deplatformed off of Twitter before. I've been suspended from Facebook and Instagram. I've had legal things everywhere I speak.
The more people that see something, the more likely it is that somebody is going to disagree with what I say,
and have a different opinion, and a lot of people will try to weaponize social media,
their social media followings, whatever fake persona that they built on social media,
that they show to the public.
They weaponize these things to try to take people down all the time.
And that...
For me, it's not so much that threads, I feel, should take down Twitter.
I feel that there should be more than just Twitter.
There should be multiple Twitters.
I am glad that Twitter is no longer going to be dominating the...
the entire conversation and the entire public square.
It's unfortunate that Zuckerberg owns threads.
That's not ideal.
And it made me laugh that they even tried to call it a federated network
because I don't really buy the idea of the federation thing
even with Blue Sky, Noster, all the rest of that stuff.
Let alone do I buy into that with Meta being the one who is doing it.
it. Brian, I like your core
point. Maybe it's users who win, right?
And by the way, I'm going to go to the other Brian R,
by the way, now there are two Brians, and
some others, but Brian R, maybe to tee it up,
I mean, Brian P makes a great comment.
You know, competition is great for
everyone except the competitors, and maybe
we the users stand to gain. Do you agree with that?
I agree, Eugene.
I think it's basically
an opportunity for a
wider marketplace
as far as
om vi kommer til å se yngre nivåer av kensurskap.
Eftersom vi har flere valg, tror jeg vi kommer til å se litt mindre,
men det kommer til å finnes en nivå av kensurskap,
for det er trist at under Twitter-situasjonen,
But on a grand scheme, I mean, you're going to have MySpace and you're going to have all these other types of scenarios that people are going to want to play out in their mind.
I think both are going to, Mario made a good point, and I think both platforms are going to play out quite well.
I think it is probably a necessary thing to have multiple platforms that are going to do similar things.
And I think ultimately Twitter is going to get strengthened by what we're seeing with threads rather than weakened on a long term basis.
Men det er en større problem her som jeg tror vi alle er litt forstått. Hvorfor vil de ultimativt ha disse plattformene?
Jeg tror ultimativt det er for AI-training, og vi ser det spille ut med de rakelimitter som Elon måtte impose.
AI training data is going to become phenomenally more valuable.
And question-answer pairs is kind of what you're seeing within this AI training model.
And Meta knows this, and they know it very well.
And just working with the LAMA model and seeing what they had to slurp up
to try to get some of the fine-tuning from Meta's data,
you can see that they were suffering.
Fine-tuning really requires a lot of the things that we see take place on social media.
The same thing is taking place in Reddit.
Now, we can argue over the quality of the information that's being transacted on these platforms.
Vi kan også fordå om den typen av kontent, men vi ser ikke nødvendigvis på kontent så mye som vi ser på formatet og hvordan mennesker interagerar med einandre, og det spiller ut i store språkmålmodeller.
Så meta-frågan til dette er...
What is the real endgame?
And I think if we're going to ever have this discussion over and over,
you have to keep in the room partly at least the elephant's tail,
if not the trunk, that this is an AI play for both companies.
I think that's so important to emphasize.
We are obviously in an AI space, so perhaps we're biased,
but if we think that Meta's backward-looking and saying,
oh, we just want to get advertisers and Web2-type money, no.
I think this is just very, very different.
And John, I want to go to you next.
We'd love to get your perspectives.
Do you agree? Disagree?
How do you feel about all this?
Well, good day, everyone.
I always agree with Brian.
His insights are always spot on.
But a couple of interesting things for me.
Number one is that I think that the on-ramp,
the facilitation of enrollment on the platform
is what gives you the bump.
And I think it's a big misperception to look at threads in the context of Twitter as Twitter the monolith.
Because Twitter is moving forward, and it's evolving, and it's making big changes.
So that's number two.
Number three, who the hell used the phrase Twitter killer?
Right?
So I went back, I read the New York Times article, and this is what they said.
Some techies refer to it as a Twitter killer.
That is absolute PR bullshit that was planted probably by Zuck and the guys.
Now that's a smart move, because what it does, it positions them not as a feature-based positioning,
but as a comparative-based positioning.
So they want to say that we are better than.
So that's interesting to me,
but I think that's extraordinarily problematic.
And in the final analysis, not too interesting.
Finally, the point I would say about Thread,
Thread is public school.
Thread is coddled and protected. Thread is, a la Mark Zuckerberg, a safe space. Twitter is homeschooled. Twitter is rebellious a little bit by very nature. And I think that's the defining point of differentiation between these two platforms. I'll hand it back to you guys.
Yeah, fascinating, fascinating. AJ, I do want to get your perspective as well, and then we'll go to GP also, but before I do that, I want to tee up, so we're talking about AI and data scraping, but actually going to the censorship part, which is interesting.
You know, I mean, I'm curious if anyone here knows how Meta censors people in VR, VR users, right? I mean, VR users are small, but it's really interesting to look at how they do that.
I mean, I've never been censored by Meta, I used to work there, I play a lot of VR games, but
vr-usere er rapportert at de blir sensert for all kinds av ting, fordi det er ingen merke til personale data
enn alle dine bevegelser, alt du sier, vr kan trackere deg og dine händer
og alt og alle bevegelser, og de kan faktisk forvandle det du gjør baserende helt på bevegelse
ikke engang se hva du gjør, fordi de, du vet, de eie forskerne er så så god
i å gjøre dette in-house, de har noen av de beste computer vision-ekspertene i verden, så
You know, some people like have their $400 meta headsets bricked for saying something,
purportedly some, you know, kid or whatever playing a game,
you know, like they're all trash talking and suddenly they're bricked.
