Latest / TFTC: A Bitcoin Podcast / #796: Own Your Intelligence Stack with Tommy Eastman

Transcript

#796: Own Your Intelligence Stack with Tommy Eastman · TFTC: A Bitcoin Podcast

0:00
You've had a dynamic where money has become freer than free.
0:10
If you talk about a Fed just gone nuts, all the central banks going nuts.
0:16
So it's all acting like safe haven.
0:18
I believe that in a world where central bankers are tripping over themselves to devalue their
0:23
currency, Bitcoin wins. In the world of fiat currencies, Bitcoin is the victor.
0:28
I mean, that's part of the bull case for Bitcoin.
0:31
If you're not paying attention, you probably should be.
0:37
I've been very much looking forward to that.
0:40
I was very excited when we first talked a couple of weeks ago because I've been using your product for four months now.
0:48
And it has changed my life.
0:49
You saw me when you walked in here.
0:50
You thought I was on the phone.
0:51
I was talking to my Hermes agent.
0:52
No, talking to your Hermes agent.
0:54
Better than talking to somebody.
0:55
I like the human.
0:57
I like the human element.
0:58
okay i do like talking to my wife greg who's off camera here he's okay to talk to every once in a
1:03
while but uh it's uh it's changed how i work pretty pretty quickly i mean i'm 35 i've been
1:12
doing tftc for almost a decade now and within six months my full workflow has changed and hermes is
1:19
critical part of that yeah what let's take a step back and talk about you personally how you ended up
1:29
at news research which built the hermes agent um you were in the bitcoin industry the crypto
1:36
industry at foundry before this and that's i mean that's a whole nother conversation i think it's
1:40
really interesting how a lot of these up-and-coming ai companies are seated with people that have been
1:45
in the Bitcoin and crypto industry before.
1:48
Yep.
1:49
But how did you end up at Noose?
1:52
What drew you to it?
1:54
And what have you been doing there since you got there?
1:57
Yeah, so I was at Foundry,
1:59
touched the Bitcoin stuff a fair bit,
2:01
but was really focused on distributed
2:04
and decentralized AI there.
2:05
I just thought that AI,
2:09
obviously an incredibly powerful tool,
2:11
very important that
2:13
we get the ideological part of AI shaped correctly.
2:21
That was my first real exposure to that,
2:24
thinking about the importance of AI for people,
2:27
for the individual, the sovereignty that AI enables.
2:32
I didn't really like a lot of the kind of first manifestations of that,
2:38
like the first projects that kind of tried to champion that.
2:41
I wasn't that impressed with.
2:45
I found Noose eventually.
2:47
They were focused on post-training open models to make them more pliable, more human-like.
2:55
This was when the Meta models were state-of-the-art, right?
2:58
The Lama models.
3:02
I quickly found that the team was just incredible.
3:06
One, very ideologically driven, right?
3:08
At Noose, we truly believe that the ability to own your intelligence stack is the most important technological innovation that we've ever faced as humans.
3:24
But also that the team was just incredibly sharp.
3:28
Everybody at Noose is just a killer.
3:31
In love with AI, very driven.
3:35
It's just an incredible, incredible experience.
3:38
And yeah, for me, I focus on a handful of things.
3:41
We all kind of wear a bunch of different hats, right?
3:43
We're still pretty small, pretty grassroots.
3:46
But so I touch a bunch of stuff on the product side, onboarding enterprises,
3:51
and manage all of our compute relationships as well.
3:55
So we have our managed product, which is News Portal, which powers,
3:59
which can power Hermes Agent and powers Hermes Agent for a lot of users.
4:03
So managing all the inference and tools behind that.
4:06
So let's take a step back and just explain Hermes.
4:11
Let's focus on Hermes specifically.
4:14
And let's, let's, because we've had this conversation off air a couple times now,
4:19
but speaking to the person who thinks that the creme de la creme of their interaction with AI is opening chat GPT and hitting chat and thinking that they're getting inexperienced.
4:31
what does hermes bring to the table and how should people view the hermes hermes harmus and how to
4:36
utilize it yeah well that's the that's the incredible piece right is that we're so early
4:42
we feel like ai is everywhere and everybody's talking about ai but there's only a small slice
4:47
of people actually using ai and then the vast majority of that slice of people are just using
4:51
ai basically as advanced google right google on steroids um like my mom loves perplexity she uses
4:58
perplexity all the time, but she's basically just, she likes it
5:00
because it's Google, but a little nicer.
5:04
So where you start where you make that jump from AI chat bot to agentic AI is when you start to enable actual autonomy So these agents start to actually be able to make some decisions for you and carry out tasks without you having to sit there and monitor them 24
5:24
You can go and have an agent book reservations for you.
5:28
You can have an agent look at data that you have and build dashboards.
5:32
You can start to really automate real tasks.
5:35
at Noose. We use it for a bunch of different things ranging from managing our CRM of enterprise
5:41
customers to taking ideas that we have for content generation and working through a whole
5:48
creative pipeline to create our end product. Why Hermes Agent itself is, I think, very special.
5:57
Two major reasons. One being that, I'll say three major reasons. One, the first being that it's very
6:04
simple to use it really just works um i think a lot of open source projects end up with this issue
6:11
where the repo gets bloated they want to be inclusive of everything um but we've taken a
6:17
stance where yes we want hermy's agent to be able to bend to your will and be completely pliable but
6:23
that it's really important that we maintain the sanctity of that repo and are somewhat opinionated
6:29
and what kind of exists in that core repo to protect the user experience.
6:34
So working out of the box is a huge reason why it's been successful.
6:39
Two is the memory piece, the self-evolution piece.
6:42
So as Hermes Agent, as you use Hermes Agent, you'll notice market improvements in the quality of output over time.
6:51
And this is a really important piece because I think where people end up frustrated
6:56
and what prevents people from using ChatGPT in autonomous ways is you can't do tasks reproducibly with the same results, right?
7:07
So you ask Chat to go do a task.
7:10
It tries to do the task.
7:11
It might figure it out the first day, and then the next day you have it do the same task, and it can't figure it out, right?
7:16
And it's a super frustrating experience.
7:18
Hermes Agent stores the primitives of that path where it found the right solution and then recalls them every single time.
7:25
So if you figure something out once with Hermes agent, you're going to be able to do it again and again and again and again.
7:30
And that's where you can truly say, OK, agent, go do this.
7:32
And you don't have to sit there and monitor and worry.
7:34
And you can have other tasks be dependent on that task.
7:37
Right. The third the third piece of why Hermes agent, I think, is so, so beautiful.
7:43
And this is like the the ideological holy grail. Right.
7:47
Is is it's able to be fully model agnostic.
7:50
And that is important for a lot of reasons. And we thought it's been important for a really long time at Noose. You've seen more and more attention come to this issue now, though, of data sovereignty and provider lock-in.
8:05
Nobody wants to be locked in to a single big lab.
8:09
Nobody wants these labs training on their data.
8:13
People want to own that stuff because that's the secret sauce, right?
8:16
That's the sauce of my company is this data.
8:20
And if I'm sending all of that to a lab that's going to use that to train in the next model, I'm giving away the secret codes.
8:29
So Hermes Agent enables you to own your full stack.
8:32
You can choose to have stuff managed if you want, your cloud instances managed or whatever.
8:37
You can basically craft whatever solution meets your sovereignty and privacy needs, and you can go implement that.
8:46
And if things change, you can pivot it as well.
8:48
So there's no provider lock, and you can use any model in the world.
8:52
You can use much cheaper models.
8:53
There's so many open source models that are incredibly performant for the vast majority of tasks now and at a fraction of the price point.
