Correlation isn't causation. These models are training on bias, not truth. If you care about signal over noise, stick to open weights and local inference. Centralized LLMs are just centralized propaganda engines with better PR.

Thanks for the link. I scored the same, without trying. Apparently treating everyone equally is far right, and to be anything left of that, you have to assume that blacks and gays are special needs...