Manus: The AI Tool That Works While You Sleep

What if your next employee was an AI who never sleeps and costs less than your daily coffee? 

Joe Downs and Peter Swain pull back the curtain on Manus…an AI powerhouse that doesn’t just chat, it works. 

Forget asking questions. Manus takes complex tasks, runs live research, coordinates agent teams like a digital orchestra, and delivers high-quality outputs while you're at your kid’s basketball game. 

From real estate reports to marketing assets for landscaping businesses, they break down when, how, and why to use Manus over ChatGPT, Claude, or Gemini. 

Along the way, they explore how AI is actually changing how we communicate…not just with machines, but with each other.

Listen For

:25 What’s the difference between an AI assistant and an AI employee?

4:41 What are the four layers of AI and where does Manus fit in?

10:19 Why AI struggles with the word “don’t”—and what to say instead

21:49 When should you use ChatGPT, Claude, or Manus?

32:25 How can a landscaping business use AI to launch a new service?

Connect with the hosts: Email 

Joe Downs 

Website | Email |LinkedIn |YouTube 

Peter Swain 

Website |Email |LinkedIn |X 

Chapters

• 0:25 What’s the difference between an AI assistant and an AI employee?
• 4:41 What are the four layers of AI and where does Manus fit in?
• 10:19 Why AI struggles with the word “don’t”—and what to say instead
• 21:49 When should you use ChatGPT, Claude, or Manus?
• 32:25 How can a landscaping business use AI to launch a new service?

Transcript

Peter Swain (00:00):

Manus is running right now, creating nine research reports on nine different frameworks for social media generation, because that's something I'm delivering in an hour and a half. So I've got tens of thousands of dollars worth of research happening, and I'm here with you talking about podcasts.

Joe Downs (00:25):

Last week, I gave an AI a job, not a question, a job. Go research this topic, pull the data, build me a comparison document, and have it ready when I get back. Then I closed my laptop and I went to my son's basketball game. When I came back, it was done. A finished work product sitting in my inbox. That's not a chatbot. That's not an assistant that needs handholding. That's an employee that works while you go live your life. If you've ever said, "I need to hire someone to handle all this, " stick around. What we're about to show you might change that math entirely. I'm Joe Downs. With me is Peter Swain. One of us teaches people how to buy self storage. The other consults with businesses on AI strategy. Together, we're just a couple of successful idiots. Hey, Peter, let's step inside the Hustle Lab and show these people how AI can put more horsepower behind their hustle than a British kettle has behind it during a national crisis.

Peter Swain (01:22):

I love it. There's a dam in the north of England where they push water up the hill every day and then they dam it at the top and then they release it and they release it in the ad breaks of soccer games because everybody turns on their kettle at the same time. Seriously, we've had to build infrastructure for the electricity surge from the kettle because we built the British Empire on a good cup of tea and the sun never sets.

Joe Downs (01:51):

Infrastructure design behind the kettle, that's amazing. The use of the kettle. That's a whole nother...That's an outtake podcast. Yeah. All right. Peter, today I want to get into something that frankly blew my mind when you first showed it to me, and that's Manis. Now, I'm guessing that most people still have yet to hear about Manis at this point. At least that's what I find in my travels. Everybody's familiar with chat and a little more clawed than I thought. Obviously, Gemini, the big names, Grok a little bit. Grac and Manus are probably the two that I find people know the least about. But for the most part, as amazing as they are at the end of the day, all of them, the chats, the Geminis, the Grock, the really ... I hate to Claude if you're listening. I don't mean it this way because I know you get me.

(02:51):

But they really just answer questions. Manis though, it does something completely different. It's a different animal altogether. So walk me through what actually happens when I give Manis a task because I see it. All right. Let me set the stage differently. I don't just ask it, or I can ask it some questions, obviously. But for instance, Peter, I'll give it a task to create a PowerPoint presentation for me on a subject. And then when you give it that task, Claude or Chat, it just ... I forget exactly what it does. You could tell it's thinking, but it thinks quickly. It might even say it's

Peter Swain (03:29):

Anything. Well, let's put it an even different way. Manice is running right now, creating nine research reports on nine different frameworks for social media generation, because that's something I'm delivering in an hour and a half. So I've got tens of thousands of dollars worth of research happening, and I'm here with you talking about podcasts.

