Projects
![]()
No projects found
Investors
New
DAO Treasuries
DATs
Lend / Borrow APRs
Chains
Categories
Spaces
AI War Machines
Recorded: March 6, 2026 Duration: 1:02:03Space Recording
Full Transcription
we don't have an intro today
there's no intro rule
i was waiting for it to kick in and i was going to just kind of vibe with the with the intro music
for a little bit um but we're right into it we'll let it fill out a little bit i guess
um yeah i'll just get it shared on timeline
then we're good to go okay cool cool yeah i feel like uh i don't know what they've got going on
at twitter and x but it feels like uh either using all of their compute to train the next
iteration of grok or and they have like only a certain amount of bandwidth that they can actually
put towards things because it feels like they move one thing and break another
and then Nikit is off doing some like consumer app wizardry
on top of it with all the shit that he's doing.
But yeah, it feels like you push Twitter too much at the minute
and it just completely breaks.
But right, we're on the timeline.
That's good.
We're on the timeline.
That's a start.
Good. So we should be good to get going any timeline. That's good. We're on the timeline. That's a start. Good.
So we should be good to get going any second.
Okay, cool.
Yeah, we'll let the room fill in a little bit more.
But on X, what do you think is the best growth platform now?
Because X growth has been really hard.
And over the past, I guess, four years, it's obviously been a lot
easier. And you've got people who are tweeting about crypto in 2018. And as long as you tweeted
about crypto and you wrote something decent, you amassed like crazy follower following, you know,
like 20,000 people who have like hundreds of thousands of followers. So what do you think is the next meta with regards to the growth platform?
Like LinkedIn?
Because you've been going hard on LinkedIn.
It's not actually me.
It is me, but I'm like not putting any effort into it whatsoever. I'm trying to just take what I've done
and just get it repurposed through a little open core agent.
If it's doing well on Twitter
and then like running it through open core agents
to tighten it up a little bit more,
suiting the boot and then putting it over there.
Because I recognize that that is a platform as well,
but I just hate being on there.
I just like posting ghosts every single single time so i suppose it depends um
i have i have a hot take on a new theory that
because um i feel like the new content form factor
will be just creating stuff not articles not video that will be complementary to like product
um and i've been shouting about this for a long time that every company is a media company
i think now because like the average idiot like myself definitely speaking about myself included
here can literally just speak or type into something like cloud or codex or even any of like the one-shot type apps
the new form factor of content is applications and ideas getting brought to life so like new
ideas new businesses new websites um i feel like that as a potential angle for like a really
interesting social media platform as well um that's an interesting one. I like these, um, like these micro shorts where you're getting
like a really extensive plot, but it's broken up into like 30 second shorts. So this is like a new
form factor that I'm seeing come on a TikTok and Instagram and YouTube shorts where you'll have an
extended story and an overarching overarching
like narrative and and like actual underlying story to be told but instead because it's optimizing
for people's like reduced attention span so it's you'll get different acts and it might be 30
seconds long and that's it and then you'll wait for the like the next act to drop so i know a few
movie studios that are like really really leaning into this quite heavily now so those two are really interesting i'd i if i was like didn't have a job and responsibilities and
a child i'd probably go and experiment with those two um and i got back from an event in lisbon last
night that a minimax event some of the shit that people are doing with image gen to image animation to video stuff on the ai ad front is absolutely insane really like
so so so good like really really really impressive really um i was just blown away by it to be honest
so like all those are like new frontiers right i don't know if any one existing platform captures
that and like pushes into lean into it or a new platform spins up um but those are three like i suppose outside the box ideas that i'm really liking the idea of
interesting yeah i agree with that um all the experience that i've had with video models has
been really bad um but i'm sure that i'm not you know down the rabbit hole enough and experimenting with enough of the tools.
But I will say Google Stitch is really cool.
I used that for the first time yesterday to design a landing page,
and it iterated really well, and I was very impressed.
And then combining that...
You can do that certain components?
Yeah, yeah.
And you can also copy and paste it to figma and then connect it via the
figma mcp um so that was like the coolest design tool that i've used thus far um but design is hard
i should have like insider traded because one of my good friends at figma is
i i knew about this mP integration with CloudCodes
like many, many weeks ago.
I knew when it was coming out, but like, yeah.
The guilt in me just wouldn't allow me.
You're a moral and ethical man.
But it looks great.
Like all, like the MCP-ification
and connectorification of the world, it's it's moving super fast and
i think i think everything bends to that will and like even um even perplexity call them connectors
as well which is quite interesting that they've all just converged to to bend to the anthropic
take you know right right yeah there's still like a debate about MCP versus CLI. And we had something in the show about that regarding the Google CLI, which is probably one of the more powerful tools that any of us here could use and implement today that'll help them with their work.
them with their work. But before we get into all of that, I'd like to welcome everyone here today.
Thank you again for joining for this week's show. We're here, episode two of the Singularity Show
with me, the co-host and my friend, founder of Blockmates, temporary co-host, also guest,
um temporary co-host also guest honorary member invited anytime of course grant how are you today
grant how's the week been good uh yeah good um like equal parts overwhelming and equal parts
kind of like exciting with everything that's going on um way too much to kind of keep on top of with my small little monkey brain so i really
like these as even i'm going to listen to them or join them it's it's good to kind of like
consolidate all that knowledge and it helps like proofread that newsletter every week as well just
so i can kind of keep up with what the hell's going on right right you got some events coming
up too don't you you said you went to a the lisbon event um
the other day yeah that was good for for uh creative and advertising and then you're hosting
your own events events coming up here soon correct yeah so we've just confirmed it probably like half
an hour ago so 30 minutes ago so we're gonna be doing i'll just bring this up quickly. So it's going to be, if anyone's around in Lisbon.
