While I tend to agree, I hope we don't get lazy with the law and prosecute on DNA alone. DNA matching is not 100% accurate, and juries aren't the best at appreciating the nuance of statistics.
It is also a problem when the DNA isn’t obvious that it is connected to the perpetrator of the crime. Semen found inside a victim is very different than DNA found inside a room where a crime occurred, which is different than DNA found at a crime scene in a public or semi-public place. If you can just collect up a bunch of DNA and run it against everyone, with no initial justification of probable cause, you can get innocent people that are guilty of no more than leaving their DNA somewhere and not having an alibi being victims themselves of lazy police investigations.
What got me was, if it's not 100% accurate, how do you explain their ability to hunt down the golden state killer using his family tree? They narrowed the possibilities to 1,000 or so, then checked all 1,000. Straightforward police work. Then a final DNA match confirmed their hypothesis that they'd found him, which turned out to be true.
I don't know. I'm vividly aware that if the matching isn't accurate, innocents will go to prison. But the truth seems less simple than that.
It's the same thing with IAFIS. If you think you're going to trial and winning with just a fingerprint match in the database you're gonna be sorely disappointed. These tools are better understood as "lead generation" than any kind of proof.
When I was a stats major, Peter Donnelly's TED Talk was required viewing on the matter: https://www.ted.com/talks/peter_donnelly_how_juries_are_fool...