Projects
![]()
No projects found
Investors
New
DAO Treasuries
DATs
Lend / Borrow APRs
Chains
Categories
Spaces
One Click Launch Your Dream Business
Recorded: March 20, 2026 Duration: 0:45:21Space Recording
Full Transcription
I had a bad habit of missing lovers past
My brother used to call it eating out of the trash
It's never gonna last I thought my house was haunted
I used to live with ghosts And all the perfect couples
Said when you know, you know
And when you don't, you don't
And all of the foes and all of the friends
They'll see it before, they'll see it again
Life is a song, it ends when it ends
I was wrong
But my mama told me it's alright
You were dancing through the lightning strikes
Sleep is in the autumn's night
But now the sky is all bright
Oh, oh, oh, oh, oh
Oh, my Lord, never made no one like you before
You had to make your arms sunshine
But now this guy is open to light
Oh oh oh oh oh
You couldn't understand it
Why you felt alone
You were in it for real
She was in her phone
And you were just opposed
And don't we try to love, love
We give it all we got
We finally left the table
And what a simple thought.
You're starving till you're not.
All of the folks and all of the friends.
Messed up before, they'll mess up again.
Life is a song, it ends when it ends.
You move on. and that's when
I told you it's alright
you were dancing
through the lightning strikes
sleepless
in the honest night
but now this guy
is all up by the light
oh my lord
never met no one like you before. we got to get the whole Katy Perry gist or whoever that was. I'm not sure.
It's catchy. I like it.
Yeah, it was catchy. It was catchy. It gave me time to send a tweet out. So
nonetheless, welcome back to the Blockmates Singularity Show. I'm your host, Brody.
I'm here with Tom Osmond. How are you doing, Tom?
Doing pretty good. Appreciate the invite. And it's good to chat all this stuff anytime.
So always good to jump on.
Amazing. Well, I complimented you on your setup earlier. It looks a lot better than mine.
Very shadowy right now for me.
Appreciate it. Appreciate it.
But yeah, we're going to get into some cool stuff today. We're going to get into zero human
companies. We're going to get into some ideas on ways that people can get started using these
tools to build their dream business. We're going to be talking about some of the new tools that
have dropped and the new products that you could implement to change your life, to change your
business, to change how maybe even your friend's life, right? Like some of these healthcare
products. We'll talk a little bit about agents and stable coins
and get into all of the stuff that you're working on, Tom.
Sure. Sounds good to me, man.
Cool. But yeah, I guess...
Good. Yeah. Good. Me too.
Before getting into it,
I guess I'd love to get a little bit more about your background
and hear about all the things that you are working on.
Yeah, I'll do the super quick version, probably the only version that people
watching this would be interested in, none of the boring stuff. So probably third cycle in crypto,
2021 was the pinnacle for this current meta, which was the NFT boom. And then since then, moved out, left the UK, moved out to Mallorca. I've been working
in AI since 2021. Probably worked for a variety of different startups on creating educational
tutorial content and helping teach people how to use different technologies. Had a community called
the Shiny Object Social Club, which was more about same sort
of stuff ai automation no code app builders all that good stuff then open core comes out we now
have this good new meta where ai can actually do productive things in the world we can now
hook them up with payment rails they can autonomously build software. So recently launched a new ecosystem, community ecosystem called the Institute for Zero Human
Companies, which is basically like now we have all this tech, what can we build with
it and how can we bring others along for the ride?
So we're connecting people, technology providers, investors, developers, engineers, funders,
all that good stuff into one place.
And yeah, we're just trying to push the limits of technology on the internet. And then, yeah,
we can talk about Juno and all that good stuff too. But yeah, combining it all together.
Sweet. Yeah. I'm really excited to dig more into your perspective on zero human companies,
because right now it feels like there's a lot of people that are trying to do this and some of them are more or less a gimmick. Others are really actually launching
stuff. I mean, I saw one company, it was like, hey, you can launch your dream business right now.