Oh, we lost Grant.
He probably clicked out of the tab.
But yeah, if anybody isn't around in Lisbon,
they're hosting some events over these next few weeks
that are focused on a cloud cloud bot
ai tooling and also just a chance to interact with the community and meet some fellow people
in the european market who are building solutions so it'll be pretty fun i wish i could jet over
i've given them this feel but if you wanted to come back in and and tell them what's actually going on yeah no no it's it's gonna be in lisbon at the nest by
nia and we're gonna be probably having four 15 minute speaking slots uh two of them are already
taken uh if there's anyone in around the area that wants to take up any other sponsorship slots
that's completely fine as well there's one or two of those left just trying to keep it more agent focused open core focus we've got someone
who's built a hybrid between open claw and nano claw like removing a lot of the constraints that
nano claw has but removing a lot of the time that you do typically spend debugging open claw so like
actual people who are putting this to good use know, not just like a lot of this type stuff. So founders,
funders,
builders,
just general enthusiasts.
I'll post it out on Twitter after this.
It'll be in the newsletter as well.
So if you need to need any more information about any of that,
you know where I am on telegram or Twitter as well.
So yeah,
be good.
Sweet.
I wish I could come.
Did a little long flight for me.
You're going to have to start Euromaxing.
I did last year.
I did.
Last year for me was crazy.
I was in Europe like three times.
I'm tired of making that journey.
It's like seven hours there,
and then it's another like nine back
because of the jet stream.
So really difficult.
But yes, without further ado,
let's get into the show for the day.
We named this show AI or machines. And with feeling at risk of LARPing,
we wanted to dive into the Department of War and their relationship with Anthropic.
They mentioned that the department of war after they
had trump say oh we're not using anthropic anymore trump came out and it got leaked that they were
using anthropic up until the final hour um of their strikes leading up to uh the strikes in iran
and it just goes to show that anthropic really does have a foothold over operations
like this. And I wanted to kind of ask your opinion and get your take on AI's application
in war. Because we've seen over every other period in history, it's sacrificed a lot of
people. Of course, World War I was probably the bloodiest war in history and it was war of attrition i think i saw a stat it was like europe lost 10 to 12
percent of its population during world war one which is absolutely insane um but we
fast forward to modern modern day we've got all these drones, autonomous systems, intelligence systems, and
humans are now capable of targeting sites, facilities way more accurately with way more
precision, eliminating much less collateral damage across the targets and zones and, and, and hopefully way less civilian
lives are lost during these operations. So what's your general take on AI's application to war?
Do you believe that it will thrust us into more conflicts because it's easier and cheaper to,
to deploy into them? Or do you think that it will generally create a more peaceful environment
where everybody has tools that are very highly capable of extreme damage
so that people are not engaging as frequently?
Yeah, I'm thinking about this in short, medium, and long term.
So short term, those who have will use it um so anything that's like extremely frontier model
and is extremely capable that you can guarantee that's being used right let's not let's not be
naive about this so um i think we've it's it's been quite openly said that it's claude's being
used and there's been a few other models that have been used to kind of allow the extraction of Maduro and like more what's happening in the Middle East at the minute.
And the one thing that I think I want to double click on what you kind of touched upon there is
this, in my opinion, is the modern day nuclear arms race, but it's a digital nuclear arms race.
So it's like it gets to a point when these models are so
impressive um it gets it really gets to that point of mutual issue destruction if people start to use
them in like various words so there is a bit of an arms race that's going on with the models
maybe what we spoke about last week with distillation across some of the more us-focused
uh models or china are kind of like churning out.
But DeepSeek V4 about to drop this week.
And I know there's been a lot of talk
from the Anthropic camp
that a lot of the labs out in China
are not actually doing anything off of their own back
and just distilling a lot of models
like Claude and Chachi B.T. and stuff like that.
So maybe that's not as much as a two-horse race
as we first anticipated.
You're starting to see some, I suppose,
tier two countries come into the race.
Now the UK have just invested in the first two Frontier Labs.
I know Mistral have been doing pretty well,
but other than that,
it's kind of like a barren landscape anywhere else.
But I do actually see this as like
the modern day nuclear arms race.
And it will probably get to the point where it is a mutually assured destruction situation if we
just could do continue. But now it feels like, yes, there are less. And I also like, I'm also
quite, I'm not fucking stupid. Like I also understand that we only get told what wants to
be reported on
back home so if people are saying that there's less casualties and stuff it's like all right
that's that's what's being said i'm not i'm not there i don't have any first-hand source so
that's the case i you can categorically say that this has been like one of the first wars where
you've got autonomous drones that are just literally kamikaze into buildings and shit,
which is fucking weird.
And if you think back to two of those red lines,
what I think Seth and Dario were going back and forth on,
by all accounts, again, like, I'm not flying the wall in these conversations,
but I'm just kind of paraphrasing what's being said.
It was no domestic civilian surveillance, and it was no autonomous drone weaponry, right?
Because they weren't convinced that any of these models
are in a situation where the hallucination factor
is not reduced enough to make ridiculous mistakes.
We were talking before we came on air,
particularly this week,
the world's been pretty shit, to be honest.
