Mockingbird Shares | https://www.economist.com/finance-and-economics/2025/02/13/the-danger-...

https://www.economist.com/finance-and-economics/2025/02/13/the-danger-of-relying-on-openais-deep-researchTLDR do not rely on AI, it's still stupidMain issues with OpenAI Deep Research, raised by the author:- struggles with tasks that demand creativity - "tyranny of the majority": you will get a consensus view, not that of the cognoscenti ©- "idiot trap": AI's writing might make you look stupidFirst and third were actually obvious, at least for me (I used GPT to help me write paragraph-length fillers, because its vocabulary is superior to mine)The second one is hard to spot, though. I have never relied on OpenAI's models for topic research - I personally find Perplexity a better way to Google things fast. GPTs are good for summaries, structuring, outsourcing programming and fancy rephrasing, at least in my perception. I prefer to do the research part on my own when it exceeds the "public knowledge" domain. Share your experience if you rely on AI in these cases, I'd like to know.#shared_link@mockingbird_shares