You just have to have some product sense, some engineering chops, and the ability to take 6 months off (not everyone can afford this, and that sucks) to build the thing and see if you can get people to look at it and buy it.
You'll learn all the other stuff along the way (taxes, marketing, and eventually: hiring). Just give it a shot. If you run out of savings and have to go back to work, in interviews you'll get to say you "did a startup."
My company is bootstrapped. It isn't growing fast enough to fit PG's definition of a startup, but it's 5 years old and makes 7 figures.
Realizing after a few months that I didn't have to do a "startup" really helped. Whew, no more putzing with decks, trying to figure out how to get into an incubator, trying to raise funds, trying to convince one of my friends to quit their job and be my cofounder, "hustling", committing myself to never sleeping again. I could just focus on building a great product.
And I think my company worked because I was product focused. In the early days I just thought a lot about what the /best possible/ solution to the problem I was solving would look like, built what I saw, put it in front of friends (and eventually strangers from the internet), and made tweaks based on their feedback (and things they didn't say but that I noticed about how they used it).
This is contrary to the "sell first" idea--when I tried to cold pitch friends on my product idea they all said "eh, I don't think I'd really use it over the existing solutions." Those same friends asked me to comp their accounts a couple years later.
It worked for me. "Sell first" might be more important if you're selling to businesses? But like, getting the "x first" part wrong for a few months probably isn't going to kill you. Just start doing stuff. Wiggle around in your space until things start to catch.