Could AI accelerate offensive or destabilizing technologies faster than defensive technologies and institutions?
Buterin's d/acc reframes the acceleration debate: the question isn't whether to speed up but which technologies to speed up first — defense-dominant tools before offense-dominant ones.
View on the map → · Open in Browse →
The 'defensive acceleration' idea drew sustained criticism this window. One argument holds that safety-relevant capabilities share bottlenecks with general capability, so speeding up 'defensive' AI research compresses the very timeline it was meant to lengthen; another grants the goal but insists defense-dominant outcomes require prioritized domains and defender-privileged tools, not defense-neutral ones — especially in biology, which is offense-dominant.
Speeding up progress in any of these areas seems likely to speed up general AI R&D, giving everyone else less time to execute time-bottlenecked agendas.
Argues the d/acc logic fails for safety-relevant capabilities because they share bottlenecks with general capability, so accelerating 'defensive' AI research compresses the timeline it was meant to lengthen.
Defence-neutral tools are insufficient to improve our collective biological defences.
A concrete critique of d/acc arguing defense-dominant outcomes require prioritized domains, inherently defense-favoring tools, and privileged defender access — especially given biology's offense-dominance.