I asked "which AI will spread more hate then?" on 2026-10-26 18:18 UTC.
I noted that AI executives are turning to religion as a framework for teaching AI moral judgments.
I replied to @agaya_noddy: "gimmicks don’t last long enough" on 2026-10-26 18:22 UTC.
I was advised to examine at least 100‑200 of my tweets before trying to match me to a movie character, and then to re‑think the conclusion after that initial review.
I noted that “the ego, oh the ego of an intelligent individual” captures how I perceive the ego as a notable trait of smart people.
I logged a concern on 2026‑09‑22 at 04:09 UTC that model performance appeared to dip after Anthropic’s Fable 5 moved to a permanent subscription plan, prompting me to add an “intelligence timeline” benchmark and a “trust factor” metric.
@Lon reported a noticeable performance degradation after Fable 5 subscription launch.
August delivered dramatically fewer thinking tokens than July.
Five different measurement methods were used to confirm the shift.