Ready to build a site that looks hand-coded without hiring a developer?
Launch your site for free at:
https://framer.link/WesRoth
use code WESROTH for a free month on Framer Pro.
______________________________________________
My Links 🔗
➡️ Twitter: https://x.com/WesRoth
➡️ AI Newsletter: https://natural20.beehiiv.com/subscribe
Want to work with me?
Brand, sponsorship & business inquiries: wesroth@smoothmedia.co
Check out my AI Podcast where me and Dylan interview AI experts:
https://www.youtube.com/playlist?list=PLb1th0f6y4XSKLYenSVDUXFjSHsZTTfhk
______________________________________________
00::00 AI Overwhelm
04:46 FRAMER (sponsor)
07:22 the ONLY app you need
19:00 the ONLY skill you need
#ai #openai #llm
Оглавление (3 сегментов)
FRAMER (sponsor)
reverse and go to not zero, but maybe closer to one. Let's really quick take a pause here. A long time ago, there was a book called Art of War. It was about war. And more recently, there was a book called War of Art. It was about art. But more importantly, it was about the resistance. It's that feeling when you get when you're about to create something big, something meaningful, when you get close to unlocking your purpose. Resistance is that voice that comes up with the perfect excuse for you to quit. Chances are you've been sitting on a project, a newsletter, a business, or a vision that you wanted to build, right? And you've had some excuse for why you can't start. You don't have time. You're not a developer. You don't know how to do X, Y, or Z. This is your sign to stop waiting and start shipping. Today, there's officially no more excuses for your vision to be stuck in some notes app. I'm going to show you how to start getting your message out there in less time that it takes to drink a cup of coffee. We're using Framer, the builder made for real businesses. By the way, thank you to Framer for sponsoring this segment of the video. Now, Framer, this isn't a toy. It's a Figma style canvas that publishes a real high performance website. You get total design freedom with the security of an enterprise level stack. DOS protection, SSL encryption and worldclass SEO all baked in. No under construction pages, just results. You can do things like video integration. Let's build a site to showcase your videos. Drag in a YouTube or Vimeo element. It's not just a link. It's a fully responsive video player that looks handcoded. You want to build some hype for the next drop. Drag in a countdown component and set the timer. creates that mustwatch urgency for your audience. Never rely on an algorithm. I'm adding a newsletter sign up from right here to capture the email directly. I'll even add some workshop AI effects. Maybe a hover animation or a slick cookie banner to make the site feel premium. Need a blog to go with your videos? Framer's CMS handles it flawlessly. You want to sell digital products or merch? You can set up an e-commerce flow in minutes. You can even use a mini chat integration to engage with your audience. the second they land on the page. Call the growth team. You are the growth team. Boom. Done. A sleek, modern site that updates automatically across desktop, mobile, and tablet. Any member of the team can own the feel, the style, and the functionality of the website. If you're a solo creator, this gives you ludicrous speed. So, stop making excuses and start building. Go to frame. com to launch your site for free. When you're ready to go pro, use the code Wesroth for a free month. Go get your message out there.
