REA Can Reverse Engineer Anything. Software Moats Are Over.
A product designer breaks down the viral 'Reverse Engineer Anything' repo, the 'software is dead' debate it sparked, and what's actually still worth building.
Posted
yesterday
Duration
Format
Essay
educational
Views
5.7K
65 likes
57 · 43
Big Idea
The argument in one line.
Reverse-engineering any piece of software doesn't make software worthless, because the value was never the code itself, it's the brand, distribution, and product judgment built around it.
Who This Is For
Read if. Skip if.
READ IF YOU ARE…
A builder or indie hacker trying to figure out whether AI reverse-engineering tools make their software idea pointless before they even start.
Someone with domain expertise (coaching, a niche skill, a small business) weighing whether to turn that knowledge into an AI product.
A non-technical founder who wants a framework for deciding what's actually worth building once AI can clone almost anything.
SKIP IF…
You're looking for a hands-on tutorial on how to actually use the Reverse Engineer Anything repo, this video is a reaction and think-piece, not a how-to.
You already have a clear point of view on software moats in the AI era and don't need the Maslow's-hierarchy framing to get there.
TL;DR
The full version, fast.
A viral GitHub repo called Reverse Engineer Anything can analyze almost any piece of software, binaries, runtime behavior, even closed-source code, and explain how it works, reigniting the claim that AI agents will soon build whatever software anyone needs on demand. This video argues that's backwards: cloning software has no value on its own, because a clone still lacks the brand, distribution, and trust built around the original. Using Maslow's hierarchy of needs, it splits what agents can build (safety and security tools) from what still needs human judgment (belonging, esteem, self-actualization). The conclusion: stop chasing trends, build around what you already know, aim for real quality instead of generic AI output, and build for a specific person you actually want to help.
Free for members
Chat with this breakdown — free.
Sign in and you get 23 free chat messages on us — ask for the hook, quote a framework, find the exact transcript moment, generate a markdown action plan. Bring your own key when you want unlimited.
Cold open framing the viral Reverse Engineer Anything repo and the 'is software dead' debate it triggered, plus the host's AI-product background.
01:02 – 04:05
02 · Explaining the Reverse Engineer Anything repo
Walks through what REA actually does: analyzing binaries, applications, and runtime behavior to reverse-engineer closed-source software, with viral examples of game mashups and a cloned Adobe Creative Suite.
04:05 – 06:55
03 · Why the 'software is dead' argument is flawed
Responds to Peter Levels' 'personal software' tweet by arguing a perfect clone still lacks the original's brand, distribution, revenue, and talent, so cloning alone creates no value.
06:55 – 11:46
04 · Applying Maslow's Hierarchy of Needs to AI products
Maps AI's current strength, auto-building personal software, onto the bottom (safety and security) tier of Maslow's hierarchy, arguing the top three tiers still require human judgment and intention.
11:46 – 14:31
05 · The crucial difference between builders and consumers
Contrasts high-agency builders who care about method with low-agency average consumers who just want the job done, framing the gap as the real opportunity for builders.
14:31 – 18:02
06 · Why value is in the product, not just the technology
Uses the Steve Jobs market-research quote and the iPhone's brand-driven success to argue the builder's job is to synthesize value from multiple sources, not just find and build the valuable thing.
18:02 – 20:22
07 · A realistic vision of the future software landscape
Compares the 'agents eat everything' fantasy to an old city skyline where legacy and new technology coexist, predicting web apps, agents, APIs, plugins, and SaaS will all keep existing side by side.
20:22 – 23:39
08 · Shifting focus from software to products
Argues people buy products, not software in isolation, and that a hundred times more builders means a hundred times more businesses, not the death of software as a category.
23:39 – 27:16
09 · Four principles for building valuable AI products
Lays out the four-part checklist: build for an outcome, build from what you already know, care about quality over 'AI slop', and build for a specific person you want to help.
27:16 – 28:49
10 · Final encouragement for builders in the AI era
Closes with a direct call to keep building even through doubt, and points viewers to his AI product community for more help.
Atomic Insights
Lines worth screenshotting.
A GitHub repo that reverse-engineers almost any software crossed 100,000 stars in about a week, and people used it to recreate the entire Adobe Creative Suite as free, open-source clones.
Cloning a piece of software has no real value on its own: a perfect clone still lacks the original's brand, distribution, revenue, and talent.
