Rich Schefren Sharing AI Breakthroughs 6-3-2025

Hello, everyone.

Yikes.

All right, let's see here. I'm just getting set up. I apologize for this. I thought I was going to have it all good to go, but apparently I don't.

Hold on.

And I am in Orlando today at what's known as an Aspire event. So if you happen to be in the Orlando area, that's where I'm at. And that's why I am in this weird background where you only see about half my face. So with that, I'm going to dive in today, just like every day, every morning at around this time, what we're doing is we're sharing breakthroughs in AI as they relate to improving human performance. And so that's what we're all about here. And we talk about this every day at 9:30. And my name is Rich Sheffrin, and I created a program called Zenith Mind os, which is all about teaching your AI how to know you better than anything has ever known you, including how well you know yourself. So it knows you better than you know yourself. And so what we do in this live stream is look at just every area that AI can be used to improve your own performance. So with that said, let's dive into the live stream itself. And let me tell you about something I read that just made me laugh out loud this morning as I was reviewing all the research that I have made for me each and every night. So each and every night, we use about 7 to 10 different deep research LLMs to find out what are the biggest breakthroughs as it relates to AI improving human performance. And this morning when I came across this, it just made me laugh out loud. 83% of knowledge workers are now using AI chatbots daily. And guess how much time they're actually saving? Anyone want to take a wild guess as to how much this. 83% of knowledge workers, how much time they're actually saving? Anyone want to hazard a guess? Doesn't look like it. Okay, well, 2.8%. 2.8%. That's it. For less than 15 minutes a day, you have the most powerful technology in human history, and everyone is using it Slightly better than a spell checker. Look, you know, I spend hours every morning going through research from Claude Perplexity, Grok, chatgpt, genspark, all of them. Because I'm obsessed with one question. How do we actually use AI to enhance human performance? Not just to write emails faster or make fancy or pretty pictures, but to fundamentally upgrade how we think, how we work, and how we live. And here's what's wild. The gap between what's possible and what people are actually doing is insanely massive. It's like we've been given a Ferrari and all we're using it for is delivering pizza. You know, the real game isn't about having AI do tasks for you. It's about having an AI that knows you so well, I mean, knows you in and out, that it amplifies everything you already are. That's what we're building, and that's what I want to talk more about today. So here's the breakthrough that really changes everything. The most comprehensive workplace study ever done, 25,000 workers across 7,000 companies, just revealed that despite near universal AI adoption, right, almost everyone using it, productivity gains are basically nothing, zero, zilch, nada, nothing. That's on one side, right? But meanwhile, you have other companies like McKinsey, right? And their internal AI assistant, which they call Lily, is being used by 75% of their consultants, and they're projecting a $4.4 trillion in increased value creation. Big disconnect, right? You see the difference? The disconnect. And it's not about the AI. It's about the relationship with the AI. Hey, Eli, think about it like this. If I gave you LeBron James's basketball sneakers, would you be able to play like LeBron? Of course not. The shoes don't know your stride, your weakness, your potential. But what if those shoes could learn how you play the game, learn your style and adapt to your style and help you become the absolute best version of your athletic self? And that's really the shift that's happening right now. We're moving from generic AI tools that have no data, no context, no information about you, to an AI that does actually know who you are and considers you part of the context, the most critical context that there is. Now, three other breakthroughs from the past 24 hours prove that this is already happening. First, AI is learning to help us think better, not just do more. There was a fascinating study from UT Austin that just dropped, and this is kind of interesting. They gave students something called metacognitive prompts, which basically are prompts that just ask questions that make you think about your own thinking. That's why it's called a metacognitive prompt, because it has you thinking about your thinking. Simple stuff like, why do I already feel this way? What do I already know about this? What perspectives am I missing? Why do I believe in this so strongly? Et cetera. And the result was that students went from passive Googlers to active critical thinkers. I don't know how they measured that, but their critical thinking skills developed and grew. They explored broader topics, they asked better questions, which at the end of the day, asking better questions and being able to formalize your intent is really what prompting is all about. They actually learned these students instead of just consuming passively. And so huge difference after the fact in regards to what they were capable of doing. Those that used AI versus those didn't. So why does this matter? Well, the future, we see this over and over again with all this research. Why this matters is that the future isn't all about AI giving you answers. It's about AI helping you ask better questions. It's about AI helping you think more clearly, more strategically, more leverage. It's a cognitive partnership, right? Not cognitive outsourcing. Just like you wouldn't want to outsource, what is the core competitive advantage of a company. You would never want to do that because then you'd be giving your advantage away, right? You don't want AI to do all your thinking. You want it to co think with you. So second, AI is getting so good that it's becoming invisible. So 11 Labs just released Conversational AI 2.0, and for the first time, you literally cannot tell it's not human. So we've crossed this uncanny Valley. And Claude 3.5 now can use your computer so you can tie it in with Zapier to an mcp, a virtual mcp, and it will then click buttons, fill forms, run entire workflows, all while you sleep. And that's what I have going on here, right? This, this research that we're generating, it happens overnight, right? All this research that's happening is all overnight. And David Huff had an interesting question where. We'll get to that after we're done. Okay, so third health