Latest / Crypto Coin Show / Large Action Models: The Next AI Revolution Beyond ChatGPT | Action Model Interview
Transcript
- 0:00Speaker 1: I'm Ashton Addison from the Crypto Coin Show and today
- 0:03Speaker 1: on Blockchain Interviews with Sina Yamani, founder and CEO of
- 0:06Speaker 1: Action Model, here to discuss AI, large action models, why
- 0:12Speaker 1: it's important to have community owned AI, and where AI
- 0:17Speaker 1: is headed if we don't have decentralized versions of this scene.
- 0:22Speaker 1: Thank you for taking the time.
- 0:25Speaker 2: Pleasures with mine as then thanks for having us.
- 0:28Speaker 1: Yeah, excited to dive into this because it's a super
- 0:30Speaker 1: important topic. Not only does almost everybody I know use AI,
- 0:35Speaker 1: they might be wiped out by it as centralized giant
- 0:39Speaker 1: companies start using it to take over people's jobs.
- 0:44Speaker 3: And we're already seeing it happening a little bit.
- 0:45Speaker 1: But I feel like the acceleration as we get into
- 0:48Speaker 1: automation and actions that the AI can take. It's no
- 0:52Speaker 1: longer ask a question, copy paste, I do it manually.
- 0:56Speaker 1: It's like the AI does it all and I've been
- 0:59Speaker 1: taken out of the equation. So I know that you've
- 1:02Speaker 1: been deep in this for many years and in Crypto
- 1:06Speaker 1: I would love to start out with what you and
- 1:08Speaker 1: your team have built at Action Model. What is the
- 1:11Speaker 1: difference between LLM and large action models and why it's
- 1:15Speaker 1: so important to have community based AI.
- 1:18Speaker 2: Yeah, absolutely so. So if you think about AI and
- 1:23Speaker 2: where it's been over the last few years, we've had
- 1:25Speaker 2: LM's large language models. These are essentially the chat lots, right,
- 1:30Speaker 2: and the way they work is you you speak to
- 1:33Speaker 2: it and then it gives back an answer. So we've
- 1:36Speaker 2: been in this advice era for the last few years. Now.
- 1:40Speaker 2: You tell it, I want to do this, rewrite this
- 1:42Speaker 2: for me, do that, and it will. It will write
- 1:44Speaker 2: that structure out for you and then you have to
- 1:46Speaker 2: go and do the work right. So that's kind of
- 1:49Speaker 2: the chatbot style. We've obviously got images and videos now
- 1:53Speaker 2: as well, which are all frightening me good. And the
- 1:58Speaker 2: next stage is that the actually does things for you, right.
- 2:02Speaker 2: It's not the fact that you can ask it to
- 2:06Speaker 2: draft you a contract or an email. It will actually
- 2:09Speaker 2: go and do that work too. You know, it will
- 2:11Speaker 2: read through your emails, make decisions because it understands your
- 2:14Speaker 2: business and it will respond. It will then follow up
- 2:18Speaker 2: do whatever work was required, draft the actual contracts, literally everything.
- 2:23Speaker 2: And so this next wave of technology is very disruptive
- 2:28Speaker 2: because it's pushing past the I'm a cool tool for
- 2:33Speaker 2: the person to use to speed up their work too.
- 2:37Speaker 2: Does the employer actually need the person at all anymore?
- 2:40Speaker 2: You know, there's the does the company need all of
- 2:42Speaker 2: this staff when you have these incredible AI employees that
- 2:47Speaker 2: can do everything.
- 2:50Speaker 3: Yeah, I know.
- 2:51Speaker 1: It's it's getting crazy, and we're starting to see hints
- 2:53Speaker 1: of it in these large language language models. I'm sure
- 2:56Speaker 1: they're going to adapt with you know, some automations I've
- 3:02Speaker 1: seen instead of CHATGBT, I think Claude has things that
- 3:06Speaker 1: they can now sort of start taking actions for you.
- 3:08Speaker 3: We've seen the Claude bought come up recently.