So I think Silliman and others have brought up like this idea of censorship
and the ethos around censorship and how that plays in.
Some early thoughts in VR that, hey, you know, maybe some of this is very aggressive.
How's that going to play out in threads, a much bigger platform?
How's it playing out on Instagram?
So that's kind of the big question.
AGI, I'd love to get your thoughts and perhaps go to GP afterwards as well.
My thought about censorship? Well, I think about Thread to get data for Meta, I think maybe, but I think Meta has already a lot of data from, of course, Instagram and Meta.
So I think the kind of data that they can get from Thread is insignificant. I mean, they have more than enough data to align Thread.
langskt språkmodell, og så videre.
Så jeg tror ikke at det er den principale grunnen.
Men jeg tror at om du vil være,
du absolutt
trenger å bli
i minst 10x bedre
enn din kompetent.
Så de kommer til å gjøre noe som blodkjene med kjøp av ditt slags trekk og så videre.
De kommer til å inkludere noen slags AI-kapabiliteter.
Men de absolutt trenger å bli 10x bedre enn kompetenten til å være spiller.
Om ikke, så ser jeg ikke det som en succes.
Så, Asia, what's the reason? So, you bring up a fascinating counterpoint to Brian's thought around data. What do you think is meta, what's going through Zuck's mind? And I'd love to hear, you know, maybe back and forth here. What's going through Zuck's mind about that? Like, why start threats? Is it just so he can get the octagon and have something to fight about? Or what's the underlying, you know, capitalistic reason for it? Mercantile reason, so to speak?
Well, I just think if you already have the network and it's very easy to develop the technology, then why not do it? It's kind of low cost. The opportunity is there. So I will do it. So I think it's as simple as that.
Maybe Buildit will come. Brian, do you have some feedback to that, to what AJ said?
Yes. So, for general alignment, of course, it's very clear that Facebook has AI training data.
But for fine tuning, that's not the case. And doing so much work with Lama and building local models,
We can always go back to the no-mode scenario with Google.
The larger and larger parameter models don't necessarily give you better outputs.
Orca is another type of technique that is out there that is proving this quite well.
The Orca model of 13B is doing phenomenal without an internet connection and 8 gigabytes on a local hard drive without a graphics card.
You know, so how do we get there?
We get there through fine-tuning, and fine-tuning requires a lot of the type of interplay that you see that takes place in deep discussions within a Twitter argument,
or a Reddit type of ongoing scenario.
That does not take place in the main Facebook environment.
dataset, it does to a certain level, but a lot of that's been curtailed. They have some ancient data where that was more likely, but more current data, it's more like, oh, that's a great cat, or that's a great baby picture. There's no real interplay. You don't see that type of interplay at all on Instagram. Instagram is, you know, wow, you know, great duck lips or whatever.
The type of data that really gets you great training,
and this is after you build your basic model,
is the interaction that you're seeing in Twitter.
That's why Twitter's data became immensely more valuable,
and that's one of the reasons
why Elon paid so much for Twitter.
So many people making judgments, he overplayed,
everybody's projecting their belief system,
mostly politically,
very little to do with science or technology.
Men i realitet, Tesla er en AI-verkåpning.
Både utenom alt, det er en AI-verkåpning.
Og han skulle ha hatt tidligere tilgang til forståelse av hvor kraftig dette data er.
Nå blir det mer i området av den vanvarende personen, eller alldeles den vanvarende AI-scientist,
I've been dealing with AI scientists and researchers for decades,
and when you were talking about,
when I would talk about like Quora,
like Quora has another really great basis of data
for building models and for fine tuning those models,
because there are a lot of debates and interactions
that take place.
That's the structure we need
to get really phenomenally powerful AI.
And I think this is part of the reason.
There's a lot of other things.
It's not the only reason Zuck is doing this.
Let's be obvious about it.
One is to get media attention.
It's a balance off the plus to the minus.
And you're always going to see this in the world.
There is always a negative and a positive.
There's always this balancing off.
There really wasn't a balance off to Twitter.
Brian, du har definitivt forstått noen kraftige reaksjoner av dine kollega-sprekere.
Jeg vil gjerne få en tilbake og en forståelse.
Kanskje AJ og Meta, og det ser ut som om du har noen kraftige tanker på dette.
Så AJ, jeg vil gi deg en chans å respondere.
Jeg tror du har mer enn nok data allerede.
Men om du ser på strukturen av dataet,
om du ser på Facebook-kommandos,
og for eksempel på Facebook,
Facebook groups and so on,
you have that kind of back and forth argument also.
So I don't think it's adding anything.
But like I said...
That's an agreement interaction that you see in groups,
because groups are self-segregating, right?
You have people of like interests and like minds,
and there's very little hot and heated sort of debate type of scenarios.
A vast majority, and I've done some research on this,
a vast majority...
of the interactions you see within forums and groups of like-minded individuals,
are very, very finite differences.
And that doesn't give enough deterministic ideas, if you will, within an AI model.
So that data is not there.
They don't have that data, and I can tell you just by looking at LAMA.
And LAMA was one of their big data sets.
It's just not there.
Meta and GP.
I agree with that.
Oh, yes.
Fascinating discussion.
Meta and GP.
You guys go ahead and jump in.
Yeah, so I'll be pretty quick. I love what Brian's saying here, and I just kind of want to articulate this a little bit. I think people misjudge the value of extremely intelligent kind of high density node accounts that are on Twitter.
right there's a very big difference between kind of the average user who's kind of here and you
know maybe they're liking and resharing content versus someone who you know and
brian's actually like a perfect example of this right like brian is in the forefront
of this area he really knows it's and it's not just his knowledge but it's also
the network that he's cultivated, the people he
can interact with, right? If he has an idea
he puts that out, like, there's going to be a lot
of other smart people who are kind of digesting
that idea with him, and so his whole point
on, like, Facebook's quality
of data is not as good.