AI, right, is now past clinical, right? There's a company called Sleep Cycle and they're running FDA trials for AI that diagnoses sleep apnea using just your phone's microphone. Right? So I've never been in a sleep lab, but my girlfriend Kim has. And I know plenty of people that have had to go into sleep labs. And this will really replace it. It'll replace your wearable no sleep lab, just. Just an AI listening to you breathe. And get this, there's a new AI therapist that is now surpassed human therapists and treating depression in clinical trials and showing a greater symptom reduction than human beings. So this isn't wellness hype anymore. This is actually medical grade AI enhancement. And we're seeing it in Field after field, especially in the medical field, but in specialty after specialty of AI doing a better job than humans. You know, last week I. Every week I try and do at least one of these where I basically feed one of these new amo, if I could speak English. I feed these AI models everything, right? To see really about how well memory is developing, how well it can assess me, how well it understands me. And, you know, so what I'll do is I'll take every single, like strategic session recording that I had from the past six months, every coaching call, every planning session, every random voice note, even the ones I sent myself at 2 o' clock in the morning. And then I'll ask it based, I'll ask the LLM based on all of this. And now everything you know about me, how I think, how I speak, what's important to me, my values, et cetera, what am I missing in my current strategy. And when I do that, right. If the system learns me the way that it should. Right. And that's why I'm constantly testing. I did this recently with Claude and it identified three blind spots that I've been dancing around for months. And not because it's smarter than me, but because it could see my own patterns right in my own thinking. And because it's a pattern in my own thinking, I couldn't see it myself. So it's like having a conversation with a version of yourself that has perfect memory and zero ego. And you know, when I was working on it, and you know, this is part of like my daily routine at this point, that's when it, it just gets clear and clear to me that we're not building AI to replace human intelligence, we're building it to unlock our own intelligence, but intelligence that we often don't have access to. And I don't know if that, like, if that difference is clear to you, but we're building a copilot to help unlock the intelligence that we have. We have, but we don't have access to. Right? It's like the difference between your working memory, I mean, your working vocabulary versus your overall vocabulary. And, you know, you can spend the next year trying to cobble together 15 different AI tools, teaching each one right who you are from scratch every single time. I don't recommend it, but you could certainly do that. That's one of the things I do to constantly be testing these LLMs. But if you are starting to see already, right. How a system specifically designed to create that deep AI partnership that I'm talking about, if you understand how Valuable it is, then that's what you should be working towards. Getting AI to know you better than anything has ever known you. Where your AI doesn't just help with tasks, it helps you see yourself more clearly, think more strategically and execute with precision. You know, that's why I built Zenith Mind os, not another productivity tool, but actually a cognitive partnership system that has you performing at your best in any area that matters to you. So before you ask AI for help with anything, maybe next time do this before you ask it, right? Why don't you first tell it what you already know about the problem, Right. Second thing, what would be your typical approach? And third, what outcome would surprise you and then prompt, then ask for help. And if you do that, just even that, you'll see how different your response will be. Because you're not using AI as a crutch now, you're using it as a mirror that shows you angles that you couldn't see before. The idea of having this thinking partner is so valuable. And then the other two elements that you really have to start getting better at if you want to get the most out of AI is one, like thinking through outcomes and thinking through your. The intention that you have to get that outcome and what would be the best way. And second is getting really good at asking questions. Because if you're tired of using AI like a fancy calculator, when it really could be your cognitive copilot. I'm doing a webinar on Thursday. I'll drop a link in here tomorrow on it for Zenith Mind os. And in that webinar what I'll do is I'll show you exactly how to build an AI system that doesn't just know what you need to do, right? Doesn't just know what you need to do. It knows who you are while you're doing it. And that's the revolution is happening right now. So a couple of other big deals that came out recently is that Chat GBT co co partnered with Johnny Ives and they're going to be creating some kind of wearable, some kind of device. They also Chat GBT also just not copyright, not trademark, whatever the other one is. I forget that for an EEG device. So I have a couple EEG devices. I have the sense AI, I have the muse, I have the nurable. I'm waiting for the nurable. And anyway, so EEG is a. It measures your brain waves alpha, beta, theta and Delta. And so ChatGPT has trademarked or whatever, whatever you do when you make technology, I forget it, what a Patent patented an EEG device so that actually what can start happening is based on how you're feeling, based on your brain waves. Like, what are you capable of doing? And having AI this is. This is already in the works. Having AI re format. Reschedule your calendar in real time as it relates to your eeg. So in other words, chat GPT. We talked yesterday about them having some kind of wearable that has eyesight and hearing, right? So now there it's ambient AI. It can pick up everything around you. And now as it picks up everything around you, it's also measuring you internally, your brainwave states and can assess what you're capable of and what you're not capable of. And then based on that, alter your calendar, reschedule meetings, etc. That's already here. So with that said, let me see now if I can actually show you. Pull up one of the. One of my research things here, and then we'll call it a day. How's that? And thanks for those of you who are asking questions and saying hi, I appreciate that. So let me just quickly find something here that I want to share with you guys and then we will. I don't remember which one was the best.

