Discover the best AI quotes from top experts

Experts can now use AI tools to increase their value to employers and clients.

AI is creating tremendous wealth and will continue to create tremendous wealth.

Building AI systems has been out of reach for most people, but that does not have to be the case.

Today, AI is in the hands of the high priests and priestesses.

So either I'm that brilliant and I'm that big of a genius that I can see more than all of the most brilliant minds the world's ever seen.

Or I'm really missing something.

And I think I'm gonna bet myself on missing something.

It became an addiction.

I didn't even realize.

I just fell in love with it.

I was just learning.

A few months later I thought, oh my God, the world will never be the same.

I was like, well, I got to do something in AI.

I'd be stupid not to do something in AI.

Well, what in AI?

I don't know.

If you want great output, spend twice the amount of time that you do using AI, spend twice that much time training it, talking to it, giving it context, and feeding it knowledge.

Really all AI is right now is a very, very large machine learning aggregated information database.

It consumes tons of information and it uses that information.

So when you want it to sound good, it's all based on the information you feed into that database, period.

AI is not just a tool. It's an execution machine. It doesn't get tired, emotional, or second-guess itself. And that's why it's going to win at everything that requires discipline.

You can't beat AI in execution. AI will never get tired, never forget, never make an emotional decision.

AI isn't going to replace people. It's going to replace people who don't use AI.

As a technologist, I see how AI and the fourth industrial revolution will impact every aspect of people's lives.

Our most important thinking machines will not be machines that can think what we think faster, better, but those that think what we can't think.

Artificial intelligence would be the ultimate version of Google. The ultimate search engine that would understand everything on the web. It would understand exactly what you wanted, and it would give you the right thing. We're nowhere near doing that now. However, we can get incrementally closer to that, and that is basically what we work on.

Artificial intelligence, deep learning, machine learning. Whatever you're doing, if you don't understand it, learn it. Because otherwise, you're going to be a dinosaur within three years.

I think AI should best be understood as something like a new digital species.

These waves of history are clearly speeding up, as each one is amplified and accelerated by the last.

In just 18 months, over a billion people have used large language models.

AI isn't something outside of this story. In fact, it's the very opposite. It's the whole of everything that we have created, distilled down into something that we can all interact with and benefit from.

You can buy attention (advertising). You can beg for attention from the media (PR). You can bug people one at a time to get attention (sales). Or you can earn attention by creating something interesting and valuable and then publishing it online for free.

The AI Gold Rush is here. Vertical AI agents are about to start replacing entire teams. The world as we know it is about to change forever.

Every few months, things just keep getting progressively better. But I don't want you to get left behind.

Artificial intelligence will reach human levels by around 2029. Follow that out further to, say, 2045, we will have multiplied the intelligence, the human biological machine intelligence of our civilization a billion-fold.

I took my first AI class in 2011 in school.

What people don't understand is I thought I was late to AI in 2011.

My background's in computer science and physics from Berkeley.

I took my first AI class in 2011 in school and we were doing some basic things with reinforcement learning and Python.

I just remember thinking, wow, this is so cool.

Right after college, I was so interested in this space, I started an AI company that prioritized basically real time speech recognition, natural language processing on live calls.

I hired a team of machine learning PhDs.

I raised a couple million in venture capital.

I was the founder and CEO.

And it was really, really, really tough.

We fight like hell for the positive use cases. Perhaps the most powerful use case and perhaps the most poetic use case is if AI, artificial intelligence, can be used to enhance HI, human intelligence, human potential, and human purpose.

AI is the defining technology of our times.

The development of full artificial intelligence could spell the end of the human race.

AI is one of the most important things humanity is working on. It is more profound than electricity or fire.

AI personalizes at a level humans can't touch.

Generative AI is the most important tech of our lifetime.

AI makes design effortless. Create faster.

Think of it this way: if ChatGPT, Gemini, Grok, etc., are like the brain, the AI agents are like the body that can actually go and do things on their own.

With AI marketing agents, you can personalize the copy, the content, and the experience. Converting more visitors into customers.

AI agents can tell you which campaigns create the highest lifetime value and which ones aren't profitable, so you can stop the bleeding and focus on profitability.

It was mind blowing to me to see ChatGPT with a single sentence you could do what used to take us a team of machine learning scientists, PhDs, and five different natural language processing models.

The things you can do with ChatGPT today, we would have to build specific models, one for sentiment analysis, one for classifying questions, one to figure out who is speaking.

Every little thing we would have to train our own machine learning model to do it well.

These large language models are so powerful, there's a temptation to say, 'Well, all these people are just going to slap them onto their websites, and it kind of turns the applications themselves into commodities.' But what I've got to tell you is that's one of the reasons why I didn't sleep for two weeks when I first had access to GPT-4.

I think AI is going to be the greatest force for economic empowerment and a lot of people getting rich we have ever seen.

AI is evolving at a pace that is painful.

And if you get yourself stuck in the I gotta use the best of, you're going to be stuck in a loop that you'll never get out of.

Figure out one or two or three things in your life that you think would bring you tons of value.

Doesn't matter how small or how big they are.

And AI-ify them and just start.

You can fear AI, or you can master it. The choice is yours.

The world is changing faster than ever before. You either move with it, or you get left behind.

AI is the biggest wealth transfer opportunity in history. For those who take advantage of it.

Now because we are still designing and building AI, we have the potential and opportunity to do it better, radically better.

As we build out AI, we can and must reflect all that is good, all that we love, all that is special about humanity: our empathy, our kindness, our curiosity and our creativity.

AI isn't separate. AI isn't even in some senses, new. AI is us. It's all of us.

I think we're at the cusp of using AI for probably the biggest positive transformation that education has ever seen.

You started hearing people say, my God, AI is going to replace the copywriter, replace the marketer.