9:00
um so yeah it's i think it enables it enables you to just have a much more sovereign ai experience
9:10
and really own the full the full stack from the floor up and it's uh i mean we the so my my hermy
9:18
story my name is marty ben i'm a hermy's user yeah so i i open claw over christmas break people
9:27
like open claw i'm like one it snowed like 30 inches in philly one sunday and the kids were
9:31
watching a movie i was like all right i'll try this out so like first i've ever set up a uh vm
9:37
set up a cloud server set up the hetzner server like went through this youtube store set up open
9:42
club and connected it to anthropic at the time when you could oauth in before they uh before they
9:50
prevented that from happening and was immediately blown away like holy crap this thing is insane so
9:57
for about four months i was building out that open claw uh instance and integrating more of
10:03
our workflows uh into into the agent set up in the harness in the cloud and like seeding the file
10:11
system in the cloud with everything that i wanted this this computer to be able to access and
10:18
along the way like april may um i was talking about open claw and i heard people it was probably actually March April people began to be like hey you using OpenClaw you should try out Hermes and then another similar situation
10:30
where it was just a Sunday afternoon
10:31
and I was sitting there
10:34
and I was like okay, I've had
10:36
10 people reply to
10:38
some of my tweets like you should try Hermes
10:40
and I was like okay
10:40
we literally went to my OpenClaw agent
10:44
on Telegram and I was like hey
10:46
spin up a Hermes agent
10:48
in your server and let me know when that's finished.
10:51
And then I hooked up Telegram.
10:53
And then within like 24 hours, I was like, okay,
10:56
Hermes agent shut down the OpenClaw agent
10:58
and we're going to begin working.
11:01
And it's been insane what we've been able to do,
11:05
what I've been able to do on the team at TFTC
11:07
over the last four or five months,
11:09
just really seeding the Hermes agent.
11:12
But to your point, I think that's what I really want to get through to people
11:17
when we discussed this on our call a couple weeks ago
11:19
is I'm interested to get your thoughts
11:22
on how you view this.
11:23
Like I think there's three legs to the stool
11:25
of an agentic system,
11:27
particularly if you're doing it for a business.
11:28
You have the file system or the second brain
11:32
and there's ways in which you can
11:34
make that second brain stronger
11:38
with things like semantic search for QMD
11:40
or an obsidian vault for a knowledge graph
11:43
with associations.
11:44
But I would bucket those into one thing.
11:47
uh second brain essentially a file system the harness uh the agentic system in our case hermes
11:55
and then the models the two most important things are the second brain and the harness and the
12:00
models like you said you can be model agnostic and that's what i love about hermes is it makes it
12:06
extremely easy to switch models and i'm hitting slash model changing for different
12:11
tasks when i don't want to burn tokens or burn expensive tokens or if i want to use a more
12:16
private open source model uh i don't think most people understand like i think we're at this
12:24
it's not that i don't think i know most people don't understand that this uh this sort of three
12:28
pronged approach and what is the value what is the relative value of the different prongs from
12:33
that three-pronged approach and i think that's where the opportunity lies and why i love hermes
12:39
i think you mentioned like the self-referential learning as part of the memory system and i think
12:46
there's a ton of people out there saying all right ai's here um i've been convinced it's not going
12:52
away i need to implement it in my business so just focusing on businesses specifically like how would
12:59
you like is that is my inclination of that three-prong approach right in your mind and if so
13:06
how do you articulate that particularly to like enterprise clients walking through this yeah yeah
13:11
I think you're completely right.
13:14
There's so much value.
13:17
The returns are compounding.
13:19
The quality of intelligence is compounding as the agent gets more and more intelligence.
13:26
That file system is incredibly valuable.
13:30
The agents are only as good as the data that they are able to access.
13:34
It's not some magic.
13:36
And so the ability to give your agent access to as much context as possible and basically just let it go to work is the secret sauce.
13:49
That's the real value in these systems.
13:53
And I think it's almost hard to articulate sometimes to enterprises, like, why would I use Hermes agent?
14:01
It's like, well, you can use it for everything.
14:03
and that can that's it's like it's almost too abstract right or too good to be true yeah yeah
14:08
yeah it's like okay what does that mean and it's like the reality is though you really can and
14:14
you've seen it i'm sure you really can use it for anything any monotonous task you may have
14:19
anything you want done just throw it at your hermes agent and see what happens
14:24
right i think the the low-hanging fruit where a lot of people immediately say wow this is incredible
14:30
is working with big data.
14:34
If you have, whether it's your Stripe data,
14:36
your customer data,
14:39
the ability to create, to organize that,
14:41
to create visualizations, to do modeling
14:43
is stuff that these agentic systems excel at, right?
14:50
They are just so much better than we can possibly be at.
14:54
And so I think that's where you see a lot of the time
14:57
people immediately go, okay,
14:59
this is a game changer. When they have some big data problem and they make visualizations out of
15:04
it, it works right away in one shot. But yes, to your original point, the structure is your company,
15:13
in its entirety, it should serve as your company brain. Those three pieces of the stool, like you
15:20
said, are the company brain. I think I agree with you that it's context or file system, harness,
15:28
and then model.
15:29
I don't want to discount
15:30
the model too much
15:30
because model quality
15:31
does matter a lot But they very additive But they becoming commoditized That where I Oh in the long tail for sure i think in the long tail you can be like over time the ability
15:43
to be model agnostic is only going to increase yes yeah i mean it's a reaching parity nine
15:47
nine months ago you couldn't use an open model for anything like they were basically worthless
15:53
right um and now you i like to say i think about you know 90 of users can do 90 of their work
16:02
with open models right like there is definitely stuff that you want the the biggest guns for but
16:08
tons of work can go to open models and yes that gap is continuing to close and it's only going
16:13
to become more and more commoditized yeah and at what point do you like do you like what is the uh
16:19
the minimum ai iq level that you need to do all the tasks like do you really like
16:24
a year from now are you going to need a frontier model to do your completely do your modeling or
16:29
or your crm yeah the vast majority of people are never going to do tasks that require these these
16:34
top tier models it's just not part of part of what they do no and the so let's take a step back and
16:41
talk about ai broadly like where do you think we are um what trends are you most interested right
16:50
now is it the data center compute side is it this model race is it the harness uh what are you most
16:58
interested right now in the world of AI outside of Hermes and news?
17:02
Yeah.
17:02
I mean, I think it's really interesting to follow how the open source models have changed
17:10
the market, right?
17:11
Like I said, nine months ago, you really couldn't use open source AI.
17:16
Now, if we transport back to January, February, you saw a bunch of prominent CEOs talking
17:23
about, oh, our employees need to token max, token max, token max, like unlimited spend.
17:28
Very quickly, they were like, whoa, way too much money and want to find cheaper solutions.
17:36
That timed well with open models reaching the point where they can be these cheaper solutions.
17:41
I think there's a lot of stigma around them for some people still about being mostly Chinese models.
17:46
China is completely championing the open source model, the model race.
17:51
But you're seeing numbers in Hermes Agent on open models go up and up, on open router, open source models going up and up.
18:02
Companies are recognizing that AI spend is out of control, that being locked into OpenAI and Anthropic is prohibitively expensive to power your business.
18:14
And open models are really the only solution.
18:18
And there's such an incredible free market of these open models.
18:24
There are so many different GPU providers in different locations, different compliance requirements.
18:30
So you can tailor whatever you need.
18:32
If you need SOC 2, if you need GDPR, if you need them in the US, there's all these different things that you can tailor to fit your needs.
18:40
But it's a much just more freer market.
18:43
So you're seeing that price discovery.