- 3:11Speaker 1: But maybe we can dive into, you know, why it's
- 3:15Speaker 1: not such a good idea for chatgbt's automation to be
- 3:20Speaker 1: taking over and to look at the decentralized or community
- 3:25Speaker 1: owned perspective of this and what happens if the same
- 3:29Speaker 1: sort of fang companies of AI just control the whole
- 3:32Speaker 1: AI space.
- 3:35Speaker 2: Well, I mean it's very obvious why it's a bad
- 3:39Speaker 2: thing that can happen. I don't need to really explain that.
- 3:41Speaker 2: The situation is that there's about a billion people around
- 3:45Speaker 2: the world that are employed to work behind a computer, sprint,
- 3:49Speaker 2: some kind of white collar job, and these aren't necessarily
- 3:53Speaker 2: very technical people. In fact, a lot of it is
- 3:56Speaker 2: repetitive tasks data entry that can easily be automated, and
- 4:02Speaker 2: AI can do all of this stuff today. Right. The
- 4:06Speaker 2: next stage in the next few years is enabling the
- 4:09Speaker 2: AI to be able to understand what is the business
- 4:13Speaker 2: and their needs, and then it will go and do
- 4:15Speaker 2: all of that work. So now, if you're the employer, right,
- 4:21Speaker 2: the majority of the world and employers, they all want
- 4:23Speaker 2: the same thing. They're just trying to make more money
- 4:26Speaker 2: and be more efficient. And when these tools pop up
- 4:31Speaker 2: and the tool can do ninety five percent of the
- 4:34Speaker 2: work for ten percent of the price, it's a no
- 4:39Speaker 2: brainer for these businesses. So unfortunately, unless the government acts
- 4:43Speaker 2: right now and puts deep and strengthy regulations in place,
- 4:51Speaker 2: then this is inevitable. It's going to happen. And as
- 4:55Speaker 2: it stands, almost all of that value transfer is going
- 4:59Speaker 2: to go to these tech companies. We're talking about a
- 5:01Speaker 2: handful of trillion dollar companies that will grab the value
- 5:08Speaker 2: of nearly a billion people over the next five to
- 5:11Speaker 2: ten years. You know, even this SEO of Anthropic said,
- 5:15Speaker 2: I think it was a few weeks ago, he said
- 5:17Speaker 2: that half of white collar jobs are we gone within
- 5:20Speaker 2: five years. And I personally, I think that's an understatement.
- 5:23Speaker 2: I think if he said what he really believes, it
- 5:26Speaker 2: would be a much more frightening statement. So, you know,
- 5:32Speaker 2: we know what happens when individuals, we're talking about the
- 5:36Speaker 2: wealthiest people in the world, these billionaires or these trillion
- 5:40Speaker 2: dollar corporations own and control so much of the world.
- 5:44Speaker 2: And this is you know, this is like twenty percent
- 5:48Speaker 2: of the planet we're talking about here. So it's clear
- 5:53Speaker 2: that this needs to be either regulated in some way
- 5:56Speaker 2: or it needs to be in part owned by the people.
- 6:00Speaker 2: And that is what the action models focused is around.
- 6:03Speaker 1: Yeah, no, I agree, And obviously being in blockchain for
- 6:09Speaker 1: ten plus years and yourself as well, that's sort of
- 6:12Speaker 1: been the goal is move our finances owned by the people,
- 6:16Speaker 1: for the people, for yourself. When we translate that to
- 6:20Speaker 1: AI in what the team's built at action model. So
- 6:23Speaker 1: from what I understand, these llms are also implementing action
- 6:28Speaker 1: as well. So what's the difference here in how action
- 6:31Speaker 1: model can stop you from losing your job or you
- 6:36Speaker 1: stop this transition from at least take a piece of
- 6:39Speaker 1: the pie from these corporations laying off a billion people.
