A lot of the public divide on AI capabilities comes down to 1) what you use (free-tier chatbots vs frontier models + agentic tools) and 2) what you use it for (emails vs coding, math, etc). We shouldn't be surprised ppl using free-tier chatbots for emails are unimpressed
Timnit Gebru doubles down on the “stochastic parrots” framing, saying you “cannot expect LLMs to be factual.” As evidence to support this, she cites errors in... Google AI Overviews. We need to start a GoFundMe to pay for these people to have access to Opus 5.5 and Astra.
The devastating thing for Timnit isn't that she's dumb and wrong. It's how little she has progressed, even as chatbots evolved from parrots to Alien Minds. That whole paper is a task worth 1-prompt parody with Sonnet now. But she's got nothing else to go off. Obsolete.
“If I stand to get lots of money from a particular thing, why do I seem to be the person saying this particular company might build something that kills us all? It sounds counterintuitive.”