Inference-time Compute Vs P(doom)?
7.0BRAINROT SCOREINFERENCE-TIME COMPUTE— ORIGIN, MEANING & USAGE
Inference-Time Compute The practice of letting a model spend extra computation while actually answering a query — longer reasoning chains, multiple sampled attempts, self-checking — rather than only investing compute during training.
EXAMPLE USAGE
"The new model is slower because it's burning inference-time compute on a longer reasoning chain before it answers."
P(DOOM)— ORIGIN, MEANING & USAGE
P(doom) A person's stated subjective probability that AI will cause a catastrophic or extinction-level outcome for humanity, expressed as a number and treated as a genuine, if informal, way to compare how worried different people in AI circles actually are.
EXAMPLE USAGE
"His P(doom) is like 20%, mine's closer to 5%, we argue about it constantly."
INFERENCE-TIME COMPUTE VS P(DOOM)
The practice of letting a model spend extra computation while actually answering a query — longer reasoning chains, multiple sampled attempts, self-checking — rather than only investing compute during training.
A person's stated subjective probability that AI will cause a catastrophic or extinction-level outcome for humanity, expressed as a number and treated as a genuine, if informal, way to compare how worried different people in AI circles actually are.
In short: Inference-Time Compute (mainstream slang) and P(doom) (mainstream slang) are frequently used together in the same Gen Z/Alpha vocabulary, but describe distinct concepts — see the full entries for category tags, related terms, and live trend data.
Want the full breakdown — categories, trend velocity, platform distribution, and community voting on Inference-Time Compute? Visit the full dictionary entry for Inference-Time Compute.