I mean this is just speculation but I don't think that the spam rn is happening because Twitter is free. It's because it's profitable and as long as it's profitable it will happen. The only way I think Twitter can reduce spam by asking for user verification.

ML also has the benefit of scaling well. If you build a ML system to detect spam, I wouldn't say there's much difference in development costs if you 10x the user base and I do not see Twitter not having such a ML system. The model isn't going to be 1/10 as cheaper to train and the size of the engineering team isn't going to be 1/10 too.

As per ML in general, I doubt we are going to see a decline, all these layoffs in big tech and ML teams are basically the only ones that are still hiring over the board. I think that's it is pretty clear that investment in ML saves money by now, it can automate processes for a fraction of the cost. My experience is only in research tho, so I'd like to hear your thoughts on this.