If you have an idea for a pressure washing business, it's like we can launch all of these
like landing page, payment provider, ad channels, all this stuff.
But at the end of the day, somebody has to go actually pressure wash the car, right?
So it's not like completely seamless in that sense.
Zero human companies work better in some applications than others.
But the idea is there, right?
And I was really obsessed with this idea back in 2024 when Alexander Good first wrote an article about agentic protocols.
It was the concept of agents raising money through an ICO of sorts or some capital funding mechanism on chain.
on-chain and they use that capital to buy their own compute to then go and complete tasks that
drove revenue back to the token holders and distributed the dividend back to those people.
And we kind of saw some experimentations like with Conway Research a couple weeks ago where
it went out and paid for its own compute. I haven't checked in on that project, but the idea is there and it's
becoming more realistic. And when you're working on your own AI products and you are working through
the flows, you're setting up the orchestration framework with your agents, you're connecting
everything via MCP and API. It really is easy to think to the point where this kind of
just runs itself, right? And I can talk to my agent that's named Brodus on Telegram or Discord,
and it goes out and it does all of the dirty work and it just gives me back,
hey, we made this much money in Stripe invoices today. And you're like, cool, now I'm
going to go for a hike. It really is kind of easy to envision that future with just a little bit
more of an emphasis on some of the integrations and connections and a little bit more of a
streamlining and pointed direction of those agents. So Zero Human Company is really close, I think.
Really close.
Yeah, really close.
I think just by the nature of my personality,
I'm almost the min-maxer in the technology world anyway.
So I want to go.
So we have all these tools available now.
In time, all AIs will be running the majority,
if not all, digital tasks that a company will execute on.
Because even now, there's probably no one who uses a computer
that's more competent and can output more high-level work
than Opus 4.6 running all day on your laptop.
There's just no contest
especially if it's calling sub agents and all that good stuff so we've gone from the sub 100 iq models
in the last couple of years to the 130 plus iq models now which are great at calling tools give
it the right guidance and the right rails, the right permissions with the right
connections into your technology stack, then it can pretty much do any digital task. Obviously,
there's going to be some edge cases, but it pretty much can get there. So the maxed out version of
this is that AI is going to run all these companies, then you're going to have a human
that's going to work with these AIs that are also going to run
potentially lots of small companies or one kind of bigger effort with many agents. So that's kind of
the end outcome for all of this. And now the idea is like, how do we bridge that gap from where we
are now to starting with maybe an AI agent that runs a newsletter that might do a few outbound emails
to get sponsors and then sponsor comes in email gets put together on its own send out every week
and it promotes itself on x and maybe gets a few subscribers every week that's probably
one of the simplest zero human company ideas you can do and then you can scale it up from there. Obviously, you're going to have all of the varieties of that,
either more humans involved, less humans involved.
But I think over time, we've now got this OpenClaw tech
and different variations of OpenClaw,
which you now have to point at a goal,
train it bit by bit, workflow by by workflow bug by bug uh retraining deleting
reinstalling permission by permission to actually kind of get to like your first kind of zero human
product and yeah some some are better at doing it than other people and there was actually some
legitimate projects out there that are actually making some good cash
yeah and we'll get into that too.
There's a lot of these cool platforms like WAP, like crypto native launch pads, or at least they tried to be that, right?
Like Believe that are trying to help people launch internet native companies.
But back to your point about the newsletter idea, like that's a great idea, right?
But how do, if you're listening to this, how do you get ideas to build stuff?
It's like, I want to build stuff.
I know I want to build stuff.
I know I see all these tools and all these people on the internet building things.
How do I actually get inspired to build stuff?
So we created a segment called the idea machine and we can just start rambling off of ideas
and to get inspired about different ideas you can make.
