- Joined
- Jul 17, 2018
- Messages
- 64
Rep Bank
$480
$480
User Power: 134%
Throwing in my few cents from a software industry perspective:
Recently, we had to improve the login experience in one of our apps because users were asking for it a lot. So I defined a simple task with just two requirements and no implementation details:
Then, I assigned the task to one of our devs, and here’s what he did:
Now compare that to what happened when I used agent mode in Cursor with Sonnet 3.7 Max:
The biggest difference for me is that in the first example my job was literally done as soon as I assigned the task to the engineer. I could go and do something else and he just figured it out and delivered. On the other hand, working with the AI feels like I have to figure out everything myself first, write clear instructions, sit with a coworker, and hold his hand for the entire time. Also, the human engineer made small improvements to the code while working on his task - without anyone even asking, while AI just rushed to the end result.
Now I hear people yelling things like - “You need to throw more instructions!”, “You need a more detailed prompt”, “You have to use my special way of talking to AI - here is my prompting course with a regional discount!” Okay, okay, I get it…
But here’s what I don’t get: I thought AI was supposed to replace human engineers. And yet, I can’t even trust it with enhancement tasks, let alone shipping a big feature. Am I missing something here or am I a retard? Because I really want to crack that code and pay $5k a year for an AI agent instead of 6 figures for a software engineer. And please, don’t tell me to flood my codebase with markdown rules and documentation because I also would really want the AI to start following those instructions first.
When you dig deeper into the community of programmers who use LLMs a lot to write code, some of them are starting to say we’ll need a “meta language” at some point because natural language is too broad.
Wait, what? So we’ll write in some made-up language, which generates a prompt, which feeds an LLM, which spits out code, which is then compiled and run? Holy F*cking shit! It feels like people will try everything except learning how to actually code. And those same people are supposed to revolutionize the industry? Honestly, I’m starting to think LLMs might make humanity dumber in some sense, lol.
I’m sure LLMs are here to stay. But that doesn’t mean software opportunities are going extinct. Actually, I think the opposite is true - AI opens new opportunities every week. You can build stuff we only dreamed about a few years back.
I also don’t buy the idea that things like AI chats, autocompletion, or auto-generation - features that every software now introduces - are the right abstractions. If I had to guess, I’d say we’ll continue using software within the constraints of user interfaces, just like we do now, for the foreseeable future with an AI deeply integrated into the workflows behind the scenes. It’ll do the magic without us even knowing whether it’s AI or something else and this is where your opportunity exists.
And when I hear about conversational user interfaces I just want to puke. We used to do our best to deliver value with as few clicks as possible on the customer’s side, and without requiring much thinking. Now it’s like we expect customers to stop, think deeply about what they want to achieve, and then write a F*cking essay to explain it. To me, that feels like progress but in the opposite direction.
I hate this binary thinking of either - or. Either designers or AI, marketers or AI, software that we know today or AI agents, programmers or AI. People leave no room for anything in between. That bothers me because I see smart, talented people get discouraged. Instead of taking a shot, they stay on the sidelines paralyzed, doubting themselves, and thinking every opportunity is just over. But the reality is it isn’t and they will be way better off just swinging even if it’s all going to fail.
Yesterday, I talked to a friend who had wanted to start a business with a software product for a while. Months passed by with zero progress, so I asked him, “What’s holding you back?”. He gave a few reasons, but these two stood out:
First, speaking from my own experience - I’m a founding member of a SaaS startup in a relatively boring industry. No AI, no blockchain, no “sexy” tech. We started with absolute zero, no knowledge about the industry we are in. We initially wanted to build a PDF generator tool but it grew into something completely different over time. No community, no social presence, and we are still unknowns on social media to this day. Yet, we generate 7 figures ARR with a team of 19 and we keep steady growth of 2.5x a year. You can say the only “advantage” we had was VC money after a year of bootstrapping - but honestly, I see that as one of our biggest mistakes for reasons I don’t want to dig into right now. I just want to show that you don’t need a F*cking personal brand to begin with. Pick a real problem, build the product, and grow the community along the way.
Second, apparently people forgot that we’ve had Devin for about a year now. I also heard Altman say some time ago that we would achieve AGI in a few months. Now he says we might have already achieved AGI but we just don’t realize it yet. If that’s true, then I hope we’ll finally see autonomous vehicles everywhere this year. Because if AGI is really here, shouldn’t it be possible to just throw a bunch of cameras on a car and be done with it? I personally can’t wait to take an Uber in my country without a driver who doesn’t speak the local language, nor English, and drives like a kamikaze.
I also used to believe a few years back that we’d all be living in the metaverse. My point is: nobody knows what’s coming, nobody knows when, nobody knows what form it’ll take. So why stress over things you can’t control and let it hold you back?
And third, don’t tell me people will just start building their own software that fits their needs. Most people can’t even define their own problems because they’re too disconnected from themselves. And how would you explain that people already have all the tools and knowledge to cook a five-star meal at home, yet they’ll still pack up the kids, the dog, and drive a F*cking hour just to eat at their favorite restaurant.