AI agents can plausibly build on-demand personal software for safety and security needs, but the upper tiers of Maslow's hierarchy, love, esteem, self-actualization, still require human judgment and taste.
Most people are bad at clearly communicating what they actually need, which breaks the idea that an agent can just auto-build the right software for them.
Builders are high-agency and care about method; the average consumer is low-agency and just wants the job done, which is exactly the gap a skilled builder can fill.
The iPhone's value wasn't the touchscreen technology itself, it was the Apple brand, the intuitive interaction design, and the paradigm shift it represented.
The real job of a builder isn't finding a valuable idea and building it, it's synthesizing value from multiple sources into the most intuitive product possible.
Software alone was never that valuable: brand, experience, mission, and outcome are what people actually pay for.
The realistic future of software looks like an old city skyline, legacy and cutting-edge technology coexisting, not one category of software eating every other category.
A hundred times more people can now build software, which means a hundred times more businesses and successes, not fewer.
In 2026, 'pretty good' AI-generated software is no longer good enough, the bar has risen, and mediocre output reads instantly as AI slop.
The four principles for a valuable AI product: solve a specific outcome, build from what you already know, obsess over quality, and build for someone you actually want to help.
Takeaway
The Code Was Never The Moat
PRODUCT OVER TECHNOLOGY
AI can now reverse-engineer and clone almost any piece of software, but the clone is worthless without the brand, distribution, and product judgment built around the original.
02Explaining the Reverse Engineer Anything repo
Reverse Engineer Anything (REA) is a GitHub repo, not a product: it analyzes binaries, applications, and runtime behavior so an AI agent can explain how closed-source software actually works.
The repo crossed 100,000 GitHub stars in about a week, and the examples people shared ranged from recombining game code into new mashups to rebuilding the entire Adobe Creative Suite as free, open-source clones.
Reverse engineering commercial software this way carries real legal exposure: you need legitimate authorization to analyze tools you don't own the code for.
03Why the 'software is dead' argument is flawed
The 'software is dead' claim, popularized by a Peter Levels tweet, argues personal AI agents will just build whatever software a person needs on demand, making paid software obsolete.
Cloning a piece of software has no inherent value on its own, since a perfect clone still lacks the original's distribution, brand, revenue, and talent.
If cloning has no value, the real question isn't whether AI can copy software, it's what makes the copy worth less than the original.
04Applying Maslow's Hierarchy of Needs to AI products
Maslow's hierarchy of needs splits into a bottom layer (safety and security) that an AI agent can plausibly spin up personal software for on demand, and three upper layers (love and belonging, esteem, self-actualization) that still require human judgment and taste.
An agent can build you a health tracker or a job-search tool, but it can't build you friendships, confidence, or a sense of purpose, those require a human actually doing the work, even if software supports it.
People are also bad at articulating what they actually need, so expecting an agent to infer and auto-build the right software for the upper needs levels is unrealistic in the near term.
05The crucial difference between builders and consumers
Builders are high-agency and want something done the right way; the average consumer is low-agency and just wants the job done, with no interest in the method.
Most consumers aren't using AI agents yet, even as tools like personal assistants become more common, which is exactly the gap a builder can fill by translating AI capability into something a non-technical person can use.
The opportunity is building things that are easy to use, deliver a specific outcome fast, and hide the AI/agent machinery entirely rather than exposing it to the end user.
06Why value is in the product, not just the technology
The iPhone didn't succeed because touchscreens were new technology, it succeeded because of the Apple brand, the intuitive interaction design, and the paradigm shift away from physical buttons, the value sat around the technology, not in it.
The real job isn't finding a valuable idea and building it, it's synthesizing value from multiple sources and building the most intuitive product around that synthesis.
A reverse-engineered clone of Photoshop can't clone Adobe's decades of brand equity or the millions of creative professionals who already trust and depend on it.
Software alone was never that valuable; brand, experience, outcome, and mission are what the market actually pays for.
07A realistic vision of the future software landscape
The fantasy is that AI agents eat every category of software; the reality looks more like a city skyline, old buildings standing next to new ones, because legacy and cutting-edge technology coexist rather than one replacing the other.
Expect web apps, mobile apps, MCP servers, agents, APIs, plugins, skills, and SaaS to all keep existing side by side, just with their shapes evolving to become more agent-native over time.
08Shifting focus from software to products
People don't buy software, they buy products, meaning everything that surrounds the code: the experience, the brand, the outcome it delivers.
