(Bloomberg Opinion) – Artificial intelligence (AI) is evolving in ways that are difficult for the human mind to comprehend. Nothing happens for a long time, and then suddenly something happens. The current revolution of large language models (LLMs) like ChatGPT was the result of the emergence of “transformative neural networks” around 2017.
Artificial neural system
What will the next half decade hold in store for us? Can we judge their quality based on our current impressions of these tools, or will they surprise us with their development? As someone who has played with these models for many hours, I think many people will be surprised. LLMs will have a significant impact on our business decisions, our portfolios, our regulatory structure and the simple question of how much we as individuals should invest to learn how to use them.
Just to be clear, I’m not an AI tabloid. I don’t think there will be mass unemployment let alone the Skynet goes live scenario and the ensuing destruction of the world. I believe it will provide a lasting competitive and learning advantage for the people and institutions that can benefit from it.
I have a story for you about chess and a neural network project called AlphaZero at DeepMind. AlphaZero was founded in late 2017. He began training almost immediately, playing hundreds of millions of chess games against himself. In about four hours it was the best chess playing company ever created. The lesson from this story: Under the right conditions, AI can improve very, very quickly.
LLMs cannot keep up with this pace as they are faced with more open and complex systems and also require ongoing business investments. Still, recent advances have been impressive.
Continue reading the story
I wasn’t thrilled with GPT-2, a 2019 LLM. I was intrigued by GPT-3 (2020) and very impressed with ChatGPT, sometimes referred to as GPT-3.5, which was 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 of oddities have become an integral part of the everyday working life of many people I know.. This semester I will be teaching my students to write an article using the LLMs.
ChatGPT, the model released late last year, received a D on an undergraduate economics exam taken by my colleague Bryan Caplan. Anthropic, a new LLM that is in beta and expected to go on sale this year, took our final exam passed in Law and Economics with clear, well-written answers (in case you’re wondering, it was scored blindly). ). It’s true that the current results from LLMs aren’t always impressive. But consider these examples and that of AlphaZero.
I have no prediction on the rate of improvement, but most normal economic analogies don’t apply. Cars get a little better every year, as do most things I buy or use. LLMs, on the other hand, can make leaps and bounds.
Still, you may be wondering, “What can LLMs do for me?” I have two immediate answers.
First, they can write software code. They make a lot of mistakes, but it’s often easier to edit and fix them than to write the code from scratch. They are also often more useful for writing the boring bits of code, freeing up talented human programmers to experiment and innovate.
Second, they can be guardians. These LLMs already exist and will soon be greatly improved. They can provide very interesting answers to questions about almost anything in the human or natural world. They’re not always reliable, but they’re often useful for new ideas and inspiration, not for fact-checking. I hope that they will be integrated into the checking and search services soon. In the meantime, they can improve their writing and organize their notes.
I started to divide the people I know into three camps: those who don’t know anything about LLM; those complaining about their current LLMs; and those who foreshadow the amazing future that awaits us. What is fascinating about LLMs is that they do not follow uniform and continuous development rules. Rather, they are like a larva about to become a butterfly.
It is human, if I may use the word, to worry about this future. But we also have to be prepared for it.
Original note: AI is improving faster than most people realize: Tyler Cowen
For more stories like this, visit bloomberg.com
©2023 Bloomberg LP