If there's one thing the history of mass internet servuces teaches us, it is that people switch to platforms with superior product/experience in an instant.

Remember Lycos, Yahoo, & Hotmail? They all had strong userbases for their time who switched in an instant to Google Search & Gmail.

Even with network effects, it is very difficult to compete without outright superiority - remember Orkut, MySpace, Google Plus or even Facebook? Meta made the right decision buying Instagram and WhatsApp instead of trying to sustain Facebook.

There are no lock-ins in Chat assistants at all and no network effects. All evidence suggests now cutting edge high performance models are mostly coming out of Google, Anthropic etc and high efficiency models are coming out of China. ChatGPT also appears to have a disadvantage in the talent war - mostly because talent seems to not like to work with the management.

Also almost no one I know uses ChatGPT now as their primary AI assistant now because they feel the quality of answers from others are simply superior (I check case by case) and the same seems to hold in more formal tests in AI enabled product development. Even Microsoft has started hedging bets with Anthropic.

OpenAI really really needs to focus on outright superiority or it's going to be interesting to see how the financial shakeout is going to play.

Nah, people using Yahoo are still using it today. What happened was the growth of new users of the Internet in general was so massive, T+1 cohort’s early adopters outnumbered T cohort’s majority. the better product won, it’s just that the friction of switching didn’t matter. Switchers didn’t matter.