18:45
you're seeing model prices decrease and you're seeing people be able to access intelligence much
18:49
much cheaper which is obviously just incredible for for the economy for business for small
18:54
businesses especially no and i think i just discussed this with jordy visser who was on
19:00
before you rolled in but i think the the latham and walk-in story from a couple weeks ago the
19:04
law firm basically said we're not using anthropic or open ai we're going to buy our own gpus and
19:10
self-hosting model that i mean that's a massive signal not only the models are good enough to do
19:15
that but also i think a trend that many people are not um not appreciating enough and i think alex
19:23
carp's been beating the drum about this but like particularly if you're a enterprise the size of
19:28
latham and walk-ins or any fortune 500 company like the feeding of your company ip to anthropic
19:37
and open ai is i think it's going to get to a point where shareholders or board boards of
19:43
directors gonna say this is like a breach of fiduciary responsibility i think you're just
19:47
handing over the secret sauce to yeah these model providers who are uh proving or who are literally
19:55
launching products that they're training off of companies that are doing this like that they're
19:59
going to try to compete with you and abolize your business while it's providing you value in the
20:03
interim and so to your point about open models like i think there will be um a massive inflow of
20:11
usage for an increased usage of these open models from like a fiduciary perspective just to preserve
20:17
your ip yeah i mean people are just throwing their trade secret it's it's actually a really interesting
20:22
psychological phenomenon right like yeah businesses are doing it but also that the average user the
20:28
the chat GPT user, the Anthropic user,
20:30
people share so, so much with these models
20:35
and almost completely unaware of privacy
20:39
or how that obviously could backfire
20:42
and the different ends that that could go to.
20:43
It's happening at the individual level.
20:46
It is happening at the business level as well.
20:47
But I think at least now, the more
20:50
more businesses are very aware of that that's something that really resonates um with businesses
20:55
that we talk to about about hermes agent is they want to ever since that alex alex carp um spiel he
21:02
went on businesses want to own this they're very very aware um they're very very concerned about
21:08
where their data is going i think the general public is lagging behind but i think there'll be a
21:12
rude awakening for them as well agreed agreed what's up freaks this was brought to you by our
21:19
are good friends, the Bitcoin team at Block. Bitcoin at Block. Bitcoin was originally introduced
21:23
as a peer-to-peer electronic cash system. We all know that. The idea wasn't just to buy and hold,
21:26
it's to use Bitcoin in many different facets. Holding is one facet. Spending, receiving is
21:30
another facet. I've been working towards this for 17 years and Block has put together a whole
21:35
ecosystem of companies to really bring about the reality of Bitcoin as everyday money. So with
21:39
Cash App, you can buy, spend, send, receive, earn, get paid in Bitcoin. When you're trading over
21:45
$2,000, you're not going to pay any fees on that. So then there's Square, which is helping bring
21:49
Bitcoin into real world commerce. Eligible businesses can accept Bitcoin payments over
21:53
lightning with settlement in seconds and 0% processing fees. Customers can pay with Bitcoin
21:57
or even dollars over lightning through Cash App and help local businesses keep more of every sale.
22:01
It's making Bitcoin payments feel as easy as tapping a card. And when you're ready to hold
22:04
Bitcoin, there's BitKey. It's a Bitcoin wallet built for real life. No seed phrases to two is
22:09
free multi-sig, so no single points of failure. It has recovery and inheritance processes and
22:14
and capabilities built in.
22:16
And the new BitKey device includes a screen
22:17
so you can verify important actions
22:19
directly on the device before approving them.
22:21
So you know exactly what you're authorizing.
22:22
So together, Cash App, Square, and BitKey
22:24
are helping make Bitcoin easier to earn,
22:25
save, spend, and secure.
22:26
It's a simple idea.
22:27
Bitcoin works when people can actually use it
22:29
in the way that they want.
22:30
That's Bitcoin App Block.
22:32
What's up, Freaks?
22:32
This episode was brought to you by Simple Mining.
22:34
I'm sure many of you listening to this right now
22:36
own Bitcoin and own it the same way.
22:38
You go into an exchange, you buy it,
22:40
sweep it to the cold storage,
22:41
maybe keep it on the exchange,
22:42
and you pay full price.
22:43
but there's a smaller group of business owners,
22:45
family offices and investors with sharp CPAs
22:47
that are quietly getting theirs a different way.
22:49
They own the machines that make it.
22:51
Here's how it works.
22:51
You buy the miners at Simple Mining.
22:53
Simple Mining hosts them,
22:54
runs them at their facilities in Iowa
22:56
on industrial power rates you can never get on your own.
22:58
The Bitcoin they mine goes straight
22:59
from your machine to your wallet.
23:01
Two reasons I find this interesting right now.
23:03
The second one is the one that nobody's talking about.
23:05
One is the economics, cheap power.
23:06
This lets you convert your electricity into Bitcoin
23:08
at a discount to spot.
23:09
You're acquiring the asset below the sticker price
23:11
everyone else pays.
23:12
two and listen closely if you've had a big income year making a lot of money out there i know a lot
23:17
of you freaks are making a lot of money miners are classified as computer equipment with the big
23:20
beautiful bill equipment means depreciation qualifying buyers can write off the entire
23:24
purchase and year one it's the same playbook real estate investors have run for decades pointed at
23:29
bitcoin your cpa will want to see this before the end of the year and with attractive financing
23:33
available you can deploy without selling assets to do it this is not a garage operation simple
23:37
mining runs over 40 000 machines which just ranked the fastest growing company in iowa
23:41
on the Inc 5000 chart.
23:43
They put together a valuable resource
23:45
called the 2026 Bitcoin Mining Blueprint.
23:47
It covers the five mistakes investors make
23:49
when allocating to mining
23:50
and how to avoid them before deploying capital.
23:52
You can access it for free at simplemining.io slash tftc.
23:57
That's simplemining.io slash tftc.
24:01
It's funny, I was taking the MetaMuse agent
24:04
for a spin this morning.
24:05
I was like, wow, this thing's powerful.
24:06
But then, how much of this information
24:08
do I want to give to Meta and Zuck?
24:10
and i think that's a question that a lot of people are not to grab yes but i mean history shows
24:17
particularly in the digital age most people don't care um that's why i think like focus on enterprise
24:21
like where you have to care because money and shareholder value for lack of a better term is on
24:25
on the line um but then for the individual like running individual agents like i think uh
24:32
hopefully uh team at news research and others can get the the ux uh to parody with something
24:40
like amuse agent or just works out of the box which i think to your point like that was one of
24:44
the most impressive things like i literally told my open call agent set up a hermes agent and had
24:49
it working within like 10 minutes yep yeah i think that um it's really important to make the product
24:56
easier and easier for people to use right because if it's you know hours of setup versus instant
25:02
setup and the privacy is the only differentiator that's that's been proven time and time again
25:07
that that's not enough. I do think on the individual privacy aspect, I think it probably
25:15
needs some sort of massive event to catalyze people growing aware of this concern. But I do
25:25
think we're at a stage where, yes, it was bad when you shared all of your information in this
25:31
digital age previously, but now the ability for agents to access it, to compile it, access it,
25:38
spread it, it seconds. I think there are way worse potential outcomes on the table now that
25:46
agentic AI is as powerful as it is if people continue to be so free with their data.
25:53
I mean imagine what percentage of chat GPT users would be horrified if their chats all got leaked Well we see wasn there the ability to like Google like if you were sharing the output
26:08
and you create like a share link that was getting cached in Google search.
26:11
Yeah, yes, yes, yes.
26:13
Something similar happened to Meta as well.
26:16
Yeah.
26:17
Where they thought they were like having a private conversation with the AI or they're
26:19
posting it publicly.
26:21
Yeah, it's horrifying.
26:22
Horrifying, horrifying.