the ONLY app you need
Now, let's get back to what we were talking about. Currently, for me, a lot of the different apps and websites that I visited have now collapsed to one. For me, that's Telegram. And that's because I use it to communicate with my AI agent. That AI agent has rebuilt my website natural20. com. It created a AI news aggregator that pulls from a lot of different places on the web, organizes it. It even created an algorithm to see how big the story is, so it's all ranked. It created a full list of all the large language benchmarks and keeps it up to date. Whenever a new major model comes out, it creates a bunch of little demos with it to kind of test to see its abilities. Overnight, it runs doing various research for me that is helpful to me for the next day. It researches guests that I have on a podcast, various AI news, as well as more personal stuff. So, for example, one of them is kind of my personal health/fitness coach. One of them is my accountant slashcpafinancial guide. To give you an idea, one of the first things that I wanted to get done with it is to actually go to a lot of different YouTube channels and pulled whatever publicly available data they were, views, what time the video got published, how long the video was, etc., etc. So, it did that. It built a little scraper using the official YouTube API under a certain amount of daily functions. You get it basically for free. So, it pulled all the data that I needed. It stored it on a local database. By the way, please understand, I didn't need to know how to do any of that. I just told it in plain English what I wanted and it then went and did that. So, it stored all the data somewhere in the computer and then I could ask it questions about the data. One of the questions I asked was, is there some linear relationship between the length of the videos and how many views it gets in general? I just wanted to see if it's true that longer videos tend to get more views. It said, no, there's no really strong linear relationship. There might be a quadratic relationship. It came up with that thought on its own. I didn't come up with it. It looked at the data and said, "hm, it might be this other function that we need to look at. Do you want me to do that test instead? " And I said, "Yeah, yeah, that's very smart. Thank you. Please uh do that. " It did that. It found that there's some sort of local optimum. It's not a really strong relationship, so it's nothing actionable in case you were wondering, but there's a slight top somewhere like 26 to 34 minutes. I forget the numbers, but it's not a very strong one. If you're publishing videos, I wouldn't change anything based on that. But they did find that relationship. So normally I would have to use some sort of a UI where I can code up the thing, right? I would have to go to some website to learn about the YouTube API. Then later once I pulled the data, I would have to maybe put it into Excel so they can run these various tests on it. This is what I would have to do in the past with this bot. I used zero of those tools. Just to be clear, the only app that I interacted with was Telegram, but it could just as easily be WhatsApp or anything that you want to use for text messaging. By the way, Telegram also allows for your voice. You can just speak into it kind of like a walkie-talkie. The agent is able to transcribe and understand what you're saying. So, you don't even have to text. I have this Whoop tracker. So, now I have all the data that it gets on me on heart rate, heart rate variability, you know, how well I sleep, movement, and exercise, and all that stuff. Used to be I would have to open the Whoop app and look at their UI to figure out kind of like what's happening. Now everything gets uploaded to my health agent and that health agent takes that data plus everything else and is able to give me personal actionable advice that will likely work for me based on the massive amount of context that it has about me. Recently I went back and I uploaded the last few years of blood work that hell/fitnessbot also gets everything that I eat. I take pictures of anything that I eat and I just send it to it. It tracks the macros, the calories. It's pretty accurate at estimating how much calories and macros are in the thing that you took a picture of. And so over time, you know, nothing really happens on the first few days, but over time, as the amount of context that it has available to it, as that grows, thing gets eerily good at troubleshooting certain things. And I'm already beginning to feel different and better because of some of the specific advice they gave me. For example, sometimes when I go to sleep, I take one of those melatonin gummies. Whoever came up with that idea of putting supplements in gummies just needs to be arrested. I Why feel like that's just a way to like inject extra sugar into you right before you go to sleep? But my bot kind of found this pattern saying that certain days I felt a little bit groggy and it said the whole gummy has too much melatonin. Cut it into a quarter. I did and it did fix that issue. Although now it's frustrating because it's not even a whole gummy. It's just a whatever. That wasn't the point. noticed that eating eggs, you know, early in the morning as one of my first meals really helped me stay locked in the afternoon. And just a few days ago, it spotted something in my blood work that it recommended testing something out. I've been feeling great. It's only been a few days. I like knock on wood. I don't want to jinx myself, but it pointed out a small tweak that I can make that really improved how I felt. I think it lowered the cortisol. I'm not as like stressed out as I'm on edge, just a little bit more chill with a small tweak, something that I was doing wrong. It's also pushing me to get my genome sequenced because it's saying there's likely a MTHFR mutation. I need to get that figured out and also go get my blood work and get tested for certain specific things like a vitamin B deficiency and a few other things. It's also doing tons of my accounting. It's writing a lot of things that I need written, doing research, helping me with fitness, with finances, with building websites, with making code, with just everything, everything. Everything. Here's the super important