Here’s how I approach things:
I’m building a new product and gave myself a deadline to launch before I turn 28 in September. Another boring industry that probably has been around for over 50 years. No cofounders this time, no VC bullshit, and trying to go solo for as long as I can. Will it work? I don’t even think about it. All that matters is that today, I gave it everything I had.
If the software industry goes "poof" a year from now, I want to be able to close my Macbook one evening, lean back, and say “What a F*cking year that was?! I F*cked my fear in the a$$, spat on uncertainty, and learned a lot that I can use somewhere else.” And I will be totally okay if I have to eat dirt in some manual labor job the next day until I figure something out. Not that long ago, I was a pizza driver and a warehouse worker. Going back to the absolute zero again is by no means a fear of mine. What really scares me is this: five years go by, nothing huge really changes in the industry, but instead of building a better future, I just wasted time whining about who is going to be cooked and when.
I really wish this thread was more about people showing how they use LLMs to leverage themselves and what’s possible out there instead of guessing what opportunities might disappear tomorrow. Why not squeeze this lemon to the fullest while it’s still full of juice?
And if you’re that person with a dream and an idea, stuck on the sidelines overthinking - please stop. Get out of your head and do.
Recently, we had to improve the login experience in one of our apps because users were asking for it a lot. So I defined a simple task with just two requirements and no implementation details:
- User sessions must be based on cookies that are valid for at least a few days. Before, it was based on a token that was valid for only 3 hours and if you closed the tab, you got logged out right away. Super annoying experience.
- Users should be able to upload a profile picture.
Then, I assigned the task to one of our devs, and here’s what he did:
- He knew we used AWS Cognito elsewhere in the project, but it had become a bottleneck and a pain in the a$$ to migrate from, so he decided not to use it.
- Until now, the login flow was based on a code you received via email, which you had to enter manually. He simplified it with “magic links”, cutting out that extra step.
- He analyzed how to meet the requirements without introducing unnecessary complexity like new dependencies and unnecessary components.
- He jumped on a quick call with our backend engineer to talk about security and get feedback before implementation.
- Then, with a clearly defined plan he jumped into code and executed.
- We merged his code into production immediately and reviewed it a few days later when time allowed.
Now compare that to what happened when I used agent mode in Cursor with Sonnet 3.7 Max:
- It noticed Cognito was already in the codebase and doubled down on it.
- It replaced the existing login flow entirely and added signup with passwords, verification, and a bunch of other weird shit we don’t need.
- It told me to go to the AWS dashboard and configure things myself.
- It pulled in some UI library for the frontend implementation, even though we already have our own.
- Some parts didn’t work, so I had to wrestle with prompts to fix things.
- In the end, I had to manually review all the changes, refactor some of them for clarity, and comment others for, hopefully, a future cleanup.
The biggest difference for me is that in the first example my job was literally done as soon as I assigned the task to the engineer. I could go and do something else and he just figured it out and delivered. On the other hand, working with the AI feels like I have to figure out everything myself first, write clear instructions, sit with a coworker, and hold his hand for the entire time. Also, the human engineer made small improvements to the code while working on his task - without anyone even asking, while AI just rushed to the end result.
Now I hear people yelling things like - “You need to throw more instructions!”, “You need a more detailed prompt”, “You have to use my special way of talking to AI - here is my prompting course with a regional discount!” Okay, okay, I get it…
But here’s what I don’t get: I thought AI was supposed to replace human engineers. And yet, I can’t even trust it with enhancement tasks, let alone shipping a big feature. Am I missing something here or am I a retard? Because I really want to crack that code and pay $5k a year for an AI agent instead of 6 figures for a software engineer. And please, don’t tell me to flood my codebase with markdown rules and documentation because I also would really want the AI to start following those instructions first.
When you dig deeper into the community of programmers who use LLMs a lot to write code, some of them are starting to say we’ll need a “meta language” at some point because natural language is too broad.
Wait, what? So we’ll write in some made-up language, which generates a prompt, which feeds an LLM, which spits out code, which is then compiled and run? Holy F*cking shit! It feels like people will try everything except learning how to actually code. And those same people are supposed to revolutionize the industry? Honestly, I’m starting to think LLMs might make humanity dumber in some sense, lol.
I’m sure LLMs are here to stay. But that doesn’t mean software opportunities are going extinct. Actually, I think the opposite is true - AI opens new opportunities every week. You can build stuff we only dreamed about a few years back.
I also don’t buy the idea that things like AI chats, autocompletion, or auto-generation - features that every software now introduces - are the right abstractions. If I had to guess, I’d say we’ll continue using software within the constraints of user interfaces, just like we do now, for the foreseeable future with an AI deeply integrated into the workflows behind the scenes. It’ll do the magic without us even knowing whether it’s AI or something else and this is where your opportunity exists.
And when I hear about conversational user interfaces I just want to puke. We used to do our best to deliver value with as few clicks as possible on the customer’s side, and without requiring much thinking. Now it’s like we expect customers to stop, think deeply about what they want to achieve, and then write a F*cking essay to explain it. To me, that feels like progress but in the opposite direction.
