AI safety and countermeasures are lagging so far behind, and advancing so much more slowly than the models themselves. There seems to be little to no hope, given we are now in an global AI prisoner's dilemma that we are going to slow down to catch up, here.