26:25
Yeah.
26:25
and it's so how would you like that's what i'm i've been trying to discern because things are
26:31
changing so quickly like last year at this time like opus 4 6 wasn't even out yet correct so you
26:39
didn't have like there was that frontier model or breaks or memories like oh wow now this stuff can
26:45
like really cook and then a month later it was like they were working on open claw you guys
26:50
been working on hermes but like people didn't become aware of like this harness thing until
26:53
that happens we're only like nine months into this what where does it like how do you think
26:58
this looks uh even three months from now a year from now and what trend should people be paying
27:05
attention to particularly around user experience and in workflows that they can incorporate using
27:11
this stuff yeah i well what you said is exactly right right like at this time last year opus 4 6
27:18
wasn't out. I think the best way to describe that change, that clear step function increase in AI
27:25
capability was at this time last year, the top 10 developers I knew, 0% of their code was written
27:33
with AI. And it wasn't better than them. Too many issues, too many hallucinations, no ability to do
27:40
long-running tasks. Now, 100% of their code is written with AI, right? Almost nobody. There's
27:47
almost no reason to be writing code except with AI now. And people are doing incredibly long
27:52
running tasks with agents. And that was less than a year ago, like you said, like nine months ago,
27:59
that change happened. And so to think about where we are from the next nine months,
28:05
it's really, really hard to be predictive. I think that
28:08
I think what's very clear is we're early on the adoption of AI.
28:18
And we're going to start to see just those basic levels, like engineers all using AI.
28:23
We're going to start to see that all be integrated into every enterprise.
28:26
This year from this time now, this time next year, I think every enterprise, all code is written by AI.
28:34
There will be really no excuse at that point.
28:38
I think it will be abundantly clear that you have to do this to keep up, to ship, because the companies that choose to do it are going to, it won't happen overnight, but it will happen fairly quickly, will drastically outperform and be able to cut expenses and be able to ship faster than companies who do not choose to use AI.
28:55
I think as far as like novel, novel behavior that emerges, I think we're just going to start to,
29:03
I think we're getting close to starting to see more and more autonomy be dropped into the hands
29:10
of AI, right? Like there's always the examples of my AI shopping for me. That stuff, it really
29:16
isn't happening yet for the most part, right? Like you can kind of manufacture some text,
29:20
but it's by no means a smooth user experience.
29:24
And so that plumbing has to continue to improve
29:28
for you to actually make that transformation
29:33
where it is as simple as text your Hermes agent,
29:35
like, oh, I need this purchased
29:37
and the Hermes agent goes and purchases it.
29:40
The next iteration of that, which is even cooler,
29:42
and this is where we're heading,
29:44
is Hermes agent is always on
29:46
and actually being predictive, right?
29:48
So if I, for example, and this kind of can happen now, but I think it's going to become more and more prevalent is I have to go to SF next week.
29:58
It's on my calendar.
30:00
I forgot to book flights.
30:02
I forgot to book hotel.
30:04
Hermes agent just goes and does all of that for me, and it's handled.
30:08
And I don't even have to worry about it.
30:09
And then I get in a routine where, yep, oh, Marty texted me and said he wants to hang out in Philly.
30:14
Hermes agent's going to go.
30:15
It's going to book my flights.
30:17
It's going to book my hotels.
30:18
It's going to do it all correctly.
30:19
It's going to send Marty the information that he needs to know.
30:22
It's going to send Marty's Hermes agent the information it needs to know.
30:25
Right.
30:25
So I think all of these minor tasks that you have to do in your life can be pushed off
30:33
to agents.
30:34
Right.
30:34
It can be foisted onto agents and clear you up.
30:36
So I like I hate booking hotels.
30:37
I'm terrible at it.
30:38
I book wrong hotels all the time.
30:39
I book wrong flights all the time.
30:41
Hermes agent should do all that for me.
30:42
And I think we're getting really close to it being able to.
30:45
And I think like that three, six month window of where it's in an always on state, it's being predictive and it's being autonomous.
30:52
I think that's that's very, very reasonable.
30:54
I'm a massive flight, hotel, train procrastinator.
31:00
Yeah, same.
31:01
It's horrible.
31:01
I've got to go somewhere in a couple of weeks.
31:04
I should have booked the flight months ago.
31:06
Yeah, I booked it yesterday.
31:07
I was like oh I paying the premium for this because it been on my calendar and it what are you seeing internally So this is one thing selfishly I want to get out Obviously you interact with individuals and enterprises
31:23
leveraging Hermes agent.
31:25
What are some of the most creative or mind-blowing ways
31:28
or workflows that people have incorporated using Hermes?
31:32
Yeah, it's a good question.
31:34
Or another way to frame it is like,
31:36
who's utilizing it the best
31:37
and why is the way they're utilizing it the best way?
31:40
Yeah.
31:41
Well, I think Noose Research utilizes it the best.
31:43
But in all seriousness, I think it's really, really important for us, right,
31:49
if we're going to deliver this product to enterprise,
31:51
that we are the number one power user of it.
31:55
And so we use it for everything.
31:59
I couldn't even try to quantify the productivity increase
32:03
that Hermes Agent has applied to Noose.
32:08
But it's funny when people ask us how many,
32:10
you know, how big is the team?
32:11
And we say, you know, 30 people, 35 people.
32:14
They're often blown away.
32:15
Just because if you've watched the rate
32:17
that we've shipped on the Hermes Agent repo
32:20
and the cloud products and the enterprise offering,
32:22
it's something that would, you know,
32:24
take an order of magnitude more people in the past.
32:29
So we use Hermes Agent for, you know,
32:31
managing our full our full sales pipeline um you know no need for these you know the standard kind
32:39
of you know crm crm sas hermes agent can manage that for our content creation pipeline um we
32:48
haven't i don't know if you've seen our i'm sure you have our our design work and the the videos
32:53
we put out aesthetically please the aesthetic we want things to look pretty um the the stuff we put
32:58
out all of that is generated in hermes agent pipeline it's it's very reproducible it allows
33:04
our designers to have a ton of um a ton of control over the end product and it takes a ton of you
33:12
know manual you know graphic design work and labor off of their plate um one of the most interesting
33:19
tools we have is uh an agent internally that all our employees can use and it basically has a bunch
33:26
of key data sources of ours um so if it you know if a user complaint or if a um a customer complains
33:34
about a billing issue or oh some something isn't working correctly a news portal or my hermes agent
33:40
there's some bug we can just ask we can just ask our internal hermes agent hermes agent debugs it
33:46
can push fixes um and and basically is able to access all of our data backends at once right
33:53
right, whatever it may need. And it becomes over time really, really capable and competent at
33:57
accessing this data and turning it into an actionable item for the user, right? So rather
34:02
than somebody having to manually sift through, you know, infinite logs basically and find issues,
34:07
Hermes agent can find it in seconds. And that's actually a good way of describing what Hermes
34:12
agent is good at. It can do any task that a human could do, but would require infinite patience,
34:19
right like nobody's actually able to sift through millions of pages of logs like you could but
34:24
you would go insane but hermese agent does it instantaneously and homers agent doesn't complain
34:29
either so it's a pretty it's a pretty fun solution yeah that's what i'm wondering like i would love
34:35
again selfishly like give your team access to my hermese agency like where am i up like
34:41
because that's what i'm like continuously like trying to figure out is all right how can i better
34:45
utilize this okay i always feel like i'm not utilizing just in ai generally it's like i'm not
34:51
utilizing everything the best of its ability and it's uh trying to like explore yeah where the
34:57
boundaries and the edges are and i feel like every week i'm an aha i'm like holy crap like i just
35:03
like uh one example this will be the second time you're hearing the story uh in two two consecutive
35:09
episodes if you listen to the geordie viscer episode but i'll tell it again i was at the bar
35:14
we publish this podcast via fountain um and so like our rss or video or audio file video file
35:23
description everything goes up there trans trans uh transcript as well they uh they have a it's not
35:31
publicly available yet i hope they don't get mad at me for disclosing that they have a built
35:36
It's an MCP that I have beta access to.