- 6:45Speaker 2: A good question with a pretty technical answer. And essentially,
- 6:49Speaker 2: these lms they output text, that's the main thing they
- 6:54Speaker 2: can do. And the way they work at the moment
- 6:57Speaker 2: with regards to kind of connectty with applications is all
- 7:02Speaker 2: through APIs And fundamentally, around one percent of the Internet
- 7:07Speaker 2: has APIs available. That means that ninety nine percent of
- 7:10Speaker 2: the Internet is only accessible to people who use the
- 7:13Speaker 2: mouse and the keyboard, and so that's these lms. The
- 7:19Speaker 2: next generation of AI actually uses the mouse and the keyboard,
- 7:23Speaker 2: so it can use the computer just like a person does,
- 7:27Speaker 2: and that means that it doesn't need to integrate with anything,
- 7:31Speaker 2: doesn't need access because the platforms, applications, websites, they have
- 7:36Speaker 2: no idea that this isn't a person because it acts
- 7:40Speaker 2: and navigates through the Internet just as a person does.
- 7:44Speaker 2: So they're becoming just as capable as people are.
- 7:50Speaker 3: Yeah, that's a great point about the APIs.
- 7:53Speaker 1: If the AI is in the operating system and it's
- 7:55Speaker 1: moving the mouse, it doesn't need to work on APIs
- 7:59Speaker 1: and that might have to shift the business model of
- 8:01Speaker 1: some of these centralized AI platforms. So can you explain
- 8:07Speaker 1: the community based part of action model? Can I go
- 8:12Speaker 1: and on action model, and you know, it's very similar
- 8:16Speaker 1: in appearance to an LM.
- 8:17Speaker 3: But I own the data. How does it exactly work?
- 8:22Speaker 2: So the Action model is a community initiative. It's basically
- 8:27Speaker 2: around So this change is coming, this disruption is happening,
- 8:32Speaker 2: and individually there's nothing we can do about it. But collectively,
- 8:35Speaker 2: in the millions, we can definitely do something about it.
- 8:39Speaker 2: We can fight back, we can resist, we can own
- 8:42Speaker 2: a stick. And the whole part of the community is
- 8:46Speaker 2: that we own a piece of it together. So it's
- 8:48Speaker 2: fractional ownership distributed across millions of people who have come together,
- 8:54Speaker 2: all facing the same direction against a common enemy. I
- 8:57Speaker 2: guess right. And and it's all about contribution to that ecosystem.
- 9:02Speaker 2: So there's different ways that you can contribute, and you know,
- 9:06Speaker 2: the easiest way is by downloading the Browser extension. You
- 9:10Speaker 2: don't have to do anything in the background. You essentially
- 9:13Speaker 2: turn it on and you're helping to train the Action model,
- 9:17Speaker 2: and that training improves the AI that knows how to
- 9:21Speaker 2: use every single website. And you know, when you have
- 9:24Speaker 2: enough people using that, you end up learning how to
- 9:26Speaker 2: use every single website.
- 9:29Speaker 1: Very interesting, and so the Chrome extension, you know, I
- 9:33Speaker 1: see some of these lms. They've just come out with
- 9:35Speaker 1: a browser. They now have desktop apps for Claude further
- 9:40Speaker 1: integrating it Chrome extensions. Is that available right now or
- 9:44Speaker 1: how where's the team at with the development of it.
- 9:48Speaker 2: Yes, we have all of that already. So if you
- 9:51Speaker 2: go to the documentation, you can see actionists and you
- 9:54Speaker 2: can see the demos. We have got dozens of different
- 9:58Speaker 2: use cases across mark sales, operations, HR. It literally covers
- 10:05Speaker 2: any kind of business department and we've got it in
- 10:09Speaker 2: beta with a number of different test clients right now.
- 10:12Speaker 2: So it runs locally on your on your map or
- 10:16Speaker 2: Windows or also in the clouds if you don't want
- 10:19Speaker 2: to leave your computer running twenty four to seven, because
- 10:21Speaker 2: that's what these new era of AI employees will be doing.
- 10:26Speaker 2: It's not I'll speak to it and it sends me
- 10:28Speaker 2: a message and it's finished. It's these AI employees. They
- 10:31Speaker 2: have their own calendar, they control their own computer, and
- 10:36Speaker 2: they can work twenty four to seven. So you know,
- 10:39Speaker 2: a single one of these can work probably at the
- 10:42Speaker 2: same speed and velocity as five members of staff. That's
- 10:46Speaker 2: what that's what makes this transition so inevitable because businesses
- 10:52Speaker 2: will one hundred percent adopt it when given the chance.