It's not...
It's always super easy to appreciate that.
If you don't have that appreciation for this kind of Pareto distribution,
super high density, like high level accounts,
and kind of what they bring to the space.
And I think a lot of people misunderstand Elon's intentions
when it came to Twitter.
They kind of thought, oh, he's buying a social media platform.
It's about free speech and stuff like that.
And to some degree, absolutely.
But I think Brian's point that this is about training AI data
and that the quality of the data is,
is extremely important, is one of the most important points
that I think people can take away from this space.
So power law, always important, and things that affect things a lot.
So yeah, that's a great example. GP, and then Xavier. GP, go for it.
Hey, as usual, I'm sort of saying what everybody else isn't saying.
Why are we allowing a Web 2.0
provably abuse of power
and provably unfavorable to the people,
migrate federated identity and centralized identity,
walled gardens and gated silos,
just so we can train our AI models quicker with larger data.
Why can't people be more patient?
Why can't people stop the discussion being about Mark Zuckerberg and Elon Musk,
and being about a new...
målet til å se på disse tingene, som er bare dette.
Selvsavverens identitet og proxiering av kredensialer av denne bruker
til organisasjonen, er det eneste som er behov av.
Det er kraft av dataanalytikken og alle data som er produsert
av deres aktivitet på plattformen, som er sammenlagt med denne selvsavverens identitet,
som er vendt tilbake.
to the organizations that want to buy it.
Web 3.0 and the intersection of blockchain
and artificial intelligence is the future,
and there will come,
regardless of Mastodon
and other intermittent failures,
disruptors,
the same as Twitter and Facebook,
where they'll take a decade, maybe,
But they will come and they will destroy the Palo Alto and the Silicon Valley global setting of the agenda.
Especially when you have people like Hoffenlacher, Schmidt and Kissinger saying that all global AI should be based on relativistic western values.
The only solution to that is an intersection of a global, flat, non-hierarchical, self-sovereign identity driven world.
Clean data, trained AI models using source of truth
from the sensory, emotive and geospatial web.
This is a discussion about how we massage to people
who have clearly abused their power in web 2 to web 3
for the simple expediency of training AI models
in the rush to grab a market share.
The market share will be there in 10 years.
What won't be there is the laws that allow you
to say the things that you can today, because across the western world,
they are being eroded by central power in the form of our governments,
which takes away the very notion of free speech.
I'll wrap up with this. Hypocritical censorship on an adult game that has guns, blood and gore,
where you can torture, maim, mutilate, kill people and other actors, is worried about awards,
I'd understand maybe if it was words that were obscene, but words are being censored while people run around the place cutting heads, mutilating, torturing, doing whatever it is in these player games, and they're worried about words.
gp is the pen mightier than the sword is always the the age-old question you also bring up another
age-old question i'm going to go to xavier next but you you bring up plutocracy as an issue brian p
brought up mobocracy as an issue so just some really great asial questions that
are definitely worth discussing in the current uh ai context xavier what's going on real quick
to answer you the pen used to be the pen used to be mightier than the sword
Now the pen must travel with the sword in the society that was created over the last 20 years.
And anybody who feels that the pen or the power of words alone will dethrone the global tyrannical globalist regime of the WEF, the WHO, the UN and your national governments with the techno dictatorships of Elon Musk.
And Linda, yeah, it's not a political statement, it is simply a statement of fact.
These people do not have the good of people at hand, and these people have colluded at scale with intelligence agencies for the deconstruction of civilizations and states.
And that is a provable fact, and let's not speak about free speech and censorship as the trunk and the tail, as Brian said.
The elephant in the room is that do you want to allow the migration of people who are part of the complex that is suppressing speech globally
to go forward into Web 3, or do you want to do it for yourself and maintain your freedom in at least one place in your democracy?
Thank you.
let's get savior in yeah i i i always love gp when he he lays it down um i i think i would
after all after my years in in psychology and um doing being a therapist and and behavioral health
counseling i i realize there's two things that human beings function off of no matter what
You know, I can, and if I say this is going to sound controversial, just feelings just aren't really there.
It's very cognizant, but there's two things humans function off of.
It's fear, or you can call it adrenaline.
And then they also function off something called dopamine, or we can call it euphoria.
Those are two things.
And I think Twitter plays on adrenaline.
And I honestly think threads is going to play off of euphoria.
Pleasure. Positivity. And they're not too far from that. What's already doing that is TikTok. You know, the death scroll? Well, that essentially could be threads. They just did that in Twitter format. So now we have the two ultimate things that translate.
that truly exist in human beings
is only the anxiety, or the adrenaline,
what we call anxiety, fears, and anxiousness, right?
And depression, and all those things,
well, not depression, but just all those heart-racing things.
And then we have the other side, which is the dopamine,
the pleasure, the oxytocins, all those things, right?
And that's going to be, I feel it's going to be threats.
If he...
And Xavier, does Twitter give you that sense of exhilaration?
It kind of gives anxiety too, doesn't it?
Oh yeah, it does both, but even then, there's anxiety.
What is that?
Again, that's a cognitive positioning, right?
It's really just adrenaline, what we really have.
And what we associate with that adrenaline, well, that's a feeling.
So you can call it anxiety, but it's a rush.
And that's what we get from Twitter.
And I feel like when we have that rush,
well, I guess it's like, you know,
getting a cigarette after getting a drink or something, right?
It's just that it balances you out.
I feel like Threads is going to be that.