Stop treating any single technology shift as a death sentence for a category of software; instead treat it as a prompt to evolve the shape of what you're already building.
Software is becoming more varied, not less, because building is no longer limited to trained developers, which also means a lot more businesses and successes will come out of this shift.
09Four principles for building valuable AI products
Build a product that gets someone a specific outcome, not an app stuffed with hundreds of features, aim at the upper levels of Maslow's hierarchy where human judgment still matters.
Turn what you already know, your expertise, passion, or business knowledge, into a product; that domain knowledge is something AI cannot replicate and people already pay for it in other forms.
Care about quality: agents can produce decent-looking software by default, but 'pretty good' no longer clears the bar in 2026, aim for the top 1% of execution instead of generic AI output.
Build for a specific person you actually want to help, not just wherever the current hype says the opportunity is; chasing trends over genuine fit is a common and costly mistake.
Glossary
Terms worth knowing.
REA (Reverse Engineer Anything)
A GitHub repo and toolkit that lets an AI agent analyze an application's binaries, runtime behavior, and code to explain how it works, even for closed-source commercial software.
Moat
A durable competitive advantage, something a product has that's hard for competitors to copy, such as brand trust, distribution, or an existing customer base.
Agent-native
Software designed from the ground up to be used and operated by AI agents rather than humans clicking through a traditional interface.
Maslow's Hierarchy of Needs
A psychology model ranking human needs in five tiers, from basic safety at the bottom to self-actualization at the top, used here as a lens for what AI agents can and can't build for people.
Personal software
The idea that instead of buying apps, people (or their AI agents) build small, disposable software for themselves whenever a specific need comes up.
Resources
Things they pointed at.
01:02toolReverse Engineer Anything (GitHub repo)
05:07linkPeter Levels tweet: 'software is dead, it's now the era of personal software'
“People don't know what they want until you show it to them. That's why I never rely on market research.”
Steve Jobs quote, instantly recognizable, sets up the whole thesis→ newsletter pull-quote↗ Tweet quote
16:01
“Your job is not to find the valuable thing and build it. Your job is to synthesize value from multiple sources and build the most intuitive product around it.”
the core thesis stated as a clean, quotable line→ IG reel cold open↗ Tweet quote
12:08
“The average builder wants to build something that does the job the right way. Whereas for an average consumer, they just want the job done.”
sharp contrast line, easy to screenshot→ TikTok hook↗ Tweet quote
24:26
“We do not want to be building software that looks like AI slop in 2026 and beyond, because the bar for software is so much higher now.”
“Cloned software has no value. So what's the point in reverse engineering anything if the value of reverse engineering it and cloning it is kind of worthless?”
states the counterintuitive core argument directly→ IG reel cold open↗ Tweet quote
The Script
Word for word.
Read-along
Don't just watch it. Burn it in.
See every word as it's spoken — crank it to 2× and still catch all of it. The same dual-channel trick behind Amazon's Kindle + Audible.
17px
metaphoranalogy
If you're watching this, then chances are you've seen this repository called Reverse Engineer Anything. go pretty viral recently and you might have also seen conversation around software being dead because of it but i have a little bit of a different perspective on this now people often use this as an excuse to not even try and build anything but in this video i want to help you to understand why reverse engineering is important and what you should actually be building instead because there is still a huge amount of opportunity in the ai space and we are only just at the beginning of what is possible with this technology and so we're going to be breaking down what is reverse engineer anything so you can understand a little bit about how it works and what it is, answer the question, is software dead, and at least give you my take and my perspective on it, and also help you to understand what you should be building with AI agents instead.
If you don't know me, my name is Chris, and for the last 15 years I've been designing apps and advising startups on product and design, and I've helped hundreds of people inside my community to turn their knowledge into real AI products, to build it with AI tools, and then launch to real paying customers. With that said, let's jump straight into the video and break this down in a little bit more detail.
If you haven't seen it already, then this is the GitHub repo that's been going viral. This has reached over 100 ,000 stars in just a matter of about a week, I think. And basically, this repo allows you to reverse engineer anything but...
It's mainly for reverse engineering pieces of software. So reverse engineering applications, binaries, and runtime behavior. Now, I don't necessarily understand the details of things like runtime behavior, binaries, that sort of stuff.
My background is not a developer, but I want to help you understand why something like this is important, even if you don't code or you're not a developer, even if you're a non -technical person just wondering how you can build a business or how you can make the most out of this AI opportunity that is in front of us right now.
