Generative AI is a Solution Looking for a Problem

Generative AI and Large Language Modules have failed to live up to the hype and companies are becoming increasingly desperate.

When generative AI first cropped up in the early 20s, the hype engine was intertwined in the rapid rise of discussions surrounding the technology. It was sold as a technology that can do anything. Do you want to commission a painting? Let AI do it. Need to write an e-mail? AI can do that. Want to write a legal brief? Not a problem as far as AI is concerned. Building a website? Just let AI do it. Coding a new app? That can be done almost instantly with AI. If it can be generated by a human, AI can handle it. At least, that according to the people leading the charge to hype AI up.

While this all sounded believable at first, disappointment in the technology quickly followed. People who wanted to be the early adopters of the technology under the mistaken impression that not learning it now means that you would be left behind quickly learned of things like AI hallucinations and generative AI slop. They were sold on the idea that AI can perfectly deliver on anything with a simple prompt and that just wasn’t the reality. That’s when the cracks started to form on the image of generative AI.

As the years progressed, the promises became even more outlandish. Forget just being an assistant or just being a fun little toy that can boost productivity. No, AI was going to take over everything. You didn’t need teams of people to keep a company running, just let AI handle it all because that’s what it is capable of doing. Additional hype became the norm such as “agentic AI” and “vibe coding”. The former is seemingly to fix all the problems of AI not understanding prompts while the latter would have you believe that there is a whole lot of people using AI to handle coding tasks to the point that a general term had to be created. The purposes were the same, though: to pretend that AI is the new reality and everyone was just using it.

This hype didn’t match reality as people who were suckered into just letting AI handle everything continued to get burned by it. It got to the point that we just started creating a massive list and tacking onto it whenever the latest sucker got burned by AI. The examples of this include lawyers getting in trouble for fake AI inserted citations in legal briefs, the CNET scandal, the Gannet Scandal, bad “journalism” predictions, fake news stories, more fake stories, Google recommending people eating rocks, the 15% success rate story, bad chess tactics, the Chicago Sun-Times scandal, a Canadian team submitting fake legal citations in their legal briefs, other attorneys submitting fake citation filled legal documents, the 91% failure rate story, AI deleting user data, the lawyer who got fined $10,000 over a bogus AI written legal brief, AI killing workplace productivity with workslop, AI having an 81% failure rate in summarizing news content, AI Overview giving out bad health advice, AI only being able to successfully complete 2.5% of commission work successfully at best, and AI slowing software developers down by 19%.

While all of this would be little more than a technological humiliation for developers, there is a much more serious side to this. That has to do with the economics of it all. The reason why there is so much hype is because large tech companies are investing billions of dollars into the technology. They are investing massively in things like RAM (which is why RAM is so expensive), data centres, and in the development of the technology. They are constantly courting potential investors with massive promises that this is the next industrial revolution and that getting in now is little more than taking advantage of a once in a generation opportunity. Sometimes, the promises amount to AI companies being worth multiple Amazon companies or whatever other obviously exaggerated claims they are coming up with.

While a number of people have been roped in on this, the missing piece of the puzzle is public buy in for the product. The claims are that everyone is going to be using it someday if they aren’t using it already and everyone will see what a magical piece of technology this will truly be. The original idea was that the companies would charge money for AI and people will be paying those fees as it has become essential for operations. When everyone on the planet is paying you, the payoff is almost unimaginable.

Yet, when the technology hit the market, the public buy in was nowhere near what the hype clearly over promised. As the technology continues to fail the people that did decide to jump into the technology, others (such as myself) looked at the technology and said, “no thanks, I’m good.” You were either getting burned by AI or simply avoiding the headaches for the most part. With less than expected public buy in, some investors started growing nervous. Where was the return on investment that was promised? They said that everyone would be heavy into AI by now, yet that hasn’t happened.

It’s thanks to the lack of public buy in that AI companies have seen a few pretty bumpy moments in the stock market. People, such as myself, recognized that AI is increasingly looking like it is in a massive bubble. With so much of the stock markets tied to tech companies who are heavily involved in AI development, that is leading to worries of just how bad the global economy is going to suffer whenever the bubble finally bursts.

