OpenAI: Soon to see human-level reasoning.
Is that a good thing or bad thing?…
I think it is good. ChatGPT 4o already does pretty good reasoning when we decompose complex questions using Chain-of-Thought (COT) prompting in our classes at Pratt Institute or Rutgers Business School.
What I would love to see is more for these sample use cases (there are more):
– I want to be in control of the media selection process, not some bot I do not control.
– As a small business owner, I want to get actual reports on my tax liabilities as I am spending, not in a report after it is too late.
– I want to hand-off all of my marketing efforts over to a system that I can control that places ads and measures responses without my involvement.
As Ethan Mollick has stated, I want to be able to build a $10B company with only a handful of necessary employees and this will finally be possible with AGI.