- 10:57Speaker 3: Yeah, yeah, I know. It's it's moving quickly.
- 11:01Speaker 1: And you know what, I'm guessing that the early adoption
- 11:05Speaker 1: of this is going to be cryptonatives or you know, libertarians,
- 11:10Speaker 1: but more people that are in the blockchain industry and
- 11:13Speaker 1: then you have to break out from there.
- 11:15Speaker 3: Is that sort of the case right now?
- 11:17Speaker 2: No, I don't think so. I think that. I think
- 11:20Speaker 2: that with the action model in the community, I think
- 11:23Speaker 2: that ninety five percent of our community will be Web
- 11:27Speaker 2: two you know, normal people that might not have any
- 11:31Speaker 2: involvement with with crypto. You know, in my last company,
- 11:35Speaker 2: I ran a fintech in the UK and we were
- 11:39Speaker 2: one of the first companies in the UK to bring
- 11:42Speaker 2: QR code payments in. So we had a banking payments
- 11:45Speaker 2: license and that essentially enabled us to connect to banks
- 11:50Speaker 2: directly to move funds without Visa MasterCard. So, you know,
- 11:56Speaker 2: our our whole narrative in that company was when you
- 12:00Speaker 2: pay with Visa MasterCard, they take up to five percent
- 12:03Speaker 2: from the business, but when you pay with us, they
- 12:07Speaker 2: don't have that fee, right So through word of mouth,
- 12:12Speaker 2: that grew to millions of users within a year, just
- 12:16Speaker 2: because people wanted to support the resistance.
- 12:19Speaker 1: Right.
- 12:19Speaker 2: People like to resist against something, they like an uprising
- 12:24Speaker 2: David versus Goliath, right, And in that case it was
- 12:28Speaker 2: Visa MasterCard. Even in a scenario when people didn't actually
- 12:32Speaker 2: lose anything, it's the business in that scenario that lost
- 12:35Speaker 2: the transaction fee, not the millions of users. But here
- 12:40Speaker 2: we have a billion people that are at risk of
- 12:43Speaker 2: losing their jobs to AI, and so not only do
- 12:49Speaker 2: we all have something at stake and something to lose here,
- 12:52Speaker 2: but the way we've structured the action model is you
- 12:55Speaker 2: actually own a fractional share of it as well. There's
- 12:58Speaker 2: an incentive to tribute and participate. So we actually launched
- 13:02Speaker 2: two days ago and we hit sixty thousand users in
- 13:10Speaker 2: the first forty eight hours. The platforms invite only, so
- 13:14Speaker 2: you can only get access if you've been invited by
- 13:17Speaker 2: someone else, and really blew us away, well, the team
- 13:22Speaker 2: and the developers, because we didn't really expect those numbers.
- 13:26Speaker 2: But it really shows how powerful that narrative can be.
- 13:31Speaker 2: And we haven't done any real marketing in Web two either.
- 13:35Speaker 2: We haven't really pushed to anyone. It's mainly been a
- 13:40Speaker 2: very small handful of early access communities that have grown
- 13:43Speaker 2: this so far, so I think that you know, this
- 13:47Speaker 2: will reach millions of viewsers because our couse is so important,
- 13:52Speaker 2: you know, our ethos as a project and as an
- 13:54Speaker 2: ecosystem is that everyone should be able to park in
- 14:00Speaker 2: AI and oh and a piece of it, not just
- 14:03Speaker 2: a handful of billionness.
- 14:05Speaker 1: Definitely, I agree, And that's some nice growth coming out
- 14:09Speaker 1: of the gate with invite only. I would love to
- 14:12Speaker 1: get an invite and maybe if we have some for
- 14:15Speaker 1: the audience, we can try and get some links to that.
- 14:18Speaker 1: And you know, it's a great example with the visa.
- 14:21Speaker 1: You know, yeah, it's huge to be saving five percent.