I feel like it's going to be almost like Lyft and Uber.
I feel like people are going to have both.
I feel like when they're having that rush,
I feel like Threads essentially might be that.
And I feel like because of those two things,
I think it's going to burn society out, hands down.
Wow, that's a bold claim.
I'm going to get, I do want to bring in the new folks
who just joined Farrell and Strangeloot,
but we'll go to meta.
But before doing all of that,
there is a purple button.
This is the hot, contentious topic.
Please do feel free to comment,
and we've got a team in the back end.
If you, you know, want to comment, that'd be great,
and we can bring you up.
We often do when we review,
and I do actively review them.
Also, Mario's company, IBC, incubates and accelerates
both AI and Web3 companies,
partners with VCs and funds,
Work with portfolio companies in return for equity, zero cash.
If you're interested, DM Mario and his team.
We've got a call organized.
There are these Shark Tank-style pitches.
I was actually just in a crypto one. It was great.
We're going to do some AI ones also.
So if you're a startup or portfolio company, hit up Mario, and we'll begin the process.
And also, don't forget to subscribe.
All right.
Meta, go for it, and then we'll go to some of the new folks.
Farrell and Strangely, who haven't spoken yet.
So, Meta, what do you have?
Ja, så jeg ville bare reagere på GPs fråga om pen er mindre enn hård, og så trodde jeg at jeg også skulle ta med det temaet av spesiet, som er skraping.
Så når det kommer til pen er mindre enn hård, jeg tror at meme, den vare av meming og internettkultur om memes,
og memes som blir brukt til å influere politikk, til meg er en av de mest...
powerful arguments for the pen being mightier than the sword.
I don't think we would have gotten Trump back in 2016
had there not been such kind of this meme energy around what he was doing.
And I think that that continues to be a really big thing.
And I think that the value of having Elon, you know, get on Twitter
and try to open this up and try to make it more free
is that we are going to see more content, right?
We're going to see more memes, more...
Intellectual discussions, more things that are closer to the truth, right?
Because I think that the best way to get at the truth is to have an open conversation,
what you guys are doing here in this space,
and what I see that happens between a lot of great accounts on Twitter
is this kind of no-holds-barred throwdown of what is actually the best idea,
and I think that's really, really valuable,
and going back to what Brian had said earlier,
really valuable for the AI to be training that data off of.
To bring up the idea of scraping, as I see it as the topic of the discussion, and maybe something to touch on here, there's a lot of people that were very unhappy about the whole rate limiting that Elon did, I think it was like July 1st or so, beginning of the month, and
Min reaksjon når Elon tar beslutninger er alltid å forstå at han tar en smart beslutning,
i stedet for å forstå at han tar en dårlig beslutning,
fordi jeg tror det er mange folk som vil reakt på det som en nederk,
å, min konto er utvalgt på nivå, dette er dårlig, hvordan kan Elon ta dette?
and just assume he has a good reason that he was kind of playing,
he was kind of weighing all the different pros and cons,
and definitely there's a really good thread,
or the person just kind of used Twitter Blue to just dump a bunch of words,
but a really good reaction that went viral,
and it was basically explaining like,
look, this is a reaction to the fact that a lot of companies are scraping Twitter data,
Særlig siden de slått API-en som en måte å tråde deres LLMs,
å tråde deres modell.
Og en av de running themes i diskussionen vi har hatt så fort
er den ideen at data i seg selv er en så viktig ting.
Og så det tar mye senere, i min mening,
for Elon å vilje å slå det av,
særlig til andre kompetenter som vilje å bruke det,
right, because that then again strengthens
his model if he's the only one that has access
to Twitter data, and Twitter is like the
epicenter for all these really intelligent
and smart takes, and so I think
that when it comes to scraping,
and it comes to the rate limiting and stuff like that,
I think Elon will continue to evolve, that's always an
arms race of, you know, innovation
on both sides, but
you know, I haven't really seen,
since like the
A little bit initially it was kind of weird and wonky for me,
and there were certain things that I couldn't send.
But since then, everything's been fine,
and I'm kind of back to using Twitter as normal.
And so I think a lot of people who reacted very negatively to that,
I think they weren't giving Elon the benefit of the doubt.
They weren't kind of acknowledging that that could change over time.
And I think that kind of understanding
that there's this larger kind of positioning at play
that goes on in a lot of these things,
and kind of giving credit to Elon
that he is trying to make good decisions for the good of the platform.
Man, it looks like we got some reactions.
Maybe we go to the Bryans, the double Bryans,
Brian squared.
Brian P., then Brian R., and Alan,
welcome to the stage.
So, Brian P., let's go for it.
So, when you were talking,
one of the things I was thinking about
was that the last, I mean, three,
tre presidenter vi har hatt i USA,
fra Obama til nå,
har stort sett vært av social media.
Og jeg har vært kjørende på spesier,
spesifikt,
DeSantis og noen av de andre kandidater
som har startet,
som er to år siden,
på spesier som har startet
de gråtskjørende tingene.
And so, the idea of the pen being mightier than the sword, the conversation that even the people that are going for the leadership roles and win in our country, the presidential seat, they all know the value of communication and the value of social media and politics.
And the value of all of this stuff.
And my prediction for this upcoming election is that this exact topic that we're talking about, the data scraping, is going to be a huge factor in who is going to be elected president, who's going to be elected, every federal office.
Because I know for a fact that there are already very powerful politicians and candidates and teams that are already...
very deeply entrenched into Twitter alone,
let alone Facebook and all the rest.
And Brian, to specifically talk on that,
I think there are some AI thought leaders
that are so concerned about what AI is going to do
in the next upcoming presidential election, right?
