Let's Chat

Watson reminds us technology cycles may still be relevant for ChatGPT

Happy Thursday, 👋

Over past few months we have met several companies discussing opportunities to leverage the platforms being developed by OpenAI, the parent company for various AI (artificial intelligence) concepts including ChatGPT. Over a hundred years ago science fiction authors begin writing stories about computers and robots able to hold meaningful conversations with humans, but in the early 1950’s such technology begin transitioning into reality.

More recent developments within the AI space have led to significant increases in user adoption metrics, indicating the potential for AI to materially influence numerous industries over the coming years. User adoption rates provide a metric to measure the potential economic impact of new platforms such as the AI platform ChatGPT which achieved 10 million users within 40 days, much faster than the social media platform Instagram which took 355 days to reach the same daily user count.

What Is It?

Chat Generative Pre-Trained Transformer or ChatGPT is an artificial intelligence program generating natural language responses. The platform is capable of passing the Turing Test, a benchmark for AI developed in the 1950s to measure if humans could differentiate between speaking with another person or a computer. As an example of the content ChatGPT is capable of producing, we asked it to provide a short self-description and received the following response:

ChatGPT is an AI-powered chatbot technology developed by OpenAI which enables the user to converse with a computer in a natural language conversation. The technology is based on the GPT-3 language model, which can generate text based on contextual input. ChatGPT can be used to create interactive chatbots for customer service, task automation, and more.

The more ChatGPT is used the better it becomes as it incorporates human feedback for continuous improvement. The platform was initially trained by scraping internet data which limits its knowledge base to information generated before 2022 and it relies primarily on English based content. When using ChatGPT, one of the disclaimers states it “may occasionally produce harmful instructions”. The disclosure is a rather ominous reminder that despite an ability to develop well-crafted responses we are still interfacing with an emotionless data driven algorithm.

Adoption Cycles

The pace of user adoption for ChatGPT has been remarkable but we still look back on the history of technological advancements to provide a degree of caution for investors. Emerging technology platforms often follow a hype cycle where user expectations increase rapidly as adoption rises, but slowly users start to reframe a technology’s usefulness as its limitations are discovered.

In the early 2000’s, IBM begin developing an AI program called IBM Watson. In 2011 it became famous for competing against human contestants on the TV show Jeopardy. Watson was successful in defeating two of the show’s champions which greatly increased the anticipation for what else Watson could achieve. At the time it was quickly approaching the “Peak of Inflated Expectations” in the hype cycle

Soon after its Jeopardy appearance, medical research began predicting ways Watson might discover new medical technologies. Watson was tasked with digesting a vast number of research journals, but scientists slowly began to realize most of the output was similar to academic literature reviews instead of groundbreaking research ideas. The “Trough of Disillusionment” became a reality as research institutions started canceling contracts to deploy Watson’s technology. Over time Watson’s capabilities became better understood as today it continues to function as a powerful AI tool for businesses to streamline and automate basic tasks but never regained its level of initial expectations.

Investor Excitement

The technology behind ChatGPT is far more advanced than Watson as ChatGPT appears to have achieved a level of human interaction capable of revolutionizing several industries including customer service, education, legal, search, literature, art, and software coding. Microsoft also recently invested a rumored $10 billion, which is ten times the total amount previously raised by the company.

The size of Microsoft’s investment will accelerate discussions around the AI industry as we already have met several emerging startups leveraging the ChatGPT platform as a foundational technology. Remaining mindful of the historical risks of technology cyclicality are increasingly important within the current environment as ChatGPT and the artificial intelligence space become more widely discussed.

Wishing everyone a great weekend.

-The Caymont Ventures Team.