I feel like I’ve been in a state of brumation since ChatGPT. I haven’t been writing any articles at all. Instead, I’ve been upskilling and silently keeping up on the recent going ons with ChatGPT and other LLMs.
See, I knew then that it was going to be revolutionary whenever GPT-4 comes out. What I didn’t expect was the wide berth popularity and adoption of LLMs once they announced ChatGPT.
A while back, I asked a startup founder what they think about data and data science in general especially when it comes to their company. Their answered surprised me. They said they weren’t concerned or thinking about it all. They further elaborated that ChatGPT is so usable by non data practitioners that it rendered data scientists irrelevant and obsolete.
This caught me off guard and left me speechless. The whole time I was thinking about how ChatGPT wouldn’t even exist if we’re not for the efforts of data scientists.
But then I started thinking, what if they’re right on some level?
As some of you may know, I’m a practitioner of NLP and I’ve made a living using spacy and other SOTA software to make sense of unstructured data. With the advent of ChatGPT, most NLP tasks can now be done with ease using ChatGPT’s natural language understanding. The days of heavy lifting code has gone away and been replaced with natural language. Text reigns supreme and is hailed as the next hottest programming language.
Are practitioners of NLP really obsolete?
This question has been in the back of my mind ever since. I remembered the Willy Wonka movie, “Charlie and the Chocolate Factory” where the boy’s father lost his job due to assemble line automation. The movie has a happy ending and the father ended up with the job of maintaining the machines that powered the assembly line. I then thought, if we are to take the moral lesson of this story, what’s next for the practitioners of NLP? Can we take a step back and work with the machinations that rendered our favorite technologies obsolete?
Yes, of course.
And this brings us full circle with ChatGPT.
Who’s going to bring ChatGPT into the hands of businesses and enterprise organizations? Who’s going to automate prompt engineering? Who will maintain the data pipelines?
We do, with the help of LangChain and other similar frameworks!
So, this new year, I resolve to write about LangChain and share what I’ve learned with the world through Medium.
Voila! Existential crisis solved.