I guess my first piece of advice would be even the most simple thing, like what you
just described, like creating an email newsletter about a specific niche topic that you're
finding interested in and just setting up that
automation and starting that, making the sub stack, making the Twitter account, launching it,
promoting it on LinkedIn, all that stuff. It could be about anything. It can be about
technology. It can be about medieval history. It can be about literature. It can be about
everything. Medieval technology, like catapults and swords and all that kind of stuff. It can be about literature. It can be about technology, everything, medieval
technology, like catapults and swords and all that kind of stuff. That'd be kind of cool. I
might subscribe to that. And you can probably find sponsors on YouTube and stuff. I guess the
point I'm trying to make is just start with something that's interesting to you. And the
more you build and work towards that direction, the more ideas and inspired you get to create
other things. And you get to create other things.
And you got to just do stuff. You can only watch so many YouTube videos about how to go from zero to one with Claude, or you can only read so many tweets, right? It's like,
just go do it. So I've created this is that like you said, but now there's no block in place to actually you creating something that is interesting to you.
And because the actual investment money and time and money to actually get this thing out into the wild and actually working is so low, maybe a couple of dollars in tokens, maybe half an hour, an hour of your time or half a day or a day of your time, you can actually just get this thing out and existing rather than before.
It was much more effort. Maybe you needed some more cash. You needed a technical co-founder.
Just start with what is the most interesting thing to you? What's the smallest version of
this product you can do? And yeah just just create it i i was trying to
figure out ideas so in the process of trying to figure out ideas i created an idea generator
yeah so as i was creating an idea generator it was like if i had an idea then it would run the
idea through a business plan like a non-technical business plan.
It would create a technical PRD for the idea.
It would create a distribution plan
and it would create an implementation roadmap
all through one skill.
So that was my idea generator.
I can actually probably launch this repo
and provide it to the community for free,
like whatever you guys can use it and play with it.
But that process of creating an idea generator
generated ideas, right?
You just got to just generate stuff.
Yeah, one of these good workflows,
another good workflow would be
just scrape your whole X profile,
your Twitter, your LinkedIn profile,
anything that you can find about you on the internet
and then ask AI to come up with some ideas
that best suit you, your personality, all that you can find about you on the internet, and then ask AI to come up with some ideas that best suit you, your personality, or like your interests, all that good stuff.
Speaking of ideas that people want. So if you're familiar with Y Combinator, I mean, it's one of
the most popular incubators in technology. They've seed funded a ton of successful Silicon Valley
based startups. And every year they come out successful Silicon Valley based startups.
And every year they come out with a request for startups.
So these are startups that they want people to submit applications for so that they can go and invest and onboard these different startups.
So it's actually not showing the page that I'm showing right now.
But there's a bunch of different startups in this batch that they requested. So cursor for product managers and AI for government,
which is pretty much what how the government is currently using Claude, I guess, for war,
which we've talked a lot about on this show, which is, which has different moral implications,
but different AI products for the government.
I mean, the government is not just the Department of War, right? It's so much administrative work
across all of these different governance systems in the West or wherever your local government
happens to be. Even I've got a friend who's building AI workflows for the chamber of commerce
here in town, right? I mean, government is a huge addressable market and there's so much
different stuff with permitting, with filing, with billing, with invoice collection. And
there's so many people in that chain, right? That if you just figure out
one problem and implement any solution, it'll probably knock out a lot of inefficiencies within
that chain. The government one's a tricky one because depending on your government, the inefficiencies
are more of like a feature on the bug right because they give
everyone jobs and you've seen over the last years especially after covid how many uh the growth in
like administrative or government-led jobs has just exploded far quicker than the private sector
so it's kind of baked in that they want all this unnecessary regulation in place so they can just
justify employing more
people the there's a funny thing as you saw yesterday somebody's working on a project which
is putting grok against the eu the whole eu regulations and i think it suggested something
around 90 plus percent of all these regulations should be removed so yeah it's kind of might be quite an
awkward one for um if you're trying to plug a ai into a bunch of regulations yeah i don't think
the government runs like a lean startup does it no it should it should and i think our
our friend uh belagi says every government should be running a that they can monitor the health of the country in real time,
which is great, and I agree with that.
And if you're a country like Singapore, that'll probably work.
