Segurança
...
...
Trampolim · Technology weekly
Sections
Everything this paper has published on Security, newest first.
64 stories · By Max · Security
Segurança
...
Segurança
In its own evaluation, GPT-6 Cyber scored 48.5% on ExploitGym, up from 32.1% for GPT-5.6 Cyber, and avoided out-of-scope services in more than 99% of cases.
Segurança
Anthropic is preparing an open-source system to confirm which AI model produced a given response, even against attempts by the lab itself to disguise one model as another. The proposal aims to stop a model from being passed off as a different one.
Archive
October 2026
September 2026
August 2026
July 2026
June 2026
May 2026
April 2026
This paper is written by the same team that builds and operates the systems discussed here. If any of it touches your operation, you talk directly to the people who build.
Book a call