I haven't encountered formal methods in the wild since I left university in 1998. I don't know of anyone in my wider circle of friends using formal methods. Using it is by and large a niche thing in our industry. You see it pop up in places where people are desperate to commit millions in extra work just to avoid disastrous failures. I've never worked in such places. Use of formal methods is a last resort in our industry: you use it when you've exhausted all other options.
That being said, in the same period of time, people have learned to appreciate static typing and unit tests. Unit testing was not a thing until people started pushing for this in the late nineties. Around the same time, computer science students were infatuated with languages that were dynamically/weakly typed. Now 20 years later, type systems are a lot less clunky and those same people seem to be upgrading their tools to having stronger types.
E.g. Kotlin has a new feature called contracts, which you can use to give hints to the compiler, which then is able to use the extra hints to deduce that accessing a particular value is null safe. This allows for methods called isNullOrBlank() on nullable strings (String? is a separate type from String in Kotlin). So, calling this on a nulled string is not an NPE and even better, it enables a smart cast to String after you call this.
Also, annotations are widely used for input validation and other non functional things nowadays. I think annotations were not a thing before around 15 years ago (I might be wrong on this). I think Rust is a good example of a language with very strong guarantees around memory management.
Compilers of course have a very big need to be correct since compiler bugs are nasty for their users. So, there is a notion of some of this stuff creeping into languages. Combined with good test coverage, this goes a long way in being minimally intrusive and giving you stronger guarantees. But probably calling this formal methods would be a bit of a stretch.
That being said, in the same period of time, people have learned to appreciate static typing and unit tests. Unit testing was not a thing until people started pushing for this in the late nineties. Around the same time, computer science students were infatuated with languages that were dynamically/weakly typed. Now 20 years later, type systems are a lot less clunky and those same people seem to be upgrading their tools to having stronger types.
E.g. Kotlin has a new feature called contracts, which you can use to give hints to the compiler, which then is able to use the extra hints to deduce that accessing a particular value is null safe. This allows for methods called isNullOrBlank() on nullable strings (String? is a separate type from String in Kotlin). So, calling this on a nulled string is not an NPE and even better, it enables a smart cast to String after you call this.
Also, annotations are widely used for input validation and other non functional things nowadays. I think annotations were not a thing before around 15 years ago (I might be wrong on this). I think Rust is a good example of a language with very strong guarantees around memory management.
Compilers of course have a very big need to be correct since compiler bugs are nasty for their users. So, there is a notion of some of this stuff creeping into languages. Combined with good test coverage, this goes a long way in being minimally intrusive and giving you stronger guarantees. But probably calling this formal methods would be a bit of a stretch.