Joe Downs (03:55):

Which is basically my opening, right? I went to a basketball game, you're sitting here on a podcast while Manis is working for us. Where I was going with this, and it's incredible, and I'm not trying to undermine that. What I wanted you to break down for the listener is when you are in Manis, you actually hit enter or send or whatever. ChatGPT says, thinking maybe. I forget what Gemini does. Claude, it's got its logo and it starts making a little dance or whatever it's doing. And you know it's thinking. Manus, it looks like I'm playing with the Whopper and I just started World War III. It's like all of a sudden code's on the screen. What's going on?

Peter Swain (04:41):

So let's just layer these up because there's four layers of AI and we're kind of on layer three at the moment. So the first one is responding, which is the stereotypical ChatGPT use case when you're saying, "Hey, I want to go to Disneyland, what five restaurant show goes when it says you should go here, here, and here." And I don't want to undermine or underplay the power of that capability because if people haven't got that when they get there like, "Whoa, this is amazing." And it is. The next part up is reasoning, which is when you see the Claude or the ChatGPT saying thinking, what it's actually doing is trying to understand your request. So for example, the top five restaurants in Disney, if you have it on responding mode, it's going to say, "Here they are. Here's the top five." If you have it on reasoning mode, it's probably going to say something like, "Well, what do you mean by the best?

(05:36):

Are you on your own? Are you in a family? Did you mean breakfast? Did you mean lunch?" But there's a lot of reasoning that needs to happen to answer that question with effectiveness. Now we're on the next one, which is researching, which is what Manis does. So when you say to Manus, "I need a report on this, " or let's try and use an example that lots of people might understand. So if you're a realtor, and if you're not a realtor, then you certainly know a realtor. A realtor will quite often need a market report, like a house price is up or down. How was the average time on market for a property in 33301 in Florida? What is the breakdown of the demographic of the people that are there? How many people have kids? How many people don't? So like a market report on what's happening in that zip code as of right now.

(06:29):

That's the perfect use case for Manus. So you're going to say to Manus, "I need a comprehensive market report on this zip as a realtor, I want to know things such as this, this, this that I just said." What Manus is going to do is it's going to go, "Got it, understood." And then it's going to spawn. You're not going to see it. You actually, as Joe said, you kind of see it kind of happening on screen. But what's happening in the guts is it's spawning what's called a virtual machine. So it's creating a mini PC that it can now start saving files on, working with files, et cetera. So it's got a file storage and then it's going to create one to N amount of agents and it's going to be like a composer and an orchestrator, a conductor with an orchestra. And the conductor's going to say, "Right, you, agent one, go and find me the year-on-year statistics of growth in the market.

(07:27):

You, agent two, go and find this, you, agent three, go and find this. " So all these little workers are now out there doing live searches of the internet to find those pieces of data.

(07:42):

And by the way, that's a capability that Call and ChatGPT both have. They have something called agents, but this is why I'm such a fan of Manis. What Manis then does is when agent number one that was going to find the average time on market gets an answer, it will inform agent two, three, four, five, and six of what it's found whilst it's doing it. So you've got this swarm of PhD researchers all going and finding pieces of information and data out and all sharing it with each other. So within about three or four minutes, Manus goes, "Here it is. Here's every restaurant in Disney and a sample of their menus and the price point and the analogies that they all have broken out into tables for you for you to be able to go and research."

Joe Downs (08:33):

So that's why now that makes more sense to me. I'm that person that's using it and is just amazed and I kind of care. That's why I'm asking these questions, how it's doing, what it's doing. But at the end of the day, I'm like, "I don't really care." It's magic and it's working. So that's why sometimes it'll stop, which frustrates me and ask you questions and wait for an answer.

Peter Swain (08:55):

Now, okay, so you can solve that by the way.

Joe Downs (08:57):

Okay. You

Peter Swain (08:58):

Can solve that by telling it that you're going to be using this unsupervised. So you can tell it not to ask you questions.

Joe Downs (09:04):

So I have to say, I'm using it unsupervised because I have to tell you, I have said, "Here's the prompt that doesn't work." Don't ask me any more curse. There's a podcast so I'm trying to show symbols of or like me

Peter Swain (09:22):

I'm cursing at it.

