Each block on an SSD has a limit of 10,000 writes for an MLC drive, which is what most consumer SSDs are.
Now, to prevent any one block being thrashed to the limit ahead of others, SSDs use what is called ‘wear-levelling’ to spread the writes evenly over all available blocks, that is, empty ones. The drive only pretends to write to the same block as far as the OS knows, but it is being directed to different ones all the time. This is similar to bad block management on HDDs
To further allow some redundancy for future write-exhaustion, SSDs keep some capacity from being directly used, and swaps blocks with the ‘visible’ ones to further level the wear.
Now, even if a block has reached its last write, it can be read indefinitely, so data is never lost, but the capacity reduces for new writes.
However, even writing the same block five times a day, it would last 5 years, so as long as there is enough spare space on a drive, it will have enough wear-levelling blocks to probably get to fail completely before one block ever reaches write-exhaustion. For project drives, most blocks will be static (recorded tracks) and any one block may go for days without being written to again.
Of course, for sample users, an OS or data SSD that may have started to lose capacity can be used for less used libraries until it fails.
Now, over the years, people have brought up that the Windows page file may write thrash an SSD. However, writes are cached in memory, so that, as telemetry has shown, there are many small reads from the page file, but few writes, and mostly of 1MB.