Oh, yeah, and on every end, on just deep fakes like crazy, that's what usually gets talked about, the deep fakes, because that's the front end of it, but the back end of it that we're talking about, the data scraping, is also super important, because never forget that Cambridge Analytica was like a huge factor in politics, and it's these very subversive things that make little tiny snowflakes basically falling that make the entire snowball at the end of the day in the avalanche, and
end up really changing the power just as powerful as the sword can be.
So I do think that the pen can be just as mighty as the sword,
and social media is where we're going to see it.
I mean, you have been seeing it.
You guys have been having some presidential candidates
from all over the world on your podcast, or space, whatever this is.
But I really do think that this is the thread that needs to be followed
through the next election.
Yeah.
absolutt legit concern for sure um you know and also by the way maybe the pen i like i like the
fact that this uh this meme is starting to uh you know circulate through our discussion perhaps the
pen becomes the sword right as llms and stuff emerge from from written words uh so brian
brian r what do you uh what do you think about all this
Wow, Eugene. Brian, right on. I agree with him. And GP, absolutely beautiful poetic insights.
Kind of tying this all together, I think this is the first time in history where I think humans have begun to realize just how valuable their context, insight and wisdom is becoming.
Vi tenger å tenge på det skraping som en kompaniens egen verktøy. Det er det ikke. Det er egenverktøyet du har.
Du er åpner det på en transaksjon av noe gratis, denne plattformen.
Men vi må ikke minst den fakt at dette er en kontekst, en innsikt og en vissdom fra en menneske.
And I have this little project I call the SaveWisdom.org project, and the idea is to save wisdom, personally, locally, on your own AI device, or devices.
And we show you how to do that, and we're going to continue to show you how to do that.
Eight billion insights are a lot more valuable than one or two history books.
So talk about federation, that is the ultimate federation, your own context held by you and saved for generations.
So now that we see that our context is valuable, how does that monetize?
What does it look like in the future?
And if we were to reinvent that,
Social media, and I know everybody's been trying that with Web3 and all these different tokens. Let's just kind of put that aside for a second. What if we really realized from day one that everything that we do, everything that we put out, I call our context.
har en dollar-nummer på den.
Du kan enten givet det til en advertiser direkte,
der de betaler deg for den bit av kontekst,
å kutte ut alle middelpartiene,
eller du valgjer ikke det.
Verden ser utrolig annorlunda ut.
Og om du går på en par enkle kalkuleringar,
bare i dagens dollar,
din kontekst for en avgjord livsperiode er omkring 1,6 million dollar.
Det er en utrolig møte av pengar som hver individu har om det ikke er mediativ, det er ingen middelparti.
Så, i slutänden, vi kommer til å begynne å begynne å begynne i denne direktionen, om noen liker det eller ikke.
Vi kommer til å bygge våre eget AI-modeller på en måte.
Vi kommer til å forstå at vi ikke vil ha det i kvaliteten, av tydelige grunner.
Så det kommer til å være en logisk konklusion. Hvordan vi kommer dit kan være litt svart.
Men det er det vi ser med disse plattformene. Hva gjør vi når vi er på sosiale medier?
Det er mange ting, men en av de tingene vi prøver å gjøre er å prøve å dele vår visskap,
vi prøver å dele vår mening, og potentiellt noe vissdom.
And if you were to keep that locally, and it was federated throughout the entire quote-unquote new social media platforms that AI can form, it is going to look dramatically different. And, by the way, censorship of that type of system with 8 billion views...
is going to be a lot more difficult.
So that's kind of where I'd like to kind of focus myself personally,
is what's the value of this data if scraping has become such a concern.
And by the way, one quick thing.
What's transpiring in Reddit is a really good example,
if you really examine it.
From all sides.
The rebellion that's taking place there is based on the idea
that you have community leaders that start their subreddits
and they become sort of the fiefdom,
the ability to be, let's call it a mini Elon Musk
in their little community.
You can see what's kind of taking place in there.
It's like, who owns this data?
And it's funny because everybody is, you know,
a lot of...
chiefs and no workers going on here.
The reality is it's those individuals that own their data.
We just don't realize it yet.
Brian, by the way, Alan Boyd, welcome to the stage, we're going to go to you, but again, another, I mean, just really interesting analogs, right, to the historical world.
Brian P. brings up mobocracy, Brian R. brings up the concept of revolution and rebellion, and it's all happening in a digital space.
So, really interesting and also exciting times. Alan, what do you have for us? Welcome to the stage.
Thanks so much, guys.
Thanks, guys.
Så, you know, I believe that Twitter has the brightest minds, that's why I'm here, it's the most nutrient-dense information available.
But as someone who runs a marketing agency, I do want to just make a couple quick points and make a case for Threads.
We've already seen that it's the fastest platform to hit a million users, even faster than ChatGPT.
A lot of the limited functionality that everyone's pointing out, like it doesn't have spaces and it doesn't do this or that, I mean, they can always extend the functionality, so I don't really see that as a problem.
Unlike other platforms like Truth Social, Mastodon, Blue Sky, Nostra, etc.,
It feels like this platform has that it factor that's kind of hard to explain.
And the last thing I think that Threads has going for it is that psychology.
What is that X factor, right?
Previously, Karim talked about Gen Z and its adoption of it.
You know, TikTok clearly has an X factor over, for example, the failed Quibi, right?
By Meg Whitman and Jeffrey Katzenberg.
Tons of money, but didn't have the X factor.
So, Alan, what gives the X factor to Threads?
I think it's really hard to pinpoint what it is.
I mean, one thing is obviously going to be the distribution of it
and the easy on-ramp to it.