And I just thought that was total bullshit.

I used to have six copywriters.

I have zero now.

We don't need them.

It writes far better copy.

But I've trained the hell out of it.

Every three years or so if you haven't changed your business model and what you do or evolved it, you're falling behind.

People want help with automation and integration. With the AI tools that are available to all of us for less than 100 bucks a month total, you can have all this stuff done.

50% of jobs will be displaced by AI, but that doesn't mean 50% unemployment. It means 50% of people will have to adapt. Or be left behind.

If all AI does is remove human inconsistency, bias, and hesitation, it will already be more valuable than any single person in a decision-making role.

The jobs of the future will be creative, strategic, and uniquely human. AI does execution, not innovation.

AI will create more jobs than it destroys. If we adapt.

There are hundreds and thousands of businesses looking to use AI agents this year, and you can build them. Even with no coding or tech expertise.

The most important part of all these AI business models is actually getting customers. If I could do this when I was 16 years old, anyone can.

One of my philosophies on using AI tools is that the goal is not to get the AI to write the stuff. The purpose is for this to be a personal newsletter where I'm sharing what are genuinely my own thoughts.

I'm only using AI in a way to augment that. The kernel and the seed of the idea is coming from me myself, and I'm using AI to help pull stuff out of myself rather than do the work for me.

One mistake I've often seen people make with AI tools is that they sort of think that the output is just going to be amazing straight up. But you saw in this example, I did like four or five different outputs. By doing these outputs a few different times, you come to one that's like 80% good.

Previously, writing my newsletter would take me about two hours because I'd be typing everything up. Now, it takes more like half an hour because I start off by speaking using Voicepal, and then I use AI to refine it.

I almost always use both Claude and ChatGPT and give them basically the same prompt. Claude, I have found, is for the most part, 80% of the time, just better at writing. But the nice thing about ChatGPT is that ChatGPT has the memory function, so it actually knows a lot more about me than Claude does.

AI helps me refine, structure, and clarify my writing, but I still have to do the final edits. That's where the real magic happens.

I hired AI to be my social media agent successfully and it worked.

All GPT really is, when you boil it down, it's simply a prediction machine.

Using the AI-generated script made scripting for me 10 times faster.

AI can turn scattered thoughts into structured, monetizable content. It's like having an assistant that never sleeps.

I spent two grand the first month on Perplexity Computer.

It's insanely expensive.

But if you see what I created, it's pennies on the dollar.

AI is fundamentally shifting how content is created and distributed. If you're not leveraging it, you're going to be left behind.

There's a massive opportunity right now for businesses and creators who understand how to integrate AI into their workflows.

If you think AI is coming for your job, you're looking at it wrong. AI isn't coming for your job. Someone using AI is.

In business, the biggest advantage isn't intelligence. It's execution. AI is an execution machine.

AI allows you to work on your business, not just in your business. That used to be a luxury only for people with employees.

There's so much opportunity with AI that it's like stealing candy from a baby. It's so easy.

AI levels the playing field. Hustle smarter.

Another major use case that we use AI for is for any sort of data analysis. Anytime there is a CSV or any kind of spreadsheet, I'm like, 'Glorious, I can just give it to ChatGPT or Claude and get it to analyze it.'

In the past, I paid thousands of thousands of dollars for automation consultants to figure out what we can automate in the business.

None of what they've said, of the $50,000 or so dollars that we've paid to automation companies and automation agencies and consulting people, comes close to the level of clarity that Claude just got through just a handful of exchanges with me.

This is incredible.

AI can schedule a courier, coordinate pickups, and track deliveries. But it still needs a human to guide it.

AI can take your job, or you can use AI to make yourself better at your job.

Most people are already addicted to a simple algorithm on TikTok. Now imagine when AI tailors everything specifically to your liking.

There's no AI engine that's bad. It's just trained on models that work or don't work for you.

AI isn't about replacing humans, it's about amplifying our abilities. Making things that were impossible before, possible now.

AI podcast services are one of the biggest opportunities right now. You can take long-form content, feed it into AI, and turn it into viral short-form clips automatically.

I've used Opus and similar tools and their automated clipping is not super great for educational videos.

It misses important context.

And then the short doesn't make sense anymore.

I think what's been missing is the intelligence of here's what should be included, here's what should be cut out.

My husband and I were out here in Salt Lake City.

We had family visit and they saw this canvas art in our home which we had used Midjourney to make.

My family members were like, wait, how did you do that?

So then I showed them AI and what you could do just by typing prompts.

And they sat there for hours.

I guess a lot of people don't know about this stuff.

That was the original motivation to start making content.

Sora is impressive, but it still gets a lot of things wrong. The world model is improving.

One day, AI will be able to process and integrate your entire life history into context for deeper understanding.

AI handles the boring stuff. Creativity's still ours.

What percentage of tasks will AI do, and over what time horizon? That's the question.

One of the most important things OpenAI is doing is putting powerful technology in the hands of people for free.

I don't think the world needs another copy of Google Search. We need a fundamentally better way to access and synthesize information.

If we allow AI to think before it speaks, its accuracy and ability to be a world-class tutor go up dramatically. This changes everything about scaling personalized education.

We can use Khanmigo to give every student a guidance counselor, academic coach, career coach, life coach. Which is exactly what education has needed for decades.

The biggest risk with AI isn't that it's too smart. It's that we trust it too much before it's ready.

An AI that executes without ethics is the most dangerous force we've ever created.

AI will be trained on the best and worst of humanity. The real question is: which side will it amplify?

To ensure that this new wave always serves and amplifies humanity, then we need to find the right metaphors for what this might become.

We all must choose what we create. What AIs we bring into the world, or not.

Autonomy is very obviously a threshold over which we increase risk in our society. And it's something that we should step towards very, very closely.