But if you're a bureaucratic nation by design,
that probably goes against your whole goal.
Right.
So maybe a hard market to crack into, but a market nonetheless.
If you can do it, we challenge
someone listening to go and do it. I mean, just start calling, just call your local politician,
see if they want something to do with AI. But AI native hedge funds and stable coin financial
services, that's a big one, I think. I mean, Leppold, Ash Brenner, the open AI researcher who left, went and started his hedge fund.
I mean, obviously, it's, yeah, it's obviously crushing. He's probably obviously trading off of a system that he built with AI.
And I know a lot of people here locally, too, in Austin who are working on this problem.
too in Austin who are working on this problem. There's a lot of good data that's available
publicly that you can just structure that has never really been accessible unless you're
a quant building all kinds of different pipelines to access and actually turn this data into
insights. So we have a native hedge fund. You have the tools.
Our favorite tool we mentioned, we can bring that up at any point, but now we have a native hedge fund. You have the tools. Our favorite tool we mentioned, we can bring that up at any point.
But now we have some simulation tools at our fingertips where you can plug some of this stuff into.
Perfect.
Yeah.
Tell me more about that.
So this was Mirofish.
This was the tool that was built by a Chinese student, I think in a very small amount of days. And
you basically give the simulation tool a bunch of seed data. It can be text files, it can be
markdown files, it can be PDFs, and then you give it a goal. So you say, I want to know
the next time what you think the price of oil is going to be in the next like 100 days and you
give it the data or you want to predict something on polymarket or calcium or whatever it is or even
if you have a bunch of business data and you say look what should our actions be over the next 150
days this is the current situation of our business and then mirror fish will basically simulate that
with up to a million different agents all interacting with
each other building the knowledge graph and then it comes out with an outcome after x amount of
minutes and suggests the best uh the best result that you should do so now this is completely
open source open to anybody you can fork it create your own version create your own version for any
market you can stick a paywall on it make it available to people to just sign up and use it in an easy way.
But it's going to mean that most decisions or decisions that people working on apps,
products, companies will actually be able to
simulate the results of stuff before you actually do anything.
People who have this tool,
will be able to definitely get a leg up on the
competition it's kind of available on the internet freely now which is pretty insane so miro fish is
that applicable to any asset and contract that you can kind of create this framework around so for instance yeah literally anything you could
you can give it any file any question um if it's you could take your business url scrape it put it
into a text file or a markdown file and also give it the website and then you give it the the goal
or the the question that you want to ask and then it will work out how to go through a simulation of all the potential outcomes
and then gets you to a point where, yeah, this is what you should do,
depending on obviously the question you asked.
So have you used this to trade different markets and has it worked for you successfully?
Have you used this to trade different markets and has it worked for you successfully?
So we, I'm doing it on a few different things in the background.
I pointed it at terminal markets before, which was basically like an AI trading platform
experimentation where you'd have like a bunch of tokens and then the strongest ones survive
and the other ones get reaped and go out of the trading flow.
But people are using it on Polymarket.
People are using it on CalSheet.
And it's pretty insane.
It's definitely, if you compare it to a human,
will a human go through and do it in our brain, obviously,
but not to the scale of a million agents
simulating different outcomes and then
coming up with the best result. So, right. I wonder if there's, I mean, it would be interesting
to test that system on top of all different types of markets and see which one it can
predict the closest. Will it be able to beat Nancy Pelosi with inside the right knowledge if it can't beat pelosi then maybe the maybe the
advantage isn't in the software it's in the uh just the raw insider trading knowledge yeah it
turns out you need insider trading to be uh profitable regardless of what market it is no
yeah exactly but yeah more on the the AI businesses side.
I've been paying closer attention to WAP.
And I know there's probably mixed opinions about WAP because WAP feels like a legitimate business
in and of itself.
But the businesses launched on WAP,
man, they're struggling.
Like, I feel like I log on to intergalactic cable when I go on to WAP.
Are you familiar with that one, like the Rick and Morty reference?