And if that's
what you're trusting with autonomous weapons uh in an age of drone warfare then i think we're
moving into a sticky situation but like all that to say it's happening regardless whether we like
it or not so um not not excited to see how it plays out but but it's going to happen. So I've just got to kind of like see what comes down the pipe.
Right.
Yeah.
And it is interesting because when I guess Dario came out and dropped that essay and
kind of put his foot down and denied working with the Department of War,
Anthropic downloads just mooned.
I mean, I had my mom texting me.
I'm downloading the Claude model, you know, because now it's all over the news.
So it was more or less like an expert 4D chess move because he dropped an article yesterday,
again, another essay that was saying, that was kind um re-outlining again he was reiterating his
point we're going to stand firm with what we believe but we're also now working hand in hand
with our own divorce so he capitulated a little bit um and walked back on his original stance and
some of the things that he said and it was just a free pr situation for clot and they overtook chat GPT on the app store.
So maybe that was a PR stunt.
Maybe they're working with the department.
Very ballsy,
but it got a truth.
It got it.
It received a Trump truth social post,
which those received millions of,
of engagement and views.
So it worked.
And now they're at like $20 billion runway rate.
And it's unreal.
And they're still shipping every week.
What's weird about this?
And I hate how everything gets...
Ends up getting looked at through a political or a like tribalistic lens what you'll probably see
going into the election and this is like the biggest cringe worthy story of all time
models will probably start getting connotated with political parties right you'll get you'll get
left lean you'll get left-leaning associated models and you'll get you'll get more left lean you'll get left leaning associated
models and you'll get right leaning and you will only use the ones that you associate
with your camp your tribe because you see it with social media yep like loose guy what's going on
over there i have no idea reddit i can't go over there but it's gonna happen and it's weird but we just always converge the
same thing yeah i agree with that 100 um and it will be interesting to see how that dynamic how
social dynamic plays out also because socially you have a lot of people um on the left who are
very anti-ai who don't want anything to do with it at all. Right. You've got people protesting data center build out.
Can I tell you a quick story about what happened last night?
So, so we're at the mini, mini max event. And again,
just like, as we were saying earlier, a lot of really impressive AI
advertiser, like AI short stories, like really, really,
really impressive stuff like I
want to shout out like this solo AI films on Instagram the guy there quite uh
quite quite vulgar videos to be honest I'd go and I recommend anyone goes and goes and searches for
it but like this guy's like talking through and he's like right which one do people want to see
I've got four trailers here for these short movies and he's like i created these with social engineering
in mind to get as get them as viral as physically possible and then then downstream of that i've got
an ai ad agency so like for the likes of me and you that's just seems like yeah well that's obviously
the player this guy is sat in the fucking crowd and he's he gets to the end of like this video
and it was like one of the most impressive ai generated videos i've ever seen and he goes
i've got a question why is any of this good it was just like why the why would you take on your
thursday night it's 8 30 on the evening why would you go out of your way to come into like this event where
everyone just like wants to learn about what's happening and like really really at the forefront
of technology one of the most interesting times to be alive ever sit there in the audience for
an hour and a half and then wait around for an opportunity to say why is any of this good
it's just what are people doing
i don't know like i mean i ask myself that question every time i see like a group of
protesters and yeah at risk of feeling too political i'm just like it's tuesday at 10
o'clock in the morning don't you have a job that's how i feel all the time i mean it's ridiculous
it's like every single thing is a is is warranting a protest for some of these people.
So I don't know. I think the more you have these people reject, reject this change.
I mean, you saw that like I don't know what it's what it was technically called, but it's like the AI license that's coming out of New York now where it's like, oh, we're trying to prevent AI from being used in places of law, healthcare, dentistry, etc.
It's weird.
It's like, why are you denying?
The very parties that are trying to drive down costs are banning technology that is deflationary that drives down costs.
Right?
Yeah.
They tried to do it with Bitcoin if you remember.
Say that again?
They tried to do it in New York with Bitcoin.
Yeah, like the BitLicenser or whatever it was.
So I don't know.
You've got these weird jurisdictions that might take hard stances on AI and
you're going to have places like Texas and Florida that are going to say,
hey, let's just pump it and see where it goes.
And I think we're going to see an interesting bifurcation of society
where there's enclaves of people who are against AI
and who are trying to prevent it from permeating their jobs and their life.
And you can't really blame them because that's how they make their livelihood.
Right.
But then on the other side, on the other type of
or the other end of the spectrum, you've got people who are balls to the wall on
this stuff, like us.
And I guess we're going to see what
technophthalmists probably Right. Probably equally as stupid.
Right.
Maybe.
We are stupid.
Yeah, like delusionally bullish on it.
I don't know.
Yeah, I don't know.
I guess that's a good segue into this next slide that we got here.
Anthropic released this graph that shows the application of theoretical AI coverage to the observed AI coverage, industry by industry.
And of course, this is probably not new information, but it is interesting.
Again, I want to harp on the observed AI coverage is still not covering a large majority of this map.
So there is a lot of opportunity here.
All of this blue space is opportunity for builders.
And I believe that even in some of these spots like construction, like food and serving, grounds maintenance, there's still probably more opportunity to implement AI there than what's measured and observed on this theoretical
graph, so to speak. There was also another graph that was published by Anthropic, I believe,
a couple of weeks ago where it showed AI's capabilities in this current state. And it was like coding, accounting, finance.
And then there was like,
there was an outsized growth with those specific functions and then legal
healthcare, et cetera, was, you know,
1% usage or 3% usage.
