(Bloomberg) – Artificial Intelligence (AI) is advancing in a way that is difficult for the human mind to comprehend. For a long time nothing happened, and then suddenly something happened. The current revolution of Language Large Models (LLMs) such as ChatGPT is a result of the emergence of “transformational neural networks” around 2017.
What will the next half decade hold for us? Can we judge their quality based on our current impressions of these tools, or will they surprise us with their sophistication? As someone who has spent many hours playing with these models, I think a lot of people will be surprised. An LLM will have major implications for our business decisions, portfolios, and organizational structures and the simple question of how much should we as individuals invest in learning how to use it.
Just to be clear, I am not an AI tabloid. I don’t think it would cause mass unemployment, much less a “Skynet goes to life” scenario and consequent world destruction. I believe it will be a lasting competitive and educational advantage for individuals and organizations that are able to take advantage of it.
I have a story for you about chess and a neural network project called AlphaZero at DeepMind. AlphaZero was created at the end of 2017. Almost immediately, he started training by playing hundreds of millions of chess games against himself. Within about four hours, it was the best chess-playing entity ever created. The lesson from this story: Under the right circumstances, AI can improve very, very quickly.
LLM cannot keep up with this pace, as it is faced with more open and complex systems, and also requires continuous business investment. However, recent developments have been impressive.
I wasn’t thrilled with GPT-2, 2019 LLM. I was fascinated by GPT-3 (2020) and really liked ChatGPT, which is sometimes referred to as GPT-3.5 and released late last year. GPT-4 is on the way, possibly in the first half of this year. AndIn just a few short years, these models have gone from being a mere curiosity to becoming an integral part of the work routines of many people I know.. This semester, I will be teaching my students to write an essay using an LLM.
The ChatGPT model, released late last year, received a D on an economics exam for college careers given by my colleague Bryan Caplan. Anthropic, a new MA that is in beta and expected to go on sale this year, passed the graduate law and economics exam with clear, well-written answers (in case you were wondering, blind spots were used). It is true that the current results of the Master of Laws are not always impressive. But consider these and other examples from AlphaZero.
I don’t have a prediction of the rate of improvement, but most of the normal economic comparisons don’t apply. Cars improve modestly each year, as do most things I buy or use. LLM, on the other hand, can jump.
However, you may be wondering, “What can an LLM do for me?” I have two immediate responses.
First, they can write program code. They make a lot of mistakes, but it’s often easier to modify and fix than to write the code from scratch. They are also often more useful for writing the boring parts of code, freeing up talented human programmers to experiment and innovate.
Secondly, they can be guardians. This LLM already exists, and it will be improved a lot soon. They can provide very interesting answers to questions about almost anything in the human or natural world. It’s not always reliable, but it’s often useful for new ideas and inspiration, not a fact-check. I hope they will be integrated into the scan and search services soon. Meanwhile, they can improve their writing and organize their notes.
I’m starting to divide the people I know into three camps: those who don’t know about LLM yet; Those who complain about the current LLM; And those who feel the amazing future that lies ahead. The interesting thing about LLM is that it does not follow uniform rules of continuous development. Rather, they are like a caterpillar about to become a butterfly.
It is human, if I may use that word, to be concerned about this future. But we must be ready for this as well.
Original Note: Artificial Intelligence Is Improving Faster Than Most Humans Realize: Tyler Cowen
More stories like this are available at bloomberg.com
© 2023 Bloomberg L.L.C.