Because you may be wondering why is it even important to be able to reverse engineer things and surely we've been able to do that already anyway. Well this is slightly different because this tool reverse engineer anything actually allows you to go into an application to figure out all of the details of how everything inside that application works even for applications which their code bases may traditionally be kind of private.
Now there is a ton of legal issues and sensitivity around this. And essentially you must legally obtain and be authorized to use these commercial tools in order to reverse engineer commercial software. But there is a bunch of examples of people doing crazy things with this reverse engineer anything repo in order to build their own software and essentially recreate some of the biggest software companies that exist at the moment.
So here is an example of the sorts of things that are being shared after this reverse engineer anything. repo kind of came out and it's basically showing you how you're combining these multiple different sources and their code into create one single thing and most of these examples here are kind of taking video games and then combining kind of two parts of it essentially using reverse engineer anything to understand these individual parts of the code base and then combine them together into something completely new And this other post came out recently as well, basically showing someone who had cloned and rebuilt the entire Adobe creative suite of software.
So you have all of these tools which are basically copying Photoshop, Illustrator, Premiere Pro, Lightroom, all of these Adobe products and rebuilding them as open source software that would be completely free to use. so on the surface i think you can see that there might be interesting applications of something like the reverse engineer anything repo where you take a piece of software that you want to implement some part of into your own software reverse engineer how it works and then build it into your piece of software or at a further level kind of take these disparate parts of different pieces of software and combine them together into something new i think one of the actual opportunities that comes out of this is being able to combine different parts of software into a single software tool that you use in side of a business I do have another video which I did on this on how I'm going about building a custom software platform to run my own business combining different features of things I use every single day like notion
And buffer just into one piece of software that I can use. I want to take this a little bit further and talk a little bit about this argument that software is dead. And this is kind of where people have taken this repo into this discussion around software being dead.
And is software even worth anything anymore? And there has been a bunch of podcasts talking about this. And I think the kind of prophecy in the AI space is very much that like AI agents will take over all of software.
basically because they'll be building kind of generative pieces of software for users on demand as they chat with their AI agent. But I think this is a very short -sighted and narrow view of the implications that AI agents and coding agents and repos like Reverse Engineer Anything actually have on the landscape of software.
And I think personally that it introduces a ton more opportunity than people think. And what this means is that more than ever, if you have knowledge, if you have expertise, or if you have a passion in something where you can help someone, you can essentially build your own products with AI that will help you to either launch a business or make some money or just genuinely help some people, even if you're just building something that can help yourself.
And this was kind of the start of it really from this Peter Levels tweet where he said software is dead, it's now the era of personal software and everyone just builds what they need whenever they think of it. Which again I think is a little bit short -sighted and he actually added this comment on here which says all their AI chat app or AI agent will just build it whenever they need it.
But I honestly kind of disagree with this take a little bit and I want to explain why. And so first let's look at what does reverse engineer anything actually mean for software. Well, it means that AI can understand how any software works, but actually cloning the software has no real value.
So if cloning the software has no value, if you clone something like the Adobe Creative Suite, for example, and then give people access to those tools, you've essentially just rebuilt something that already existed. All you don't have is all of the other assets that exist inside of that business already. You don't have the distribution.
You don't have the brand. You don't have their revenue or their talent. You don't have so many things beyond just the software that go into making something like Adobe work.
And so cloned software has no value. And so what's the point in reverse engineering anything if the value of reverse engineering it and cloning it is kind of worthless? And I think what people like Peter Levels are kind of insinuating here is that if you can build anything with AI agents and your agent can kind of spin up software on demand as and when you need, the actual value of paying for software is kind of Pointless, you don't need to pay for software because you can get your agent to build it at any time.
But I think there are some massive flaws in this because the idea that everyone just builds what they need, even if it's their agent building what they need, relies on you being able to clearly communicate to your agent what you need, or your agent being able to assume what you need based off of the context and the information that it already has.
And for this, I actually want to talk about Maslow's hierarchy of needs a little bit, definition here that we need to be clear of is and and that is around everyone just builds what they need so what they need um fits into this maslow's hierarchy of needs so i think what they need when people say that software is dead and and that your agent can build whatever anyone needs is that i think this is talk just talking about this second level of needs because there are so many other levels of needs above this which you actually need to have real intention around creating or actually being able to achieve as a human being that go beyond one agent kind of actively or practically proactively build for you in a piece of software.
