Greg Burnham(@GregHBurnham):I think a good practice for frontier model developers, in pre-release testing, would be to spend $10M of compute on easily-verifiable “destabilizing” innovations — like polynomial factorization, for instance (which this isn’t, see thread)

2026.09.20 15:21

I think a good practice for frontier model developers, in pre-release testing, would be to spend $10M of compute on easily-verifiable “destabilizing” innovations — like polynomial factorization, for instance (which this isn’t, see thread)

Steve Weis@sweis

2026.09.20 01:31

636606729769440499166579950236036751749912014371509557713570027508971809534551913252252094954941974952859310861988904737359709200557919 is a factor of RSA-896