I give Gelman a lot of credit for popularizing hierarchical models, but you give him too much.

Before Stan existed we used BUGS [1] and then JAGS [2]. And most of the work on computation (by Neal and others) was entirely independent of Gelman.

[1] https://en.wikipedia.org/wiki/Bayesian_inference_using_Gibbs...

[2] https://en.wikipedia.org/wiki/Just_another_Gibbs_sampler