Tipping points and technological lock-in aren't new (e.g., the printing press, the radio, TV, phones, the internet, etc.). I do absolutely see your distinction, though, that prior technologies may have been less capable of imposing specific belief systems compared to frontier LLMs.

That's a fair point, and a key reason why competitive open models are so important. I definitely don't want to be stuck using Google's latest and greatest LLM once companies, marketers, and PACs can pay to subtly influence responses as they choose.

We've all seen this movie before, though.