I've been using 10+ years and have definitely considered that a lot over the years. It's a valid point. I used to be concerned with that, but now I think that is not a big issue for two reasons:

1. Vendor lockin when you are a startup should be a low priority in my opinion. Your priority should be moving fast and getting customers.

2. To expand on 1. It's not really vendor lockin, the only things vendor specific are your infrastructure templates and interface code. Rewriting the interface for another cloud provider takes a fraction of the time of rewriting your codebase (source: I've done it many times).

To come full circle to the question, one thing you should probably not do as a solo-prenuer is concern yourself pre-maturely with vendor lockin at the expense of getting a working product.

Edit/Example: The code for interfacing with AWS Aurora is the same as any other MySQL or PostgreSQL code except for authentication. Switching to a non-IaaS database WILL take time but no more time than rolling your own. So your choices are not "lock in" or "freedom".., it's, spend lots of time setting up databases now or later when you have proven your business model (and potentially never)

Edit 2: This is an anecdote certainly but... of the dozens of startups I've worked on and with, some of which have grown pretty large. The only vendor lock-ins I've ever found to be issues have been with software frameworks (especially ERPs). Never infrastructure. I used to be worried about infrastructure lockin, now I prioritize product market fit.