Wait 5 years. After 5 years of the hype cycle the rough edges are sanded off, or the concept has ceased to be relevant.

You could have built most (web) things in the past 20 years with Java and a SQL database. It's worth keeping an open mind to what useful things might come out of a new trend but give it some time to avoid chasing fads like Web3, NFTs, NoSQL, Microservices or LLMs.

(This is not to say any of these are entirely useless in every scenario but the useful parts will have stood the test of time and the rest is for 'thought leaders' and con artists)