Thoughts from a Sandwich Maker on Elon Musk’s Call to Pause AI
First off, I'm not Elon Musk, nor am I Bobby Flay. The vast majority of my professional experience revolves around slicing deli meats and serving sandwiches. However, my current journey in running a restaurant technology business has put me at the center of this AI drama, and I have some thoughts to share.
What's going on?
ChatGPT made headlines earlier this year as an impressive "AI language tool designed to engage in human-like conversations, understand context, and provide useful responses" (according to ChatGPT). Many companies, including ClearCOGS, use it to build AI assistants and operating systems. Google, Apple, Facebook, and Microsoft (which recently invested $10 billion in ChatGPT's parent company, OpenAI) are all launching similar systems.
Fast forward to today, Elon Musk and AI experts have called for a six-month break in developing AI systems more powerful than OpenAI's GPT-4, citing potential risks to society. They want to ensure AI has positive effects and manageable risks. The open letter also advocates for improved safety protocols and collaboration between AI developers and policymakers to enhance AI governance.
I find myself in a unique position, having spent the better part of the last few years immersed in the world of tech, startups, and AI. Every day, I listen to the tech class argue that generative AI is either (1) nothing more than a fancy user interface or (2) the greatest technological revolution since the Internet. I can't tell you which it is (after all, I'm a restaurant guy, not a technologist), but I can tell you how it feels having spent a fair amount of time building a product with it.
My first exposure to machine learning a few years back left me in awe of its power to predict demand patterns. In fact, it was so game-changing that it led us to start our restaurant demand forecasting company, ClearCOGS.
It took six months and a team of data scientists and programmers to build our machine-learning system that answers questions like, "how much bread do I need right now?". Venture capitalists and restaurant executives still balk at the idea that a team like ours could build "AI" solutions in such a short time.
What's happening right now is different.
In January, we began building our AI Ops management system on ChatGPT. In less than a month, we had 24 AIs, each designed and trained to accomplish specific tasks for restaurant operators. There's one that specializes in profitability, another in inventory analysis, one for training staff, and so on. They all work together like something out of The Jetsons.
But that was back in January before ChatGPT was available! On March 1st, the much more powerful ChatGPT was released for development. Two weeks later, the 100x more powerful GPT-4 was unveiled for testing and is already in the hands of some companies to build with!
The power of these tools and the speed of development has been breathtaking. It's hard for me to see this as anything but an inflection point, and it sure doesn't feel like we've built the infrastructure as a society to handle it.
Should there be an AI “pause”?
The fact that we're able to build things like this to help restaurants is great. But at the end of the day, not everyone has good intentions. In fact, it often feels like every tech breakthrough’s first use-case is to try and sell us products more effectively. Then, five years later, it finally shows up in the workplace to actually help us.
Is this the beginning of Skynet from Terminator… doubtful.
But, you know those constant spam calls about your car's warranty? The "IRS agent" requesting gift cards? The fake social media accounts trying to sway public opinion?
Those are AI language models, and they're already beginning to be upgraded with what ChatGPT told me are "impressive human-like abilities".
The sad reality is that we live in a world where AI language tools tend to be used against us more often than to help us. At the end of the day, we should probably be a bit frightened. Is taking a “pause” the right thing to do? I have no clue. But these new AI tools are powerful and getting ever easier to use, so taking a step back to think things through might not be the worst idea.