co-authors claim that GPT-4 can be more easily

October 18
OpenAI integrates DALL-E 3 into ChatGPT
The integration means users don’t have to think so carefully about their text-prompts when asking DALL-E to create an image.

Users will also now be able to receive images as part of their text-based queries without having to switch between apps.

October 17
Microsoft-affiliated research finds flaws in GPT-4
A Microsoft-affiliated scientific paper looked at the “trustworthiness” — and toxicity — of LLMs, including GPT-4.

Because GPT-4 is more likely to follow the instructions of “jailbreaking” prompts, the co-authors claim that GPT-4 can be more easily prompted than other LLMs to spout toxic, biased text.