This is a common trope - and has a name - "gell-man amnesia", that people raise again and again, but I don't think it's the whole story. There are different incentives at play. Journalists are sometimes investigators, sometimes storytellers, and other times just mediators. The article mentions the exagerations and unique incentives at play in the research domain. It is in universities' interests to present research in a certain way. Journalists, as well, are incentivised to create juicy headlines for mainstream consumption, but not always and not consistently. Sometimes it's necessary to simplify, but that doesn't mean truth is completely lost. Detail is not always necessary for insight. Sometimes as well, a news domain is suffiently well understood to be applied as intended; e.g. traffic, weather, crimes in general, matter-of-fact reporting of sequenced real-world events. But then we get nth-order insights: e.g. geo-political reporting from "this thing happened" to "the implications are...", and those necessarily lose some truth because consequences and causalities are hard.

I'd say, generally, things are not so simple as "news = ~lossy". Some domains suffer from bad reporting more than others, and it's worth inquiring why that occurs on a domain-by-domain basis.