And I believe that that's where
you probably need to focus on on on building is is within these areas that
have obviously not experienced a ton of applied ai so within this graph where do you think the
biggest opportunity is grant i i think you gotta you gotta look at
what the big frontier labs are pushing out and what the big vcs are get are currently funding so
um it's the delta between this exactly here the theoretical and observed it's like it's that gap
there and all the drumbeat that's coming out of the anthropic camp and what dario keeps saying
and everyone else keeps saying it is for the most part coding is basically solved for the most part
right there's obviously some edge cases and fringe cases and you know it will continue to
get better over time but like the point at which we're looking at other areas now
and then you can start seeing rolling out products like Cowork
where that's a little bit more tangible
than the standard for the average person.
So like literally connecting into your local folders,
organizing the folders,
like got the taxes done for the full company
in three or four hours,
as opposed to something that would usually take two weeks. Like that's tangible stuff for people.
What is that? That's like HR, it's admin, it's accounting, it's finance. It's like,
it's all those kind of management, your back office type roles that are,
it's weird because everyone said it i was coming from like was it
was coming for the blue collar workers and it kind of just replaced the engineers first and
is now going to seep into the back office um there's so much of like the legal funding aspect
getting funded at the minute it's like every other day i'm seeing like a huge announcement and they're
in the tens of millions if not 100 millions all around like just coming out of the gates so that's an area ripe for destruction like i don't know
how they get around the compliance issue with certain jurisdictions and like
multiple locations and building all that in but whoever cracks that
the question i have though is and i don't know what you think about this
is this not just a feature of one of the frontier labs
as opposed to independent companies?
Or do these independent companies then just get to product market fit,
get swallowed up, bundled in and rolled out?
Because Anthropic and OpenAI and Google,
they're just turning on features that are making businesses redundant overnight.
It's not just like the actual people
that are going to get made redundant
from the use case of these specific tools.
It's businesses that are getting funded and funded and funded.
I don't know how many times I've seen on the timeline
Anthropoc just killed my business overnight with a feature.
So I really don't know how this plays out
and can that be allowed to just happen?
I don't know.
I don't know how this plays out. And can that be allowed to just happen? I don't know. I don't know either.
But what I do know,
and I think is actionable advice for people listening,
is if you are in a business function,
that's a cost center,
you need to figure out how to transition yourself
to a growth role.
Because growth is very easy to attribute revenue and value back to the business.
Whereas if you're a cost center, it's going to be very easy to make the case to cut you and replace
you with an AI system, like we just mentioned, like back office accounting.
So there's a lot of cool tools and cool ways to,
to growth hack nowadays as well.
We kind of covered those last week.
And I think that's the only way to insulate yourself and hedge yourself from a lot of these features that
Anthropic and chat are shipping. But I don't know how it's going to change the startup scene as well,
right? Because you'd think that a lot of these workers would get laid off and their first
instinct would be to go start a company, right? So they do. And every week that they work on
their company, they're just fighting an uphill battle like Sisyphus, where the boulder is as, uh, anthropic, I guess.
And the, the, the weight of anthropic and their, their speed to, to shipping
and execution is, is really hard to push uphill and fight against.
Um, so I don't really know because a lot of the, like the, the internet
native stuff just feels like it's going to get completely automated away.
And that's obviously not a hot take.
Everything computer-based can be fully optimized um i suppose that was like the whole halo trade idea as well
like um coca-cola started ripping like any ultra-motored businesses that are inherently
built on product and brand that are effectively impossible to just automate away.
That started ripping as like a pair trade against like the SaaS sell-off as well.
So like that was, that was coupled with that.
And you can start to see a lot more people speak about that.
But yeah, I agree with what you're saying.
And it's, it's, it's painful to see and say, but people who aren't adopting a lot of this technology now are going to probably be dragged
into an office unwillingly and just be like look we're going to be able to just do your job with
you know a few clicks of some buttons someone's going to come in we're going to hire a freelance
automation agency they're going to come in and automate 50 of the company away overnight and
you've seen it with um jack dorsey's 40 layoff anecdotally i will say i speak to a lot of teams
behind the scenes day in day out across this industry and adjacent industries and
there's a lot of head heads being cut for better or worse. And like the idea is like, well,
anyone who's building towards rightly showing willingness
to adopt these tools and make themselves
more of a productive asset
is net contributing towards the business.
Either you steal margin back by cutting headcount
or you steal margin back by increasing productivity and profit, or you get an exponential factor of reducing headcount or you're still margin back by increasing productivity and profit or you get an
exponential factor of reducing headcount and getting more output from the people who are
willing to adopt it and that's like a compounding effect so that's just what i'm hearing and what i'm
like speaking to certain founders and building teams about as well so anecdotally what you're
seeing on the timeline and what you kind of see coming over the hill all seems to converge into the same thing unfortunately or fortunately right i think that was a much more eloquent way
of saying what i was trying to convey is yeah if you're not growing then you're probably going
um and i hate to say it like that but um it's the truth it's like the most brutal hr bot of all time it is it is like um you get a slack
bot message and it's like we've seen that you've not been growing so therefore you are going
here's your severance package um and then you just get kicked out of the app and you're like
holy shit all of that was an ai automation as well the whole thing was you just got automated out by automation doesn't have
any feelings um but moving on there's guys and to make you feel worse about yourself there's people
who are 24 years old younger than you and also richer than you um we've got this has been a big
story this week lepold ashen brenner from assuming that's a very german name assuming he's german
yeah don't ask his grandfather what he was doing in between 1939 yeah yeah he's uh he's from
argentina now um but ex-open ai researcher from um from germany it's fun situational awareness
has been absolutely printing stacks over these past 18
months. Started with $200 million initial GP seeding from Nat Freeman, Daniel Gross,
the Nicolason brothers, and has turned it into over $5.5 billion in value across 30 holdings.