Joe Downs (09:23):

Yeah, no. Stop asking me more questions. Well- I'm going to bed, I want this done and it can wake up in the morning and it's got another question.

Peter Swain (09:31):

So I think I've told this story here before, but I own a German Shepherd. He weighs 100 pounds and we had to go and get a professional dog trainer because this is a potentially dangerous dog. If he's out of control or we're not able to control him, somebody could get really hurt. So we were hired ourselves a trainer. And what became so obvious in the very first training session is that I was expecting my dog to speak human and the reason the trainer was so good was he knew how to speak dog. So he knew how to make my dog do things in seconds that I hadn't succeeded in months. I used that because one of the things that you don't know, which you're about to know, and for everybody here, AI isn't very good at negatives.

Joe Downs (10:17):

It's not good at negatives.

Peter Swain (10:19):

So when you say don't ask me questions, it's not very good at understanding the word don't.

Joe Downs (10:26):

Oh, so it's like a child.

Peter Swain (10:28):

100%. Now, interestingly, humans aren't very good at either. If you say there are not three birds on the branch, if you look at the way in an MRI, if you look at how the brain fires, the brain fires, there are three birds on the branch, not. We have to understand the positive version of the statement in order to negate the statement. AI is doing the same. So instead of saying, don't ask me any questions, be much better saying you are to operate fully unsupervised. If you reach a point where you have a query where you feel a human would be useful, operate on your best guess of what you feel the answer should be and include an addendum in this report of any assumptions that you made during this process.

Joe Downs (11:17):

The longer-

Peter Swain (11:19):

You hear the positive version of the negative statement?

Joe Downs (11:21):

I do. And I'm going to circle back to something you might've said on the first episode and acknowledge something because when you first said it, I was like, "Oh yeah, okay. Well, he's probably just speaking in the superlative." But the more I use AI and the better I get, I guess, at prompting, you're right. The longer the prompt, the better because it's just ... In fact, based on what I learned in your mastermind in the last week or two, I'm not letting my projects live in Claude even anymore. I'm creating a central source of truth document or a Rosetta Stone or mini Rosetta Stones at the very least that have rules and regulations of what I want it to do when I ask it to do something.

Peter Swain (12:11):

When you, myself and when you talk to anyone that's done this for years, we all do exactly the same. It's because recognizing ... I've said this quite a few times, but probably not here, you could use a Ferrari to plow a field and you could race laman with a tractor. It doesn't mean they're the right tool for the job. And when you start understanding how these systems fit together and where to use them and where not to use them. So for example, let's talk about something that people get wrong with Manis. Manis is not a replacement for Claude or ChatGPT or Gemini. It's LLM, it's AI thinking capability, it's interface, it's talk, it's conversation is poor because it's not optimized for that. It's optimized for, "Go and find me these things. Okay, I'll go and find you those things and I'll compile them into a report for you.

(13:08):

" So the workflow for Manis is go and get the thing from Manus and then take the output and use that in something like ChatGPT or Claude. So if you go back to, I'm using the Disney example because when I went to Disney, I did exactly this of going to Manis and saying, "Find me every single restaurant in every single Disney park and rank them for me by nutritional quality because I'm trying to stick to my diet whilst I'm away. I know that's going to be near on impossible at Disney, but I'd at least like the safe options per park and I'd like to see how that's going to correspond with my kids' preferences. And by the way, my wife is allergic to tomato, so I'm going to need options that can support that. So things like heavy stews or curries are not going to work." And it went, "Got it.

(14:00):

" But once it produced me that, I think it was like a 68-page document, I saved the document and I went into Claude and said, "Okay, we're going to Animal Kingdom tomorrow. I want to plan the restaurants. This is the information that you need to help me do that. " And Claude went, "Got it. Thanks very much for all of the information. Now we can chat and talk." And it can say, "But by the way, that one isn't open until 10:30 and you said you're going for rope drops, so we can take that off the list." So that's where Claude is going to be really all ChatGPT or even. But yeah, Manus, it's a different experience because it's not a conversation, it's not a brainstorm. And both those things are insanely powerful, exactly as you said, this thing is doing things and delivering results that you would have to pay at least a thousand dollars a time if you wanted to hire somebody to go and do the thing.

