“the GLP-1s are going to give you gigacancer in 10 years” crowd is going to have a lot of trouble with this one

You can use proxy signals in healthcare to determine if there's something seriously wrong with a drug. When you're running a study and all of the vitals look fantastic, you don't say "we know all the awesome stuff this medicine does", but you do say "there's an absence of awful things this medicine does".

@alphaenu @Bonecondor So we only magically learn good things over time after a drug is on the market? We never find out horrible things later on that might require the drug be pulled?

It's quite uncommon for a drug to be yanked after approval because of long-term side effects that weren't indicated by some metrics during trials. There's two decades of data available on GLP-1 agonism. Look at exenatide and tell me how many more years of data you need before giving modern GLP-1 RAs the green light.