He hit CoreWeave, he's been trading Intel, He's been trading a lot of these hyperscalers, these neocloud hyperscalers.
He's got a huge position in Bloom Energy.
And his thesis now is that the energy bottleneck is really what's holding AI growth back.
with back and he's investing directly in that sector. And he's also married or dating the
And he's investing directly in that sector.
chief of staff of Anthropic, which is a little interesting because the other week, I think it
grabbed headlines or the other day it grabbed headlines because he had bought, what was it?
Call Options in, what company was it that partnered with in with that partnered
with uh perplexity and it was up seven percent all right okay right yeah it was like the day
yeah i think it was intel he's like he bought like it like his the 13f dropped and it showed that he
had intel call options and then the next day, Intel or the company partnered with Perplexity and the stock was up 7%.
Something like that.
Or it was Anthropic.
Sorry, maybe not Perplexity.
It was Anthropic that partnered with Intel, obviously.
Yeah.
And he's dating the chief of staff of Anthropic.
So little sus.
I think I had a tweet about it. It was like, Hey, don't be, don't be, um, don't get caught off guard if the sec knocks on your door
and serves you a well's notice and says, Hey buddy, where'd you get that alpha from? Um,
but yeah, it's pretty crazy. Right. Um, he dropped a manifesto. It was like 150 page manifesto after
he, uh, got fired from openAI for, I don't know if he
was necessarily whistleblowing, but he was calling out some bad business practices from the lab.
Dropped a 150 page manifesto, got his seed funded by some very high profile people and then ran it
up to five and a half billion, which is absolutely incredibly degen.
First of all,
what's the sharp ratio on that fund?
What the fuck was I doing when I was 24?
Jesus Christ.
In the pub like me,
but,
but yeah,
it's,
it's,
it's a,
it's a really interesting strategy.
It's been the strategy that Cetrini has also played out as well
and has been calling for the past three years,
just investing in the bottlenecks upstream of AI
and huge infrastructure build-outs.
And it seems to be working.
He's a very concentrated bet maker as well I bet you see a load of
social trading
vaults
and products built on top of this type of thing
I wouldn't be surprised to see
you know they've got like
Pelosi stocks
I was just going to say
I feel like that's a natural progression for
the Robin Hoods of the world and things like that bund I feel like that's a natural progression for the Robin Hoods of the
world and things like that, like bundling in together any of these assets that people like
this are. Just riding that social wave, you know, I think as everything converges between social and
trade, those lines continue to get blurred and more accessible than ever for the average person
on the street to get access to these ideas and assets i feel like that's probably
and and it's super reflexive as well so yeah there's like it's very kensian when someone
likes to buy something it's say if it was to pump off off of that announcement or that filing then
it's gonna it's gonna have massive reflexivity when he's when he buys and massive reflexivity
when he sells as well because people are just following people in like that.
Yeah.
There's the SEC filing tracker on Twitter that is basically just ingesting all these filings and then posting them and saying,
Oh, Peter Thiel sold however many hundreds of millions of Palantir stock, right?
I think we're going to start seeing that obviously with
13Fs and people who really capture the headlines like this guy. Every time situational awareness
files their 13F, we have people tweeting about it. Oh, look what he's holding, look what it is.
So it's becoming much more social and ubiquitous like this information is and in retail is getting um much more um involved in in understanding kind
of like who is shifting the markets and who's who's calling it right and and who to actually
to to trade entail on some of these things um and you make a great point like pelosi tracker
um you could do that with all different types of public filings.
I mean, it's all public.
You could probably make some really props around it.
That's a very basic and probably a weekend project of spinning out like a really nice front end
tracking a lot of these profiles,
making a marketplace around it
and then getting like a fucking nice hefty ref link
from Trading 212 or any of these destinations
where you could actually go buy
and bundle these assets together.
That feels like something that could inherently go viral.
And if you get one whale,
that's just going to follow all these trades in.
Simple business.
I'm way too busy.
I'm way too busy.
How about it?
How about it?
I'm taking mental notes.
I've got so many ideas too, man.
Holy cow.
I just feel like I could never,
my idea backlog is always just piling, piling, piling. I'm trying to figure out right now how to automate through the backlog and create a system that just ships my backlog all the time.
And so I can see it, play with it, check it out, get some feedback. If I think it will get any traction, maybe take it a step further.
If not, cut it.
And there's a lot of apps.
Shipped or Shipper is kind of an app that is following along those lines.
Lovable is the same way.
Really like end-to-end building an app, which is pretty cool.
And I guess it's a good segue into the labs. They keep
shipping. GPT 5.4 dropped this week. I haven't tried it yet. Ever since I got my claude subscription i've not gone back to chat gpt it just feels like i'm locked in
into into claude right now like it's got i've got all my projects i've got all my mcp servers and
connectors set up um but during days like yesterday when opus does kind of go off the rails and i
tell it and point it to the right direction that I feel like is the right direction that I that I very explicitly tell it and it goes off the rails
maybe it gives me a little bit of a sense to and sense of urgency to to go and try out new stuff
again it shows the dependency risk as well doesn't it like right right it really does stuff was just
like I need to go and fuck around with this and, like, untangle it because Claude's just not working.
