I agree. If we don't go to Web3, data will be more restricted in the future, as we've seen from Reddit. Sure, the data will be available in LLMs, but if data are monopolized by big companies, what's to say that they won't put them behind paywalls.
More importantly,, what will be feeding the LLM training? Who will have access to it?
Might not be much of an issue today since the models are expensive to train. But what happens when the price of training comes down yet data is only available to those who can pay.
If Web3 and public data doesn't happen, then up to date information behind subscription payments will happen.