I mean, Twitter's always kind of been,
at least for me personally,
looked at as like a more intimidating platform
because you actually need to be skilled
in what you're talking about to have a real voice.
Like you can get dunked on, you know,
and exposed really quickly on Twitter,
so it can be intimidating,
but the psychology of it,
i think is a huge factors whereas like just look at crypto with bitcoin and
ethereum and you know the psychology of people uh rushing to the altcoins to be the early adopter
to to see the more upside i think this is just another play on that psychology and
people's emotions to be the first mover on a new platform so
I don't think threads is going anywhere. I think this is a real threat to Twitter. And I'm interested to see how this plays out. And I do agree with everyone that, hey, look, I mean, it's a, you know, you're selling your soul and your data on the meta platforms, but I really don't think a lot of people care.
Det er så intressant. Så snakket om det, og vi går til håndene, men David i publikum, dette er ikke verifert, men han sa noe veldig bekymrende.
Han sa, har noen mennesket den forståelse at man ikke kan løpe threader uten å løpe Instagram etter at man downloader den nye appen?
Jeg vet ikke om dette er rett eller ikke. Jeg har brukt threader. Øverigens er det en dag gammel eller så.
Men den erfaringen, jeg vil si at on-ramp, mine initiale impressioner er likaså som Alan sa.
The onramp is so easy, and you're connected automatically to the greater Facebook web.
So that is something to be said.
Far-L, what do you think of the situation?
Jeg har ingen hårdhet i denne frasen når det kommer til threader mot Twitter.
Jeg har aldri gillat å bruke sosiale medier til nære.
Mest av mitt aktivitet er i AI-spacet,
fordi Twitter har vært og fortsatt er det beste stedet
for å stå opp til datum i samarbeid med andre grupper,
men det har vært en fantastisk opplevelse på Twitter,
spesifikt i AI-spacet.
Og siden dette er en AI-stavning, vil jeg gi deg en AI-perspektiv.
Jeg vil tilgjøre Brian at data er det primære direktivet,
og det handler ikke om kontentet av data,
som er kunst som er inbredt i data,
men mer om interaksjonen, psykologi,
forståelse som du kan finne der. Det er en gulvmine, ikke sant?
Most people's experiences with LLMs at the moment is using them as knowledge repositories.
So you ask ChatGPT, what is this, what is that?
You'll get some hallucinated answer because it fills in the gaps by trying to predict
what the answer should be.
But the LLMs real power is as reasoning machines.
And what we mean by that is you can put them in a loop and chain them together to create a sort of thought process to solve a much more complex query than just what is x, right?
So...
That means that...
Farrell, I want to dive deep into that part,
and I want to let you continue,
but specifically, can you describe for our audience
what is important about this data?
I think Brian talked about it,
core data versus regular web data,
and you brought it up as well,
and I'll frame it in this context.
So when did this MetaSwap team, do we suspect,
when were they created to create Thread?
Because obviously we knew when Elon bought Twitter...
and then more recently he started to say hey you know we're gonna you know sue a bunch of ai
companies or or you know at least threaten to do that and then of course recently he did the
the rate limiting uh on the data scraping which coincided with the launch of thread right so it's
not like thread probably happened over the weekend but i guess you know the timing would be interesting
because it would also point to the motives of zuck right and zuck and co and how much data
matters right so so maybe kind of like
what do we think the motives were like you're i think you're affirming that but also the timing
as well as um you know basically who um you know like how does this data how is this data different
right like can we can we define it maybe eli five how's this data different in nature i i would say
uh from two perspectives right the first perspective is that these are let's use
the word threads of short form chains of interaction data or or thoughts across different
individuels,
and that's very rich data.
It's not all the data that they're going to use
to train models, but it's very rich data
in terms of interaction, reasoning,
psychology, and
everything necessarily. It's not necessarily
about the content of the knowledge, right?
It's not about, oh, this is
an accurate response
to this specific question.
But, to me, at least
in my opinion, the current LLMs
trained with
almost all of the knowledge of the internet, right?
Then they're lobotomized to be able to be aligned
and provide you with the right answers.
But the next generation of generative models,
in my opinion, and based on my research,
I'll just preface it with that,
they will be small, they'll be hyperlocal,
and they won't need to learn from all of the knowledge of the internet.
But rather, they'll need to be trained to learn how to reason better,
and then they can retrieve, process and learn from new data in real time.
So, as you can imagine,
the challenge becomes less about
let me collect all the high quality answers to my questions,
but rather let me collect data
that can help me train a model to reason better.
And again, in my opinion...
Short form interaction data like Twitter and threads is a goldmine.
Now it has to be processed and there's a specific architecture that you need to be able to train it on,
but that's where these models are headed, so that's where we should be looking at when it comes to data.
So let's not kind of mix the two, right?
It's good to get...
Let's say a bunch of research papers and train it on that, but that's not the objective with the next-gen models.
That's interesting. Yeah, personal models, a lot to uncover there. GP, we'll go to you and then the hands.
Thank you. All the companies that have been spoken of today call my country, the Republic of Ireland, or properly ERA, either they're global or they're EMEA HQs.
78% av vårt kåpetagstak kommer fra 10 U.S. multinationale kåperasjoner,
den største er Apple.
Vår regjering har kvatt med Apple
ikke til å aksepte 15 biljoner av kåpetagstak
for de engelske menneskene mot EU-partiet.
The data centers that hold your data for your AI training models are located in my country,
where our data protection commissioner has failed to adjudicate over 96% of the complaints received regarding breach of privacy by all of the companies you speak of here today.
Takk for ating mediet.
I've heard the words rebellion, the power of the pen versus the power of the sword,
and I have heard it talked about in the abstract,
and my dear love of AI and all things technology.