So personal software could absolutely build something where you get, you know, you have a health tracker, you can find a new job, you can buy a property quicker, you can manage your family life, or you can improve your social ability. And I think this is really what people are talking about when they're talking about how, you know, agents can just build personal software for you based on whatever you need, is really this safety and security layout of Maslow's hierarchy and needs.
Because when you get into the other levels and needs, you start to have to add human judgment and human taste and human intention on top of those kind of basic functional tasks that you need to accomplish. So if we take some things from each of these parts of the hierarchy of needs, we say friendship. AI can't build you an app to make friends.
You have to go and do that. Now, you could create an app that would help you to make friends, but the actual act of going and making friends is a very human act. And this is something that agents are not necessarily going to be able to be good at creating an application for.
And so there has to be at some point, some level of human judgment and interaction with being able to build a piece of software that allows or enables people to do that better. Again, with confidence, achievement, respect of others, these are all human needs that go beyond these functional needs that your personal agent could just spin up a piece of software for and again when you look at this self -actualization level In the Maslow's hierarchy of needs, you have things like creativity, spontaneity, experience and purpose and meaning.
And now some of these things you could argue that, you know, your agents would be able to build software for on the fly for you and they're just custom pieces of software for every single person. But I think people who have experience with building creative tools are going to be able to build a better creative tool for people to achieve something specific.
in terms of enabling them to be more creative. And the other part of this is that I just don't think people are very good at communicating what they actually need. Or even if they do, they can't communicate it clearly.
And so in order for this to be true, what we would have to rely on is an AI agent being able to assume that we need help in this kind of morality space in our lives based on the information that it knows about us, then to spin up a custom app. in order to help us achieve that specific thing. So I think at the moment we're talking about agents and software that exist very much in this safety and security level of Maslow's hierarchy of needs.
And where there is huge amounts of value and huge amounts of opportunity still is in building for these other levels of Maslow's hierarchy of needs is for building software that directly addresses these other human needs that we have. And so what else does RIA mean for software? Well, it means that AI can understand how a software achieves a specific goal.
So I can ask it, okay, how does Duolingo retain users in their app? and it will go in and understand like how they calculate all the values based on when a user was last active, what subscription plan they activated, how long they kind of did their activities for, all of that information in order to help you achieve a goal within your software.
So you can kind of reverse engineer how a piece of software achieves a specific goal. And of course, if agents can reverse engineer any of this software, that also means they can then build it. So agents can create anything on demand.
based on these user needs and again going back to Maslow's hierarchy of needs it's based on user needs but it's based on these more functional levels of the user needs and so the first my first kind of tip in terms of if you want to look at opportunity to build is to think about how can you build software and products that address these top three levels of Maslow's hierarchy of needs because there is going to be a vast amount of successful software that fits in addressing those specific needs that AI agents just won't be able to spin up custom on demand for individual people now for a moment let's just play devil's advocate let's say okay software is worthless because if we can reverse engineer anything then any piece of software can be created infinitely on demand and kind of achieve and accomplish any specific goal for any person Well, I feel like there are some differences between what people are actively doing as builders in this space and what the average consumer is doing.
And I feel like we're disconnected from these two things. And especially if you're building for consumers, you need to understand consumer behavior in order to understand where they are coming from. And so the difference between a consumer and a builder is that generally like someone who's building always...
or is kind of thinking about entrepreneurship or building their own things, they are more high agency. The average consumer is more low agency. They're not taking actions to proactively build this stuff for themselves.
They're just going about living their lives, solving any problems that kind of come up as they live their life. Generally, the average builder wants to build something that does the job the right way. Whereas for an average consumer, they just want the job done.
They don't need it to do it in a specific way, they don't need it in a specific format, they just want their job to be done. They also don't really care about the method, whereas a builder would care about the way that it works. And also, the average consumer is barely using AI, let alone agents.
