A Skeptic’s Guide to LLMs: ChatGPT and Beyond
Six use cases that highlight AI’s contribution to life and work
Justin, a Monkeyjump Labs executive leader and founding partner, was skeptical of ChatGPT when it was released in late 2022.
He felt the same way about the crypto-backing ledger database, Blockchain, too.
Even though technology is his passion and activity of choice, when Blockchain technology came out, its knee-jerk popularity drew up red flags for Justin. It seemed to be “in fashion” to try and use blockchain technology in all kinds of different projects.
Companies did strange things like releasing their own crypto currency or using blockchain as a backing database, when a normal database — such as a relational database or a document database — would have likely been a better fit.
Other outlandish cases in which blockchain was used include Thanksgiving turkey tracking, and even a Long Island iced tea company changing their name to Long Blockchain Corp.
Companies identified that a significant amount of money was being invested into the blockchain system, creating a perfect FOMO (fear of missing out) storm. Other reasons that had companies jumping on the bandwagon included:
It was of interest to potential employees that their next company understood what was most relevant. This produced high interest in working for companies who took the leap to incorporate blockchain technology.
External potential customer interest
Customers found confidence in the companies who were staying on the cutting edge.
Ignorance (this was most worrisome)
Companies missed the mark when it came to fitting the right technology to the right problem or fitting the right solution for a customer’s problem. Companies were eager to use something because it was shiny and new– not considering whether they are focused on the customer’s problem and solving it. The chase seemed to reveal that they were picking the solution before the problem.
For many, the same bandwagon mentality for Blockchain seemed to apply to the LLM (Large Language Model) ecosystem — the current largest player, being ChatGPT.
Would ChatGPT be another infatuation that would fade with time and reality checks?
According to Justin, the problem didn’t lie in the technology itself, but in the way that people wanted to use it. To expand a bit on point three above, if a client asked, how can we use ChatGPT for our product or brand? Justin would suggest asking a different question. He says that the very first questions should be; what is the customer’s problem? Does ChatGPT solve the existing problem? Instead, it seemed as though people were asking the question, are we manufacturing a problem for a solution that we are excited about?
Chatbots are a beneficial way to control common questions and navigation guidance around a website or software. This type of ChatGPT uses an interface that is a substitute for a human salary which might be too expensive for a company to purchase, solving both a hiring problem and potentially increasing revenue.
Internal chatbots like the “digital employee” that we built for our own company (named George), use ChatGPT to answer employees' company-specific questions like, “what is our time off policy?” or “How do I use the new espresso machine in the breakroom?”. The advantages include a more streamlined onboarding process for new employees.
Creative writing benefits the companies and employees who have to produce a lot of creative writing pieces like articles, newsletters, news reports, social media content, etc. ChatGPT solves the problem that already exists, auto-completing text in motion and by generating new content based on prompts given.
Software developers can use ChatGPT to assist in the agile process. When developers get a feature, they break it down into tasks, which contain descriptions. ChatGPT could assist in writing those descriptions to save on time. Developers could also inquire with the help of the action model of ChatGPT which would automatically generate those feature set descriptions when triggered by a prompt. When coding, ChatGPT can cut down on build times and by using an autocomplete feature. ChatGPT can be easily integrated into a developers IDE (Integrated Developer Environment)
Note taking, just one of many administrative duties, is often one of the most time consuming tasks across industries. ChatGPT is already solving this problem by taking notes for and during meetings, and even adding meetings to the calendar that were decided within the conversation.
Pathology is only one area of many that is benefiting from ChatGPT. Within the pathology space, AI can interpret and analyze pathology reports, getting results and answers faster than medical journals.
At the end of the day, humans are still putting AI and ChatGPT to work for us, and not the other way around. Nothing on this earth is fool-proof, but we believe that ChatGPT is a tool that helps experts in their industry move faster, when they know how to use it.
AI is here to stay
It’s perfectly fine that many people have watched ChatGPT’s entrance and performance play out without getting involved. But, the reality is that AI is here to stay. Companies are adopting many different forms of AI to increase production, streamline processes, assist employees, and so much more.
If any of the above use cases are missing from your business, we can help you! As technology experts, we understand how to determine what kind of AI is best for your business.