Joe Downs (15:04):

You're so right. It's so interesting listening to you, my journey through all of this has been one of discovery, excitement, frustration. Then I talked to you and then I realized, "Oh yeah, you're trying to race a tractor in Lamon. Why don't you just take the ... " Or better yet, you're trying to plow a field with a Ferrari. Why don't you just take the Ferrari engine out and put it in the tractor? And I don't know if that example lands, but the bottom line is that's what I've learned very recently because when I discovered the power of Manis, there was a period of time where I just skipped Claude and went right to Manis and I was having Manistee everything because it was like a one-stop shop. I was like, "Well, this is the greatest AI of all. " Only to realize it has all of its flaws, I'll say, I'll use the word flaws because of how I was using it.

(16:09):

It has no flaws or very few flaws if I was using it the right way.

Peter Swain (16:13):

Yeah, I would say they're not flaws, they're design limitations that are there

Joe Downs (16:16):

Deliberately.

Peter Swain (16:17):

It's like trying to stick a family of seven in a two seat of Mazda

(16:21):

And going, "This car's rubbish." No, it's just the Master MX5 is designed to go naught to 60 in 4.2. It's not designed to put your zoo of kids in. The thing that I think that I wish I could find new and better ways to say this sentence, and I love that we're doing this, the podcast of it, is to me, the number one mistake people make with AI is they think it's a technical skill because they're using a computer to do it. So it's a technical skill, it's a tech thing, it's a geek thing. And what you find out throughout your journey, and I think this is kind of what you were saying, but maybe in a different way, is it's actually a communication skill.

Joe Downs (17:02):

That's exactly what I was going to say. It's a communication portfolio

Peter Swain (17:04):

100% communication.

Joe Downs (17:07):

And I'm learning how to use it like that.

Peter Swain (17:09):

And what I've found is I've become a different communicator with humans as a result.

Joe Downs (17:14):

Me too.

Peter Swain (17:15):

And in some ways worse, I've got less emotional capacity for bullshit.

Joe Downs (17:20):

Yes, but your details are better.

Peter Swain (17:23):

100%, but my briefing of somebody is now way better than it was because I'm like, "Do that, but don't do it. " And because I'm so used to, I'm like, "Oh, if I give this level of detail, if I give detail prompts, I get the result I want. " And I stop making assumptions for humans at the same time. I

Joe Downs (17:40):

Just gave my kids a prompt the other day. I told them, when I started saying what I wanted, I said, "You know what? Let me tell you what the outcome is. The outcome needs to look like this. "

Peter Swain (17:51):

This is what a good outcome is for bad outcome. These tools, these are guardrails.

Joe Downs (17:54):

And these are all the tools that you're going to need to achieve that outcome.

Peter Swain (17:58):

Yeah. I'll give another good example. I spent four hours trying to fix a bug in an agentic system that I was producing the other day, and the bug was that it was an agent, so just a grand big prompt for the sake of this conversation, and I'd given it a tool, and the tool was my content library from my mastermind for it to search. And it was returning blank results in this huge, big document. And four hours later, I realized it's because I told the agent that this tool is available for your use. This tool contains all of these things and da, da, da, da, da. What I hadn't told the agent was that you must use the tool. I'd never actually told it that it had to use it. I'd done what I'd done I would do with a human. If you imagine a parallel, you asked me to come and house it for you for three days because you're away with the family.

(18:56):

You would tell me the heating switch is here, the HVAC's here, the AC's here, the thing's here, and I as a human would understand that I should probably not flip the electric switch unless there's been a blackout. I don't need you to tell me that you've told me where the AC unit is. If I want to change the temperature, you don't need to say, "Now, Pete, if you want to change the temperature, this is where it is and this is how it works. You just need to tell me it's there and how it works." I'm like, "Got it. I understand why I'm going to use it.

Joe Downs (19:29):

" Well, I think it comes back to what I just said a second ago because as you were telling that story, I was thinking, well, yeah, if you and I are having that conversation, I don't need to know, but a child, you might need to tell them, not only does this switch turn the light on, if it gets dark, turn the switch.

