You’ve been reading a lot about ChatGPT lately and that lingering feeling you can’t seem to shake is your correct assumption that the world as we know it will never be the same.
OpenAI rolled out GPT-4 today, its latest iteration of the AI tech behind ChatGPT – the artificial intelligence chatbot that’s taken the world by storm. The company showcased some of the next-level capabilities of the new version on its website, including human tests it had taken, and how it scored.
The freaking thing passed the bar exam. It scored in the 90th percentile, meaning its score would rank in the top 10% of all humans. It scored in the 93rd percentile in the Evidence-Based Reading & Writing section of the SAT, and 99th percentile in the verbal section of the Graduate Record Examination.
Check out how it scored on a bunch of other tests as well:
These are all advanced tests and it crushed almost every single one. Without prior training on them.
Bar exam, Law School Admission Test, SATs, Graduate Record Examinations (except writing – “just” average there), USA Biology Olympiad, U.S. National Chemistry Olympiad, AP Art History, AP Biology … it destroys most humans in all of them.
There are even more test results on OpenAI’s website.
Let me remind you, we are still in the very, very early stages of this technology. Like, the infant stage.
What a time to be alive.