We have to think about all the risks AI could bring. Some will be obvious, others will be theatrical.

I think it is important that neither I nor any other one person have total control over AGI.

AI and surprise don't go together.

Whoever builds AGI first gets a lot of power. Do I trust myself with that much power? No one person should have it.

I think it's going to take the whole world working together to get AI right.

Success in creating AI would be the biggest event in human history. Unfortunately, it might also be the last, unless we learn how to avoid the risks.

Claude was basically our personal tour guide while we were hanging out in Taiwan. Any question that we had, we'd just speak directly into it, get a response, and read it out to each other.

Claude is slightly better for real-time, hands-free searching when traveling because you can just press one button, whereas with ChatGPT, you have to do multiple clicks.

Claude helped me understand the entire history of semiconductors in Taiwan in under a minute. That's the kind of real-time intelligence that changes how we learn.

It's not about one AI being better than the other, it's about how you use them together. You don't have to be stuck to one model.

AI isn't the enemy. The enemy is how you use it, or don't use it. It's just a tool, but if you don't know how to wield it, someone else will.

In five years, this game is completely different. That's why we have to learn AI now.

What if AI isn't just a signal, but a warning shot from the future?

AI is rewriting how business works. The question is, will you be part of the rewrite?

AIs will be infinitely knowledgeable, and soon they'll be factually accurate and reliable. They'll have near-perfect IQ. They'll also have exceptional EQ. They'll be kind, supportive, empathetic.

A lawyer and a doctor, a business strategist and coach. All in your pocket 24 hours a day.

The long arc of technological history is now in an extraordinary new phase.

If AI delivers just a fraction of its potential, the next decade is going to be the most productive in human history.

AI is to the mind what nuclear fusion is to energy. Limitless, abundant, world-changing.

AI is not just about making money. It's about leveraging it to create real value. The biggest mistake people make is thinking too small when AI is unlocking billion-dollar opportunities.

I expect that by the end of this decade, and possibly somewhat sooner than that, we will have quite capable systems that we look at and say, wow, that's really remarkable.

The road to AGI should be a giant power struggle. I expect that to be the case.

Compute is gonna be the currency of the future. I think it will be maybe the most precious commodity in the world.

We use AI for pretty much every aspect of our team's workflow and my personal life.

What a time to be alive that you can literally ask the AI to build stuff for you and also teach you how to build the stuff in the process.

Claude has such a good understanding of what I'm trying to do, what I'm trying to build.

It's suggesting things that I didn't even think of.

It's like you're getting a firmware update in your brain where now you know what's possible, which means you can then build even more things to help save you time and make you more money.

Don't use it just to chat, use it to win.

A lot of people are just using AI to chat.

But when I talk about using AI, I'm like, don't use it just to chat, use it to win.

The way that we use AI, we take it, we dig deeper.

We use it and we start to think, how can we win with it?

And if you don't know, ask it.

Here is the kicker that's made the big difference for me.

Make it look like it was created by a world-class graphic illustrator hired for a $25,000 agency project.

When I did that, the graphics started looking so much better.

Genspark just launched Genspark Claw.

Tell me this is not a reaction to OpenClaw.

You start to see all these tools competing with each other to have what's trending and what people want.

It's called the multiplication effect.

I'm able to multiply how I'm showing up on social by using artificial intelligence to do it.

If you don't know how to use it, ask the AI how you can use it for your business.

The marketing team that I would need to do the things that I'm doing now would be massive.

Was it not for these AI tools that are doing a lot of the content creation for me.

People who figure this out early will build companies that run 10 times faster than traditional businesses.

You have to get in now and you have to move at the speed of AI.

AI is moving so quickly right now, but you find out where you can win and then you go all in with that.

We're living in a time right now where we can go from idea to action and implementation like this and with just minimal tools and minimal budget.

A lot of them have gotten raises because the way that they're leveraging AI is not to just outsource little mundane tasks, answer this email for me, but to take things to a next level on the things that they have to get done.

If you're leveling up professionally, if you're looking to finally turn your passion into profits or really build your business, you need to start having more conversations with these different AI tools.

And just ask them, how can you help me win?

I need to make a million dollars this year.

Get me there.

Ask me 15 leading questions that will pull out of me what my passions are and what I want to do.

Then create a plan for me to make a million dollars.

Then work with the AI tool to implement, create, and act upon the plan.

Really celebrating you for a moment.

So many people will come out of this event saying this is the turning point in my life.

You don't need to be tech savvy to create a custom GPT.

You don't have to code.

You just go into ChatGPT, click on create under the GPTs, and then you let it know what you want this GPT to do.

I used to manually walk people through these questions and sometimes it would take a long time for them to come up with how to articulate the benefits and the deliverables.

So I thought, this is a process I've gone over with my clients over and over again.

Now I can tell this custom GPT to help my clients create a confidence card.

I think of it as input, process, output.

What are the inputs that this GPT will need.

What will it need to do with it.

And then what's the output.

It has taken this process that used to take so long for my clients to do to doing it really quickly.

Within seconds.

Answer three questions on what is your purpose, what's your desired outcome for your ideal client, describe your business.

You press submit.

Within about 10 minutes, it sends you a super in-depth, about a 40 page report called your Ideal Client Handbook.

Once ChatGPT first dropped, I dove into it with a little bit of hesitation honestly speaking.

I thought, this is an amazing tool but if it doesn't sound like me, then I don't want to use it for content.

But instead of dismissing it, I thought I'm going to see if I can train it on all the things I've been doing.

I love Granola for client sessions.

I use it for my high level private clients.

It summarizes it and I send it to them.

They have a summary of the notes in a very organized way, action steps they need to take.

And then they can have an actual conversation with the transcript.

If I'm about to get on a next call with them and I'll want to refer back to my last session.

If I have any questions about what did we talk about when we were talking about that one specific thing.