35:44
And so Bound has video of the transcripts,
35:47
and I've connected the MCP to my Hermes agent.
35:51
And I can literally just voice the text like,
35:53
hey, Tommy and I, the podcast is published now.
35:59
Let's get meta here.
36:00
Let's see if my Hermes agent will pick this up by reading the transcript.
36:04
I want you to use the fountain MCP to clip out
36:08
the section of the conversation with Tommy,
36:12
where we're talking about how to connect Fountain's MCP,
36:16
what I use it for.
36:18
Take the most engaging quote from that,
36:21
and then a one sentence brief description tag Tommy in it and then post it on X And I did something similar to that last night within five minutes the hermese agent had used fountain to find exactly what i described in
36:33
the podcast i published yesterday clip it out send it to me to review and then like send tweet
36:38
it's incredible within five minutes it's incredible like i'm at the bar having a beer
36:43
yep and getting work done yeah yeah for a media company it's like hey that is work yeah yeah well
36:48
Well, no, right. I mean, that's a non-negligible amount of labor that you would have had to do.
36:53
It's kind of a pain and a tedious task. And it's like it's like death by a million paper cuts when you have to do all those things all the time.
37:00
And the amount of time that it frees up by you being able to just dump all of that on it is I mean, I guess I'd ask you, can you quantify the impact on business that, you know, Hermes Agent has had for you all over the last four months?
37:17
uh yeah i mean i think from it's increased our profit margins increased our revenue
37:26
one thing i've hated historically i've told this to people many times like like we're a small shop
37:33
it's me i ran pftc by myself for six seven years so it's only within like last three or four years
37:39
that we've brought on any employees i mean there's five of us and some contractors uh
37:46
I did all for many years I did like the recording the editing the publishing
37:53
but the bane of my existence and as funny is what I did at Barstool Sports
37:57
and I worked there was ad sales and like putting the other decks is something I'm
38:01
I appreciate good aesthetics but I don't know I don't know how to manifest them
38:06
in the digital world very well so like all my decks were like Google slide like
38:11
flap yeah yeah yeah trust me it's a good show but uh went and found like a beautiful slide deck
38:19
skill connected it to the YouTube API connected it to a fountain API connected it to a bunch of
38:27
other things that are pertinent to putting together a media deck and basically say hey
38:31
for example like if I was gonna try to sell you on an ad deal like hey I would like news research
38:39
to to be uh an ad partner with us uh think about ways in which we like we would be a good brand fit
38:46
for them base it off of our our demo data like our our stats on youtube and whatever and like come
38:52
up with a deck five minutes later uh it's got a html page on versell beautiful deck that can
38:58
actually sell like on top of that it's like let's make sure we're pricing this right like run market
39:04
research about podcasts of our size like what is cpm is there any premium on our audience specifically
39:10
that could do that research and come up with like accurate pricing and that has led to more closed
39:17
ad deals for us so yes it has long way away saying yes it has yeah but i like what you said too about
39:23
um you know you have a hard time personally manifesting aesthetically pleasing stuff in
39:29
the digital age but hermys agent enables you to do that yeah right and that's what's so cool about
39:33
it is if you're a high agency person right now, there's no better time in the history of the world
39:40
to exist. I can't write code anymore, but I manage our compute clusters and I can get all of the
39:48
visualizations, all the telemetry that I could possibly want from it just by interfacing in
39:53
natural language and build dashboards that help inform our business decisions and allow us to see
39:59
how we're acquiring new users and how our compute clusters health is and that you know that would
40:05
take data scientists front-end engineers it would it would take a you know a series of people to do
40:10
that in the past and now i'm just prompting hermy's agent and it's building out what a full team would
40:15
have to build historically so it empowers you to do pretty much anything in any dimension that you
40:20
want right in any domain that you want i mean it's it's yeah it's unbelievable freaks look at
40:26
at me. I'm glowing. I've got like an angel's halo going around me. You know why that is? I feel good.
40:31
I feel taken care of. I feel blessed, healthy, happy. And that is because I'm a CrowdHealth member.
40:36
My family and I have been CrowdHealth members for five years now, literally this month. Five years
40:40
ago, we joined CrowdHealth. We've had two babies. We've had multiple health events and we're never
40:45
going back to health insurance. CrowdHealth is crowdfunded healthcare. So you sign up for Crowd
40:50
Health, you pay a monthly fee, you help out with other people's bills. And it's significantly
40:55
cheaper than health insurance. We were on Cobra as a family of three. When I left my last job
41:00
before I went full-time to CFTC, went on the crowd health. Now as a family of five, we pay,
41:06
I believe, $700 a month. It's significantly cheaper. They're going to negotiate prices
41:11
lower for you. They've consistently negotiated healthcare prices as much as 50, 60, 80% in many
41:17
cases. They help out with babies. If you have a pregnancy, you pay the first $3,000 and the
41:22
crowd covers the rest. If you have a regular health event, you pay $500 and the crowd pays
41:28
the rest. Go to joincrowdhealth.com. Sign up today. Use the code TFTC. Opt out of health
41:33
insurance. I'm an insured baby and I love it. Use the code TFTC at joincrowdhealth.com.
41:37
You'll get $99 a month for the first three months.
41:39
that you're on the CrowdHealth platform in the community.
41:43
Bitcoiners, you found sovereign money.
41:45
Now find sovereign health and sovereign healthcare.
41:48
All right, freaks.
41:49
You know I don't take sponsor money from products
41:50
I wouldn't use myself, so listen up.
41:52
The Avon Bitcoin Visa card is one of the most interesting things
41:54
I've seen in the Bitcoin lending space in a long time.
41:57
Here's the deal.
41:57
You can get a line of credit up to a million dollars
41:59
backed by your Bitcoin without selling a single SAT.
42:02
No games, no annual fees, no minimum draws.
42:05
And your Bitcoin is custodied by Bitco,
42:07
one of the most trusted names in digital asset security.
42:09
AVEN never lends it out. There's no rehypothecation. You stay in control.
42:13
You can lock in a fixed rate for up to 10 years. 10 years, that's 10 times longer than most lenders
42:19
out there. Or go interest only for up to five years. Rates start at 8.99% APR. For a product
42:25
that lets you keep your stack and still access liquidity, it's hard to beat. On top of this,
42:30
that's what? You also get 2% unlimited cash back every time you use the card.
42:33
Spend fiat, keep your Bitcoin. That's the whole game. If you've been stacking for years and you
42:38
liquidity without triggering a taxable event this is worth a serious look go to aven.com
42:44
bitcoin that's aven.com bitcoin check it out i mean let's shift this is the compute side what
42:51
are you what are you seeing on the company you're talking about your cluster what uh on the physical
42:57
side of things like you actually need the gpus you need to download like what what are you seeing
43:02
there like what is it like managing a cluster maybe we'll start there all sorts of headaches
43:06
managing clusters and compute for sure.
43:11
What I think is really interesting about
43:15
the way compute has shifted, it's shifted with use,
43:17
which is, I guess, obvious, right?
43:19
But I think it's worth noting, again, we said this time last year,
43:22
Opus 4.6 wasn't out.
43:24
Developers weren't using it to write code.
43:26
Good developers weren't using it to write code.
43:28
I was using it to write code, but good developers
43:30
were not using it to write code.
