It may only take one or two further iterations of improved capability (and integration with other tools, such as computer algebra packages and proof assistants) until the level of "(static simulation of a) competent graduate student" is reached, at which point I could see this tool being of significant use in research level tasks.
At this stage Tao should disclose whether he has any financial relationship with OpenAI, including stock ownership or even access to experimental models or more computational power than the average user.
I've never seen any academic hyping up a company like that, unless they explicitly have/had a financial relationship like Scott Aaronson.
https://meta.mathoverflow.net/questions/6039/has-mo-been-red...
The question of exploitation is not a question of anxiety, but of exasperation.