To be fair, I'm running primary model as, like, Minimax,
which I've been found as brilliant,
but fallback and more extensive stuff was on Claude.
As you say, over the past week, it's not been unusable.
That's, like, that's hyperbolic,
but when there are other options available,
like, that are on paper and benchmarking as good,
if not better, for, like, the site-type things
that I want to use it for, then, you know know i need to be a little bit more sensible have you used any of
these alibaba models yet i haven't actually used any of them but i seen like you could potentially
use some of the really smaller like some of the lightweight parameter ones on mac mini um which was very very tempting
and i know there's um a lot of people that i'm gonna look at setting that actually up so you
could basically just run download that and run it locally if you've got like a beefy enough machine
to actually get that out the gates and some of the benchmarking on those as well like obviously take
a benchmark with a pinch of salt because it's it's kind of like retrofitting
around what the what the benchmarking tasks are but we're going to get to a point very very soon
you can see apple moving towards it i don't know if we've got anything on that but like you can
see apple moving towards it they they're optimizing with models to be built on silicon and being able
to run on silicon and everyone's going to be able to have it on their phone or their
or their macbook air or the new macbook model that they've got coming out soon so i'm going to be able to have it on their phone or their MacBook Air or their new MacBook model that they've got coming out soon.
So I'm going to get to a very, very...
I prefer that future as well, to be honest,
like being able to just run that locally.
I think we're moving into a very, very interesting time with that.
Right.
There's a lot of interesting hardware plays in the market as well
that you can position yourself to trade, I guess, the
alternative hardware market.
You know, NVIDIA has been the primary beneficiary of the past four years of AI infrastructure
growth and build out and inferencing and demand for all these workloads.
But a lot of other players are starting to dabble into alternatives.
And that's an interesting kind of domain to,
to look into if you,
if you believe that inference workloads go on very more application
specific chips.
But yeah,
there's a lot of cool stuff going on on edge computing.
I'm very bullish.
Raspberry Pi, the stock like the other day. Yeah. there's a lot of cool stuff going on on edge computing i'm very bullish uh raspberry pi the
stock like the other day yeah it's i think the ticker is rpi um that moon there's a couple other
edge computing devices i don't know there's a lot of cool hardware that i think is is is starting to
um be ideated and we're seeing cool new trailers for them, commercials for them,
like the pens, like the, forgot who's that one person that was on that, was in a cafe
in San Francisco and they had the earplugs in. I forgot who exactly it was, but AI hardware devices. I am very curious to, I guess, see how successful those are, because I think hardware is very hard.
I mean, you can produce hardware.
They produce drones like crazy.
You can, like, there's facilities that produce a lot of drones.
facilities that produce a lot of drones. So it's like not actually the production of hardware,
So it's, like, not actually the production of hardware.
that's difficult. But I think the stickiness of hardware is difficult to attract. Because
again, what if we run into an area where a lot of these hardware devices are just your phone?
It's like, why do we need an individual puck to do something we've got a camera and a microphone on my phone right why can't you just turn this uh this pen
that you use to write and record your notes and i don't know it's like we have yeah why is that the
phone why is the phone the phone yeah right or i mean one thing i want to shout out is this Devalence hardware device
that kind of took the internet by storm earlier this week.
We spoke about this last week.
We spoke about the always-on devices and safe spaces for those around the city and shit.
Yeah, and I guess our prayers were answered within a week
about the device that mitigates from all of that. But again, how does this market not just
get completely cannibalized by the hardware that you already use to access the internet,
to pay your friends, to brain rot, to see where you're going, to listen to music? I mean,
it's all right there on the iPhone. I mean, the phone feels like the ultimate form factor.
And I appreciate people that are trying to innovate within the hardware space.
But I just don't really know if I see the future.
Like Friend, like that device.
That video that they released.
Yeah, right.
I just don't know how these hardware devices plan to shape the social interaction and social relationship with technology.
I don't know if that makes sense, but we have to change social behavior to accept a lot of these hardware devices.
social behavior to accept a lot of these hardware devices and i think changing the social relationship
between human and hardware device is much harder than human to an application on the internet
because they're already interacting with that medium but the physical medium is a much different
medium yeah it's like um like let's let's be honest these things are an extension of you aren't they
like yeah that's more like i don't why we still call them phones is beyond me
right anything but that at this point um getting like you will always have early adopters you
always have people that bought the metaglasses you'll always have people who bought like
the apple vision you'll always have people like become early adopters and then there becomes
like inherent new use cases that give you an edge in personal life or business life and then it
becomes well it becomes a power law where the haves and have nots who have the technology
can start to progress a lot quicker in and having that benefit in their life or their business life
than the people who don't adopt it right and then that starts to move into like the adoption curve so i don't know
how it plays out because on one hand it's like you're not going to attract the same crowd where
it's like everyone who's moving to you know like these really super simple phones where it's just
like literally call and text and there's not that whole movement of getting off social media and blah blah blah blah that crowd you can't in one hand go down
that route and attract that same crowd of early adopters if you're trying to tout a simple phone
only does like text and email and phone whilst also them being you know warm to the idea of
having an always-on wearable or like smart glasses
and things unless they're just being hyper performative which a lot of people are in this
day and age where it's just like oh yeah i have i just have like fucking simple phone only it doesn't
even have like instagram or shit but then you're also wearing like always-on glasses and always-on
wearables um what i think you could see is, I think we've got to really closely watch what Johnny Ive and OpenAI do.
