Just be warned, a startup often has the promise of more autonomy, but there are plenty of startups run by ego-centric micromanaging wanna-be engineers. Be careful equating, 'being able to talk to the CEO' as autonomy, you may find yourself having to explain your decisions/code to people who's engineering skills are 'being able to convince a VC to give them money'. You'll also probably not tackle 'large problems' but hack and slash a Node/Django/React app into meeting an MVP.
Choose wisely.
That being said, I'd go for a startup if you want to eventually run your own. There's so much you'll learn from seeing things actually occur that you'll never learn from anywhere else.
My experiences from startups have been underwhelming for reasons similar to yours. I haven't been part of an autocratic ego-centric management, but I still felt like I didn't deliver any real impact. I was just told to fix bugs, close tickets, etc. Same cog in a machine feel but without the benefit of working closely with people to mentor you.
In the last startup I was in, even though we're in a team of only 3 engineers, implementing design changes was an uphill battle. The senior engineer lives in Eastern Europe so communication time was difficult, and he had a very impractical way of doing things (preferred his own hand-made JavaScript framework over third parties, no modules, no integrated testing). These things lead to making myself a harder sell for companies that follow less unorthodox software development practices.
Yeah, they can often be more freeform, but also by giving you the illusion that you can flip things around, or be a big fish in a small pond. Being that big fish is not good if the pond itself stinks.
That's absolutely true. The assumption "that you will always get more autonomy at startups" is fanciful.
In reality, many if not most startups are run by inexperienced and often immature managers and engineers who are substantially less qualified and skilled at running a team than their equivalents in more mature tech companies.
I have seen this too. Particularly when the first round of early employees don't have a lot of experience, but have a significant amount of influence and sway simply by nature of being an early employee. It can be very frustrating being someone with experience that has to sit back and watch big mistakes be made despite warnings from people who have done it before.
There's an attitude I've noticed also of, "we're not Xyz." Hate to break it, but most problems aren't really that unique. If you resist learning the lessons from other companies, you will repeat their mistakes.
> Particularly when the first round of early employees don't have a lot of experience, but have a significant amount of influence and sway simply by nature of being an early employee.
As is often the case...
Very often the first generation of employees at a startup will consist mostly or solely of folks with 0-3 years experience at most. Then if that startup survives, all these people are "naturally" promoted to senior / team lead / tech lead levels...
> It can be very frustrating being someone with experience that has to sit back and watch big mistakes be made despite warnings from people who have done it before.
I feel you, brother. I've been there too.
> There's an attitude I've noticed also of, "we're not Xyz."
It's called "young arrogance".
"Hey, we're a bunch of straight-out-of-school engineers, but clearly we can do better than Google because we're awesome!".
> you may find yourself having to explain your decisions/code to people who's engineering skills are 'being able to convince a VC to give them money'
> you'll also probably not tackle 'large problems' but hack and slash a Node/Django/React app into meeting an MVP
Those hit way too close to home for me. I made the mistake of joining such a business when I was fresh from college and broke. The only good thing that happened is that I got some savings out of it. Issues included broken spaghetti code, hacking together MVPs with enough fancy graphics to impress clients (faking it all the way), and having to explain to my tech illiterate boss why I couldn't "just fix it" on the harder problems.
Oh, and the micromanaging is real too. It can wreck your mind to the point of needing professional help.
The real kicker was that it was all on an indefinite "contract" (1099 but you sit in the office like a regular worker - I already filed the IRS contest forms) with low pay and zero benefits. Never working at an early startup again.
Choose wisely.
That being said, I'd go for a startup if you want to eventually run your own. There's so much you'll learn from seeing things actually occur that you'll never learn from anywhere else.