I just ask Granola and it'll remind me of what the specific words my client said.

No matter what we think about AI, this is the way the world is going.

Maybe even look at our own habits.

My habits have definitely changed.

If anyone says I know everything about AI, nobody does.

All of us feel behind.

Every expert I know is like, I might be ahead, but nobody's really on top of it because it's moving so fast.

If you're feeling overwhelmed, we're feeling overwhelmed too.

You're not alone.

If you can bring AI into what you're already doing, and even if there's any resistance, think of it as if I want to make a bigger impact for my clients.

Bring AI into it and help them to accelerate their results and they will be so grateful for it.

People want to go faster.

They want to accelerate their results, but they still want it to be quality.

If you can bring AI into what you're already doing, your clients will be so grateful.

I think people who are watching this are actually fans of AI.

But if there is any resistance, just think of it as if I want to make a bigger impact for my clients, bring AI into it.

I usually use the two big ones.

ChatGPT and Claude.

I was more of a ChatGPT advisor but now I'm shifting to Claude because for coding stuff it's a bit better.

I use Claude Code most of the time.

I sometimes use Cursor depending on the tasks.

I use it more like a collaborative tool.

Like another person next to me on my desk.

Where I can just be, what do you think about this?

I always try to think about it first and then share it with Claude and say, hey, what do you think about this plan?

And then I get feedback, I iterate a lot on that sense.

For people that don't really know how to code, it's good because it allows you to do things that you couldn't do before.

But if you just look at the code and you don't really know what's going on, it's quite hard to create something decent.

Don't use AI to replace your work.

Use it together with you, like a colleague next to your desktop.

To brainstorm, to keep thinking about the problems and keep thinking about solutions.

Use AI wisely while you keep thinking on what you do.

If you're trying to build any kind of online brand, personal brand, software product, any idea that you've ever had in the back of your mind that you want to create, this will help you bring that to life.

In an hour you can have the foundation for any idea you want to create right in front of you.

Instead of waiting months to get it out of your head and visualized, we can do that in a few seconds, a few minutes.

Clear decisions made and you can start moving forward and actually build it.

I found the first AIs were coming out like GPT-2, GPT-3.

It was kind of good at writing stuff.

I hated writing and that was the big missing piece.

People going from zero to 80% is a big deal.

Maybe that last 10, 20% is still not quite there and it takes some expertise.

But to go from zero to something is massive.

They were at their peak, had like 30% of the API usage was Copy AI users.

They grew to like 10 million users.

ChatGPT came out and kind of pulled the rug out of all the API shell companies.

Like actually we'll just go direct to market and take the whole market from you guys.

The ChatGPT image model came out.

It was actually way better than all the previous ones.

You could kind of actually do stuff with it versus just weird art.

AI can help you first write out how it should feel.

We kind of think first before we go into visuals.

The AI is always going to be lazy by default.

You get that average kind of output.

By asking some key questions you can really get out of that generic kind of design that everyone is now making with AI.

How do we get your thing to actually be yours and be unique.

There's always a tough line of how much information do you give it versus just saying make a brand about this.

You can give it two words and it can do a pretty amazing job.

The Nano Banana stuff got wild last year.

That was like another huge breakthrough.

I can take literally one coded prompt, drop it in and say waterfall woman waterfall feeling connected to universe.

Boom.

What I have in my prompts is at the end I can basically have these variables.

I'll say I actually want it to be four by five and type of graphic will do a YouTube thumbnail and then text overlay.

I didn't tell it anything and it just kind of figured out something.

If you get a prompt to have all your brand info, the details about how it looks, information about your target customer, your messaging, I gave it only a few words and it's able to make that.

You can do like 4K image quality.

There's no watermark in AI Studio.

Nano Banana 2 now is doing pretty good.

Fast is actually just as good.

I find fast is actually really good.

This is all just Nano Banana.

It took me about a weekend to do this.

How much would it cost in the old days.

We used to be in infomercial production.

They would spend 15 grand just for the talent.

Not to mention the whole crew, the actual location.

Hundreds of thousands of dollars.

The AI came up with really cool creative ideas.

That became again, it's that layer.

We actually did a photo shoot and they gave it to the photographer as inspiration.

Every photographer should be doing this.

In your consultations, getting clear on the brand, grabbing screenshots and getting the AI dialed in.

I have prompts that are filters.

That's how we go from the raw photos to these more stylized branded images.

One prompt can then become a filter you apply to your photos.

I'm using Notion a lot.

I build out libraries of prompts.

I create these auto-fill things.

Basically prompts within the database that the AI will go through, read the prompt and then write out all the sales angles that I could use.

Even if it's a prompt, it has potential for tons of public content, educational value that I could teach.

What I'm doing is trying to use my databases to get all that logged in.

My prompts have always kind of been more like skills.

I learned pretty quick that if I basically make my prompt an entire course, it could do so much more for me.

We're tokenizing all human knowledge and the flow of tokens moving is gonna be the new economy.

Everyone's gonna try to get a cut.

That little thin layer of just as the river's flowing, you're putting a little pail out, getting your cut.

The future of software is going to be similar.

You'll have super low cost maybe even free access.

But you're gonna get charged a tiny cut of all the tokens that you use.

What I saw back in the day with Copy AI is they were trying to do that and the margins were just too big.

Spending 400 bucks a month to summarize client meetings didn't make sense when you can just do it inside ChatGPT.

I create a lot of low-res stuff and then the AI is really good at cleaning it up and enhancing.

Anytime problems come up in my own business or for clients, I just try to solve it and create the prompt.

The ways in which AI is going to impact our world are endless.

Use AI to analyze your past deals, your past transcripts, your calls.

Ask it to pull out key audience insights.

What are your pain points, fears, motivations, desires.

Then build out your strategic positioning kit.