Now this is getting more and more commonplace now that people are having access to personal agents like Metamuse and ChatGPT Dots, but the general population is just still not there yet. for someone who's actually building they're really aware of the way that impact agents can impact life and so this actually points to the opportunity that we have in front of us because if you are a builder you have a massive opportunity to build for someone who is in this space as an average consumer because you can translate all of this stuff to fit their world so what does that actually mean that you need to build so we need to build things that are easy to use and easy to understand we need to build things that complete a job or get users a specific outcome we need to build things that do it quickly and easily and get as best a result as possible for people and we need to be able to abstract ai and agents into something that just works rather than relying on a user to understand or even know how an agent even functions and i think this is also why product will still be relevant in the next decade
okay so we can understand like how the average builder is different to the average consumer in terms of you know their approach to ai and it also and also we can understand what this means that we need to build and how we can build for this top level of maslow's hierarchy of needs and why there might still be value in software that fits in those specific areas but i want to present this kind of conundrum here as well because i think software is dead is a very attractive thing that people talk about because if you can reverse engineer any software what is what is the value in software anymore and this is all based on people thinking that value is derived specifically from time and effort that has been spent on something but i don't think this is always true and i want to mention this quote from steve jobs here because i think this is really important when we're talking about this argument of is software dead you know if ai can build any software you know immediately then what's the point of building software anymore you know software businesses software products or like selling software to people anymore is that people don't know what they want
until you show it to them that's why i never rely on market research and i think this people don't know what they want until you show it to them is really really true when it comes to ai products and when it comes to products in general if you think about the iPhone, for example. We already had mobile phones.
And the value was not in the thing itself. The value was not in, you know, the value is not in AI or AI built software. The value is in everything around the technology.
The value is in the Apple brand, the brand of Steve Jobs, the intuitive interaction of the iPhone itself. The fact that they completely changed a paradigm that was pre - prevalent in the industry of using physical buttons on phones to change it to be fully touch screen.
The value is in everything around the technology, not the technology itself. And this is why I still think there is such a huge amount of value in building AI products and building AI software, because your job is not to find the valuable thing and build it. Like we're not out here trying to just find any old opportunity and then just build where the opportunity is because that is going to lead to something that will just be very easy to commoditize and for anyone else to build.
Your job is to synthesize value across multiple sources and build the most intuitive product around it in order to help a specific person achieve a specific outcome. And this is why I believe that product is still so important in this age. of ai and that if you have knowledge if you have expertise or a passion you can really turn that into an ai product that can still turn into a massive massive business and now yes there are tons of other things that you need around that than just the software because i don't think just software by itself even has ever really been that valuable it's all of the other stuff around it it's brand it's experience it's in the outcome that it gives it's in the mission The purpose that people rally behind.
And yes, somebody can clone Adobe Photoshop by reverse engineering it, but what they can't clone is that brand equity that has been built over decades of time. They can't clone the millions of creative professionals who use and trust those Adobe creative tools inside of their business and inside of their work every single day.
That is something that you can't clone and that you build over time. And yes, I agree that software is not... that software by itself is not a moat anymore you can't just build software and expect nobody to copy you and build it the moat the moat is the same thing that has existed in products and business for the last however many decades is all of the things around the technology so your job is to synthesize value from multiple sources build the most intuitive products around it and there's this argument that well won't software just be created from user intent you know by their agent they say oh i want to do this and then your agent is going to build infinite pieces of software for you that can do any number of jobs and i think what i would look at is this comparison this is something that i use and i talk about a lot in my community is that our view of this imagined kind of utopia in the future if you think about what a city looks like and i see this very much the same as like ai software or the ai industry and ai landscape is that in the cities we imagine this future of a city as all of these incredible kind of like sci -fi buildings everything looks the same across the entire city everything is kind of high -tech
And there is this overall aesthetic that pervades everything in this city, where the reality is actually very, very different to that. This is just a picture of the landscape in London. You can see in the background there is a cathedral from hundreds of years ago.
There is modern skyscrapers in the background. There are these columns here from buildings that were made hundreds of years ago. There is cars and vehicles that exist from the last however many decades.
There are roads and road markings that reflect the most recent road regulations. And what you get is this combination of multiple eras in this messy concoction of all of these things being present. And I think what this means is that, and how this relates to software, is that the fiction we believe is that agents will eat everything.
But the reality that will actually exist is that multitudes of software will exist. A massive ecosystem of new and old technology that exists together. where people still use all of these different things from web apps to mobile apps mcp agents apis plugins skills sas all of these things are still going to be viable businesses moving forwards now the shape of them might change yes you know you might have a sas that needs to be agent native instead of a traditional sas that you click around a dashboard you might have a different shape of plugin that evolves over time with these ai agents you might have different skill formats that exist You might need to build a web app that sits directly inside the Claude or the ChatGPT interface rather than being on a web browser.