43:33
At this time last year, if you needed a cluster of whatever, 128 GPUs, I could send a text out and have 10 different providers bidding to win that.
43:48
Now it's the inverse.
43:50
I basically would have to go and text 10, 20 different people and say, I need a cluster.
43:55
I need a cluster.
43:55
I need a cluster.
43:56
Nope.
43:56
We don't have any.
43:57
We don't have any.
43:58
We don't have any.
43:58
It's unbelievable how squeezed the HPC industry is right now.
44:06
I think we're up against a lot of land and power limitations.
44:13
I think you're seeing the regulatory environment has just really squashed the ability to continue to grow that.
44:20
But also just the usage was such a parabolic increase in inference needs that we weren't prepared for it, right?
44:30
Yeah. And so to that point, who is getting access, obviously, outside the frontier model?
44:39
Because you have the training, obviously, you have the inference, and obviously, Anthropic, OpenAI, Meta, Google, SpaceX, they all have incredible access to this stuff.
44:49
But I think we've discussed like open router.
44:52
I think Stripe's acquisition of open router is a very strong signal.
44:55
And I think this neocloud sector is going to, I think, basically having open market for different models is a very smart idea.
45:05
Especially if we believe that being model agnostic is going to be a trend that strengthens in the future.
45:12
But to that point, like how do you get priority?
45:15
like how do you both on okay let's back up infrastructure side what makes you a good
45:21
cluster what do you need to do right for your end customer to make sure people come back to you and
45:27
buy compute for you because there's a bunch of stuff with like networking and latency and all
45:30
that then two how um for the for the offtaker like how do you how are these people the neoclouds
45:39
that own the gpus and will inevitably host them how do they decide whether or not you're a good
45:44
you're a good off taker that deserves access to you yeah does it even matter it's like you're
45:48
paying you yeah um so on the first piece yes it's managing managing these nodes is difficult um
45:59
there's a lot of entropy in the management of these nodes that i don't really fully understand
46:04
um but there's a lot of down time um with if you're talking about managed inference specifically
46:10
there's throughput and latency, throughput latency uptime. These providers do want to work to
46:21
have a good experience for the end user because it can be a huge headache and people are paying a
46:25
lot of money for these nodes. If their throughput isn't met, if their latency requirements isn't met,
46:29
it really ruins the end user experience really, really quickly and people are going to stop
46:33
people are going to stop paying for paying for machines from a specific provider.
46:38
Right now, the market I think is so squeezed that there's way more wiggle room and tolerance
46:44
for that stuff just because it like where else are you going to go And to your second question I know that that plays a role Like who is the end user Who is the buyer
46:58
Who is the off taker?
46:58
Plays a role in the decision making process of these companies.
47:02
I've heard people talk about how, like, the sales cycle for these AEs is basically flipped.
47:07
Before it's like, I need to go find people to sell this compute to.
47:10
Now it's like, I have all these people to go sell this compute to.
47:13
I have to build a case of why my people to internal teams, why my people are worthy of of getting the right to this compute contract.
47:21
So the thing is completely inverted. It's it's it's it's fascinating. Right.
47:24
And I think it's always company. It's internal politics and kind of company dependent on that stuff. Right.
47:31
There's different partnerships between between companies and favors had, of course.
47:36
I think a lot of these companies do try to care about open source and try to support different parts of the ecosystem.
47:47
I think people recognize that if all of this compute ends up very, very consolidated, that's a dangerous situation for a multitude of reasons.
47:59
um so i think yeah they there's you know economics play a big part but there's also a bunch of um
48:07
kind of strategic initiatives that drive where where this compute ends up and i'm happy we're
48:12
discussing this now because i meant to bring this up earlier i think it's important to clarify for
48:16
people we were talking about open source models many of which are predominantly trained uh in
48:21
china and then released uh people hear that they're like oh i'm sending all my data to china now
48:27
that's not always the case like if you're connecting directly to a deep seek servers
48:32
like yes you are but i think the understanding around like no like it's open source so you can
48:37
have american companies go buy gpus and download the model there and you're you're not serving it to
48:43
the chinese servers it's a chinese model run on american servers correct massively different right
48:48
and that's been a huge education piece that we're trying to drive home is you can have all of these
48:55
cost savings you can get really good performance at you know a fraction of the cost you know 60 70
49:00
80 90 cheaper than u.s closed frontier labs um and you can do it with yes it's a chinese model
49:07
but it's running on infrastructure owned by a u.s company in a u.s data center on u.s soil right and
49:13
those are where your data flowing is two totally different things yeah well you can say that people
49:17
like it's phoning home to china though right it's like yeah no no it's not it's definitely not and
49:22
And then, again, tying in privacy to this, too.
49:26
I know we were talking about Maple AI.
49:30
Venice has obviously exploded in popularity.
49:33
PPQ is another one.
49:39
I am very bullish on running these open source models in trusted execution environments.
49:44
Just to increase the privacy layer.
49:45
Because that, shout out to Mark Goodwin, who's been on the tip on Twitter, saying,
49:51
Like even if you're using open source, it doesn't mean to centralize.
49:53
There's still somebody hosting those servers at the end of the day.
49:56
But with like trusted execution environments, being able to run
50:00
the inference in a secure enclave and have everything end to end encrypted
50:04
from your device that's making the prompts to the
50:08
input in transit is encrypted, then you do everything on the secure enclave.
50:14
You re-crypt it, send it back.
50:15
Like I think that's a massive opportunity to I'm just rambling now.
50:19
yeah and i think i think that and i think i think local models local models as well yeah and local
50:25
models are continuing to get better hardware consumer hardware is continuing to get better
50:29
and better right and so the ability for you to run a model on your local machine and do again
50:36
it's not going to be a hundred percent of your work but a good chunk of your work especially if
50:39
you have sensitive workloads like you can run those on local models and and get get real work
50:45
done well this is actually something i'm happy you brought this up this is something i've been
50:48
like thinking like i'm not smarter i'm maybe i'm not i won't you're smart i won't denigrate myself
50:55
i'm smart but i do not have the skills to post train an open source model my own data
51:01
as like something of like her if noose could figure out a way to take what i built with my
51:06
hermes agent and be like okay i want to use this model um this local this slm like locally uh hermese
51:14
agent have all the context of what we've been working on over the last six months like can you
51:17
help me post train this model on our specific data like that i think at least for my purposes
51:23
that's one thing i've been uh somewhat obsessed with for the last week or two is like okay i've
51:28
gotten to this point like what else can i do yeah how can you squeeze more performance out
51:31
of this train a model and all of our stuff like yeah how do i do that yeah um that's what i'm
51:36
looking for anybody wants to help me let me know yeah we can do that we can get you that so that
51:40
actually was part of the reason that hermes agent was created it was it wasn't created to be this
51:46
outward facing product it was created really out of like a completely utilitarian birth our our um
51:54
co technium created this hermes agent to help as a small ai research lab right you always trying to find like these these moonshot ideas to to push you to the frontier And Technium created this self Hermes agent to help us you know become frontier model trainers
52:15
So the original birth of Hermes agent was really to do, was in part to do that.
52:20
And that's something we should do a job of highlighting, how you can use Hermes agent in that stuff as well.
52:26
yeah so what uh what's the plan with news like you have the agent you have this portal we got
52:33
this enterprise what uh what are you guys striving for um in terms of what this company looks like
52:40
um and i know you said this on the phone like who knows what any company looks like
52:44
like what are your what are your north stars and who are you building for at the end yeah yeah i
52:50
think who are we building for is everybody. And I think I answer that really intentionally that way
52:57
because I think we've built this product in a way that it meets you wherever you are.
