The article makes it clear that (as the author understands it, at least) someone who uses open source software in their commercial product is liable; the people who wrote the open source code [1] are not.
> If a user is harmed by software, the person they paid (targeted ads would count) must compensate them for the harm – unless the software provider can prove their software played no role in the ... harm. If open source resources are [used by] your code, you’re responsible for their performance too. *The open source resource licensed away their liability to you*.
(Emphasis mine)
[1] Assuming they used a license that limits liability, such as Apache.
The article is attempting to create a scare about things that have always been true. If a telco's services crash the telco has to compensate customers even if it was a postgres failure that caused it by failing to authorise handsets for a connection in a cell. For example.
It's not exactly such way.
This is only case, if you are 21 or 25 years old (depend on country/state) and if you have insurance which cover this case, or if you have some juridical document for exactly this case.
For example if you toddler/teenager, NOT accompanied by an adult, will be responsible people, who have responsibility to restrict your appearance (entrance) in this park.
So in EU, usage of OSS or products with OSS dependencies, will be effectively prohibited for teenagers. This is not very large share of customers, but approximately 7% of EU residents.