diff options
Diffstat (limited to 'content')
| -rw-r--r-- | content/blog/delusion.md | 120 |
1 files changed, 120 insertions, 0 deletions
diff --git a/content/blog/delusion.md b/content/blog/delusion.md new file mode 100644 index 0000000..63c889d --- /dev/null +++ b/content/blog/delusion.md @@ -0,0 +1,120 @@ ++++ +title = "The secret to doing crazy things is delusion." +date = 2026-09-20 +extra.image = "/assets/images/post-image/delusion.png" +extra.header_title = "delusion" ++++ + +Throw everything you know about reasonable expectations out of the window. + +When you think of an insane idea, something almost impossible, something delusional, what's the first thing you do? +Do you brush it off as insane, infeasible, or just plain impossible? +If so, throw that habit out of the window. + +When you think of an insane idea, something almost impossible, something delusional, write it down. +Expand on it. +No matter what you think, no matter how insane it sounds, no matter how impossible it looks, try it. + +Make the thing you thought to be impossible. +Write down that idea that's been in the back of your head for years. +Build the thing you thought to be delusional. + +You never know what you're capable of. + +The secret to doing crazy things is **delusion.** + +## Be unreasonable. + +That idea in the back of your mind. +The concept you thought was impossible. +The innovation that's "infeasible." + +Is it really infeasible? +Is it really impossible? + +Sure, it might be delusional, but who knows — you might do something crazy. + +## Delusional ambition. + +Ambition is usually constrained by what we think our limits are. +We set goals based on our current abilities. + +What if you just... stopped doing that? +What if you set a goal so delusional that your current skillset wasn't enough for it? + +You'd have to grow into it. You'd have to expand your skillset to reach your goal. +It might take you a month, a year, two years, five years, maybe even a decade, but who knows — you might do something crazy. + +## Romanticize the process. + +Motivation is commonly fueled by results. + +The recognition. +The money. +The finished product. +The applause. +The light at the end of the tunnel. + +Shift your focus. +Don't let the results hook you in. +Romanticize the process. +Get addicted to building for learning. Building for fun. Building for joy. + +Sometimes, projects never happen. +Life gets in the way. +You run out of money. +Someone else comes up with your idea and does it before you. +The plan you spent months working on fades away into the abyss. + +But you know what doesn't fade away? +The knowledge you gained from that experience. + +Your first project might not have taken off, but your skills definitely improved. +Keep trying, because who knows — you might do something crazy. + +## Delusion != correctness + +Last year, I designed an incredibly ridiculous embedded systems project. + +I called it quadrant-robot. It was going to be a home robot running embedded Linux on a Raspberry Pi, with voice recognition, an LLM, SLAM capabilities, a custom Linux-based operating system (Yocto/OE), OTA updates, secure boot, verified boot chains, custom partition structures, a recovery environment, and much more. + +Just to clarify, I didn't even know how to design a PCB. To be honest, I still don't know how to design PCBs. I do low-level development and embedded systems, but I've never really been a hardware guy. + +You might ask what my solution to not knowing PCB design was. It was to just use a Raspberry Pi. For a commercial product. +Oh, did I mention I tried to use Python for the A/B updates because I wasn't as comfortable with C back then? +Yeah, good luck putting an entire Python interpreter in an initramfs. + +I did not have a single thought about licensing Pis for commercial use, about the Raspberry Pi CM5 being too bulky for my project, about the limitations of verified boot on Raspberry Pis. +But I had a vision and ambition. + +Somewhere in my plan document, I wrote this: +*“Maybe we could partner with the Raspberry Pi Foundation to make it like an endorsed platform to easily build robotics on top of.”* + +I said what? +Who in their right mind looks at their ambitious project and thinks *"Maybe the Raspberry Pi Foundation would want to partner with us for this great idea!"* + +Apparently, me from last year. + +In hindsight, there were a _lot_ of flaws, but the idea itself wasn't completely ridiculous. Many of the specifications for verified boot chains were correctly done, the partition structure was clean, and A/B updates seemed okay-ish. (A/B was good, but why Python? I could've just used RAUC smh) + +Was it delusional? Yes. +Did it ever happen? Definitely not. + +The value of this project wasn't the end result. It was the crazy amount of embedded systems I learnt along the way. + +Do you know how much I learnt about Yocto, Buildroot, and Linux in general? +I tried multiple build systems, including a script to modify Alpine, Buildroot, Yocto/OE, and I even considered LFS. + +I learnt all about verified boot chains, dm-verity/fs-verity, hardware root of trusts, partition structure, A/B updates, and more. +quadrant-robot never happened, but now I'm using the knowledge I gained to build _real_ things with Yocto. + +The project died, but the skillset stayed. +quadrant-robot never happened, but I think it was the turning point that transitioned me into a true systems programmer. +It was the point where I went from "teen who makes cool stuff in Python, JavaScript, and maybe a bit of Go" to "teen who spends his spare time writing UI frameworks in C and making Linux distributions with Yocto." + +Being ambitious and delusional isn't about making everything work on the first try. It's about the cycle of being willing to learn by building, being willing to take roadblocks as a learning opportunity, and being willing to look back at yourself in a year and just laugh at how naive and delusional you were, then using those skills to make something even better. + +--- + +The secret to doing crazy things is **delusion.** +Keep building, keep planning, and stay delusional, because who knows — you might do something crazy. |