Human PHD Level LLMs Within a Year?
July 23, 2024•456 words
"In a sense, artificial intelligence will be the ultimate tool because it will help us build all possible tools." ~ K. Eric Drexler
"You can't just stop technological progress. Even if one country stops researching artificial intelligence, some other countries will continue to do it. The real question is what to do with the technology. You can use exactly the same technology for very different social and political purposes. So I think people shouldn't be focused on the question of how to stop technological progress because this is impossible. Instead the question should be what kind of usage to make of the new technology. And here we still have quite a lot of power to influence the direction it's taking." ~ Matt Haig
"Artificial intelligence would be the ultimate version of Google. The ultimate search engine that would understand everything on the Web. It would understand exactly what you wanted, and it would give you the right thing. We're nowhere near doing that now. However, we can get incrementally closer to that, and that is basically what we work on." ~ Larry Page
Elon Musk's prediction about the future capabilities of large language models (LLMs) like Grok by the end of the year is quite intriguing.
Extremely wealthy companies are investing hundreds of millions of dollars in thousands of GPUs and CPUs for supercomputer complexes dedicated to parallel computing and quantum computing. These complexes occupy acres of space and consume massive amounts of electricity, with some located in places like Memphis, Tennessee.
Considerable strides have undeniably been made in artificial intelligence, resulting in the remarkable advancement of Large Language Models (LLMs). Musk's viewpoints in his interview with Jordan Peterson on the potential for LLMs to attain human-level intelligence, and it is widely acknowledged that current LLMs show knowledge and understanding at a level comparable to that of a graduate student at a university.
As Musk pointed out, the rapid progress in AI research, with the continuous training of large language models (LLMs) on ever-increasing data and the development of new capabilities, is undeniable. However, it's important to remember that there are still significant challenges to overcome, such as common sense reasoning, dealing with ambiguity, and achieving true creativity.
The AI community is currently debating the timeline for achieving human-level AI or Artificial general intelligence (AGI). While LLMs like Grok may make significant strides by December, the question of whether we will reach Ph.D. level human intelligence by then remains to be seen. It's a topic that continues to intrigue and challenge researchers and governments.