Ask HN: Getting over Burnout with Imposter Syndrome
No synthesized answer yet. Check the discussion below.
1) The degree anxiety is understandable, but just in your head. I run a comp sci program for high schoolers as an replacement for universities. They don't get a degree but must learn how to signal they are capable. It's a skillset that most people don't learn until later in their careers. In your case, you likely have accumulated a number of valuable skills in your previous job. Since you haven't switched jobs, you haven't had much practice in communicating your skills to potential employers. Most engineers I know hate this, but it's sales/persuasion, and it's unavoidable and beneficial for you in the long-run. Take the perspective that you don't have a skill issue, but a signaling issue.
2) RE: AI - I would learn the fundamentals, only if you're interested in the engineering aspect of it, and implement some toy projects just to keep learning. Try vibe coding, it's not going to kill you haha. Learn about its benefits and limitations so you can at least comment intelligently if it comes up.
3) Shipping projects is hard psychologically unless you've developed discipline or there are other constraints/deadlines. If I were you, I'd work on projects in a more social manner. That means building with other engineers, talking to people who would benefit from what you're building, or just helping out existing communities. The socialization would help keep your spirits up. It's also the best way to build relationships and open other opportunities.
4) Hiring/recruiting is going through big changes now, partially due to the rise of AI generated applications. Building trust via other channels is going to be much more important moving forward. We've had to more or less turn off all inbound hiring and rely on our vetted professional networks due to the massive glut of applications. Job search is brutal now, but it's just as bad on the hiring side. We're going through changes were the old methods don't work anymore, and the new haven't been figured out yet.
Good luck!
2. I think they're mostly just making calls to the existing LLM's apis.
3. I ask myself why I started the pet project. If it was to learn something new, well, I probably got bored because I learned the the part that was interesting to me.
Big tech is pretty standard. You have to be impressive enough for a callback, then a possible tech screen where you write code in coderpad or something, then some combination of coding, behavioral, and architectural interviews almost always done virtually.
lots of people practice on leetcode, I recommend everyone at least practice the common easier questions and solving them quickly. The medium and hard ones are worth the investment if you're going for big tech, but maybe less important for startups that all have their own style of interviews (that vary quite a bit).
Startups and no name medium sized companies can be random. Some have collaborative coding, some copy big tech, some just talk to you, some look at portfolio projects or give take home assignments, anything and everything here.
> How important is it to have a visable portfolio of working projects?
Much less important than having recognizable names on the resume, and also less important than having contributions to known Open Source projects.
I highly recommend putting time into existing well known projects (Apache, anything done from Meta/facebook/Google etc), and trying to get a PR accepted, than hacking away on your own project.
Every senior in college has an extensive portfolio of projects on their github, it's very hard to tell how much was copied or how good it is.
A much better signal is someone who can deal with the intricacies of an existing gargantuan codebase, go back and forth in PR comments, follow a style guide, work with maintainers, take criticism, have clear communication, and see something through (That may take a while!).
> How much of the AI hype is HN nerds nerding out about AI versus actually implementing AI, versus gluing AI apis together?
Eh, for me, AI has been amazing at teaching me a new PL, and porting small bits of code.
It's been helpful for writing unit tests, getting scaffolding going, etc. but sometimes incomplete.
It's been on rare occasions able to solve a bug for me.
that's about it. YMMV.
Source - no degree, senior engineer at Google. Also spent 10+ years working at shit companies and doing open source before getting a chance at bigger more well known companies, then doing that for ~8 until I could work my way into getting FAANG interviews.
On your questions: I believe portfolios matter less than signal. A solid GitHub readme, a tiny demo with personality, or even a clean writeup can open doors. Hiring today is noisy, clarity stands out more than size.
AI hype is mostly wrappers + vibes. But that's a feature. If you can ship small, real tools that feel smart, you’re ahead. Nobody’s expecting you to train models.
On staying engaged: what helped me was shipping for insight, not polish. Projects as questions, not proof. The goal isn’t “a finished app,” it’s “what did I just learn about what I love, hate, or want to do more of?”
You clearly still care about the craft, about growth. That’s the rarest signal of all.
Rooting for you.
> I'm also suffering quite a bit of imposter syndrome due to not having a 4-year degree.
A degree isn't a finite indicator of a person's knowledge, capability, or achievements. 39 months after getting mine, I still haven't landed my first paid tech job to enter the industry, and I don't see that happening anytime soon, if ever. Granted, I haven't tried much in the past year, but it wasn't any better when I did.
I once rejected the notion outright because, with obvious bias as a college student, I felt it invalidated everything I'd worked toward. But it's become apparent my degree meant nothing in the industry.
On the contrary, you got a job and remained employed for 11 years. That piece of paper, or the lack thereof, didn't make a difference in what you experienced and accomplished.
> I'm suffering from a lot of analysis paralysis trying to select a side project for a portfolio. How do you keep yourselves engaged with pet projects? My github is a field of projects 1/4 of the way completed before I lost steam on them.
Personally I... don't. I started one of my projects back in August 2024. I made a homepage, UI, and a basic database schema... and then I blanked out on how to progress and hardly touched it for 11 months.
I got some wind in my sails again in July - almost a full year later. A whole year of new experiences and insights, inside and out of the tech domain, led to better practices and features never conceived the first time around. So even if you're not coding or even looking at your project, who's to say you're not progressing with it?
I read of this phenomenon in innovation called the S-curve, where advancements appear to occur rapidly in a short time due to a change in tech, paradigm, or process. Perhaps the next burst of progress simply requires time to find and cultivate through many seemingly insignificant insights gained from other experiences. No point in rushing through - if I'm not working on it, it's not the right time.
That's also why I have like a billion (okay, more like 3-5) projects going. I hit a brick wall on one, so I work on something else where I have a clearer vision. Eventually, something triggers a flash of inspiration, and with a fresh mind, I can return and hit that burst of progress in the S-curve again. Eventually, I'll have many projects with lots of progress due to the staggered curves.
Conceptualizing a whole project is daunting, too. Too many bells and whistles lead to a paralyzing spiral of overthinking and self-doubt. Everything has to be perfect and fleshed out. But the mantra of "done is better than perfect" rings true. Once there's a solid foundation, it's much easier to incrementally improve the project, rather than trying to get it right in one go, and eventually reach what was envisioned.
I find it important to remember why I started a project in the first place. With my own imposter syndrome, I always gravitated toward doing what was considered best practices to be a "real software engineer." I could never do right by myself, so I overcomplicated the tech stack and implementations to hell, following the words of others, trying to prove myself. If it wasn't complex and grueling, I wasn't doing anything worthwhile.
However, I eventually reminded myself that the code is the least important part of my software project.
I never started a project to have code, but to address a personally close issue or fulfill an itch. If the project addresses that goal, then why does it matter if the load times aren't perfectly optimized? Why does it matter if it doesn't scale to millions of users when I haven't even gotten 1 yet? Who cares if it's not "impressive" enough to hypothetical people I've never seen?
Implementation has its considerations, but shouldn't be to the point of blocking all progress. Certainly not to where working on the project isn't fulfilling anymore. So, sometimes, after a long paralysis, I yell at myself mentally to, quote, "Just freakin' do it already, for the love of everything holy, just write the fudging thing- [insert more aggressive self-yelling]", overriding my thoughts to do without that pesky thinking.
But yeah, those are my accounts with those challenges. Good luck wherever the future takes you!