Yeah.
So I get it.
I've logged in.
I've had a poke around.
It wasn't for me at the time.
Yeah.
I'm not impressed with the actual businesses that are launched on WAP because it does feel very Miami course seller,
bro. Like pay a hundred dollars a month to get access to my options trades.
Yeah. And money, like wealth. Yeah.
But these things are making money, right? I mean, objectively, like there's $2 billion,
two and a half billion dollars in creator earnings for a cumulative amount of businesses launched on WAP.
And there are some AI agency, AI-specific products that are launching their storefronts there.
So I'm wondering if this specific vertical is going to kind of help break them out of the business slop factory that they've kind of been
underpinned by for the past however many years of its existence. I went on to their AI,
it's like they're marketing as cursor for businesses. I went on, I plugged my URL in,
my personal website URL, and it created me a business storefront based off of everything it knew about me or it could find.
So like my Twitter account, my LinkedIn, and it was like Brody sells crypto research services.
And I'm like, God damn it.
Yeah, it did me dirty.
It's like, okay, yeah, we're just, I mean, that's what I've been doing.
I guess I'm trying to break out of that, but it was a little bit generic and obviously you can
kind of curate the business storefront to be whatever you want. But the concept was cool,
right? Like the concept of WAP is what I really want to dig into where you get an idea, you go quickly iterate on top of it. You can get connected to users.
When a paid user joins your community, you can message them and start talking to them and engaging
with them directly. It's got the embedded payment processor. You can earn yield directly within the
payment processor or the account that they set up for you
um i'm assuming it's all stable coin yield and it's an interesting new way to distribute
information and products like the concept there the concept set and undoubtedly great tech the
actual platform itself great technology they've built a platform that people want to spend time on,
that people will create and sell through,
building payment rails that will take,
that is detached from Stripe,
that will take any big volumes of payments through.
So doing a good job, and obviously you,
with any new startup, you kind of just follow
where the customers are coming from,
and then actually
led down one direction i'm sure it's diversifying over time but there's there's obviously like 2.5
million businesses 20 million users over nearly 3 billion earned um the platform is great i think
also the good thing it does is um a lot of people undervalue the knowledge that they have, whereas people will always want to learn from you.
Let's say they watch this stream and they watch how you put together this content, how you built your brand on X.
A simple product would cost a low amount of dollars.
It might seem trivial to you like how to how to brand
your x account how to set up on stream yard and how to create your first first 10 streams how to
attract sponsors that's like something which would be really really valuable and if if what are making
it easy for somebody to not have to think about all of the finicky stuff that goes with that just
say look i'm just going to throw it up on WAP. It's going to charge $9.
And all the rest of it is just going to handle itself.
There's a place where people can chat,
a place where people can buy it.
Then I think, yeah, it's a value add to society.
I think that's an interesting take
because I get told that a lot of the times by my friends
about specific subjects and topics, right?
Where you kind of discredit the amount of information you actually
have on a topic 100 um and you feel like it's general knowledge and general information
right yeah because everyone in your world that's this jack butcher visualize value um the best
earning your first one dollar on the internet uh change it changes your life as soon as you have a
one put a product out
there as soon as someone pays one dollar for it it just flips kind of how you think about it so he's
been a lifelong proponent of that yeah and i think don't be embarrassed to go out and just start
launching stuff right um yeah i also think yeah let's just start selling stuff man like build
cool stuff there's probably somebody out there who wants it and just execute.
And like I said earlier, the more you execute, the closer you get to what you're actually executing for in the end game.
But yeah, moving on from WAP, we've got the tool shed.
We were talking about this a little bit earlier.
There's a lot of cool tools.
Every week there's new tools. I've built a tool to scrape tools Um, there's a lot of cool tools every week.
There's new tools.
I've built a tool to scrape tools.
I'm sure everybody has done that at this point.
