Tinfoil is rolling out automated safety safeguards inside its Tinfoil Chat secure enclaves over the next few weeks, running open-weight models in hardware enclaves so conversations stay invisible to e…
Tinfoil, a company offering private AI infrastructure, provides products including Private Chat, Private Inference API, and Tinfoil Containers that run AI models inside secure hardware enclaves to ens…
Tinfoil announces cached prompt pricing in its Inference API, a feature that reduces compute for eligible requests by caching recently processed inputs, but the company warns that caching introduces t…
NVIDIA's confidential computing (CC) on Blackwell GPUs introduces performance overheads ranging from 1% to over 100% for inference and 18–32% for training, according to benchmarks from Tinfoil. The pr…