- 14:25Speaker 1: That's most of the margin of most companies that have
- 14:28Speaker 1: you know, thin margins. It's like it's the difference between
- 14:31Speaker 1: not making money and having your business survive. But with
- 14:35Speaker 1: the at least with the you know, the chat cheept
- 14:38Speaker 1: moment when they were launched, it made it so easy
- 14:42Speaker 1: to just like type and get the answer that you know,
- 14:46Speaker 1: grandparents were able to use it, and I think that
- 14:49Speaker 1: was a huge part of the adoption. And now as
- 14:53Speaker 1: we move to autonomy and AI agents, there's definitely more
- 14:58Speaker 1: of a technical barrier right now, even myself exploring through
- 15:01Speaker 1: different platforms like the claude bot that came out for example.
- 15:05Speaker 1: You know, you have to use lines of code, you
- 15:08Speaker 1: have to like no no JS or Python.
- 15:11Speaker 3: And even for tech.
- 15:13Speaker 1: You know these white collar workers, as you mentioned a
- 15:15Speaker 1: lot of them, they're not very technical. They're sort of
- 15:17Speaker 1: just doing low level tasks. Somebody has to do it.
- 15:20Speaker 1: They it's hard for them to set up an AI
- 15:23Speaker 1: agent that requires code.
- 15:26Speaker 3: So does Action model solve that?
- 15:29Speaker 2: Absolutely completely right, and there will always be a thriving
- 15:33Speaker 2: open source community. And I mean look at the open
- 15:37Speaker 2: source community for llms right now. We've got thousands of models. Firstly,
- 15:43Speaker 2: none of them come close to the big boys. But secondly,
- 15:48Speaker 2: look at the size disparity in user audiences. In terms
- 15:53Speaker 2: of AI users, I would say ninety nine percent of
- 15:56Speaker 2: them are paying for it via the big tech companies
- 16:00Speaker 2: and it's only the skilled technical people that are even
- 16:04Speaker 2: bothering to test the open source models because inherently it's
- 16:08Speaker 2: quite complicated. So we are solving that, like we're making
- 16:13Speaker 2: it extremely easy for businesses to use this, and in
- 16:17Speaker 2: part that's one of the functions of our development, but
- 16:21Speaker 2: also the community as well. You know, the biggest challenge
- 16:26Speaker 2: that's going to come with this next stage is actually
- 16:30Speaker 2: the onboarding. If you think about hiring an employee, you
- 16:36Speaker 2: have to onboard them for one to two months, teach
- 16:39Speaker 2: them everything about their job. You know what platforms they
- 16:43Speaker 2: have to use, how to use them, when to use them, policies, processes, procedures.
- 16:48Speaker 2: It takes a lot of time and a lot of
- 16:49Speaker 2: manpower to do this, and if you want the same
- 16:53Speaker 2: level of output from an AI employee, you need a
- 16:56Speaker 2: similar level of investment. Right, So we've built a tool
- 17:00Speaker 2: that essentially does this, and it's very simple to use
- 17:05Speaker 2: for businesses. It's called the on border and you're sent
- 17:09Speaker 2: to turn it on and it's completely optional. By the way,
- 17:12Speaker 2: This is for what I think the majority of businesses
- 17:15Speaker 2: will use because it's so easy, and you turn it
- 17:18Speaker 2: on and it starts to shadow you and it learns
- 17:22Speaker 2: from what you're doing, and you can leave it on
- 17:25Speaker 2: as long as you want, whether it's for five minutes
- 17:27Speaker 2: to show the AI a very specific thing that you do,
- 17:30Speaker 2: and then you can schedule that manually in the calendar
- 17:33Speaker 2: of the AI employee, or you can leave it on
- 17:35Speaker 2: for a month and it will learn everything that you
- 17:37Speaker 2: did in that process and by the end of that period,
- 17:40Speaker 2: you'll have an AI employee that's built a brain of
- 17:44Speaker 2: all the context, all the different variables, where all your
- 17:48Speaker 2: files are, what platforms you use, when to use them,
- 17:51Speaker 2: what kind of workflows, schedules, calendars, everything, right, And so
- 17:59Speaker 2: that's you know, one of the tools, and the community
- 18:02Speaker 2: aspect of it is the workflow creation, right, so we
- 18:06Speaker 2: have a marketplace within the ecosystem. It's one of the
- 18:08Speaker 2: ways to earn more of a fractional stake. And it's
- 18:11Speaker 2: essentially the fact that you can create workflows and then
- 18:16Speaker 2: use them yourself, or you can publish them to the marketplace.