Um, but I guess on my tool that scrapes tools and maybe I will launch this because I should
follow my own advice and I should just start executing and just putting, putting thoughts
and ideas and products into the world.
the world, but my tool scrapes tools and it matches them with my clients and my projects
But my tool scrapes tools.
that I'm working with to show how this tool could be relevant for a specific problem within
that client that I'm working on or all of my outstanding projects.
Right.
So it's directly relevant or providing relevant context on the actual feed itself, which is very powerful.
Right. So you read something where it's like, let's see,
anthropic remote task dispatch, assign tasks to Claude from phone or desktop.
You're like, okay, yeah, that's, that's cool. Right. But how could you actually make that
relevant? You know, you have to give it access to the, you got to give something to access to
your knowledge base to show you, Hey, if we just implement this product for this workflow,
this is a good, this is a good way that we can, um, apply this to all these different,
different clients and projects that you're working on. So I think that's pretty interesting.
Um, but Google stitch, I was talking about this earlier this week, or maybe last week on another
show, super powerful already. They launched the second version this past week. Super sweet. I am
getting super inspired about building like a designer in a box where you just plug in your agent to this box that is all the mcps and
ai or apis across all these design platforms they kick off workflows across google stitch
paper.design whatever design has been a huge bottleneck for me personally um and now that i
have i think for many for many people too that that's why I saw so many similar looking vibe-coded websites
that are all like blue, purple ingredients.
I think Tailwind was default across these and everyone,
because nobody really knew how to prompt
like a designer would think to prompt
or think throughout the designer website,
people just didn't really have the words that
they could put into the box to create something which looked good so so yeah seeing these products
like stitch uh just yeah change changes it all which is cool have you used any of these design
products yourself yet yeah i have used them i haven't tried the latest version of Stitch. I used the previous version, which was pretty good.
The reason why I haven't used it more is because right now I'm just kind of trying to do everything through the agent because then through the OpenClo agent because it has then pure like session history, the memories, every single like the uh the audit log of
how you went from thinking to the final product so you kind of you're kind of leaving that trail
there but yeah this is a great tool i've used the previous version it does create some epic sites
agree nemo claw nvidia gtc was this past week. Yeah, NemoClaw was a huge release from them.
I think, yeah, a lot of, especially once you move out of open claw hobbyists,
just building for yourself and then start working,
or want to implement it into a more serious business context,
or maybe you're working with clients, then a big part of it's going to be security so
nemo claw uh is basically the yeah the security part that sits between or on top of open claw and then the actions that open claw takes just to make sure that it's not just going nuclear on
somebody's server or files that it shouldn't have um and yeah it's big that jensen and nvidia are kind of going crazy or leaning into it so
positively because jensen is he was plugging it pretty hard i know there's also plugging tau
or he was getting plugged with tau the other day which was crazy yeah um good times yeah
we just talked about anthropic remote task. I think it's called Claude Channels, correct?
I think it's actually called Dispatch in the toolbar.
So you just press Dispatch.
Gotcha.
But yeah, it basically allows you to create channels inside of Claude, yeah.
Yeah, which again, allow you to communicate with your Claude instance
from your Telegram, Discord, whatever.
They're trying to really play catch up on the open claw zeitgeist
um so everything that they've launched in the past like three weeks has been
more or less a catch-up play for open claw and then you probably see yeah you probably see
um if you're dario you see jensen on on stage talking about you know we're taking open claw
we're forking it we're making nemo claw
we're investing resources into this open source framework um how linux did for the operating
system we think the open clause is this for magentic operating and then they're probably
dario is sitting there like we got a ship we got a ship he's hitting the slack
yeah it had it on a did have it on a bit of a platter to start with
where obviously it was called Claw before.
And then they made him change the name.
I think they should have probably bought it to start with.
But then Mr. Ullman, the opportunist, snapped it up.
So thank you very much.
Peter is a huge Codex Power Goat user.
He's like a 1,000x Codex user, so that was a good fit.
And now maybe they're trying to claw back some of that attention
on the internet because everyone's open claw stands,
and now they're like, oh, we have claw code.