But the reality is going to be that all of these different types of software are going to exist together alongside each other and all of the different types of them are going to be able to be successful in some way. So what does this mean if you're building with AI? Well, I think we need to stop thinking about software.
and start thinking about products and this might seem vague but i think we need to start stop thinking about software in isolation by itself like a specific piece of technology is going to kill another piece of technology or that a specific shape of software is going to die because the reality is that people don't buy software people buy products because of everything that surrounds the software and once you stop thinking about software and start thinking about products And I'm going to show you exactly what I think you should build in a second as well.
I think you'll be able to build such a better piece of software from that. So software is going to become more varied. You're going to have a set of all of these different types of software, from apps to agents to MCP to APIs to plugins to skills.
And you can combine them in a ton of different ways to create something that is still going to be valuable, even in this era of intelligence with AI agents. And so I would say for the first instance, do not be put off building things. take your passion take your expertise take your knowledge and build something from that using these ai tools even if you have to pivot and change the shape of that over time is still worth building because there is a huge amount of opportunity and fun just in building these products that can really have a massive impact on people's lives also there is going to be a hundred times more software and this is because instead of just developers being able to build software now everybody can build software and there's also going to be a hundred times more businesses and successes off the back of this as well because the value from building custom software is going to be something that everybody can reap the benefits of rather than just being confined to a few specific people who are trained in development.
or confined to a few specific companies who build software products for people to use and this is really why i think that if you're a small business or if you're a person who has knowledge or is an expert in something or has a specific set of skills i really think this is your time to shine if you can learn how to build using these tools you can take that knowledge take your skills turn it into something super valuable as a product for someone that someone is going to pay money for in order to get a specific outcome Or just build something that's going to leverage that expertise in order to have a massive impact on the world and take what you already do and you already know and turn it into something super meaningful for other people.
Also, reality is chaotic. it's time to accept the mess and build there is no perfect or final form of a product it's not like everything is going to agents everything is just going automatically to agents and then people will only ever interact with their agents ever again there's going to be a ton of different shapes of software your agents are going to interact with a ton of different types of software and this is also going to evolve constantly and also bear in mind that agents are a type of software as well so with all of this information and what we've seen in terms of the conversation around this repo of reverse engineer anything where is the opportunity going to be well i can give you four pieces of advice if you are building with ai in terms of how do you think about building something that can actually you know make money or turn into a business or whatever you might want to make that thing become or impact the lives of a lot of people that you want to help or help you to take on more clients or whatever that might be.
I think you have to think about these four things. So the first one is to build a product that gets someone an outcome and think about Maslow's hierarchy of needs and those top levels of that pyramid. Think about what outcome that you can help use your expertise in order to help someone achieve.
And we don't want to be building apps with hundreds of features. We want to be building a product. The second is to take what you know and turn it into a product.
So take your expertise, take your passion, take your business knowledge, turn that into a product that you can give to people to help them achieve something meaningful in their life. There will always be space for that and you can always change the shape of that over time, no matter how AI evolves. Because ultimately, your domain knowledge, your knowledge that you have gained as a human being across learning multiple skills and evolving over time is something that AI cannot give somebody and it's something that people already likely pay you if you're a freelancer or you run a small business they likely pay you for that anyway so you can leverage that by turning that into a piece of software turning that into a product that people will want to pay for in order to get that meaningful outcome the third thing is to care about the quality of it we do not want to be building software that looks like AI slop in 2026 and beyond because the bar for software is so much higher now
Agents can create pretty good looking software, but pretty good is not good enough. You need to have something that looks incredibly strong. You need to have something that has a slick and professional user experience, regardless of if you're building an MCP server, an API for agents.
an ai plugin or a traditional kind of sas application you need to have an experience that is aligned to that target customer it meets them where they where they are and we want to go beyond ai slop to elevate our product into the one percent that stand out and fourth you want to build for the person that you actually care about helping and who you can actually help don't just build based on opportunity and this is a mistake i see so many people make is they just think this is where the opportunity is this new tool is where the biggest opportunity is so i have to build there that is not the case at all if you think back to that image of the city with all of these different buildings from all these different architectural eras in there is going to be successful types of products in all these different shapes of software still over the next five to ten years so you do not have to pivot until whatever the newest technology is whatever the newest shape of things are you have to be aware of what's going on but ultimately you just have to build the best product to help the specific person that you care about helping to achieve a specific outcome in their life or in their work
and from that you just use the latest technology to do it in the best possible way you evolve it naturally over time and so these are the four things that you need to bear in mind if you're still building if you are still building i would highly advise you to keep building even if you've built something that you think this isn't going to work this is rubbish this doesn't work start from scratch build it again Change it to be agent native.