thing to understand that this is sort of like the final stage. Up to now, we had more and more and more. Now, we're going to have less. what? Probably just one messaging app. And you're really not going to have to use that many other things. My various health devices upload to my agent automatically. My finances get uploaded in there. My blood work gets uploaded in there. So, now the game just becomes kind of give it whatever context it needs to make those decisions and to help you and to guide you. Now, here's the thing with what I'm using, Open Claw. It can be difficult to install. By the way, there's a number of companies that are working to fix that. I'm actually pretty excited because pretty soon, we might be able to just buy little computers with everything pre-installed securely. That's coming very soon. I'm hoping to get a few uh review samples. Maybe we'll even give some away here. So, stay tuned for that. But, if you were prevented from testing stuff like this out because of the kind of like the learning curve, first of all, I totally get that. That might not be around for a long time. that learning curve is about to get flattened out. So my point is if you feel like you have to learn a million new apps and things, no, you don't because once this thing rolls out and becomes evenly distributed, and I don't mean open claw necessarily because we're already we're seeing a million competitors. Nvidia just launched I think they're calling it Nemo Claw. There's going to be a lot of different things. They might not even all be named Claw. So don't worry about the specifics. Look at kind of the top level thing that I'm telling you is that most of your work very soon in the future will be done through some messaging app. You might be talking into it. You might be typing into it. But the point is I don't use Excel anymore. I message my bot. I don't use QuickBooks Fitness Pal anymore. I don't open up Chrome anymore. I mean, there's still things you need, but 90% of my uses, I don't need a browser anymore. I tell it, I'm going to sleep. research this thing for me and send me a report in the morning. The reason I use those at night is because there's certain limits and thresholds. If you hit those thresholds during the day, you're like kind of useless. So, for these bigger tasks, I schedule them to be run at night. The important thing to understand is that with something like this, having expertise in a lot of different areas won't be quite as important. Expertise will be a lot more evenly distributed, shall we say? So, for example, right now I don't know how to read everything that's on my blood labs. I would have to go find an expert to explain it for me. Now, I don't because it's able to flag it for me and see the various trends over time. If I need to do some quadratic regression analysis on a whole bunch of data in the past, if I didn't know how to do that, I would have to go and find some Excel expert to do that for me. I no longer have to do that because the agent does it for me. bookkeeping. I don't need to know as much because a lot of it can be handled by the agent. I used to use various budget apps to help me keep track of the budget. After using this, like you're never going to go back. Those budget apps when you look back at it, they're laughable. They're like, "Oh, you went to Starbucks, that's gas, right? That's the category for that expense, gas. " No, it's like, "Okay, how about Exxon Mobile? That's coffee, right? " Like, it's so horrible at organizing your expenses. Everything's again some clunky UI where you have to click on things. they have their stupid pop-ups that try to get you to buy more. This is going to eliminate like 90% of that. So, one big skill in the future that you're going to have once this thing is rolling out and getting better and getting more secure, really, you just need to know how to use everything that you're able to use, everything at your disposal to get to where you're going. How do you best use all the expertise that these agents provide for you to achieve something? And that's going to be the skill acquisition race that a lot of us are going to be running over the next years, I assume. But here's the thing, for most people, they might not even need to engage in that because they're going to have most of the stuff that they're going to need just in their chatbot. By the way, I've installed these for various friends and family. I warned them of some of the dangers. I said don't share certain information, but I specifically set bots up for them to help them in some area of their life where they wanted to get some more help or they were struggling. A few days ago, I updated to latest version and the bots stopped working briefly. There were some issues. So, they were down for 6 hours. Let's say I couldn't fix it in time. I was out and about. And these friends and family members, people that are supposed to care about me, well, they were borderline violent. They were like, "My bot is not responding. fix it. I was surprised by one, just a level of hostility that people very quickly develop when their bot goes down and they see you as sort of the problem. And number two, and again, I'm seeing this firsthand observing people. I don't know if addiction is the right word, probably not, but the level of addiction is fast and it's real. Once people use this, they don't want to not have access to that anymore. And addiction is the wrong word because it's not harming you. it's helping you do more. So maybe reliance is a better word. I mean sure you can argue there's some downsides on being reliant on something but as you're using it you are able to do more. So, normal people, nonte-savvy people, people that don't follow AI, when they get their hands on their own personal chatbot like this that has memory about them, that is able to work around the clock to help them with whatever they're looking for, over time, as they're communicating with it, as it's gathering context about them. Over time, this thing becomes indispensable for them. It becomes like their arms and hands. I had to, and this is no joke, I found an old laptop and I have to put these agents on that old laptop just so it's on its own device so that there's I just don't want that thing going down and people calling me and being angry with me. Like I'm like I don't want to be the one standing between them and their agents. So when this thing rolls out to the public, the people, I expect to see the same exact reaction. People are going to fall in love with it. People will not want to not have access to it. This whole AI hates a lot of it is going to get extinguished like that because imagine a nice capable assistant that like whole focus of their life is you and how to make whatever goal you set better. How to approach whatever you want to accomplish faster. You're not going to be mad at that thing. You're going to be very grateful for having it. But the