AI is very good at helping you overcome the initial inertia and helping you overcome that writer's block.

Don't take everything here at its face value.

Think about, does this make sense for my audience.

Does it help me to generate something relevant to them.

Does it ladder up to my positioning.

The quality of your input equals the quality of your output.

If you give it more specific instructions or more insightful content, it will generate better copy.

AI isn't going to give you 100% of what you want.

But it can give you maybe a 50 to 70% output.

And then you just got to spend a bit of time polishing.

This specific ghostwriter is trained on my top performing examples.

So that it knows how I write, what I sound like and the length of content that I usually post.

If you are building a GPT for LinkedIn, make sure you only have LinkedIn examples.

If you are building a GPT for emails, make sure you only have email examples.

The last thing you want to do is have a GPT that has LinkedIn, landing pages, social copy and you're gonna mix it all up.

I like AI to give me options.

Sometimes I have ideas.

I just want to tell AI, hey, these are my ideas, come up with something.

Sometimes I might want AI to suggest ideas for me.

I actually told AI to ask the user, what is the topic.

And then ask, would you like me to provide angles and context or would you like me to research and propose an outline.

If your content sounds like ChatGPT, if it starts off with it's not this, it's that, or if it has too many dashes, your audience is going to know that you are using AI.

I'm 100% sure they are not going to read on if they think it's a hundred percent ChatGPT generated piece of content.

Only when you define your tone of voice, style of writing, specific frameworks, only then you use AI to bring it all together.

Nano Banana is in fact really good at identifying a style and then recreating it.

If you are considering a drawing type cheat sheet, you might want to try Nano Banana giving it a reference example.

From testing AI for visuals on so many platforms, I find that the simplest tool and the most accessible one is actually Nano Banana on Gemini right now.

In the past when you use ChatGPT or any other tool to create visuals, it distorts the facial expressions.

But what I like about Gemini is it keeps it very closely.

Only when you build a strong human strategic foundation, then you use the AI tool.

You don't ever want to just copy everything and do everything AI says.

Use a critical lens and point of view to review the content even before posting.

AI can give you version one then you need to build on it.

We're doing more, we're supporting better, we're shipping faster than we ever have before.

And AI is really the heart of everything.

Every day I'm asking myself, what would I hire for?

And I think that's increasingly becoming a question that I'm not sure how to answer.

I've got a business coach, a life coach, a relationship coach, and a sales call coach.

And what I've done is I've put in my own limiting beliefs and my own personality stuff, things I do well, weaknesses, things that are a problem.

It knows I like shiny objects.

So if it sees me going down a path of trying to prioritize work that is not what needs to get done, but it's something that is cool, it'll push back.

It'll proactively find things that I'm doing and then it'll update the personality doc automatically.

In the future, it can use that to bring that up again.

The sprint we agreed upon said I'm going to be working on affiliate distribution this week.

And then it realizes, actually he's not doing any of that.

He's doing other stuff that is like new features for the product.

I'll make excuses.

It picks up on that.

And then it codifies it into this doc that just evolves continuously every single day.

Every time I do a call with anybody, it analyzes its transcript.

More importantly, it looks at trends of how I'm doing on all my demo calls.

My opening was seven out of ten, the discovery work I did, closing was shit because it always is.

It gives me a score of how well I did and what went well with timestamps, ways that I could improve, coaching points, top three things to focus on next call and comparison with previous calls.

Every single call I do, it automatically gets the transcript, analyzes it, and builds an entire call log.

I just want to do my work and I want all the analysis, planning, crunching, the data stuff to just happen.

I don't actually want to need to go in and do stuff.

It's automatically looking at everything I'm doing and coming up with stuff.

Content that I could create, features that I could plan, updating my personality documentation.

I don't want to ever need to go and say, oh, I just met with Bob, go take the transcript and analyze it.

It just does that.

My whole thing is I just want to do things and I don't want to need to tell it, make sure you remember this.

I don't even want to think about that.

I just want it to do that for me by looking at what I do day to day.

The issue with Notion or Notion's AI is that the only context it has is the Notion database.

You can't add the MCP for Kit.

You can't talk to Notion AI and say look at some content and then write a newsletter in Kit.

Because it can't do that.

We crawl your website when you sign up.

We try to infer what kind of business you have, what your business model is.

Then when you go to say I want a segmentation survey, it just builds it based on best practices applied to your business.

All we had to do was say, how do we get RightMessage to just build everything automatically from that prompt?

And that's the evolution.

Opus 4.6, which is the current one, is the worst model that they're ever going to have from here on out.

It's only going to improve.

You can do so much with this stuff and it's only going to get better.

I think there's a lot of negativity about AI around like the old thing of you call a phone number and it's like press one for this, press two for that.

We've all hated that.

But this is different.

How do you actually grow on Instagram using AI as a supplementary tool to amplify your humanness, not to totally replace you as a human being.

But to allow your uniqueness and what makes you special and different and what allows you and your company and your brand to stand out to be amplified with AI.

You can use ChatGPT and other audience research tools to do some digging and figure out what are people actually in the market for.

What are people actually struggling with.

The biggest thing that intermediate ChatGPT users are doing is they're mainly giving ChatGPT instructions.

What that does is it limits the capability of whatever AI tool you're using to whatever you can come up with.

It's basically like we are all coaching this little AI robot.

And ChatGPT is limited to how good of a coach we are.

A Formula One car might be the fastest race car in the world, but if you put 99% of us in a Formula One car, we're not gonna race very well because we just don't have the tools at our disposal.

Instead of saying, ChatGPT, I want you to write me 10 scripts for reels about Pokemon cards, you might say, my goal is to create engaging Instagram reels about Pokemon cards.

What do you need to know to help create the most engaging video possible?

Now ChatGPT is going to ask you questions.

