The Ascendancy of AI

On December 5, 2022, Kevin Roose highlighted the “brilliance and oddities of ChatGPT” in a compelling article for the New York Times. He pointed out that segments of the Internet were bypassing the intentional limitations of ChatGPT's underlying large language model (LLM) technology, which powers chatbots like ChatGPT, Perplexity, Bard, and Meta’s BlenderBot 3. A year later, concerns emerged about the implications of such powerful technologies, especially after ChatGPT achieved the record for the fastest-growing user base ever, as reported by Ars Technica.

In 2023, Andrew Tarantola observed that ChatGPT had “transformed the world in just one year,” as the AI frenzy reached its peak. Influencers from Ezra Klein to Sam Harris debated whether AI would be humanity's savior or its destroyer. Reports highlighted disruptions in the tech industry as ChatGPT demonstrated coding abilities that could replace some software developers. There was a widespread assumption that AI would continue its relentless march forward.

The Potential Downfall

But what if both the optimists and pessimists are mistaken, and AI simply... fizzles out? What if it becomes another failed innovation, akin to NFTs, Amazon's Fire Phone, Facebook Home, or Google's series of missteps with Google Glass, Google G+, and Google Wave? This is a future that is seldom discussed, yet many of us instinctively recognize its possibility.

We sense it in our collective eye rolls when redundant chatbots are introduced to our favorite platforms. We cringe when Bumble announces “AI concierge” services for a dating app—implying that romantic prospects won't even need to converse. The notion that technology could evolve to the point where love could be facilitated by AI dating AI feels excessively ambitious.

Hubris is nothing new in the tech sector. Recall Mark Zuckerberg's 2021 prediction that the metaverse would profoundly influence every product they created. He envisioned a profound sense of presence, yet what has transpired since?

Consider when Amazon opted to sell Echo devices at minimal profit, banking on the belief that consumers would eagerly use Alexa for shopping. They overestimated the public's inclination to buy invisible products, resulting in substantial losses—nearly $10 billion on Alexa devices in 2022 alone. These examples illustrate that the tech industry often succumbs to an overzealous optimism, where monetary resources and corporate influence cannot compel us to embrace products we don't value.

A Question of Sustainability

ChatGPT's rapid growth is often cited as a sign of its inevitable success. It surpassed every major social media platform, including TikTok, reaching 100 million monthly active users in less than a year. For context, here’s how long it took various platforms to hit that milestone:

• LinkedIn: nearly eight years
• Facebook: four-and-a-half years
• YouTube: four years and one month
• Instagram: two years and six months
• TikTok: nine months
• ChatGPT achieved this in just two months.

These statistics are frequently used to highlight the technology's robustness, regardless of whether the speaker believes AI will be good or evil. However, the shortening timelines raise a crucial question: does rapid adoption truly indicate quality?

The speed of new technology adoption tends to accelerate as more people engage with the Internet and spend more time online each year—not necessarily because the technologies are improving in usefulness. Meta’s Twitter alternative, Threads, crossed the 100 million monthly active user mark in just five days.

Meanwhile, ChatGPT’s growth has stagnated, while Threads—hardly perceived as a potential harbinger of doom—has quietly surpassed it, boasting 150 million monthly active users. After experiencing an 11% decline since last year, ChatGPT has returned to its initial mark of 100 million monthly active users.

The Connection Disconnect

One possible explanation is that competitors like Google, Meta, and Perplexity are claiming larger segments of the generative AI market. Alternatively, it could be that users are trying these technologies and finding them unfulfilling. It’s entirely possible for impressive technology to be functionally inept.

The success of social media is clear—our need for connection drives it. However, we lack that same sense of connection with AI. There are legitimate concerns about AI's impact on the music industry, as it could lead to a flood of soulless compositions, produced solely for profit. Even drum machines require human input. The prospect of AI-generated music conjures a nightmarish vision of a lifeless void, far removed from human experience.

AI-generated writing often falls short, and even if it were perfected, few would rush to read it. Image generators briefly entertained us with amusing memes, but now they elicit eye rolls as they clutter our feeds. Memes do not pay the bills, and the fact that they currently dominate the use of AI image generators, a year and a half in, highlights the difficulty of monetizing these technologies, let alone forming a genuine attachment to them.

While tools like Adobe have integrated generative AI into their offerings, these enhancements function within established frameworks and are limited in scope. We’re not crafting the next Mona Lisa with generative AI in Photoshop; we’re merely tidying up imperfections.

A year and a half into the generative AI experiment, no one—not even an army of content creators incentivized to profit from AI—is producing significantly profitable work. The more I contemplate this, the more it seems likely that AI may become another failure, akin to a sophisticated NFT bubble with marginally higher quality than previous trends.

The Future of AI

Both AI optimists and pessimists often echo a common refrain: “You're focusing on today's AI, but think about what happens when these technologies improve. Consider the AI of tomorrow.” We tend to assume that generative AI will follow a familiar trajectory of technological advancement. When we look around, we see the successes. But could it be that our bias toward survival is clouding our perception?

Let me pose a question in response: have Facebook, Google, Gmail, or Instagram genuinely improved over the past decade? What about Zuckerberg’s metaverse? As of 2023, Meta has poured over $40 billion into the metaverse, yet it remains unpopular, and we certainly don’t adore it.

ChatGPT was a novel and engaging innovation at its launch, breaking the monotony of the past two decades of social media. However, that initial excitement has waned. With a clearer perspective, it's evident that we haven’t perfected any AI technology to date.

AI technologies, while distinct from LLMs like ChatGPT or image generators like DALL-E, are not a new phenomenon. We've had rudimentary AI-driven phone services since the 1970s. Personally, whenever I need assistance beyond simple billing, I instinctively press zero to reach a live representative. I have little faith in these flawed systems to manage my phone plan. Despite decades of corporate investment, these technologies have not improved.

If you think I’m cherry-picking, let’s consider Google Translate, which celebrated its 18th anniversary last month. Google has had nearly two decades to refine this tool. While it performs adequately for brief translations of minor significance, it quickly becomes unreliable with more complex sentences.

Input a moderately sized sentence in a language with grammar more intricate than English, such as German, and it falters, producing nonsensical output. You wouldn’t trust it for legal documents.

Even recommendation algorithms are wearing thin on us. How many hours have you spent scrolling through Netflix, surrounded by endless suggestions yet unsure what to watch?

While it feels like we are on the brink of a technological revolution, it’s plausible—even likely—that generative AI ends up being another Google Translate: a technology that succeeds in about 20–30% of tasks but fails at the rest.

That said, I do have concerns. I worry about the ramifications of generative AI, but not because it might be overly effective—solving or destroying our world. My apprehension lies in its potential shortcomings. Given the current incentives and the vast number of individuals eager to make a quick profit online, I fear we will inundate the Internet with subpar content that few enjoy. It could become an incessant barrage of low-quality material. However, spam has existed for decades, and it hasn’t brought about the apocalypse.

Get my emails here. Visit The Science of Sex on Substack, where we explore human sexuality through the power of science.

Visit us at DataDrivenInvestor.com

Subscribe to DDIntel here.

Featured Article:

Join our creator ecosystem here.

DDI Official Telegram Channel: https://t.me/+tafUp6ecEys4YjQ1

Follow us on LinkedIn, Twitter, YouTube, and Facebook.