The river itself dried in around ~5000BC.
The text attributed itself as being authored in 3000BC.
The idea that text was written in 1500BC (3500 years ago) was proposed by Max Muller which was contested by his contemporary western and Indian scholars alike ( they proposed earlier dates)
In his late life, Max himself admitted that the dates he gave were hypothetical, they were to considered minimum and it could actually be more than that.
"Scientific data and new evidence converge on the understanding that between 3000 and 2000 BCE (about 4,000 years ago), the river's demise began when Sutlej diverted from the present-day Ghaggar-Hakra valley to join the river Beas. By around 3792 BP (c. 1800 BCE), the river had completely dried up."
I encourage you to read research literature from the last century. Given that all of our linguistic estimates put Proto-Indo-European itself at around 3000 BCE, it makes no sense to claim that the RigVeda was composed earlier at that time. Indeed, Sanskrit likely did not exist at that time.
I can't trust the linguistic estimates fully as they are vague and are prone to many assumptions and approximations in their calculations. Just like change of stance of Max Muller, such estimations change depending on whether much older antiquities have been excavated or not. Like RigVeda described Saraswati during its time as alive.( I was mistaken that river dried in 5000bc, it dried around 3000bc to 2000bc ). It's only a matter of time since new evidence are discovered that pushes such estimates further back in time.
What they do very well imply that certain developments happened a 'minimum' of years ago.
Authorship's mention of time as in rig veda cannot be completely discarded out of consideration.