- 18:20Speaker 2: And when you publish them, anyone can go and find
- 18:22Speaker 2: that workflow and within one click they can attach it
- 18:25Speaker 2: to their own AI employee and it onboards the workflow
- 18:29Speaker 2: within a second. Right, So it's super super simpletant It
- 18:32Speaker 2: makes it easier for businesses to be able to use this.
- 18:36Speaker 1: Wow, that's very cool, Sena and the onboarder. Yeah, that's
- 18:41Speaker 1: a huge time drag for most of these traditional companies,
- 18:45Speaker 1: especially anybody who's an employer and they lose an employee,
- 18:48Speaker 1: it's like, oh, we've got to train someone else that's
- 18:50Speaker 1: going to be. You know, the documentation you have to
- 18:53Speaker 1: go through to set that up and then they have
- 18:54Speaker 1: to follow it all.
- 18:56Speaker 3: It's very laborious.
- 18:58Speaker 1: But at that point, if the AI has onboarded, you know,
- 19:03Speaker 1: done on the documentation for onboarding, they could just do
- 19:06Speaker 1: the job themselves at that point exactly.
- 19:11Speaker 2: That's that's what's coming like, and that is how easy
- 19:16Speaker 2: it's supposed to be for businesses. You know, businesses are
- 19:19Speaker 2: not going to be able to manually configure these things.
- 19:22Speaker 2: If you look at you know, the rise of NA
- 19:25Speaker 2: ten and make ZPR and these tools, I mean they're fantastic.
- 19:29Speaker 2: They have their limitations, don't get me wrong. You know,
- 19:31Speaker 2: the biggest one being that they can only use APIs
- 19:34Speaker 2: therefore ninety nine percent of the Internet's functionality isn't available
- 19:38Speaker 2: to them. But the bigger challenge is the fact that
- 19:42Speaker 2: a there it's RPA, so they're brittal workflows. And secondly,
- 19:48Speaker 2: they have to be built almost entirely for a specific
- 19:53Speaker 2: use case. You know, a business can't just one click
- 19:56Speaker 2: in store and get them up and running. They have
- 19:59Speaker 2: to be configured one by one and it either requires
- 20:02Speaker 2: the business to spend a lot of money on someone
- 20:04Speaker 2: that can do that or they have to learn with
- 20:09Speaker 2: like a one month learning curve, how to do it themselves.
- 20:14Speaker 3: Yeah.
- 20:15Speaker 1: I've been through that, going through zaper and the automations,
- 20:18Speaker 1: and there's a lot of limitations. It's great once you
- 20:20Speaker 1: get it set up, but yeah, if the AI could
- 20:23Speaker 1: just control the computer, keyboard and mouse, then you would
- 20:26Speaker 1: save all of that. And I think those automation companies
- 20:30Speaker 1: are also realizing that, and they're integrating further AI and
- 20:34Speaker 1: they eventually they'll probably have to shift their entire business model.
- 20:38Speaker 2: I don't think they're going anywhere. You see these these
- 20:42Speaker 2: new automation AIS. They can use the computer. It's not
- 20:46Speaker 2: going to necessarily replace a Zapier or any tent. It's
- 20:50Speaker 2: going to know how to use them. Right. So these
- 20:55Speaker 2: these workflows, through these workflow tools like Zapier, they're great
- 20:59Speaker 2: for a specific use case that you might need to
- 21:03Speaker 2: run a thousand times a second. Right and instead of
- 21:06Speaker 2: you knowing how to build it, the AI can actually
- 21:10Speaker 2: go and use Zapia to build it, test it, configure it,
- 21:13Speaker 2: deploy it, and maintain it. So that's the difference that's coming.
- 21:20Speaker 3: Definitely.
- 21:21Speaker 1: So you talked about the onboarding, you talked about the workflows,
- 21:25Speaker 1: which I wanted to dive into on the workflow because
- 21:29Speaker 1: that sounds like like I'm trying to understand how to
- 21:32Speaker 1: own more of the community owned AI if you're the
- 21:35Speaker 1: one presenting sort of like a library of the code,
- 21:39Speaker 1: or like, here's something that the AI can do to automate.
