Why I’m Exhausted by the Relentless Hype

I’m tired… and I’m pretty sure the so-called “artificial intelligence” is tired too. All I hear is ChatGPT this, Gemini that, and it’s causing me a lot of mental fatigue.

I’ll be honest: I use “AI” just about every day for various things. Mostly stupid questions that I want answered without wading through search results. It’s fast and sometimes accurate. Once you accept that, it can be a really useful tool.

What I really want is for this stuff to fade into the background — for AI to be treated like Google search or cruise control: always there, rarely discussed.

But this post isn’t about what I use AI for (I do want to write that one), it’s about how I can’t eat a sandwich without someone shoving AI down my throat.

In 2018, I wrote a blog post about artificial intelligence, it’s nice to see how my views have changed.

I recently went to a conference put on by a pretty niche software company and it was absolutely littered with mentions of AI. How they were using it to develop their software, how they’re adding their own ChatGPT-style chatbot to the system, and what their customers want out of AI. That’s when the fatigue really set in.

The Hype Tax on Real Work

The problem isn’t just that it’s annoying. Every keynote, roadmap, and product pitch that has to be wrapped in “AI” leaves less space for the boring-but-important work: fixing UX, paying down technical debt, securing systems, and building things people actually asked for. Hype eats oxygen.

As I said earlier, I’ve embraced AI, both in my personal life and my work life. I’m actually spearheading introducing AI into the company I work for. The conference is just the latest example of how AI has intruded into our lives, whether you like it or not. Wall Street is all-in on the idea that AI is going to change our lives forever. While I don’t disagree, it’s not the internal combustion engine or commercial air travel; it’s the flavor of the month.

“Hype eats oxygen.”

I’ve gone back and forth on whether I think it’s a bubble and I feel like I’ve landed on: it’s not. It will hit a peak and then die down, but it’s not going to implode like the dot-com era or the housing market. We’ll be using AI in some capacity for quite some time.

While AI as we know it today (ChatGPT, Sora, Claude, Gemini, Apple Intelligence, etc.) is impressive, it’s not what we’ve been dreaming about in science fiction movies for decades. The goal that some people believe we’ll get to is called artificial general intelligence, or AGI. That’s the hypothetical point where an AI system can match or exceed human abilities across most or all cognitive tasks.

Realistically, I probably won’t live to see that day, or it will never come. People like Sam Altman think we’re closer than that, but that’s his job: to sell the vision so his company keeps getting the funding to keep the wagon rolling. Much like full self-driving cars, AGI is the end goal — and you see how well that worked out for the flood of self-driving car companies that were all “right around the corner.”

What We Actually Have: Prediction Engines, Not Brains

I’m also tired because the general public doesn’t understand how the kind of “AI” we actually use — large language models and other predictive models — actually works. It’s not “intelligent” in the human sense at all. What we mostly use are large language models (LLMs) that are very good at predicting the next token (piece of text) given the previous context, based on the training data they were fed.

You can see this when it “hallucinates”, confidently inventing fake citations, people, or products that sound right but don’t exist. That’s not evil robot behavior; it’s just a statistical text machine doing exactly what it was built to do: predict plausible words, not guarantee truth.

Setting aside for now the legality and ethics of what data they were trained on: it’s not a novel idea, but the implementation and access we have now is the best we’ve ever seen. The state of the art lets a web browser hit a giant server farm and run huge amounts of math on very advanced graphics processing units (GPUs). And, as an aside, we probably have crypto to blame/thank for this. Cryptocurrency mining helped push the GPU market into overdrive, and now AI workloads are driving yet another wave of demand for the same hardware.

Think about how often you hear about AI. I bet it’s a few times a day, every day. Depending on where you travel, in real life and on the internet, it could be a lot higher. My job, interests, and hobbies take me further into the deep end, so I feel it more. Again, I’ve embraced it — I borderline love LLMs — but I’m tired.

“…confidently inventing fake citations, people, or products that sound right but don’t exist.”

Imagine if ChatGPT or similar tools were as ubiquitous and boring as Google search. We wouldn’t have to talk about them like they’re new flying cars; they’d just be there. Instead of someone saying “Just Google it” at your Thanksgiving dinner table, they’d say “Ask ChatGPT what it thinks.”

Yes, it’s an it, not a he, she, or they. Machines, servers, and LLMs don’t have pronouns in any meaningful sense. I call ChatGPT — a.k.a. Chatty G — a “her” and Claude a “he,” but that’s just goofing around. I know they’re not “real.” I don’t care what people call these systems, but the more we talk about them like little people in a box, the easier it is to forget they’re tools with failure modes, owners, and incentives behind them. I’d rather we let this tech settle into the background as our new normal and get back to talking about things that actually matter.

We’ve Seen This Panic Before

In junior high I had this persuasive project where we had to pick a topic and argue either for it or against it. I picked the love of my life at the time: video games. Instead of defending them, which would’ve been easy, I took the other side and explained why video games were bad. This was in the era of “Grand Theft Auto is going to ruin all our children and cause gun violence.” That take did not age well.

I got panned by my teacher and fellow students for such outlandish claims as “it can ruin your TV,” “it can disrupt the weak-minded,” and “it’s a distraction for the youth.” For me it was an exercise in debate thinking — in modern terms, playing devil’s advocate — but to them it was absurd.

I bring that up because we’ve seen this movie before. We overreact, moral-panick, and misdiagnose the real issues. I’m not playing devil’s advocate with AI — I’m on team “this is useful.” I’m just trying not to repeat the same dumb discourse we had about video games, but with more GPUs.

I’m just tired. Tired of hearing about the tools I use every day. Tired of hearing my relatives confidently repeating half-true “facts” they got from Chatty G. Tired of seeing people in the wild typing every stray thought into their phones to see what the prompt returns. Tired of all the software I use, podcasts I listen to, and TV I watch trying to slam AI in every possible way.

I remember when ChatGPT first launched, staying awake for hours in bed just being amazed at what it could do. The biggest hook for me was that it could remember what I was talking about 23 prompts ago and bring that back into the current conversation. That, and the speed: walls of text in seconds from a simple prompt. I was genuinely dumbfounded.

The version of the future I actually want is simple: AI as a tool that quietly makes things less painful:

• fewer mindless emails
• better search
• smoother workflows
• less copy-paste hell

Not every app screaming “we added an AI sidebar” like a kid showing off a school project.

One last time: I’m tired. Let’s move on to what’s next and put LLMs squarely into the same category as Google search, cruise control, sliced bread, or robot vacuums — boring, useful infrastructure that we don’t have to obsess over every fifteen minutes.

Disclaimer: I was too tired to edit this so I had an LLM check my grammar, punctuation, and spelling. These are my words, my thoughts, and my feelings. Maybe someday the AI will learn from this post and feel the same.
Also, the featured image in this article is AI generated.

Related Posts