53:07
The open source repo, if you want to run a local model on your local network and never touch any
53:15
noose research managed product, that's great. Like the Hermes agent repo will always be able
53:20
to empower you to do that and we'll continue to improve it and maintain it. Um, if you want
53:25
certain things managed, like, great, we'll do all that as well. We will be the best place. We are
53:30
the best place to have your Hermes agent, you know, hosted and managed. If you want, you know,
53:35
different pieces managed, but you want to own certain pieces or do some work locally. Great.
53:40
We'll, we'll meet you wherever we are. So we really are building this, um, building this for
53:45
everybody. I think the two sort of key outcomes over the next handful of months for us, which
53:54
lead to where we want to be in the end, is one, decreasing the amount of time that you go from
54:00
Googling Hermes agent or having your friends send you a Hermes agent to realizing, wow,
54:07
this is a magical, magical experience, right? How can we get that time? How can we make it a
54:14
frictionless onboarding. How can we meet people on whatever device they are? How can we enable
54:20
people to send others a Hermes agent? Oh, Marty hasn't used Hermes agent. Let me send it to you.
54:25
You can click a link, open it up, and ask it one question and be like, wow, and not have to do
54:31
setup, not have to configure anything. So that's really, really important, right? Is how can we
54:35
make sure it is frictionless as possible and make it so that you as quickly as possible realize
54:39
that you have a tool that you've never experienced before in your hands.
54:44
And then the second piece is delivering Hermes agent to enterprise in a way that they are able to control their AI stack.
54:56
It is their sovereign AI stack with all the bells and whistles that enable employees to be much higher agency, much more output, and do it in a way that isn't prohibitively expensive, that leverages the plethora of models and tools out there that can power Hermes Agent in a meaningful way, delivering that product to enterprise at scale.
55:24
so yeah those those two things can we trim down that timing can we deliver to add a prize at scale
55:29
the uh meet people on their device that's been one of the the interesting observations i've had
55:35
over the course of the years uh who attracts to what specific form factor like i was telling you
55:41
like i haven't used cloud code or codex desktop in seven months people like are you doing that like
55:49
are you are you trying to get left behind i'm like no i literally it's connected to all these
55:53
these different models. And my go to form factors is like voice to text in telegram, I have a group
55:59
chat with the agent, just the agent and me, but the group chats allowed to create topics. Yep. So we
56:04
have topics that represent different sessions that are ad ops, Bitcoin brief, or the common or
56:11
newsletter builder. If we have a business meeting, okay, like biz ops, what do we need to get done,
56:18
blah blah blah uh and there's something about that form factor is just vibes with me like just
56:24
being able to like on the go on a walk around the neighborhood and you saw me literally when you
56:29
walked in i had my newsletter session open like talking about how i want to frame today's newsletter
56:34
and just that that form factor being able to walk on the go and doing my phone clicks with me
56:40
for others it's like and i think a lot of people are talking about vendor lock and people are locked
56:45
into uh to cloud code and codex specifically those desktop apps and they think like this is the only
56:52
way that you can leverage this i'm like no you you gotta you gotta get yourself off self off that tit
56:57
as quickly as possible because you're gonna get vendor locked in and that's not what you want
57:01
um but it is like erin but something like ed on our team he loves the uh he loves the hermes desktop
57:08
app that how he interacts with it and i i never frankly i never use it yeah just use it through telegram um and i think there is it becoming clear to me that there going to be different preferences and different form factors for how you interact with this and thinking about
57:24
a company like news research trying to design for all those different ways which people are going to
57:29
prefer to interact with it is it's a crazy problem to think about yeah it's a it's a it's a massive
57:34
problem right and um and you're exactly right there's not there isn't some you know golden
57:40
solution where, oh, everybody is going to use this as the way to interface, right? At least
57:45
certainly not yet. And I think the stance that we generally take is don't be like, we're definitely
57:51
opinionated on some of the best ways to use these things, but we don't want to ever lock in unless
57:55
it's blatantly obvious that that is the place that we need to lock in, right? So we allow WhatsApp,
58:02
we allow Telegram, we allow email, like wherever you want this thing to live and meet you is
58:08
is important because people use it people use it in different ways there's a lot of people that
58:12
would never want to talk want to talk as much as you right like they would never want to do that
58:16
everybody is so different and everybody finds different ways to get the performance out of
58:20
these agents that they want that you you need to build for a huge range of different um platforms
58:27
yeah yeah it's funny when we like as you're mentioning that like email the hermes agent has
58:33
an agent.to mail email address um instead like for my wife whenever she's like ah we're in the
58:42
market for a new car and she's like oh she's like doing it the old way i'm like hey just email martin
58:47
i gave him permission to email you back like tell him what you want he'll go do like a search for
58:51
you so like my wife has emailed martin the more sophisticated marty is uh is doing all this work
58:57
but it's like it's been fun watching my wife just i mean just she'll be like oh i emailed martin
59:02
today and he like did some good work for me like that's pretty incredible does she like martin more
59:05
than you no no of course i hope not not yet at least maybe at some point maybe when robotics get
59:14
here shoot i'm kidding i'm kidding the uh but um no it's just like the yeah like that's one form
59:23
factor it's like hey honey instead of like you take my phone and getting telegram just email
59:27
martin yeah and he'll get you a list of cars in the area that are within the the range of what
59:33
you're looking for yeah she's been going back and forth with them in emails right that's yeah and
59:38
it's probably a magical experience for her right yeah yeah yeah and awful and awful it's all of
59:43
those monotonous tasks like going and searching and getting quotes from six different from six
59:47
different vendors and finding the car you want and the reviews all of that is it goes from these
59:52
incredibly laborious processes to you know a few minutes texting your hermes agent no no we have uh
59:59
ed on our team he has his agent uh and our agents are emailing each other yeah like so we have stuff
1:00:06
for the business like ed in his telegram or his hermes desktop app with his agent like all right
1:00:12
email martin um and tell him that we did this thing in this part of the business and make sure
1:00:18
he updates his second brain to yeah to recognize that it's uh and i feel i still feel like i'm not
1:00:24
utilizing it like i'm like we do that and something like a workflow like that will uh emerge i'm like
1:00:29
is this the right way to do something like this like is there there are more efficient ways that
1:00:33
point being is there's so much to uh to explore in terms of all these different workflows in these
1:00:40
cases particularly teams like thinking about how teams interact with agents and i think dorsey's
1:00:46
interview with sequoia earlier this year really uh really cemented my thinking about like how to view
1:00:55
a company in the world of agentic ai the company is like essentially like a knowledge layer and
1:01:00
your job is to give it the best most robust knowledge that every employee can access and so
1:01:08
um i haven't let everybody into the group chat on telegram however we do have like discord and the
1:01:14
the agents from wild and discord and they can talk there and interact with them so that that
1:01:20
use case of creating a second second brain for a company specifically and then getting access the
1:01:25
right access to individuals on the team is a massive problem uh not a problem massive opportunity
1:01:32
uh because when you think about like permissions especially if you're like selling enterprises like
1:01:37
okay i want becky from finance to have access to this data but not this data yeah um and like just
1:01:42
building the guardrails for that. Yep, totally. Massive opportunity. Yeah, it's a massive
1:01:46
opportunity. I think there's a bunch there too, as far as how can you use Hermes agent to get more
1:01:53
out of every employee, right? And I think what you see is certain employees, it's a skill to be able
1:01:58
to prompt these agents, right? Like different people get different performance out of agents.
1:02:03
It's like very much a feel thing. And I think you have people that are incredibly, incredibly
1:02:10
like your top performers,
1:02:12
Hermes Agent lets you bring other people up to their level,
1:02:16
and that's what we're rolling out for enterprise
1:02:19
is this concept of a collective wisdom,
1:02:23
a collective brain where you can take
1:02:25
the performant employees' top skills
1:02:27
and deliver those to relevant employees.
