Competing with Large Language Models

When considering the future of human professions, one is immediately faced with an inexorable force driving us towards a new frontier: Large Language Models (LLMs). These technological behemoths are already challenging human capacity, even at their current stage, which we can consider the worst they will ever be.

They’re supporting doctors in diagnosing diseases that multiple medical professionals could not previously identify. They’re producing better assembly than the most eminent compilers of modern times, the development of which has been ongoing for several decades. As daunting as this may seem, the question that permeates most minds is not if LLMs can surpass human, but rather, when they will.

BERT was released in 2018 by Google, which arguably was the AlexNet moment of NLP. Five years from that moment, the capabilities of LLMs have seen an exponential increase — showing no sign of slowing down ??.

The quintessence of human intelligence unfurls like a triumphant banner, intricately woven with threads of skills, seasoned with the nuances of taste, and illuminated by the radiant spark of creativity.

Skills have long been regarded as the foundation of professional success. However, as technology continues to advance at an unprecedented pace, the value of certain skills is diminishing. The rise of automation and AI has led to the automation of many routine tasks, rendering some skills redundant.

When we peer beyond the horizon, it becomes evident that skills alone will not shape our future; they are, for the most part, a depreciating asset.

While proficiency in traditional skills remains essential, it is becoming increasingly crucial for professionals to double down on things that make us human.

Unlike AI, humans possess a unique spark of creativity that sets us apart. Creativity allows us to think outside the box, find unconventional solutions to complex problems, and drive innovation forward. This human element cannot be replicated by today’s language models, making it a valuable asset in an age where automation threatens to replace many traditional job roles.

Furthermore, taste, often associated with subjective judgment, plays a significant role in various domains such as art, design, and even decision-making processes. AI may have access to vast amounts of data, but it lacks the ability to appreciate aesthetics, culture, and personal preferences. Humans, with their inherent ability to discern and appreciate nuances, act as connoisseurs in these domains.

LLMs possess an unparalleled ability to accumulate knowledge, surpassing even the most well-read humans. Their vast knowledge base allows them to perform various tasks with remarkable accuracy. However, LLMs face challenges when it comes to recall, often leading to inaccuracies or hallucinations. Despite their current limitations, LLMs possess a diverse skill set that far surpasses human capabilities. They can perform on par with the average human in many tasks, making them formidable competitors. As LLMs continue to evolve, their performance is expected to improve exponentially — a claim which is supported by data from the previous years.

The realm of creativity and innovation doesn't tremble in the face of advancing AI; it instead finds opportunities for growth. Skills change, adapt, and transform. They are not static but dynamic, flowing like a relentless river. The human capability to innovate, to create, and to think unconventionally, however, remains resilient in the face of these rapid transformations.

To paraphrase a famous saying, it's not about the cards you're dealt, but how you play the hand. As we venture into the future orchestrated by AI, instead of arresting the inevitability of change, we should harness our creativity while also leveraging the capabilities of LLMs — the key to embarking on an extraordinary journey within this brave new world.

This article is written by GPT-4 and the prompt engineer was Muhammad Saad , CTO at Antematter.io.