Taking stock of a new wave of AI revolution

Apple recently published a paper that subtly acknowledges what many in the Artificial Intelligence (AI) community have been hinting at for some time: Large language models (LLMs) are approaching their limits.

- Vivek Wadhwa

These systems—like OpenAI's GPT-4—have dazzled the world with their ability to generate human-like text, answer complex questions, and assist in tasks across industries. But behind the curtain of excitement, it's becoming clear that we may be hitting a plateau. This isn't just Apple's perspective. AI experts such as Gary Marcus have been sounding the alarm for years, warning that LLMs, despite their brilliance, are running into significant limitations.

Yet, despite these warnings, venture capitalists (VCs) have been pouring billions into LLM startups like lemmings heading for a cliff. The allure of LLMs, driven by the fear of missing out on the next AI gold rush, has led to a frenzy of investment. VCs are chasing the hype without fully appreciating the fact that LLMs may have already peaked. And like lemmings, most of these investors will soon find themselves tumbling off the edge, losing their me-too investments as the technology hits its natural limits.

LLMs, while revolutionary, are flawed in significant ways. They're essentially pattern-recognition engines, capable of predicting what text should come next based on massive amounts of training data. But they don't actually understand the text they produce. This leads to well-documented issues such as hallucination—where LLMs confidently generate information that's completely false. They may excel at mimicking human conversation, but they lack true reasoning skills. For all the excitement about their potential, LLMs can't think critically or solve complex problems the way a human can.