1:02:29
and they immediately get boosts in their performance.
1:02:32
Their agents are immediately better.
1:02:34
They're shaped in the way that you want them shaped.
1:02:36
If you, however you want your top employees perform,
1:02:39
if you want other people like that,
1:02:40
like it will shape other people's agents in that way too.
1:02:43
And when you have new hires,
1:02:45
instead of having like this cold start on-ramp experience,
1:02:48
you can equip them with a skillset of tools
1:02:51
that are formulated, you know,
1:02:52
in the way that TFTC wants things written, right?
1:02:55
In, you know, all of that stuff.
1:02:57
and you can have your agents molded kind of to your voice
1:03:03
and the way that you want things done, your guidelines.
1:03:06
And you can give that to new employees, existing employees, et cetera,
1:03:10
to kind of help shape people in your direction
1:03:11
and also give them that productivity boost.
1:03:14
Yeah.
1:03:16
What do you think this does for the job, Mark?
1:03:18
Are you, the permanent underclasses is quickly approaching
1:03:22
or a world of a billion flowers are going to bloom?
1:03:25
um i think we could have a whole nother podcast on that i think i think i i think i i think i land
1:03:33
somewhere somewhere in between right i think that what this does is you know our
1:03:38
it enables you to not do all of these monotonous tasks right if you've i think my big oh my god
1:03:47
i'm gonna lose my laptop job yeah my my bigger concern is oh you like you're worried about not
1:03:52
just like, you know, making Excel sheets for the rest of your life or whatever. Like that's your
1:03:56
concern. I'm really worried about the paycheck. Yeah. Like I think what this, what this should do
1:04:01
is free people up to do, you know, more thoughtful and higher agency tasks than sitting there and
1:04:06
like button pushing. Um, so yeah, I mean, is it going to be this straight shot to a million
1:04:12
flowers? No, like there's of course going to be bumps in the job market is, you know, these
1:04:17
disruptive technologies move through their natural growth
1:04:21
and integration and flow.
1:04:25
But I think the optimistic view is that this will enable
1:04:29
humans to do things that are way more fulfilling
1:04:32
than what a lot of people have convinced themselves is foundational to their lives Yeah that sort of philosophical musing i had over the last year uh because you mentioned earlier like if you high agency you
1:04:46
can get a lot out of this and that's the the question i have is there just like a natural
1:04:51
distribution of individuals who are inherently high agency and a very large percentage of which
1:04:59
or not or is there an ability to engender high agency in people um and do ai tools inspire them
1:05:08
to become a high agency i don't know i think that's what we're going to find out in the next
1:05:12
decade we'll find that out i'm sure it's some mixture of both yeah um
1:05:20
what's your most contrarian take and what's going on right now
1:05:23
my most contrarian take of what's going on right now um
1:05:27
I don't know if it's contrarian.
1:05:31
I'm sure it's, there's a lot of people, I think, that probably share this viewpoint.
1:05:36
I never know how pigeonholed I am, or sorry, how echo chambered I am.
1:05:41
But I think that if you want to talk about existential risk and the way that existential risk of AI,
1:05:53
like AI, misaligned AI, killing all humans or whatever,
1:05:57
I actually like sometimes believe in there's some I believe there's some degree of credibility to that argument as a as a possibility.
1:06:05
But what I think is I think where we get to crazy town really fast and it's really, really disingenuous is you see large companies and lobbyists pushing for.
1:06:17
a stop of open source, right?
1:06:21
Like open source has become this sort of,
1:06:24
has become this evil and this scapegoat.
1:06:29
Whereas if you ever actually believed
1:06:31
in existential risk as a real threat,
1:06:34
open source would be the last thing
1:06:35
you would be concerned about.
1:06:36
You'd be concerned about
1:06:37
this massive consolidation of compute.
1:06:39
You'd be concerned about, you know,
1:06:41
the frontier constantly being pushed
1:06:43
by the closed labs right like distillation to create open source models like definitionally can exceed AGI unless the top labs already have done it themselves
1:06:55
So I think the. Yeah, the way that that argument has been has been pushed and is certainly kind of the controlling argument in the vast majority of the population,
1:07:07
likely not those that are very in tune with AI, but the way that that argument has been framed to the vast majority of the population, I think is just incredibly.
1:07:15
incredibly disingenuous incredibly and it's almost disgusting at some point because it's like if
1:07:24
i don't think most of the public is aware because they don't pay close enough attention it's like
1:07:29
very clear to me it's like okay you're looking for the you've raised a ton of capital you're
1:07:33
deploying a ton of capital into this infrastructure built out particularly for the frontier labs uh
1:07:39
and you have a need to get a return on invested capital.
1:07:44
And you're seeing these open source models begin to nibble at your feet.
1:07:47
Not only that, like many people are turning to them for 90% of the work,
1:07:51
but they're individual use cases.
1:07:54
And you're trying to pull up the ladder and shut the door behind you
1:07:57
just so you can save those profits.
1:08:00
And I think when you consider how profound of a shift this technology is for humanity,
1:08:06
the fact that there's literally two or three companies trying to pull up the ladder and control
1:08:10
yeah it's just incredibly uh i would say evil at the end of the day well where did all the data
1:08:17
come in the first place right like it's all of our data that i mean the fact that they were buying
1:08:21
like centuries-old books and like cutting off the bindings and yeah and then burning them it's like
1:08:26
what the hell yeah it was all our data that it came from in the first place right and then they
1:08:30
yeah it's it's just disingenuous dealing with it with the new york times right now i believe yeah
1:08:34
i think that that uh i guess copywriter that copyright case is moving forward i mean this
1:08:42
has been a pleasure um i'm sure there's more i am gonna pick your brain off uh off air about what i
1:08:48
should be doing with my agent i love it any any final thoughts final final words of wisdom before
1:08:53
we wrap up here i don know about words of wisdom i think um i think at some point it will be very obvious that you should care about privacy in the age of AI and agentic AI I think that
1:09:14
the more proactive you can be about it and the steps that you can take to empower a world where
1:09:22
you have more ownership as a business and as an individual of your AI and where the data that you
1:09:28
give to your AI flows, the better served you'll be and you can avoid a lot of headache down the
1:09:35
road. And so I think looking for obviously Hermes agent, but looking for tools that enable you in
1:09:41
the AI realm to have good controls over your data, but do it in a simple way that isn't a super
1:09:47
burdensome and a huge headache for you, I think it'll be a really, really valuable investment
1:09:51
going forward i completely agree with that um thank you for joining me thank you uh in the team
1:09:58
at news for for building hermes has been incredibly uh powerful for us here at tftc and uh excited to
1:10:04
see what you guys build next love it thanks marty thank you peace and love freaks thank you for
1:10:10
listening to this episode of tftc if you've made it this far i imagine you got some value out of
1:10:15
the episode if so please share it far and wide with your friends and family we're looking to get
1:10:20
the word out there. Also, wherever you're listening, whether that's YouTube, Apple, Spotify,
1:10:27
make sure you like and subscribe to the show. And if you can leave a rating on the podcasting
1:10:33
platforms, that goes a long way. Last but not least, if you want to get these episodes
1:10:37
a day early and ad free, make sure you download the Fountain podcasting app and go to fountain.fm
1:10:44
to find that $5 a month,
1:10:47
get you every episode a day early,
1:10:49
ad-free, helps the show,
1:10:51
gives you incredible value.
1:10:53
So please consider subscribing
1:10:55
via Fountain as well.
1:10:57
Thank you for your time.
1:10:59
And until next time.
1:11:00
Okay.