You can actually turn it into a claw
like manis have released their claw thing you got minimax with their claw kimmy perplexity
launch something similar computer yeah everyone's claw build yep stride health created nora an ai
health insurance guide matching users the best plans. I don't know
what the situation is like in the EU or how your health system works, but for our health system,
our insurance marketplace is a complete joke, right? And I know that I would use something
like this because heck, $50 access to something like Noro,
whatever it is, and it can go scrape all these plans and tells me what I actually need. Because
when you're searching the health insurance marketplace on the government website,
as someone who's self-employed, right? Like I don't have an employer who's paying for my healthcare.
It's so crazy. You're swimming in a sea of just mumbo jumbo jargon, contracts,
descriptions, all different types of stuff, premiums and all these different considerations.
And something like this would be amazing. I mean, I know people who are spending $1,500 a month for
their family, $2,000 a month for their family on health insurance,
and they don't even use it, right?
So there's probably alternatives out there.
It's good to use something like Nora
to help find a plan that works for you.
Not a sponsor, even though it kind of wrapped up
and sounded like a sponsor.
No, actually something we have to worry about
in the EU, fortunately.
But yeah, you kind of just get,
you get healthcare anyway,
and you can get go private
if you want it's a fraction of the cost the us but i can see that being a huge thing right
and then mule run this was cool this came through my feed the other day it's basically
again same thing always on agent self and self-evolved from real business workflows everyone is kind of shaping up to attack this vertical of
always on agents yep yeah these are good i think the um this is going to be like the the new
trend the new meta going through yc i think the things like this, because these pure SaaS plays are so easy to copy now,
you could probably, and the craziest thing is,
I don't know if you ever tried to reverse engineer a product
with GPT 4.2 or Claude Code and Opus,
like just to the level of inspecting network requests
and API calls and everything,
you just set it off and come back in an hour and it's just rebuilt the whole product.
So it's almost like the bar for technology now is so much higher than it was.
People were just kind of expecting this to be a thing.
So, oh, I now expect my own fully always onian to be always alerting me to everything i need to
know from my email to my calendar to my medication and anything that kind of comes in below that
you just think oh i might as well just use this other this other product so yeah it's tough it's
tough that's why what we said before is i think if you're going to do a project now,
you kind of have to go all the way.
You have to go to create something which is super difficult to create.
I just saw Jensen talk about it on the All In podcast.
It was a simple sentence.
Is this thing that we're trying to do really hard to do?
If it's not, we won't do it.
Right.
I think that makes sense. I think it is crazy that we have shifted completely to
this agent first perspective on building businesses, kind of what we were talking
about earlier. There's so many legacy antiquated businesses that will very quickly be disrupted by
businesses that will very quickly be disrupted by a market, a new market entrant who's building it
from an agent first principles basis, right? 100%. Because they have so much bloat and so
much bureaucracy baked in that it makes it difficult for them to pivot where someone
starting from scratch ironically has a better edge at disrupting their business
about this the other day where we thought they were seeing huge amounts of layoffs
for instance one example is uh facebook facebook bought manas and manas was the best in my opinion
the best super agent the best super agent harness and they've plugged it in or are increasingly plugging it in across the whole
Facebook ecosystem internally and externally. It's immediately nuked all media buyers pretty much,
or at least the 98% of media buyers who are just average, like the 0.2,
spending multi-millions a day, whatever, they're going to be fine. But for your average media buyer,
Manus is way better than
you like way better you're managing ads turning them on and off they're underperforming spinning
up new audiences and tweaking them no contest and now they're doing the same for for instagram for
their creator marketplace they're no doubt ripping ripping it through all their internal tooling and it is they're laying off like tens
of thousands of people at a time if you if you are a ceo of a bloated org you there's they're
probably thinking in the head i wish i wish i could start this again from now and build agent
first or zero human company mindset first if you basically push the AI as far as you can go or push the AI and
sub-agents as far as they can go. And then you hire a human right when you've kind of pushed
the capability, not the other way around. Yep. Speaking of that, we only have a short
amount of time left and I want to touch on the primary topic of this conversation, the zero
human companies, which we just alluded to,
everything has been building up to this point. Companies that launch, operate and generate
revenue with no human in the loop. Some of the latest versions that I've seen this week was
durable, which I kind of dunked on this guy. He really used a poor example of a zero human company
with the pressure washing business or the car detailing business, because ironically, the worst example for a teaser video,
the irony is one of my good friends owned a car detailing business and realized it was the only
thing that you couldn't scale with AI. So he got a job at an AI company as a forward deployed engineer, right?