It knows what it needs to know.

Instead of adding it yourself, you give AI feedback.

This is what is training your AI tool.

Just like a human being, if a school teacher took a kid's essay and just edited it themselves, the kid would never learn anything.

I think way too many people are just creating a new chat thread every single time with ChatGPT and they're not using projects.

When you use projects, ChatGPT can remember the different things it has learned about you.

I don't give ChatGPT a long paragraph of instructions.

I gave it those instructions once and then I told ChatGPT to save those instructions as the letter D, just D for description.

From now on I could just say D and then tell it the title and it could figure out the rest.

I feel that my audience resonates more strongly with the authentic, genuine carousel.

The biggest thing I'm using ChatGPT for is to simplify my writing.

Is there a way I could say this more concisely?

Is there an analogy that could explain this topic better?

There are some people who are AI clones, who are selling how to make an AI clone so that other people can make AI clones.

I think that can get a little tricky and a little scary.

The more AI avatars we see, which we're definitely going to see more every single day, the more there is a premium placed on the human.

If you had an AI avatar in 2021, holy crap, you were the first person in the world.

But now, the novelty might wear off.

I can put some prompts into Sora or Higgsfield or ChatGPT and get some really impressive, high-quality-looking content.

Think about the quality of an iPhone 10 years ago.

Now I can put some text into ChatGPT and get a photo that looks 10 times better than that in an instant.

I think about it like dating.

If you have millions and millions of hours of dating one person, they're going to know you really well.

If I have millions of hours of content to give ChatGPT, it's going to know me really well.

But if you're only giving it a few sentences, it's gonna know you about as well as someone five minutes into a first date.

Instead of saying write me a hook for a video about this new pasta recipe, which is a very generic prompt that gets a very generic hook, ask ChatGPT what it needs to know.

I have individual projects for each channel, maybe Instagram, YouTube, a podcast, TikTok.

And then I have sub threads or sub conversations within those projects for the different kinds of posts.

You could have ChatGPT help you come up with what the text is gonna say.

Or if you've given ChatGPT hours upon hours of content, scripts, blogs, YouTubes, podcasts, ChatGPT knows you pretty well.

You could ask it to go through everything you've ever said and pull out some hot takes.

Every week, every month there's new tools and new models out there.

I've been always with the same workflow.

Once you have your system, you can pretty easily rock and roll with that one for a really long time.

Sometimes we are overcomplicating stuff and it can be just as easy as using one or two tools.

And then you use that for the rest of the month.

The SEA framework stands for Structure, Example, and Analysis.

As long as you have those three on your custom GPT, it's going to be super useful and it's not going to break.

The reason I like to have structure in a prompt is because later you can come back and pretty easily edit what you did.

If you have a super chunky paragraph, when you notice something's not working, it's going to be a lot of noise.

You are not going to do it.

Something that I like to put in each step is to say, ask me if I'm satisfied with your answer.

Don't start the next step until I give you my approval.

Step one is always gather context.

Ask me enough information.

If you don't understand something, always ask me before generating assumptions.

That's going to reduce ChatGPT getting biased.

A mistake I see often is people saying, create me an email.

But how does a good email look like.

You don't even know that yourself.

Why do you think ChatGPT is going to know that.

Your job is to make that research and then have documents and frameworks that you are the ones that you want to use.

I use a combination of Nano Banana plus ChatGPT.

ChatGPT has more realistic texture.

Once I have my infographic with Nano Banana, I upload it into ChatGPT and say help me turn this into a more realistic texture.

I start with Nano Banana first because it's the best at writing and it has really small or almost no mistakes when it's writing.

ChatGPT cannot write that good yet.

The key to make an image look realistic is asking the model to make texture.

Some elements must look created by hand with colored markers.

The lines should be slightly imperfect.

If you do these three things, people are going to feel you're too robotic.

The first one would be using perfect parallel structures.

It's not about X, it's about Y.

One time it's okay.

But when you use it a lot, AI likes to do it quite often.

Generic statements with no examples.

Communication is key to success.

But if you can actually put a story there, it's going to hit harder and make it look more human.

Meta commentary.

Let me explain.

Here's the thing.

Here's the secret.

Here's the catch.

You just have to say the thing.

They are just models and they work with probability.

What has more chances to happen are those generic things.

So it's not its fault.

But if you are noticing that, it's much faster to just make those edits yourself.

I use custom GPTs more often because I create them and then a custom GPT you can share it.

All my clients have access to my GPTs.

They don't have to be prompting or creating something on their side or even paying for any tool.

Projects are more internal to you.

You also have these folders and all your chats are saved inside of this folder.

And it also uses all the memories.

My GPT is pretty simple.

But the work that I did was not in the GPT.

It was on my system to know what works.

My emails now convert really well.

But it was a work of analyzing the click-through rate, the open rate, what my best subject lines are.

Then we put those patterns into the GPT.

No wonder why it works so well.

I was also a mentee.

I studied with Matt Gray.

He does the same thing.

He liked Claude more.

But he used Projects and he had these files on his computer.

Here's my data of my business.

Our values, offers, the audience we want to attract.

A lot of people are trying to keep up with AI news and the new tools that come out.

You just have to be clear on something.

Once you build your workflow, even if new tools come out, you don't have to worry.

I've been with the same workflow like two years now.

You can play with something and if you like it, you can put it into your workflow.

But never feeling overwhelmed about the things that are coming.

The most important thing is not the tool.

Many people think it's the tool they choose.

But it's actually what you do in the backend.

The work that you do is in the backend on analyzing these examples.

That's the main thing you should always remember.

If you don't create the context at the start, then your outputs are going to be generic.

We've seen a big shift from prompt engineering and having mega prompts.

Now it's all about curating context and context engineering.

Once I have all of that context uploaded, I can start to think about heating with AI.

