You're forgetting the alternative solution estimated to allow billions of people to make essentially unlimited transactions while retaining node decentralisation once technology makes it safe to scale up to 150MB blocks--lightning network. Obviously 150MB blocks are a long ways off from where we are now, but they're certainly closer than the 24GB blocks which would be needed under the current design to allow a similar throughput.

The Core plan for blocksize increase is not, "never", but, "not today". It seems that they have judged the difficulties associated with deploying a hard-fork to 2MB outweigh the gains which might be had from such a small increase. Maybe after we see segwit and weak blocks/IBLTs implemented to alleviate some of the technical limitations we can consider a blocksize increase, perhaps to something more substantial than 2MB, though as the calculations regarding the bandwidth/computational resources required for any given blocksize are beyond me, I won't make any guesses as to what size we might want to target.

Or maybe a fork to 2MB will be announced after segwit to attempt to lessen complaints, though I can't imagine this will change the minds of anyone convinced that, "blockstream is evil". My understanding is that segwit will fix the biggest technical problem associated with 2MB blocks (validation times potentially getting really long for particular blocks), but deployment is still a problem regardless.