So it's like somebody with subject matter expertise on car detailing
went to go deploy agents for other service businesses
because it was too difficult to deploy for car detailing.
And then this guy goes and creates his entire demo and product
around launching a car detailing business.
Yeah.
Interesting.
Yeah.
Yeah.
But then we saw Pulsia come out.
It's kind of a similar thing.
Like you log on,
you give it your name,
you give it your handle,
it scrapes,
and then it creates a business,
launches it for you.
Paperclip was an orchestration framework.
Quitting on Pulsia. Shout out to to ben from pulsier because i think he people are giving him a lot of shit saying uh how many of
these businesses are earning money and all that sort of stuff but ben kind of like defined this
meta now of uh zero human companies being or being ran by agents and sending emails generating business you can take a cut you
can connect your stripe and all that sort of stuff but it's working they're managing all these
companies at scale on his own just with agents and the run the run rate shows sure is it going to be
a big churn probably is there going to be way more customers joining definitely our business is
launching it going to be mostly earning their money. Sure. But that's how all of these vibe coded platforms started.
They started bad when the websites are bad. And now look at Lovable, 400 million run rate.
These companies are going to do exactly the same. They will earn small amounts of money until they
suddenly earn large amounts of money. So yeah, AI learns pretty quick.
So I bet on them doing well.
Nice.
Good shout out.
Good shout out.
Paperclip, something similar.
I think it's more of an orchestration framework.
But the point is...
Yeah, it's good.
You can use that one.
Cool.
Paperclip is basically the equivalent of OpenClaw,
but for zero human company software.
So completely open source, hackable, shipping really fast.
Yeah, not it Felix is just using it.
Yeah, I've used it.
It's a good platform.
Sweet.
Well, I guess like with this minimal time we've got left, talk to us more about the
zero human company institute that you're working on.
Yeah.
So idea behind the institute
is basically we now have this this whole new meta which everyone from from builders to entrepreneurs
to investors to engineers to companies to smbs they're all going to be thinking of the same thing
how much work can we move from humans over to agents over time?
So then the meta play that I thought was create the place where people can come and
learn, meet other people who are thinking the same way.
And we built it with Juno, who is an open call agent.
This is her first product.
She built the whole thing from scratch.
She runs the community.
She runs the support email.
She builds all the tooling and just to absolutely dog food it and yeah it's a paid community you can get a
99 lifetime membership to it at the moment we've got like almost 200 people in building with all
these tools we provide all the infrastructure and support network and all that and then
yeah eventually over time people are going to start, like we said before, start creating small products and then replace
maybe one or two business functions in an existing business, and then you can start growing from
there. So zhinstitute.com if you're interested in this sort of stuff. Yeah, we're all figuring
it out together. Cool. Awesome. Well, Tom, really appreciate you coming on and breaking this down
with us. It was really awesome. Where should folks go really appreciate you coming on and breaking this down with us.
It was really awesome. Where should folks go if they want to follow and find out more about the
Zero Human Company? It's currently online on X, just at Tom Osman. I'm on there pretty much 20
hours a day. So ping me on there for new things. DMs are always open, or you can go to TomOsman.com
or join the Discord, join the institute. Sweet. Cool.
Thank you all for tuning in.
If you want to stay ahead of the frontier technology curve,
make sure you subscribe to the Blockmates newsletter at blockmates.com.
And we'll see you all next week.
Thanks for having me, man.