Peter Swain (19:49):

I have for many years thought that the best description of AI is a super genius on the spectrum eight-year-old.

(19:58):

It's wicked smart, but some of the things it does, you're like, "Why did you just do that? How have you come to the conclusion this is a good idea?" And then you look back at what you have said. I had one with my son once where I'm like, "Sam," who's now, he's a, "Have you brushed your teeth?" Yes. I'm like, "Okay, breathe on me. " And if you're not a parent, you don't understand what I'm talking about. If you are a parent, you understand exactly what I'm talking about. And I'm like, "Whoa." I'm like, "You haven't brushed your teeth." And he got really upset. He's like, "I have. " I'm like, "You haven't. I have. You haven't." I'm like, "Okay, show me what you did." And it's 15 seconds long with a dry brush with no toothpaste and no water. I'm like, Sam. And I find myself getting frustrated.

(20:41):

I'm like, "Sam, that's not how you brush your teeth." And then I realized, have we actually explained this to you properly? Have we walked you through the steps or did we give you a two-minute guide a year ago and then walk away thinking that we had done the job?

(20:56):

So it's like, no, Sam, that's not how you do it. You do this, then you do this. And you say, "Oh, thanks, dad." And it's such an AI journey because when AI gives me bad results, I'm like, "Oh, I've made that mistake again, where I forgot to tell it that it must use the tool."

Joe Downs (21:12):

Yeah. Let me bring this back to Manis for a second and you're spot on with that. But since you just said tool, it made me wonder, so if I'm listening and this is all still new to me, how do I know when to reach for Manis is the tool or Claude is the tool or Chat or Gemini, but in particular, why Mannis? What is the use case where I go, "Oh, I go to Manis for this. " And by the way, is that first or do I go somewhere else first?

Peter Swain (21:49):

So the answer is always to think about how the human interaction would work that goes with it. So let's pretend that you're interested in the book Becoming Superhuman by Steven Kotler. If you have read the book, you might want to talk about it with a friend. So as soon as we say the word talk about, then we're going to need a claude or a ChatGPT or a Gemini. If we want a conversation, we're going to need one of the standard LLMs.

(22:17):

But if you have heard about the book, Steven Kotler becoming superhuman from five or six different people, but you don't have time to read it, but you really want to understand the basic concepts of it, then what you might do with a human is tell them to go and write you a report on it and bring it back to you. That's going to be manness. So the answer to the question, and I think that's probably a bad answer, but what I'm trying to get at is what level of interaction do you want inside the process or are you seeking an output? Are you seeking a process or are you seeking an output? If you're seeking a process or interaction or a conversation, you're going to want to use Claude, Gemini, ChatGPT, Grock, one of those. If you're seeking an output, a document, a presentation, a pitch deck, a financial sample projection, if you want to find out the rules and regulations for short-term property rentals in Ohio, you're going to want to use Manus.

(23:16):

So Manus is going to give you an output.

Joe Downs (23:18):

So totally, I think the listener understands that. Let me give a use case because you corrected my use of it recently. And by the way, I'm here to report, oh my goodness, it's so much smoother this way. So as you know, I'll be quick. I'm doing my Storage 100, which is a hundred days of self-storage videos, different topics on YouTube, and I'm having Manis create the decks. And I started the project in there and said, "All right, here are the day the chapter each day, this is the headline, this is what we're going to talk about each day, the topic we're going to deep dive into." And I gave it the headlines and said, "Create the decks." And I gave it rules and it's been very good, better than I could have produced myself and certainly in the amount of time, but man, there's frustrating elements to it.

(24:11):

And I was telling you this a week or two ago, and you said you're doing it backwards, have Claude create it, the content, the style of the deck, everything, and then output instructions for Manis to do it.

(24:29):

And the way you explained it was think of Claude as your copywriter and Manus is the graphic agency, graphics arts agency or whatever that comes up that takes that copy, puts it on the slides, produces the slides for your output. And that not only from a remembering my rules and regulations standpoint and actually creating a central source of truth document for the slides, but even just having Claude create the speaker notes, the notes on the slide, but also the presenter notes, which was more important, which really was holding me up, where I was having the issues, game changer. I feel like what I did was I put ... You know how you say with employees and people, you got to put right butts in the right seats. I feel like I just put the right AI in the right seat for that two-step process.

