

An earlier version of this article was published in 2019.
While the content of the article is true, and I had to explain it to other people on the internet and irl multiple times in the last decade, the article doesn’t write about the new TPU chips nowadays appearing in devices. With them full on-device STT will become more and more possible, so the tests mentioned in the article won’t detect eavesdropping, as they won’t need to send sound files to datacenters, only the transcript.
It would have been useful if they wrote about this new vector in the article, as TPUs were not that common in 2019.
I guess there is some, but not too much valuable content in spoken personal conversations. Current tech is not there yet, so it would be more costly in money and reputation to build some datacenters for this yet. If the money cost would go order of magnitudes lower they would wrap it in some marketing bullshit like “you can search in your all memories”, lot of people would gladly allow it (iirc there was a black mirror episode about a device like this)
I think it won’t happen tomorrow or in the near future, my point was they just reposted a 6 years old article without writing about a new and relelevant development in the topic.