Think about how it fits into this AI world. And if you want to learn more about how you can do this, this is exactly why I created my community as well. Because ultimately, my goal is really just to help as many people, including you watching this, to turn your expertise, your knowledge, your passion into an AI product that you can actually use to change the lives of other people.
And you can actually use to start your own business or whatever you want to do. and so if you are building you can go over to school .com forward slash ai apps if you want to find out more about that and the link is in the description down below as well so i think with the reverse engineer anything github repo that's gone massively popular recently it's really reignited this conversation around is software dead is it worth building software anymore and i think it absolutely is worth building software but you just have to be aware of what shape software is taking now in order to be successful with it and you have to combine all of these elements around the software to make it successful as well so the kind of baseline is you build something that's agent native you build something that agents can use and you build something that's going to align with where ai is developing over time and then the second part of this is you have to focus on all of the elements around the build as well the product part of what you're building what is the experience like who is the target customer what is the outcome that you're going to help them achieve and what proof do you have that you can actually help them achieve that
Now, I think there is thousands of people out there with expertise, with skills who have helped people already to achieve something in their professional career or in their life that could very easily turn that expertise, turn that knowledge or that passion into a product that they could put in front of people and the people would want to pay for.
I am massively bullish on that, even if the shape of that product end up being completely different over the next five years. This is something that I'm massively committed to helping people with. And so if you do want help with that or you want to know where to get started, if you have expertise, if you have an interest, if you have passion to turning that into an AI product, then please do check out my community.
I would love to help you actually build something and then launch it and make it successful. If you enjoyed the video, don't forget to like and subscribe. Thank you for watching and I'll see you next time.
The Hook
The bait, then the rug-pull.
A viral GitHub repo that can reverse-engineer almost any piece of software reignited the 'software is dead' debate, so a product designer breaks down what reverse engineering actually threatens, and what still has to be built by hand.
Frameworks
Named ideas worth stealing.
06:55model
Maslow's Hierarchy of Needs for AI Products
Physiological
Safety and Security
Love and Belonging
Esteem
Self-Actualization
AI agents can plausibly auto-build personal software for the bottom tier (safety and security), but the top three tiers still require human judgment, taste, and intention to build for.
Steal fordeciding which tier of human need your AI product should target instead of competing on raw automation
23:39list
Four Principles for Building Valuable AI Products
Build a product that gets someone an outcome
Turn what you know into a product
Care about the quality
Build for the person you actually care about helping
A checklist for deciding what to build once AI can generate software on demand: aim at a real outcome, lean on domain expertise AI can't replicate, hold a quality bar above generic AI output, and build for a specific person rather than chasing trend opportunity.
Steal fora pre-build gut check before starting any new AI product
CTA Breakdown
How they asked for the click.
VERBAL ASK
27:08product
“if you are building you can go over to school.com forward slash ai apps if you want to find out more about that”
Soft, single mention near the end pointing to his paid community, repeated once more at the very close ('please do check out my community') rather than a hard sales pitch mid-video.
Add Modern Creator as a preferred source and Google shows you more of our breakdowns in Search, Top Stories, and AI Overviews. It only changes what you see, and you can undo it in your Google settings anytime.
Add to Preferred SourcesOpens your Google source preferences with us pre-loaded. Tick the box and you're done.
A walkthrough of Jev, a new AI model built to answer in scores, categories, and true/false calls instead of paragraphs, and where that actually fits inside a real project.
A product designer builds a full native iOS app end-to-end inside Claude Code, using a packaged skill system and the new in-app iOS simulator, without ever leaving the chat window.
00:32
01:15
01:37
01:58
02:20
02:42
03:03
03:25
03:46
04:30
04:51
05:13
05:28
05:56
06:17
06:39
07:23
07:44
08:06
08:27
08:49
09:11
09:32
09:54
10:15
10:37
10:59
11:20
12:04
12:25
12:47
13:08
13:30
13:52
14:13
14:56
15:18
15:40
16:01
16:23
16:44
17:06
17:28
17:49
18:33
18:54
19:16
19:37
19:59
20:42
21:04
21:25
21:47
22:09
22:30
22:52
23:14
23:57
24:18
24:40
25:02
25:23
25:45
26:06
26:28
26:50
27:33
27:54
28:16
28:38
15:41
17:19
26:24
13:43
17:19
15:09