Peter Swain (25:24):

This is literally what's happening at the moment. I'm just going to read from this. So this is my prompt interchord. "I want to create a claude project that uses knowledge files for messaging frameworks such as Donald Miller's story brand, such as Russell Bronson's XYZ, such as Stephen Kotler, XYZ. In order to do that, I need to get Manus to produce the files for me. Give me the seven prompts that I need to put into Manus for Manus to create me those context files. And then I said," Oh shit, before we do that, what frameworks do you recommend? "And it came back and went," The Neuroeconomics of Story by Paul's Act, the Storybrand Framework by Donald Miller, Curiosity and Identity by Adam Grant, the ethics of vulnerability by Brene Branson, belief stacking by Russell Bronson, and I'd also recommend a neuroscience compilation on the Raz thing and the hook ladder strategy from Liz, my wife.

(26:23):

"Like," Okay, cool. Give me the seven prompts for that. "And before we started, I opened Manus and hit new agent, copied prompt one and hit go, went to agent number two because Manus can run them concurrently, paste agent number three, paste, age number four, paste, agent number five paste, and I just looked and they're all finished. I'm now going to bring all of those back into Claude to do the work with them that I actually want to do with them.

Joe Downs (26:51):

I don't know what to say. I find myself speechless at the same time. I'm like, " Why can't you do better?

Peter Swain (26:59):

"But again, let's try and make some other use cases for people. Let's pretend you're a BDM or an SDR for a company or for yourself. You're a small business and you sell to ... This is actually the thing I did yesterday," You sell dumpsters to commercial companies in Colorado. "He's like, " Yeah, we're trying to identify companies because they just did this deal to dispose of concrete for a company and actually they're getting paid to recycle the concrete. So they're actually now earning twice over. They're getting paid for the lease of the dumpster and they're getting paid to recycle the concrete on the other side as well. "He's like, " So we're trying to find other people where that would apply. "I'm like, " Well, just put it in Manis. "That's a Manus use case of, " I am doing this deal with this company.

(27:51):

It's a great deal for me because I rent dumpsters and we're now getting paid for the recycling of the concrete as well. What other companies, types of companies and specific companies, and look on LinkedIn to see if you can find the actual person I need to contact. "And he's like, " It can't do that. "I'm like, " Just try it.

(28:14):

"We carry on having our conversation, we come back 10 minutes later, there was some very heavy expletives from him as he looked through the 62-page report of you can do the same kind of deal for this company, this company, this company, this company, the top employees in Colorado for those are this, this and this, and I have found named executives for this, this and this. I don't know how much you would pay a research company to do that job or I don't know how many hours you would have to, what, 60 companies, an hour, two hours a day.

Joe Downs (28:51):

Well, ZoomInfo would cost you 15 grand a year.

Peter Swain (28:56):

So you're at least saving 15K, but this thing is adding layers of intelligence that ...

Joe Downs (29:01):

Yeah. Well, so an interesting segue, two questions, quick answers, and then we got to get to the parlor trick. Speaking of cost, how does somebody access this? What does it cost? And can any entrepreneur, business owner, person use Mannis today or is there a technical skill required?

Peter Swain (29:28):

No, you go to manus.im, M-A-N-U-S.im. And I just had a quick look because my SaaS bill is pretty crazy. So you're looking at $60 a month for 12,000 credits with a 300 refresh. You're looking at 60 bucks to get at least one of these type of outputs a day that we're talking about. So you're talking like two bucks a day in terms of output.

Joe Downs (29:55):

So it's not a $20 a month ChatGPT or Claude member Ownership or subscription, but it's not a $15,000 a year ZoomInfo subscription.

Peter Swain (30:08):

The thing for everyone to think about here is, and you and I have discussed this a few times now, but if you are not seeing an ROI on AI, you're doing it wrong. Full stop end of

Joe Downs (30:20):

Story.That goes without saying. Yeah. It's too easy to achieve ROI.

