I much prefer AI safety being researched before we build a really powerful AI. This reminds me — before GPT was even released for the first time OpenAI made n

I much prefer AI safety being researched before we build a really powerful AI. This reminds me — before GPT was even released for the first time OpenAI made noises about safety and consequences and it was shrugged off, because back then GPT looked like a toy that could only write barely coherent Amazon reviews or blogs about cryptocurrencies, and that's a too low bar to demonstrate intelligence.