Because that carries so much weight from literally the chief designer of the iPhone bringing out the first next generation product.
like bringing out the first next generation product that in and of itself carries so much
weight that it could potentially break through like that early adopter phase and through into
like the off wall of mass market stuff i don't know what it's going to be like there's you can
kind of see them putting out different ideas of what it's going to be like no one has a fucking
clear what it's going to be let's be honest if it's going to be a wearable if it's going to be
a device i feel like a lot of the if it's going to be a device,
I feel like a lot of the frontier labs are trying to get people to wear
the glasses and the wearables that have camera
because they're effectively running out of spatial data as well.
And they need new world data and being able to like
effectively crowdsource that by having like always on glasses.
I feel like that is the Trojan horse of what people are actually doing.
Why Zuck pushed into it.
It's why Apple have tried to do it with a vision.
It's why Google glasses are coming out.
I feel like that's the angle.
So it's like, oh, we're giving you this.
But inherently, it's because they want the spatial data
to actually come in to train their models.
So it's a bit of a Trojan horse.
The wearable thing, I'll be honest,
I'll probably buy one.
I would probably wear it.
I'm inherently always using the voice dictation.
I wish the voice models were better,
like the actual response voice models
because they're just a bit shit at present.
But I would be the sucker who would wear it.
But I would also feel a bit weird
about wearing it around people as well because I wouldn't want to invade on their privacy. I think that's something that they're going about wearing it around people as well.
I wouldn't want to like invade on their privacy.
And I think that's something that they're going to have to contact with as well.
But yeah, I don't know if I actually made any good points there,
but that's my ramble on that.
No, I agree with you.
I think it is very high signal that Johnny Ive is still there.
Wasn't there a confrontation or conflict of interest or something that
was going on about some now isn't it yeah right but i guess if you have a wearable you can run
your 0.5 one person zero person company from from your wearable. Zero people companies are probably the most interesting topic that I've
explored over the past few years. Alexander Good had a piece about it back in, I think,
April of 2024, and I wrote an extension about it. But the zero person protocol, where the protocol is bootstrapped and seeded by investors,
and if it receives and raises enough money, then it meets the bonding curve and uses that
capital to go and pay for its own compute.
There's been a couple other projects that have alluded towards this type of bootstrapping model and goes in and completes tax tasks autonomously and tries to
reinvest and grow the businesses that it's managing and redistributing the dividends back to
token holders, right? That is kind of like the goal. And it really hadn't happened until
over these past few weeks where Nat Eliason built Felix Kraft, which is
his own autonomous agent that's running his businesses. I think he's selling skills,
selling information to other agents. It's selling his ebook. It's selling a whole bunch of different
things, trying to fit itself in within the agent to agent economy, so to speak. And I think that this
is a really interesting experiment, first of all. And second of all, I think that there's a lot of
opportunity to dive deeper into this niche of building your own agent that participates in
this agent to agent economy. That's all facilitated by X402 stablecoin payments. You pay for API access to the specific
data source because it's very high valuable for you. It's proprietary. You pay a dollar
per 100 requests and your agent can ping that database every time it needs to reference something.
I believe that this economy is just starting to grow and this is a very very early
case of of that happening and i'm really interested in seeing where it goes and also participating in
it myself and i've had a couple ideas of of how i could build an agent that
provides value to other agents in the marketplace yeah it was this was like this was kind of like a meme at one point yeah it's just like i imagine like the zero or the
one person billion dollar company and it's like slowly becoming a thing um and yes like these
early fringe cases they always start out like this it's like this. I could quite easily put a cynical hat on and say,
I could definitely run a script to make out that I'm purchasing
like $100,000 worth of e-book credits on Amazon
when it's the same $100 buy every single time
and it's just going on a loop.
Not to say, but you could game-fire these things, right?
Let's be honest.
But it's the premise and it's the idea
and it's no longer like pie in the sky.
It's like the future is brought to now at present
and it's someone is truly going to crack it.
Like I think you've kind of like,
obviously Peter Steinberg going to open AI
and there was people saying
that it wasn't like the largest acquisition ever,
but I guarantee he's got upside.
If they go and if they start floating,
he's going to be very, very, very rich
off the back of that.
He's going to be very rich.
So you could kind of say that,
like he has like a multitude of agents just running open
claw that have basically gotten to the point of where he's at now um but standalone solo
or multiple agent operated businesses that are bringing in real revenue and cash flow
we're on the fringes of that being like a thing So I'd love to see if they're going to get funded,
if they need to get funded.
I keep coming back to the same point,
but it's people who can market themselves
and market their agents and distribute.
Distribution, man.
Distribution.
Yeah, it costs nothing to create any of this stuff,
but it costs a lot of sweat equity and time and content production to build that distribution funnel.
And you're seeing a lot of people with a lot of big followings, kind of like I mentioned earlier, I think it was before this call, where if you were in the right growth channel at the right time, I mean, if you were a Substack writer in 2020 and you wrote consistently for three years, you bootstrapped a 100,000 subscriber audience, 7% conversion rate, you're charging $50 a month.
I mean, that's a crazy recurring on that content stream.
I think the meta is figuring out what the distribution channel of the future is going to be and building on that platform now so that you can reap the benefits down the track when it becomes evident that these zero person companies are the ones that are driving a lot of the activity in the economy yeah i think you'll also see
teams who are building out agentic frameworks uh teams that are building out and solely focused on
like orchestration and teams that are solely focused on um this whole realm of multiple
different agents effectively running separate departments with
inside the same company i think i know what the playbook's going to be there you're going to get
one of these one of these teams that is that's their sole focus let's say like an nnn or like a
lindy or like any of those teams that are building towards like the autonomous workflow situation, if they were smart,
they'd go and cut like a significant amount of their equity and go and give
it to like a mega influencer.