Peter Swain (30:25):

Yeah, because each of these documents that we're referring to that we've just gone through are thousands to tens of thousands of dollars, the key, the secret to it for the listener is now to say, "Well, what would uplevel my life, my business dramatically and substantially?" It might be on the personal side, not the professional side, you might want an unbiased report on what the effects of a GLP-1 injection up. Because if you go to a GLP provider, they're going to tell you it's the most amazing thing ever. And if you go to traditional media, they're going to tell you it's probably one of the worst things ever. So you might want a balanced review. Manus is great at things like that. Go and provide me with an unbiased review of the pros and cons of a GLP-1 injection. So on the health side, but there's so many use cases.

Joe Downs (31:09):

I don't even know if I'm just thinking about what we just said, if you're not getting an ROI, I don't even know if that does it justice. I'm thinking if you're not getting 10 to 20 to 100X ROI, you're not using it right.

Peter Swain (31:22):

I would nowadays, I'm just checking, I'm not lying. Yeah, I would lose my iPhone before I lost my AI subscription.

Joe Downs (31:32):

Wow.That's a tall statement.

Peter Swain (31:33):

I really would, because apart from anything else, if I had claude.ai, I could find out how to get sponsored and get a free phone back. If I had the phone-

Joe Downs (31:40):

But you have AI on your phone.

Peter Swain (31:42):

Yeah. If I had to choose, I'd choose AI over phone. There's no comparison.

Joe Downs (31:47):

Let me set up a parlay trick so we can close this out. Using Mannes, of course, the subject of the day. Here's the situation, Peter. A listener runs a small landscaping business. They want to add a new service, let's say outdoor lighting installation as a new add-on service. They need marketing material, but don't know where to start. They probably need an FAQ for their website, a checklist for customers so they can use to evaluate if they need the service. And let's say a video outline, they can record themselves, a video outline so that they can create the video to record themselves. How would Manis knock this

Peter Swain (32:25):

Out? So you'd go into Claude and you would voice prompt exactly what you just said, and you would ask it to produce the prompt for you to type in, put into Manus for Manus to output those assets that you need and you'd go away for five minutes and come back.

Joe Downs (32:43):

So here's what you just said. I'm coming to you for help. Peter, here's my situation. I've got a landscape company. I need help with marketing materials. I probably need an FEQ. I don't even know what else I need. Should I do a video? Should I do that? Should I do this? Is this product or service going to work? Are my competitors doing it? These are all the real ... I was being brief, but these are all the real questions I have. If I have that landscaping business and I want to add this service. And what I heard you say was, while you're telling me, tell Claude. Claude will tell you what to do and give you all the proper instructions to then take over to Manis to have it executed, delivered to you within minutes.

Peter Swain (33:25):

100%. The fallacy of my entire business is I teach smart people how to use AI. And when they get to a certain level of it, they realize that they can just ask the AI. But don't tell them, "Oh, I just did it on the

Joe Downs (33:39):

Property." No, you're killing us because I was just about to tell the listeners if you've got a situation you want us to solve in the show something you're stuck on like the landscaping business, email us at idiots@successfulidiots.com and we'll pick one and solve one live for you. Or as Peter just ruined it, just go to AI. It'll tell you too.

Peter Swain (34:02):

Yeah, either way. Well, the human touched off and not maybe sometimes kind of.

Joe Downs (34:07):

I mean, we have to add some sort of element to this, Peter. It can't just be all Claude or cloud. Sorry, Doug. All right, let's land this thing. If this episode clicked for you, please like and subscribe. You know the drill. We've now covered, folks, Gemini, Grock, Claude, Manas, and feel like ChatGPT is present in every episode. It just has to be, it's the original. But that's a hell of an AI toolkit, Peter, and I feel like we're just getting started. So here's what I want you to do. If you don't mind, share this episode with the one person in your life who's drowning in madness. We all know who they are. You might be married to one. Send them this episode. They will definitely thank you later. And don't forget if you have questions, ideas, or you have some issues with, I don't know, British.

(35:02):

Who's your football team?

Peter Swain (35:03):

Spurs. Tottenham.

Joe Downs (35:04):

Come. Totten, I'm sorry. If you have any issues with Tottenham, email us at idiots@successfulidiots.com. We'd love to feature them on the show. Don't send any hate mail about my Philadelphia Eagles. We'll figure that out, folks. All right. See you next week.