But also on my timeline, you will see the physical representation
of what censorship at scale,
and the devolution of a democracy to authoritarian values,
driven by, bought and paid for politicians
who own the companies you speak of, results in, in a country.
Namtlig Eire, Republic of Ireland, som bare er uleenskapsfamilie i Britskia for 100 år.
Vi er Sandboxen, hvor dødene som du trenger dødene dødene er lokert og konsumerer 40% av vår elektriske gridd,
hvor sin forbruk er favorit, i stedet for domestisk forbruk av våre børn.
Det er ok å snakke i det abstrakt.
Det virkelig...
er hva som går an i Irlanda nå,
og disse fyrkene du snakker om, og deres heggepån,
er et direkte resultat av devolutionen av
de personale frihetene av fem millione iriske børn,
and we are the sandbox for the playground of the lawfare
that travels with these companies and will be rolled out in the rest of the EU.
I owe it to my countrymen and women and myself to make that statement
because it dovetails perfectly with the AI town hall
because the thing that drives wealth in Irish society
as the notional sixth richest nation on this earth
is not the wealth of the people, but the per capita GDP,
which is irreparably high.
I will wrap with this.
I would appeal to people
to not close your eyes
to what is going on in my country.
I am not a political person.
So, I...
Sorry, go ahead.
Ja, jeg var bare på å putte noen nummer rundt, og jeg tror vi trenger spesialitet til det, men Irland har det siste høgste GDP per capita, 114 000 eller så, av IMF, men bare 5 millioner mennesker, som er små i dagstiden, og tax breaks og incentives, som GP har snakket om, som har sitt eget spesialitet, og vi vil åpne det i en annen tid, men det er veldig viktig. Xavier, la oss gå til deg, kommentere på alt dette, og så går vi til Brian etter det.
Ja, jeg elsker data.
Og jeg er glad at alle er til å ta fram dette nå.
Vi snakker om freder, og om data-konneksjon,
om hur fremtiden, og om denne personens kraft.
Jeg føler at fremtiden kommer til å være en masse
av fyrkene som kan tilføye personer, UUIDs.
Essentially, they're phones.
They're essentially the specific numbers that connect you and identify you with that phone or that computer.
And I see the future as, especially if you think about each of these social media platforms,
they're their own individual databanks.
And within those databanks, we have our individual accounts, which those are databanks.
Well, I see the future going beyond that, and just turning our phones, or using our UUIDs with the connection with our IP addresses,
and each of us may have a special IP address, almost like social security numbers,
where that's our individual IP address, with an association with the UUID, with our phones,
and I could see us storing, essentially, and this would save companies millions of dollars, personally, and the world.
that our phones connected with our own personal IPs,
because those are expensive too.
And you're talking about the models actually running on the phones, right?
Not just like accessing to the cloud, like at GPT?
Is that what you're saying?
Hands down.
I see that people, I see the future of the phones,
the future phones or whatever we're going to use,
we're going to be the databanks.
No longer will the latency and all issues,
I feel like that will solve the latency,
where every individual is their own databank.
I hope that the future you're talking about is bright, but this is an AI space, so just to get into it though, right now the NVIDIA H100 reigns supreme for things like 16-bit training and 8-bit inference.
The consumer NVIDIA 4090, which is a gaming GPU with 24 gigabytes of VRAM monster, it does about a third as well as...
as the H100 for 16-bit training,
but it's not yet on phones, right?
I mean, you don't need 16-bit inference, for example.
But I'm curious, what is the timeline, right?
Like, to maybe break it down for folks,
when can a normal phone, right,
in our pockets run AI models, right?
How soon is that coming?
I personally think it's just that we already have somewhat of that structure.
If you think about it, we have clouds.
And if you think about it, companies are charging you for clouds.
Well, no longer the cost of that storage is on the company no longer.
Now you just tell a person you have to add more storage,
so you have to pay for extra storage for your phone.
So now the expense is on the user and no longer on the company.
And I feel like that will solve the whole latency issue.
Det er interessant. Før vi går til Brian, Kee, AGI, vil du kjøpe inn?
For det hører ut som om AGI eller Farall har en syn på hvordan vi alle har vårt personlige AI på våre telefoner.
Jeg er bekymret for tidspunktet. Hva tror du?
Jeg var i trening av en AI-modell som vann i gamen av Go, slik som DeepMind.
Og jeg var i trening av den AI-modellen på en iPhone 5, 6 år siden.
så vi kan allerede trene AI agent på telefonen.
Jeg tror vi kan allerede rulle noen modeller på telefonen.
Hvis vi vil ha kraft av denne typen av GPU,
NVIDIA RTX og sånt på telefonen,
så tror jeg vi kan se med den progres som jeg ser,
I think one year and a half from now we could have that kind of computing power on the phone.
I cannot guarantee that, but I see a lot of progress with the design,
using AI to design that kind of AI chip.
I think it's going to go very fast, but we can already do a lot of things with the phone.
It's just people are just not leveraging that yet.
Men, before we get... Yeah, go for it.
Yeah, I'll just add that there's already a lot of indie developers
that have managed to run models, open source models,
like Alpaca, Lama, and so forth, on iPhone 14s, right?
So, inference is already here, right?
Training is a different story,
but it's not very far away for us to run...
Almost, especially if the models get better, and they will get better, they are getting better.
We're getting models that are a hundredth of the size of ChatGPT, or GPT-3 specifically,
that are performing almost equally.
So, it's getting there.
It's going to probably...
be there within the next
year or two. We're going to have LLMs on the phone.
They might not have everything
that you have with GPT-4,
imagine being on an airplane
with no Wi-Fi, and
you can ask your phone
anything, and it will be able to
kind of answer it. Of course,
there's also the novel
architectures I was speaking about earlier.