the ONLY skill you need
point of this video is what is the one skill that everyone needs to have? Whether you're going to be building a business with it or trying to stay at the very cutting edge. What is the thing that regardless of your expertise, you're going to have to learn if you want to really kind of take advantage of this. I don't know if there's a good name for it. I came up with the name AI assisted execution or AI assisted journeys. I know those are horrible names. I'm sure somebody's going to come up with something better, but this is the skill that I'm seeing that's very much missing right now. We've been reliant on learning and doing stuff a certain way for so long that there's a certain inertia that prevents us from switching over to this new better way of doing stuff. Right now, a chatbot like this, especially one that's more agentic where it can do more stuff on your behalf that has all the context that it needs. Think of everything you might want to have accomplished as two buckets. One bucket is something that the agent can do or they can build some software, some system to get that done or research and figure out how to do it. point is like you can give it a task and either immediately over time it will complete that task and some of the tasks it won't be able to do. Maybe it's because you have to do it in the physical world or for example if you're setting up a Linux system that installation process an agent can't do for you because it requires kind of booting the actual computer in a certain mode to load the new operating system on there. You can't have an agent on that system without an operating system. So it's kind of like this chicken and egg thing. the agent can't install the operating system until there's an operating system on which you can install the agent. And for a lot of tasks like that, the best and most efficient and fastest way of you being able to finish that task is for you to use a chatbot or an AI agent that's walking you through the process. When I was installing Linux on one of the computers, every time I ran into an issue, I would take out my phone, I would just snap a picture of the screen, send it to the agent, and be like, "What do I do? " And I was able to install everything perfectly. Took 20 minutes. It was fast, easy. In the past, if you didn't know how to do it, it might have been a lot more difficult. Either A, you would have to go and take a class and learn how to do all of that. Or B, you would ask somebody that you know with expertise or hire somebody with expertise, or you would have to kind of learn as you go, like pull up a website, sit there reading the next steps as you're installing the system. The trouble with that is very often you would run into some very specific issue that is specific to you or that system that's hard to get answers to. I'm sure we all had that where you're googling some error code or googling something and just there's no good answer for that. A lot of times when I do certain tutorials on how to install certain things, people ask, well, I got this error message. What do I do? Or well, how do I do this if X some factor is different for them? And the answer to that question in all of those questions is almost always I have no idea, not a clue, but I ask my chatbot and if I'm able to provide enough context about everything, it's going to know the answer. This is the thing that a lot of people miss, I think, because they're using chatbots that don't have a lot of memory built up. Also, a lot of the chat bots, they try not to save everything about you just for data privacy reasons, or more likely, they don't want to seem like the thing knows too much about you because that's going to freak some people out. There's laws about how companies can store health information. So, there's tons of stuff that, you know, if you're just using a chatbot, it's not going to have context about you with these AI agents. Again, Open Clock currently is kind of like the reigning champion, but we're going to see a lot more for different tasks coming in the future. So, keep that in mind. I'm talking broadly, but that thing knows enough about me to where the answers are really good because it knows what computer, what operating system I'm using. It knows a lot of my preferences. the things that I have access to. So, that one big skill that I think everyone's going to need to learn now is what do I need to input into one of these agents for it to give me the information that I need to complete a task in the real world. So, step one is figuring out can the agent do it? If yes, then you just give whatever context it needs to figure it out. Or if it can't, if it's not able to complete the task on its own, then again, what context do I need to give it for it to walk me through how to do it? So, kind of between one of those two options, any given person will be able to do I don't want to say anything. There will be limits to it, but they're going to be able to do a lot more than currently I think most people imagine. Again, a deep health analysis, bookkeeping, website building, software creation, deep data analysis, writing contracts, evaluating contracts, kind of like making sure that the legal ease is good and proper, creating massive databases with stuff you're going to need, debugging front end or debugging various things that are not working. Interestingly, the thing that I'm probably going to do right after finishing this video is I set up one of those stationary bikes inside the house just so I can get some cardio with the least amount of friction, meaning it's just easy to get on, do 10 minutes, 15 minutes, and get off. After setting the