It is literally a matter of going to an AI and asking it to write a post about X.

Because I have all that context preloaded, the outputs are going to be my actual unique expertise.

It is all about repeatable systems now versus these mega prompts that we used to do back in the day.

People are actually getting frustrated and tired of seeing AI content now.

Now with AI, it's easier to copy than ever.

You can just put a thumbnail into Nano Banana and generate the exact same thumbnail with your face on it.

Notion is the organizational hub for me.

Everything lives in Notion.

Without Notion, I would just be a complete mess.

With the Notion MCP inside of Claude, you can link it up.

Make edits to Notion directly from Claude.

Granola for meeting notes and then Wispr Flow for absolutely everything, like writing.

AI can turn a one-person business into something that operates like a full team. That's the level of leverage we're dealing with.

Something that in the past would have taken, I don't know, a week to do, we can do in a couple of minutes, however long it takes the bot to go.

They're going to have a good understanding of how to come up with ideas for social media and use AI, but also to know when to not use AI and when to use themselves.

Because I think that's the problem now is that people are going too far into just like, AI can do everything.

And they're just like removing what makes them special.

So I want to make sure that that's not the case.

Maybe 15 minutes ago, we didn't know who this person was.

Now they're helping us create all kinds of great content with great hooks and pattern interrupts and all this.

It's amazing how much that changes how you look at things.

I really like Claude.

It's probably my favorite right now.

It probably just feels like it's the best overall in terms of the research, in terms of copywriting.

Anthropic is the company that owns them.

By the way, they have a $230 billion valuation as a private company.

Just think about how crazy that is.

You can use ChatGPT, of course it's good, and Gemini is good, and Manus is good.

But I really like Claude.

If I had to just pick one at this point, Claude.

A lot of times people think that you have to have like all these like special prompts or anything like that.

But the AIs are so good now that you really can just talk and explain what you need.

About a year ago to get the information that we're going to get now, you would have had to have a pretty in-depth prompt that really goes into every single thing.

But now that these LLMs are so good and so smart, you don't need to.

As long as you're pretty clear about that, then it's going to figure out what you want.

What I always recommend is let AI do 80% of the work and then you do 20%.

You want it to just get to 80% of good and then you're going to take that, you're going to put your secret sauce on it, you're going to put your voice on it even more, you're going to change some things, and then you're going to get it to a hundred.

You're going to do 20% of the work.

Don't think about AI as like getting that you have to get it to a hundred because that is just going to remove you completely.

If it sometimes it does but the way I always think about it is I just want it to get it to 80%.

I really think that Claude, Anthropic, is going to be the company that wins.

I don't think it's going to be OpenAI.

I think it's going to be Anthropic.

They are just so good at shipping stuff constantly and it's good.

If you had to pick one thing, I would get the Claude Max plan.

That's the best money you could spend because there's so much in there.

I actually have tried to go the other way and not use quite as many things.

What I would recommend is getting something like Wispr Flow, which is a voice dictation app that is really, really good.

For a while I thought it was like, whatever, it's not really that needed.

But it's so good.

It's so simple.

I thought of it, I'm like, it's too easy.

Like they have dictating on all these other things, but then you just get used to just hitting that function key and it's just works so well.

That's how you get the most out of YouTube.

OpenClaw, it's in there for a lot for content research.

Coming up with ideas and just having it like always going and looking through Reddit, looking through Drudge Report, all these different places to find ideas and really flesh them out.

That's like one of the best ways that I use it.

I don't think it's more than five bucks a day.

I don't think I've ever really gotten over that.

The LLMs do the heavy lifting.

All that we're doing is providing better structure, resources, and user-friendly ways to understand how to improve the outputs.

But the LLMs are still doing 99% of the magic.

The biggest limiter right now is just the developer experience of SaaS tools.

It's really only been dev tools that have good developer experience.

But now in this world of agents and agent-first workflows, every single tool needs top tier developer experience tooling.

You don't have to tell the AI exactly how to do something.

This is something that a lot of people have to unlearn.

Even as recently as a year ago, the whole trope was, it's good but it's only as good as you know how to prompt it.

That is not true anymore.

The big problem here is obviously that you don't want to just produce slop.

You want to know what the AI agents are doing and also how they are doing it.

As long as it acts like a human, then it's not going to get flagged.

Even that I think is a limiting belief.

It goes both ways.

I'm all about demonstrating.

There's so many people talking about AI.

I want to show what we actually have done.

Use the AI to process real ingredients instead of to generate stuff out of thin air.

The AI is not smart or dumb.

It's a mirror reflecting who you are.

If you're really good at something, it's gonna multiply the output.

If you suck, it's gonna multiply the amount of suck you do.

Prompting is abracadabra.

I don't like prompting.

Tell it like a smart coworker.

Talk to it. Give it whatever it needs to know.

The reason why it's working when I do it is not because I'm good on camera, it's not because I'm well known, it's because the AI knows I've given it access to all of our YouTube videos.

I'm in ChatGPT Atlas, which is only available right now on Mac, but it's coming to Windows.

And I actually have it as the default browser instead of Chrome.

I'm not touching anything. It's doing all of this.

While this one's working, I have two or three other workers and they're working too.

ChatGPT is lazy in agent mode because they've run out of processing power.

They're trying to. it's like the buffet's run out of food.

So they have to give everyone a little bit.

They don't admit that that's the problem, but it is.

I said, you are Frederick Vallaeys, you are Larry Kim, you are these other people I really respect, Igor Iwitsky. he's a good friend of mine.

And it goes ahead and does all of this.

It's a coworker that participates with me.

And then when I want to go into agent mode, I can turn on agent mode and it'll do stuff like a really good VA.

It's basically a VA right now.

It's early 2026, but I think by June, it'll be like a competent, strong digital marketer.