Those are a bit
farther off, but you never know with open source,
right? Like, everyone's
innovating at such a high rate,
and I think we'll get there much sooner
than we believe.
That's incredible.
Let's go to the Bryans.
Bryan squared, so Bryan P., go for it.
I know Bryan R. has his hand as well,
so Bryan P.,
I feel like this went down a whole rabbit trail of the the technicalities of
gpus and the uh llms and i'm going to kind of pull it back a little bit
I think it's a good thing to do both right?
So I mean people come I love this specifically to try to get a little more technical
But also kind of do the eli 5 thing. So the balance is nice. But yes, please please dive in brian
So the I wanted to make one quick point going back to the threads versus twitter
I made a list just from looking at at them for like a ux experience of what?
threads needs to actually compete with Twitter.
It's better integration with Instagram,
DMs, third-party app integration,
statistics and data, keyword search,
topics, spaces, and Google integration.
If they can get all that done, and I'm sure they're working on different variations of that, I really think that they can compete.
And then I raised my hand because we were talking about what data they might want.
And I wanted to give everybody kind of... I'm 42.
I'll give you a perspective of what Facebook was like when it launched when I was in college.
I'm right at the generation... I'm at the cusp of Gen X and Millennials.
And...
At my age range, it's basically cut off of where when social media first came out, we were already kind of like teenagers or in our 20s and things like that, even older.
And so it was a different experience to the people that are younger.
So Facebook, at the very beginning, I can tell you one thing is that I have a lot of friends who still don't have Facebook.
My brother doesn't even have it.
And I remember a time in the early or late 2000s where they basically asked everybody, if you uploaded a group photo of somebody, you would usually tag people's faces in there.
And they started asking, who are these people you didn't tag?
And they started gathering information and asking us to like kind of narc on our friends and family who did not have Facebook so they could also start to track them through our Facebook accounts.
And this was like crazy.
20 år senere, nesten.
Og så når du spør meg om hvilken slags data de får,
Facebook har for en lang tid relatert på dette for å anvende aftaler.
Det er deres største kjøpende,
å anvende aftaler og å sälja dette data til andre mennesker,
og leverere det så de kan gi deg det riktige på riktig tid.
Og forskjellen mellom Facebook og
Twitter har alltid vært som Facebook er hvem du er for din familj og dine vänner,
og Twitter er det du virkelig tror,
og det er der du virkelig begynner å prøve å, det er der revolutioner, Occupy, BLM,
mye av de andre ting, Arab Spring, de var alle støttet av Twitter,
det var der folk ikke virkelig talte på Facebook, de var det i en slags måte,
men på andre måter, og så når du går nå in threads,
is now targeting two different things.
There's the younger generation of Gen Z,
who are the kids of my generation, essentially,
and they don't really say what they think
in front of their family, their elders,
their grandparents, and their parents who are on Facebook.
So they go to other places to actually say what they really think.
And so Facebook is trying to recapture that same thing that got it there in the first place with this young generation of Gen Z,
and really digging into what it is that they really think, because that right now is the most valuable data.
That's really what I see them as doing.
creating threads, like the business purpose
behind it is to capture that
other bit of connection
and those other conversations
and things that Facebook won't
really show because people just aren't really
who they are in front of their family,
basically. That's a fascinating historical
perspective. Metaprime, you're going to be the last comment
because we've got to wrap. Do take us
home. It's been a lively discussion, but yeah,
what do you have to end the space?
Nei, jeg elsker bare hva Brian sa om den ideen av
at det beste data kommer fra trådløs uttrykk.
And I think that that is a very key element when we judge,
because I think one of the greatest,
the great things about this space, right,
is not that we have the answer of if Threads is going to make it
or beat Twitter or whatever.
It's highlighting the aspects that we need to pay attention to,
to see which one is going to be successful.
And I think, you know,
Brian and other Brian have talked a lot about the importance of the data mining
and how that's really kind of a big,
motiverating factor behind getting into this space, right?
But I also think that the, that idea of quality of data is really, really key.
And again, I would just highlight the idea that it's these big key accounts
that have access to huge networks and kind of the cutting edge
and articulating their ideas well.
That's another thing is you're going to have a lot of people that are talking about
subjects, but if they're not articulating those ideas, if they're not supporting them, if they're not drawing in from other facts, it makes it a little bit harder, right?
They're not as robust arguments. And I think that quality of data is a really key thing.
I mean, Xavier even pointed out in a comment to the space thread that, you know, Twitter or data is gold, right?
And I almost think of it as like, well,
You can almost think of
the Twitter network, or
social media networks, as sheep.
Right? And not in a
demeaning sense, but in the sense that
they can be fleeced with no harm
to the sheep for their wool.
And so you want to get these really high
quality sheep, right? The finest
wool, the best wool, but in order to
to build that flock, you really have to have a reason
to bring those people in.
And so I think that, in my mind, is
the question that Threads will have
to answer. What can you offer
that's going to be better than Twitter?
And I don't think Elon's going to roll over
and just let Threads be better. I think he's
going to be constantly innovating, and
I think that competition will be healthy,
as Brian Romello was pointing out. It's good competition,
but I think that really is
the big factor, is that
quality of platform,
The meta-metaphors and the quality,
the meta-metaphors are just rampant today,
and I love it.
It's been a great discussion.
We do have to wrap, though,
so we do these every Tuesday and Thursday,
12.30 p.m. Eastern, 9.30 a.m. Pacific,
so please do join us
and get more of this raw authenticity
and the great data that we all have here.
So thanks, everybody.
Have a good day.