Let's go to.

I guess we'll go to. Let's go to chat GPT. I guess we're going to chat GPT.

Key research.

And then we will go here.

All right.

And I want to.

Sorry, guys.

I'm going to get better at this. I know. And I am getting better at this. I'm just not getting better as quickly as I would like.

All right, so let's see here.

Where is the.

Where are the scenes? Is this the scenes? There we go. Now there's the cameras.

One more second, guys. I know this can be boring and.

Damn it. All right, let's see. Where are we at here? All right, where are the scenes? Scenes? This one was this overlays? All right, is it scenes? No.

All right, let me pull up this research. Oh, I can just do it in Streamyard.

Duh. All right, let me just share screen. Let me shoot Chat GPT. Share. All right, now let me see. What am I seeing here? All right. Oh, perfect. All right.

And I can actually do it like that. And I can go here and make this 150%. All right, so now you should see it. Let me pull this up here, I guess here. Now when I pull this up, will it go?

Let's see. Yeah, it will. Oh, sweet. All right.

All right, so when I did, this is just a Deep research. This is before like I started consolidating it all. And here we go. Right, so OpenAI Super Assistant Vision for ChatGPT Category Productivity Optimization Cognitive Enhancement OpenAI's plans revealed via court filing OpenAI aims to transform ChatGPT from a single chatbot into a super assistant that knows you intimately, manages tasks proactively and serves as your primary interface to the digital world. Weak documents outline a chatgpt that can see your screen, understand context and take actions on your behalf. From drafting emails and scheduling events to ordering dinner, the AI butler would be ubiquitous across devices and modalities, essentially becoming an ever present cognitive aid. Okay, why it matters. If realized, this could dramatically boost personal productivity by offloading routine chores and complex multi step tasks to AI. It's like having a tireless executive assistant for for every aspect of your life and work. A super assistant work could eliminate tedious business work, letting people focus on higher level decisions, but also raises questions about job displacement. Right, here's some information about ambient computing, right? No need to open an app or browse. Imagine interacting with an invisible assistant anywhere and how that could change daily routines. Right. Like that's what we're talking about. We're talking about ambient AI. The second one, right? AI that turns prompts into projects. If you haven't, if you have a Perplexity account and you're a pro, you now can have it create slides, write documents, you can do all the research and then create it. And you know, Perplexity I'm having create a deep research for me like this as well. Right. Also Perplexity helped me create a Python script that would run into Substack and find prompts for me. So that was kind of cool. And then another information about another health AI coach.