- 21:42Speaker 1: Here's an onboarding process for my business that could be
- 21:45Speaker 1: applied to other businesses. People can use your code per
- 21:48Speaker 1: se and then you're getting a share of the revenue
- 21:51Speaker 1: from that.
- 21:54Speaker 2: Yeah, that's exactly how it works, right. So as an individual,
- 21:57Speaker 2: you can create workflows like I said, yourself, and if
- 22:02Speaker 2: you want, there's a nice shiny button it says submit
- 22:05Speaker 2: to marketplace. Once it goes for a review process, it
- 22:08Speaker 2: gets published. The action model team as well can also
- 22:13Speaker 2: feature certain workflows, especially if it's a really good workflow
- 22:17Speaker 2: in a particular domain. So the marketplace is split by platform,
- 22:23Speaker 2: So you have workflows for Gmail, workflows for Figma, workflows
- 22:28Speaker 2: for Salesforce. Any platform that exists, you can create workflows
- 22:32Speaker 2: for it. You don't need any integrations, you don't need APIs.
- 22:35Speaker 2: This is the other thing that our engine is really
- 22:40Speaker 2: really good at Zapier and nat and as you know,
- 22:43Speaker 2: they are still technical. You need to learn with a
- 22:47Speaker 2: significant learning curve of at least a month. If you
- 22:50Speaker 2: want to build anything you know reasonably complex. They're technical
- 22:56Speaker 2: and ours. You just need to be able to speak.
- 23:00Speaker 2: If you can read and write, then you can create
- 23:02Speaker 2: workflows in our ecosystem because our workflow engine is mainly
- 23:07Speaker 2: about telling the computer what to do, so the nodes
- 23:11Speaker 2: are mainly navigation notes. You know, if you want the
- 23:14Speaker 2: AI to open the browser, click on the url bar,
- 23:17Speaker 2: type in Twitter dot com and then go and click
- 23:21Speaker 2: post new tweet, you would literally explain that step by
- 23:25Speaker 2: step inside of our workflow engine. You know, step one,
- 23:29Speaker 2: open Chrome, step two, click the url bar. You know,
- 23:34Speaker 2: you were speaking to it in natural language, and that
- 23:37Speaker 2: is what basically helps the AI be flexible, but provides
- 23:43Speaker 2: it with the FLEXI provides it with the capability to
- 23:46Speaker 2: be able to do almost anything.
- 23:49Speaker 3: Yeah, it's very cool.
- 23:50Speaker 1: You mentioned earlier about how there's thousands of other open
- 23:55Speaker 1: source AI models that people are experimenting with, but they
- 23:58Speaker 1: really don't have the processing power.
- 24:02Speaker 3: And the libraries that you know, the fang of AI
- 24:05Speaker 3: have is how will action model?
- 24:09Speaker 1: You know, is the community based efforts enough to try
- 24:13Speaker 1: and keep up with the growth of you know, we're
- 24:16Speaker 1: seeing every month, like there's a new Gemini model, there's
- 24:19Speaker 1: a new Claude, and like they're getting.
- 24:21Speaker 3: Past PhD level.
- 24:23Speaker 1: How can open source and decentralized ais compete with that?
- 24:28Speaker 2: So lms, I mean, you know, I think it will
- 24:31Speaker 2: be it will be a tighter difference between top models.
- 24:38Speaker 2: The main difference might end up just being time to
- 24:42Speaker 2: first token, you know, and the speed of the result,
- 24:46Speaker 2: maybe at the actual cost because those companies have such scale.
- 24:50Speaker 2: I mean, Google has their own one hundred infrastructure, you know,
- 24:54Speaker 2: they own their own gyms, and so their economies of
- 24:58Speaker 2: scale are much better as well in regards to what
- 25:00Speaker 2: we're building. There is a significant difference. You know, if
- 25:05Speaker 2: you try and use a traditional LM to navigate the
- 25:09Speaker 2: Internet and use the interface of a browser, it struggles
- 25:14Speaker 2: to do that for a number of different reasons. One,
- 25:17Speaker 2: the trending data of an LM is about a year old, right,
- 25:22Speaker 2: so if there's been any changes to the UI to
- 25:25Speaker 2: the platform, then it doesn't even know the correct sequence
- 25:28Speaker 2: of events. It doesn't know the functionality that's available, right.