Obviously, the developers are pushing back against this. In addition to trying to continue to fuel the one engine that has fuelled all of this (hype), the companies have started making massive investments as well to try and give the impression that there is buy in. Smaller companies associated with AI get headlines of billions of dollars being poured into their company. Often, it’s just other AI companies shuffling huge sums of money with the hopes that the headlines that they’ll get more investment dollars. What they ended up with is a giant financial circle jerk that really only exacerbated the fears of an AI bubble. One such image that is floating around that really paints a grim picture of the situation really highlights just how self-financing this whole situation is:

To be clear, nothing about the above image is financially healthy. It’s the kind of things that make financial stock bubbles, well, bubbles, because the circular financing will eventually break down. In a bid to make the situation seem more viable, the companies really only made the situation worse. This goes back to public buy in.

So, in recent years, companies have started flooding the market with references to AI. If you are buying a cell phone, you are inundated with large signs saying how this cell phone is fuelled by whatever AI they are trying to pawn off onto customers. If you are using creative software suites such as Adobe, Adobe will be pushing AI hard onto the those customers. Even buying a car is increasingly becoming a case of navigating all the advertisements about AI. Even here on Freezenet, AI is being pushed hard to supposedly make my articles better (seriously, screw off with that noise). Windows has Copilot forced onto consumers and Microsoft has pretty much made it one of those impossible to uninstall programs as well. Anything that is useful has AI being crammed into it as companies practically beg users to use their AI solutions.

Even Google is struggling to find public buy in with their products. Once upon a time, their AI Mode for Google search was one of those “select few users” who want to try it. A pop up appeared asking me if I want to test run it (to which I declined). After a few months, the pop ups started cropping up at an increasingly frequent rate. At times, every time I go to search for something on Google, I would get those annoying pop-ups to which I keep repeatedly hitting the decline button. AI Mode is less of something that select few people are having the privilege of trying out and more something that Google is spamming the ever living fuck out of in a desperate bid to get people to use it.

So, faced with users like me, Google is also now trying to shove its crappy AI into Gmail. A few weeks ago, Google announced that they are cramming their AI into Gmail in a desperate bid to get customer buy-in there as well. In the weeks since, numerous AI “features” are being squeezed into as many corners and spaces within the e-mail.

Forget getting users to pay premium prices for AI responses, AI companies are struggling to even give AI away for free. Whether users think using AI sounds too complicated or users realizing that AI is not even close to what was promised and simply opting out, AI companies are struggling to get the numbers they promised investors. What good are multi-billion dollar server farms if only a handful of users end up using them?

The stark reality is that AI, at most, is capable of making small suggestions for things like writing. Instead of putting a comma there, maybe it should be placed there? If one word in this sentence is past tense, maybe the other word in that sentence should also be past tense. What about changing the word “glisten” to “shine”? While that is substantially more realistic to what AI can do, it is by no means anywhere near as sexy as “press the button and AI will handle the rest”.

AI was sold as a cure-all solution to all of life’s problems. When it failed to meet the technological threshold of being all that, AI became a solution looking for a problem. What problem can AI solve that makes life easier? That is becoming an increasingly short list – and a list that is full of asterisks that have the explanation that it is much slower and less efficient than just handling the problem yourself. You could spend the 20 minutes to write the article or you can spend hours writing prompts to accomplish the same thing. What is a better use of your time?

Almost no one in the public asked for AI. Even after 2023 when the AI hype began, there was a large contingent of people who basically said “nope” to AI. No matter how many times people pushing AI insist that the technology is getting better and better or that people who don’t use AI risk falling behind (a very common way to try and stoke Fear of Missing Out or FOMO), not enough people out there are buying into AI that justifies the promises and hype. With companies now thrashing about struggling to even give AI away, we might, mercifully, be one step closer to getting that bubble popped so we can move on with our lives.

Drew Wilson on Mastodon, Twitter and Facebook.