OpenAI, the former nonprofit (it’s complicated) tasked with shepherding ethical A.I., has released GPT-4, the most advanced version of its A.I.-powered large language model. The impact on the tech industry (and at least some technology jobs) could be huge.
If you’re interested in A.I. and machine learning (or your job depends on it), it’s worth checking out what exactly OpenAI has to offer here. “We are releasing GPT-4’s text input capability via ChatGPT and the API (with a waitlist),” the organization wrote in a statement breaking down the technology. “To prepare the image input capability for wider availability, we’re collaborating closely with a single partner to start. We’re also open-sourcing OpenAI Evals, our framework for automated evaluation of A.I. model performance, to allow anyone to report shortcomings in our models to help guide further improvements.”
OpenAI claims that ChatGPT-4 not only outperforms existing large language models, but can perform well on several types of exams meant for humans, including the LSATs, the SAT math section, Leetcode, various AP tests, and even the theory knowledge portion of the sommelier test.
While the day your A.I. chooses a wine for you is likely far off, tech pros are probably wondering if GPT-4 might accelerate industry-level changes to how they write code and otherwise do their jobs. To that end, OpenAI’s statement calls out the ways GPT-4 isn’t fully reliable: “GPT-4 poses similar risks as previous models, such as generating harmful advice, buggy code, or inaccurate information.”
In other words, you probably shouldn’t trust GPT-4 to code a largely bug-free program for you, and it might not be a good idea to integrate it into an automated chatbot for customer service—at least not without a good deal of human supervision. However, it’s clear that the technology is iterating rapidly, and OpenAI seems determined to make future versions as “safe” as possible.
It’s not inconceivable that, within a few years, machine learning could streamline or even eliminate a range of coding tasks, but it doesn’t seem likely that even the most sophisticated model will successfully replicate human attributes such as creativity, effective teamwork, and intuition. For those seeking career security within this brave new world, “soft skills” such as empathy, communication, and leadership can prove useful. In a similar vein, jobs such as project or program manager, team lead, and creative director will also likely endure well into the future.