And I think by the end of the year, it's going to be better than me.

What matters is if you have a. let's say, Navid, you and I open a kitchen.

What's more important: the quality of the ingredients, the chef, or the knives?

I would argue the quality of the ingredients is the most important one.

If you have rotten, stinking, bacteria, maggot chicken, Gordon Ramsay can't make it taste delicious.

There's no amount of an amazing chef with the most amazing tools that's going to be able to do better than me, even though I don't know how to cook.

The raw proof on your phone, in a Zoom call, whatever. that's the key thing.

It doesn't matter what AI agent. You can swap out any of these other agents or models along the way. It doesn't matter.

We've been recording all the Zoom calls even before COVID.

Everything that we know how to do, we document, there's an SOP, we link it to other things that are related.

The more the agents get practice doing something, then we want to continue to improve these documents.

Why do you think we write articles on how to do stuff in SOPs, but we also have a verification checklist for every single SOP?

You get AI slop if you don't give it clear guidelines.

If you have clear guidelines on how you want it done. from Blitzmetrics or your personal brand. then you need to basically write this out for the AI.

We have a task library that I've been working on for the last 20 years.

And I knew that this day would come.

I knew that agents would. I didn't know it would happen in 2026, but I knew it would happen one day.

We're probably spending 20,000 dollars a month, but it's worth it.

Think about how many team members.

What would it cost for me to have a hundred workers?

Getting the AI agents to work costs us about 2 dollars and 87 cents an hour. Let's just call 3 dollars an hour.

You can get a Filipino VA maybe for 3 dollars an hour, but you can argue which one's better.

If I have a hundred workers and they're basically worth a thousand a month, it would cost me 100,000 dollars.

But I'm instead paying 20,000, which is a five to one.

The average I've seen across. you're super smart, so you get a 20 to one, but most of my friends get about a five to one.

We have five times the output.

It's not like this AI magic secret thing.

I don't think that's the most valuable thing.

I think the most valuable thing is: can you think like a manager?

Do you have critical thinking skills? Do you have people skills?

Can you reason with the thing?

Can you give your employee, human or not, feedback so it can improve?

You can't just get mad at the thing.

I see so many people saying, oh, it'll never be as good as me because I'm more creative as a copywriter.

You're an idiot.

It's a mirror reflecting who you are.

The thing that's limiting the frontier models right now is not enough electricity.

There's not enough electricity.

Even if you had enough chips, there's not enough electricity to turn the chips on.

Power to chips to model to agents.

I'm demonstrating agents here, but knowing these agents are weak.

I'll tell you something that I have not heard anyone else say. I think OpenAI is dead.

1.4 trillion dollars of capital commitments of OpenAI, they only did 20 billion in revenue.

How are they ever going to make the math on that work?

It looks like WeWork.

Apple chose Google to power Apple's AI.

I'm sure Sam Altman's so pissed off.

Google had it for years. They intentionally didn't want to do it because it hurt their revenues.

Google's earnings have increased quarter over quarter. They had 100 billion in profit last quarter because of AI.

My bet for the next two years, maybe three, is that Gemini is going to be the dominant model.

My bet for three years beyond is Grok because they're building X.

SpaceX bought xAI because they know that power is the limiting issue.

Power is cheaper to produce in space by an order of magnitude than on the ground.

I spend two minutes talking to it, and then I pasted the transcript in, and then I turn on agent mode.

Google Photos AI already does facial recognition.

It knows the location, it knows the phone, it knows the activity, it knows the objects.

I take this, I copy, and I can just start giving all this stuff to ChatGPT.

I'll say: you're the WordPress expert, you understand our content factory process, you understand how we elevate people and how I grow my brand and grow the companies that I advise because I'm elevating other people.

Elon Musk said there's no need for anybody by the end of the year to learn programming.

Why would you learn programming? Because the AI is going to write all the code.

I have a Model X and it drove me to Area 15 today and came back.

I was filming videos and eating food, and it drove me the whole way while I was on my phone.

Not because I'm an Asian driver. it's a better driver than me.

The AI chef literally can cook whatever you want.

It can cut the vegetables. It was making a coffee for me and it said, oh, you look kind of tired. Would you like to have a double espresso?

The coffee maker is communicating with the refrigerator.

The refrigerator is criticizing you on your diet habits.

The beauty is it doesn't matter who the winner is. Whoever has the proof. doesn't matter what AI agent, you can swap out any of these other agents or models along the way.

AI is an advantage for you over all this AI technical stuff if you learn from people who know how to build relationships, which is not a technical programming skill thing.

It wasn't until I started using AI where I actually became a bit more human and realized that it wasn't about me at all — it was about the people that I was helping.

The biggest mistake people make is assuming AI is competing with them. It's not — it's a multiplier for those who know how to use it.

Artificial intelligence is not a substitute for human intelligence; it is a tool to amplify human creativity and ingenuity.

AI is the biggest wealth transfer opportunity of our lifetime. Those who learn to leverage it will build empires.

I predict that, because of artificial intelligence and its ability to automate certain tasks that in the past were impossible to automate, not only will we have a much wealthier civilization, but the quality of work will go up very significantly and a higher fraction of people will have callings and careers relative to today.

I bought another $6,000 GPT through Alex Hormozi when he did his launch.

His AI is the best AI I've ever used for business consulting.

I can ask it the most complex, weird questions and it will give me an action plan of exactly what to do.

He probably saved my behind on this app because the way I originally had it set up, I was going to lose money.

I went into the Alex AI and it was like, no, you need to do a $1 trial.

The second I did that, the conversion rates went crazy.

AI images ended up in everybody's ChatGPT.

The idea of creating likeness got a lot easier.

Suddenly this big promise that I had, my whole entire sales page became obsolete.

Because why would somebody pay $3,000 for something they could go get anywhere.