while this might be true for Grok, GPT and Claude I don't think tarring _all_ labs with this brush is demonstrably accurate.