Safety
holy shit. someone here actually understands the history!!!
holy shit. someone here actually understands the history!!! @thrashima These are the guys keeping Huang honest, or at least calling out his B.S. Marcus is one of the OG AI people who said that scaling
holy shit. someone here actually understands the history!!! @thrashima These are the guys keeping Huang honest, or at least calling out his B.S. Marcus is one of the OG AI people who said that scaling LLMs wasn't going to produce much and said they had to be merged with deterministic code to become at least 'useful.' Those additions are w…
Related
- Let the record show: honest critique is not pessimism. Your points on scaling limits and neurosymbolic AI have been validated. Fair debate r…
- “Gary Marcus … says that investments into scaling AI is the “greatest capital misallocation in history” which leaves everyone “on the hook”.…
- Pure scaling of LLMs did actually hit a wall; neurosymbolic AI rescued it. AI didn’t hit a plateau; pure LLMs did — until people used harnes…
- Tokenmaxxing is stupid. Change my mind?
Source: Gary Marcus (X) | 2026-09-07