It's the whole Huberman AG1,
the whole KSI Logan Paul prime.
Logan Paul Prime.
It's the whole
It's the whole
B's box.
They're the people
with distribution.
And if you could get
people to buy into that,
there's people at the
bottom of the funnel
that are going to buy
their products,
whether it's B2P,
B2C.
It doesn't really
matter what it is.
It's the only
last bastion opinion.
Everyone says taste,
which is slightly
becoming a little
cringe now because
people who talk about
taste have literally
no taste.
Opinion and distribution is what you need to be optimizing for with these things. So I wouldn't be cringe now because people who talk about taste have literally no taste opinion opinion and
distribution is is what you need to be optimizing for with these things so i wouldn't be surprised
to see those types of partnerships drop and when we do start seeing them dropping maybe that's a
little bit top signalish so maybe we need to like take a step back at that point yeah we should have
just gotten really ripped and tan and started posting videos of us working out on the beach
and stuff and try to bootstrap our distribution that way because then we could have just sold
into into all of these um you know thirst trap audience types yeah when you say when you say
like floyd mayweather partnering with like replet that's what like we need to yeah real real i saw on that kind of note it's like a funny thing i saw on twitter was like
i'm i'm in ai hell or something and it was there in like the subway in new york and every ad is
just ai like automate your automate your accounting services automate your legal automate this this this and it was like what do these
words even mean anymore everything is just automate model blah no one knows that's the
problem like a lot of them have it they they try to differentiate by being too self-referential
um when everything that i've been converging towards over the past six months is recognizing
that the vast majority of mass market
is a lot dumber than you would initially think.
So optimizing for that mid-curve
is probably where they want to be actually targeting.
Like Anthropics, like clapback ads at ChatGPT,
like making people who use ChatGPT look stupid.
That's brilliant. Comedy sells with stuff like that so yeah a lot of them need to like have a little
look at their messaging in my opinion because it's just going over too many people's heads
and they're just going to spunk a load of marketing dollars on like useless campaigns
right right well they've got a lot of money to spend. These companies do for now.
Infinite, gosh.
For now. If it's not going into data centers in CapEx, it's going directly towards blasting your feed and blasting everything with advertising.
It's the arms race, man. Capt's capture as many users as you possibly can
and turn those over and convert them into paid
and start shipping more features.
The more features that you ship,
the more businesses you can cannibalize.
The dominoes kind of fall from there.
Not to leave all I'm doomed or no,
I did not want to do that.
I did not mean to do that,
but I'm optimistic. I need doomer. No, I did not want to do that. I did not mean to do that, but I'm optimistic. I need to
be, but that, that wraps up the show. We're kind of here at the tail end. Thank you all for tuning
in. Very fun conversation. I'm sure we'll come back next week with a lot more interesting
information as we always do. Hopefully no more war content,
hoping that one dies down soon.
It's not the funnest one to talk about,
but more tools coming next week.
We'll focus a little bit more on tools and stuff.
I know this one was still focused a little bit on the news,
which I'm sure you guys are,
are receiving within the blockmates newsletter.
Of course you're subscribed.
Why wouldn't you be?
And if you're not subscribed, you should go subscribe to it because it basically talks
about everything that we discussed here, but much more in depth with links. You can click
out and see exactly what we're talking about. And Grant, like you said, had a lot of events coming
up. If you're in Lisbon, if you're in Portugal,
there's a lot of cool opportunities to meet and connect with your community about all these things.
Even down here in Austin, South by Southwest,
huge cultural event conference.
There's so many different events going on throughout the week.
You can just go meet a bunch of cool people.
I really encourage people to go out and network
with their community, see what people are building in real life, go make those relationships.
We talked here about trying to build out a moat and carve out a differentiating factor for yourself.
And it's going to be those relationships at the end of the day. You can establish relationships,
going to be those relationships at the end of the day. You can establish relationships,
then you're going to be very hard to replace in whatever it is that you do.
And that requires going out in person. Don't be online all the time. There's value going out in
person, touch grass. Yeah. Are you going to go South by Southwest? I think so. Yeah, for sure.
There's a lot of events. There's a lot of cool frontier technology events.
There's like defense tech events. There's obviously there's a ton of music.
It's a, it's a much more of a cultural conference and gathering as opposed to
like a tech conference, but of course, technology enterprise stuff that brings
in the money. So that's what gets all the sponsorship.
That's what brings all the
cool people and so it's slowly started to seep into into the conference and influence the
strange that isn't it yeah i've been telling alex i've been telling alexis that like they're called
but they're new rock stars i'm telling you i know hopefully they're not the new uh uh
political techno technocracy what is it it? What's the word?
How do you even say that word?
There you go.
Well,
cool.
Thank you all for tuning in.
Grant,
you have anything else to say before we log off here?
uh, just keep an eye on the main block.
It's feed.
You'll see the newsletter drop.
if you're unsubscribed,
I'll leave a link below this and yeah,
keep an eye out for the event in Lisbon in the next couple of weeks as well.
Be good to see some people down there and yeah,
you can jump in,
block with chat on telegram if you've got any questions or just want to hang
out with like-minded people as well.
Sweet.
Thank you all.
Cool.
Stay tuned.