- 25:32Speaker 2: And secondly, it's not trained on the UI of a website,
- 25:37Speaker 2: so it doesn't know where the buttons are, it doesn't
- 25:39Speaker 2: know what the buttons look like, so it makes it
- 25:43Speaker 2: hallucinate much more when it comes to actions than it
- 25:47Speaker 2: does with normal text, and hallucination with text is already
- 25:50Speaker 2: around twenty to thirty percent depending on topic. Right, So
- 25:55Speaker 2: if you want an AI as a business to do
- 25:57Speaker 2: the work for you, it needs to be closed to
- 26:00Speaker 2: one hundred percent success rate, right. And that's where the
- 26:04Speaker 2: actional model ready prospers because we get live data. We
- 26:08Speaker 2: have already tens of thousands of users within a couple
- 26:11Speaker 2: of days that have downloaded the browser extension, and that
- 26:14Speaker 2: browser extension is tracking and learning from all the different
- 26:20Speaker 2: platforms that our community is using. That means the AI
- 26:24Speaker 2: is learning live on all of these platforms exactly how
- 26:29Speaker 2: they work, What is the functionality, What is the sequence
- 26:33Speaker 2: of steps required to complete a specific workflow, what do
- 26:36Speaker 2: the buttons look like the mouse coordinates? All of this
- 26:39Speaker 2: information is live and accurate, not one year old, not guessing, right,
- 26:44Speaker 2: and so that drastically increases the accuracy, which is exactly
- 26:50Speaker 2: what businesses need.
- 26:53Speaker 3: I agree, I'm excited to check it out.
- 26:55Speaker 1: I can leave a link to the Chrome extension in
- 26:58Speaker 1: the show notes below, and what's the best way to
- 27:01Speaker 1: follow along with this onboarding, find out the workflows as
- 27:06Speaker 1: more come in, and just the progress of action modeling
- 27:09Speaker 1: to just dive into.
- 27:10Speaker 2: It, well, depends on how involved you want to be.
- 27:13Speaker 2: If you want to follow us and the journey, you
- 27:15Speaker 2: can go on x actional Model AI. Follow us there.
- 27:19Speaker 2: Do you want to get involved, you know, actively contribute
- 27:23Speaker 2: out a discord channel, But but everyone can go and
- 27:26Speaker 2: sign up to go to action model dot com and uh,
- 27:30Speaker 2: and you know, go and search for referral codes. I'm
- 27:33Speaker 2: sure you'll publish, you'll want to within this this video
- 27:37Speaker 2: as well sounds great.
- 27:39Speaker 1: I will definitely. I can add all that in the
- 27:41Speaker 1: show notes. And I appreciate your insights into where AI
- 27:45Speaker 1: is heading. And you know, automation is exciting. It's just
- 27:49Speaker 1: a matter of getting it in the right format and
- 27:52Speaker 1: UH and not being thrown out of your job for
- 27:57Speaker 1: for potentially a billion people by the automation. So I
- 28:01Speaker 1: think there's a lot of people that they understand what's coming.
- 28:03Speaker 1: You and I both know and getting involved in something
- 28:07Speaker 1: like this will help yourself and.
- 28:11Speaker 3: Others as well.
- 28:11Speaker 1: So I appreciate the initiative from Action model, and I'm
- 28:16Speaker 1: looking forward to following along. I would love to follow
- 28:19Speaker 1: up again, as you know, it's only been two days
- 28:22Speaker 1: since the launch and it's already considered pretty successful, but
- 28:25Speaker 1: there's a long way to go to get to save
- 28:28Speaker 1: a billion people. I'm looking forward to following up in
- 28:31Speaker 1: the near future again with more updates and as ai
- 28:34Speaker 1: continues to grow quickly.
- 28:37Speaker 2: Thanks for having me, Austin, it's a pleasure.