>>40599831
>>In the actual study PhD level experts who took the same test questions averaged ~66%
>Ok and chatgpt got 39%, what's your point? That doesn't imply parity with PhD.

Sorry, that was the initial score on the benchmark. Models now surpass 66%:

https://developer.nvidia.com/blog/nvidia-llama-nemotron-ultra-open-model-delivers-groundbreaking-reasoning-accuracy/

https://artificialanalysis.ai/

>This was published in early 2023 meaning BY DEFINITION this isn't reasoning
Your focus on "reasoning" is semantic and aesthetic focused. Whether it's an LLM or LRM, the ability to answer correctly is what demonstrates human parity.

>>The entire point is to showcase parity in human intelligence. Cope.
>Ok and that doesn't. It literally just shows it's really good at pulling from textbooks basically.
False.

>>believing that because AI doesn't "think"
>I know it doesn't, it's impossible on a fundamental level for LLMs to think.
Humanity does not have a consensus framework for what "Thinking" is. Again, you're focusing on semantics surrounding human-like aesthetic of thought. This is irrelevant if these models are capable of FUNCTIONALLY replacing humans.

>If they achieve human parity in intellectual tasks required for human work (they already have for many)
>[citation needed]
I just provided multiple examples of human intelligence benchmarks of which AI have passed; some of which are the basis for allowing expert-level practice in their respective domains. If you can't put two and two together you're a lost cause. I challenge you to pass the USMLE, or score >66% on GPQA (you can even use google).

Why do you think we are in a literal geopolitical arms race with China re: AI advancement? Do you think Microsoft, Google, NIVIDA, Meta amongst others are investing 100s of billions $USD collectively in R&D, infrastructure just as a public facing money grab gimmick?

Let me guess, you're not even aware of physical AI and the pending robotics revolution