Right.

And they're just like popping up left and right and helping people more than personal trainers can. And character AI adds video and social features for creative chatbots. And let's see what other research here. Anything else interesting from chat? I don't see anything that. That is that new, that exciting. So with that, I have one last thing I want to share with you today and that's just I want to share another Zenith video that I got. So that's my goal here. Let me just pull that up real.

Quick.

And we'll be watching it for the first time together because I haven't watched it and that should be fun.

Cool. All right, so let's see here. Where are we? This was yesterday. There we go. All right.

And we're going to go to this.

I guess let's go to version 2. I haven't seen this one yet.

I have a YouTube link. Even better. And let's have a look and see what it's like.

All right.

Play a YouTube video.

This is gonna be interesting. All right, let's see.

Gonna get rid of this and let's see if I can just add this and then we will call it a day.

So let's see here. Present video file. Oh, even better. All right. Oh, no.

But that would be if I downloaded it.

Okay, that's not going to work. Okay.

You know what? I will save this for tomorrow, I think. I don't want to waste anyone's time. So with that said, make sure you tune in tomorrow. I'll be back home. Right now, I'm in a hotel in Orlando. But I appreciate you being here and the. Think about today as you. As you do whatever it is you do and you use AI in whatever way you use AI, think about in your interactions with it, what it would need to know about you for you to be able to do a much better job and have to tell it less. And then along those lines, let me share a. A suggestion that I've given a lot of people that seems somewhat obvious to me. And I think it's one of those things that after you've used AI for a while, you kind of like these kind of hacks kind of just fall into the way that you work. So one of the things that we've probably all had the experience of is that, you know, you get a prompt, you have a prompt, you create a prompt, whatever, to create an outcome. And so now you're interacting with whatever LLM that you're interacting with to get that outcome. And you put that prompt in and then you're massaging it. And until you finally get it the way you want it, most people then stop there. What you really should do. And this is like how you start becoming a bunch better prompter and therefore how you're able to get much more, much better results out of AI is that when you finish and you've gotten the output the way you want it, you should always end by asking Chat or asking Claude or asking Deepseek or asking Manus what would have been the prompt that would have generated this output from the very beginning. So that this way, every time that you actually are prompting and doing anything with AI, you're actually also improving your tools, your prompts, your ability to get the. The output that you want with the least number of steps, the least amount of feedback necessary and so on and so forth. Oh, nice. Michelle said. I love Zenith. My best decision in 2025. I'm using Streamyard to stream. So anyway, I'm here in Orlando. I'm at this event called Aspire. I don't know how much time I'm actually going to spend in the room, but I'm happy to be here nonetheless and support my girlfriend, Kim. And so with that said, I will see everyone tomorrow. I'm just super proud of myself for schlepping my stuff here and actually doing this. It just says to myself, it's a way of me telling myself that I'm serious about this and that even when I'm traveling, I plan on doing these. And I want to stay updated with what's going on in the world of AI to improve human performance. And I also want to plug that back to you, and I want you to be informed, because this is the doorway into our future. And I think the people that jump on these things early are the ones that are going to seize an advantage at a time when an advantage could translate into massive leverage. And so if you're performing at your best at a time when the world is in transition, odds are you're going to end up in a much better place than before the world went into transition. And so that's my goal, is to show you how you can, in increasingly improve your performance by leveraging everything that's out there as it relates to AI to improve your human performance. And so along those lines, to higher profits beyond. And I will see everyone tomorrow at the same time at the same place, to higher profits beyond. Rich Chefron, over and out.