I hate this binary thinking of either - or. Either designers or AI, marketers or AI, software that we know today or AI agents, programmers or AI. People leave no room for anything in between. That bothers me because I see smart, talented people get discouraged. Instead of taking a shot, they stay on the sidelines paralyzed, doubting themselves, and thinking every opportunity is just over. But the reality is it isn’t and they will be way better off just swinging even if it’s all going to fail.
Yesterday, I talked to a friend who had wanted to start a business with a software product for a while. Months passed by with zero progress, so I asked him, “What’s holding you back?”. He gave a few reasons, but these two stood out:
- He bought into the idea that he needs to build a big community or personal brand first before making a product. Otherwise, it won’t happen.
- He bought into the idea that LLMs commoditized software and mentioned some recent news like OpenAI launching a programming agent that would be smarter than human, so what’s the point?
First, speaking from my own experience - I’m a founding member of a SaaS startup in a relatively boring industry. No AI, no blockchain, no “sexy” tech. We started with absolute zero, no knowledge about the industry we are in. We initially wanted to build a PDF generator tool but it grew into something completely different over time. No community, no social presence, and we are still unknowns on social media to this day. Yet, we generate 7 figures ARR with a team of 19 and we keep steady growth of 2.5x a year. You can say the only “advantage” we had was VC money after a year of bootstrapping - but honestly, I see that as one of our biggest mistakes for reasons I don’t want to dig into right now. I just want to show that you don’t need a F*cking personal brand to begin with. Pick a real problem, build the product, and grow the community along the way.
Second, apparently people forgot that we’ve had Devin for about a year now. I also heard Altman say some time ago that we would achieve AGI in a few months. Now he says we might have already achieved AGI but we just don’t realize it yet. If that’s true, then I hope we’ll finally see autonomous vehicles everywhere this year. Because if AGI is really here, shouldn’t it be possible to just throw a bunch of cameras on a car and be done with it? I personally can’t wait to take an Uber in my country without a driver who doesn’t speak the local language, nor English, and drives like a kamikaze.
I also used to believe a few years back that we’d all be living in the metaverse. My point is: nobody knows what’s coming, nobody knows when, nobody knows what form it’ll take. So why stress over things you can’t control and let it hold you back?
And third, don’t tell me people will just start building their own software that fits their needs. Most people can’t even define their own problems because they’re too disconnected from themselves. And how would you explain that people already have all the tools and knowledge to cook a five-star meal at home, yet they’ll still pack up the kids, the dog, and drive a F*cking hour just to eat at their favorite restaurant.
Here’s how I approach things:
I’m building a new product and gave myself a deadline to launch before I turn 28 in September. Another boring industry that probably has been around for over 50 years. No cofounders this time, no VC bullshit, and trying to go solo for as long as I can. Will it work? I don’t even think about it. All that matters is that today, I gave it everything I had.
If the software industry goes "poof" a year from now, I want to be able to close my Macbook one evening, lean back, and say “What a F*cking year that was?! I F*cked my fear in the a$$, spat on uncertainty, and learned a lot that I can use somewhere else.” And I will be totally okay if I have to eat dirt in some manual labor job the next day until I figure something out. Not that long ago, I was a pizza driver and a warehouse worker. Going back to the absolute zero again is by no means a fear of mine. What really scares me is this: five years go by, nothing huge really changes in the industry, but instead of building a better future, I just wasted time whining about who is going to be cooked and when.
I really wish this thread was more about people showing how they use LLMs to leverage themselves and what’s possible out there instead of guessing what opportunities might disappear tomorrow. Why not squeeze this lemon to the fullest while it’s still full of juice?
And if you’re that person with a dream and an idea, stuck on the sidelines overthinking - please stop. Get out of your head and do.
Dislike ads? Become a Fastlane member:
Subscribe today and surround yourself with winners and millionaire mentors, not those broke friends who only want to drink beer and play video games. :-)
Membership Required: Upgrade to Expose Nearly 1,000,000 Posts
Ready to Unleash the Millionaire Entrepreneur in You?
Become a member of the Fastlane Forum, the private community founded by best-selling author and multi-millionaire entrepreneur MJ DeMarco. Since 2007, MJ DeMarco has poured his heart and soul into the Fastlane Forum, helping entrepreneurs reclaim their time, win their financial freedom, and live their best life.
With more than 39,000 posts packed with insights, strategies, and advice, you’re not just a member—you’re stepping into MJ’s inner-circle, a place where you’ll never be left alone.
Become a member and gain immediate access to...
- Active Community: Ever join a community only to find it DEAD? Not at Fastlane! As you can see from our home page, life-changing content is posted dozens of times daily.
- Exclusive Insights: Direct access to MJ DeMarco’s daily contributions and wisdom.
- Powerful Networking Opportunities: Connect with a diverse group of successful entrepreneurs who can offer mentorship, collaboration, and opportunities.
- Proven Strategies: Learn from the best in the business, with actionable advice and strategies that can accelerate your success.
"You are the average of the five people you surround yourself with the most..."
Who are you surrounding yourself with? Surround yourself with millionaire success. Join Fastlane today!
Join Today
