It depends how you define MVP. I don't think releasing a product that has all of the features you want, hacked together, and kind of working is a viable option. There are too many half baked products released every day.
However, I think building the most differentiating feature extremely well and releasing it to see if it gets traction can work. If you can't get traction with your most important feature, adding a bunch more isn't going to help you stand out from what already exists.
If you are not releasing fast feel free to read (re read) advice from YC. I have been an engineer mindset entrepreneur all my 14 years of professional life. In 2020 I broke that mindset and I am, for the first time, seeing customers in a B2B product that I have not even created.
Absolutely it is still valid. Releasing an MVP so that you can engage with early adopters is still a good way to test alignment with your customers. It’s still helpful in checking product/market fit and for prioritizing future product development tasks. The tricky part is getting your MVP to be noticed when so many alternatives exist in most use cases. Good luck!
My opinion is that it is 100% valid. MVP != poor quality. I think if you have an idea you need to realise it and building an MVP is the best way to do that. Most ideas centre around a single idea/theme. Make your product execute that idea well then add all the bells and whistles after.