Are you sick of AI yet?

I'm exhausted by AI, and I reckon I'm not alone.

Here's the thing – I use AI every day. Hell, I was just using Warp's AI assistant to remind me how to find files modified in the last 24 hours. The command? find . -mtime -1. Dead simple. I've typed it a thousand times. But there I was, asking an AI for something I could have recalled in two seconds because it's become the default. It's like we've trained ourselves to Google things we already know.

It's like we've collectively decided that every piece of software needs an AI sprinkle on top, regardless of whether it makes a lick of sense. Remember when everything had to be "in the cloud"? Or when blockchain was going to revolutionise your morning cuppa? Well, now it's AI's turn to be shoehorned into every conceivable corner of our digital lives.

The Hype Machine Has Gone Into Overdrive

The numbers don't lie. A recent survey found that 76% of young Americans say AI is talked about far more now than five years ago. But here's the kicker – only 37% actually want more discussion about it. That's a proper gap between what we're being fed and what we're actually hungry for.

I've been in this industry long enough to see several hype cycles come and go. Remember when everyone was going to be a "data scientist"? Or when every startup needed to be "Uber for X"? But this AI wave feels different – more pervasive, more exhausting, and frankly, more annoying.

Even Stack Overflow's latest Developer Survey shows the cracks forming. Trust in AI accuracy has plummeted from 40% to 29% year-over-year. And why? Because 45% of us are spending our days debugging "almost-right" AI-generated code. It's like having a junior developer who's read every programming book ever written but has never actually shipped anything.

// What I asked for: "Simple function to validate email"
function validateEmail(email) {
  const re = /^[^\s@]+@[^\s@]+\.[^\s@]+$/;
  return re.test(email);
}

// What AI gave me: "Advanced email validation with ML-powered domain verification"
async function validateEmailWithAI(email) {
  const basicValidation = /^[^\s@]+@[^\s@]+\.[^\s@]+$/;
  if (!basicValidation.test(email)) return false;
  
  // AI-powered domain reputation check
  const domainScore = await mlModel.predictDomainTrustworthiness(email.split('@')[1]);
  
  // Sentiment analysis on email prefix
  const usernameVibes = await nlpEngine.analyzeEmailUsername(email.split('@')[0]);
  
  // Quantum-inspired probability calculation
  const validityProbability = quantumEmailValidator.calculate(email, domainScore, usernameVibes);
  
  return validityProbability > 0.87; // 87% confidence threshold based on GPT-5 recommendations
}

See what I mean? Sometimes a regex is just a regex.

The Social Media Echo Chamber

Scroll through tech Twitter (sorry, X) or LinkedIn these days, and it's AI this, AI that. Every other post is either breathless excitement about the latest model or doomsday predictions about our impending obsolescence. It's exhausting.

I saw a post on Threads that perfectly captured my mood: "Use AI to replace the boring, repetitive work, and let humans do what we do best – creating unique pieces of art." Spot on. Instead, we're using AI to write poetry whilst still manually updating spreadsheets. We've got it completely backwards.

AI-powered computer terminal showing excessive AI integration My LinkedIn feed on any given Tuesday

The Hacker News crowd – usually a reliable bunch for sensible tech discussion – has been particularly vocal. One post titled "I Am Tired of AI" garnered over 1,100 comments. That's not just a few grumpy developers; that's a proper uprising.

The Backlash Is Real (And Getting Louder)

Remember Duolingo's pivot to being an "AI-first" company? Users didn't just complain – they deleted the app en masse. When your language learning app decides that human teachers are passé, people vote with their uninstall buttons.

The gaming industry's been copping it too. Square Enix released a game with AI-generated cover art, and the community absolutely savaged them. "AI slop," they called it. Not exactly the reception you want for your latest release.

Even market research is showing the tide turning. Nearly half of both Gen Z and Baby Boomers oppose AI in advertising. When those two generations agree on something, you know you've properly mucked it up.

The Developer's Dilemma

Here's the thing – I'm not anti-AI. Far from it. I use AI tools daily – GitHub Copilot, ChatGPT, Claude, Warp's AI assistant, you name it. They're genuinely helpful for the right tasks. Hell, I've even built a few ML models in my time. The technology itself can be brilliant when applied thoughtfully.

What I'm tired of is the relentless pressure to inject AI into everything, regardless of whether it adds value. And the fact that I find myself reflexively reaching for AI even for trivial tasks I could do in my sleep. It's like being told every meal needs truffle oil, and then finding yourself putting it on your cornflakes out of habit.

// Actual code review I had last week:
// Original PR:
fn validate_email(email: &str) -> bool {
    email.contains('@') && email.contains('.')
}

// PM's requested "improvement":
use ai_sdk::{GPT4, PromptBuilder};
use quantum_ml::EmailProbabilityEngine;

async fn validate_email_with_ai(email: &str) -> Result<bool, AIError> {
    // Step 1: Basic regex check (5ms)
    let basic_valid = email.contains('@') && email.contains('.');
    
    // Step 2: Ask GPT-4 if this "looks like" an email (2000ms)
    let prompt = PromptBuilder::new()
        .system("You are an email validation expert")
        .user(&format!("Is '{}' a valid email? Respond with JSON.", email))
        .build();
    
    let ai_response = GPT4::new()?.complete(prompt).await?;
    let ai_thinks_valid = ai_response.parse::<bool>()?;
    
    // Step 3: "Quantum-inspired" probability check (why???)
    let quantum_score = EmailProbabilityEngine::calculate_email_entropy(email)?;
    
    // Step 4: Combine all three with "proprietary algorithm"
    Ok(basic_valid && ai_thinks_valid && quantum_score > 0.7)
}

// Me: "This turned a 5ms function into a 2-second API call that costs money"
// PM: "But think of the investor deck!"

Finding Balance in the Chaos

So where do we go from here? I reckon we need to take a collective breath and remember a few things:

  1. Not every problem needs an AI solution. Sometimes a simple algorithm, a basic heuristic, or (shock horror) human judgment is the better choice.

  2. Focus on genuine value. If you're adding AI to your product, ask yourself: Does this meaningfully improve the user experience? Or are we just chasing investor buzzwords?

  3. Be honest about limitations. That "almost-right" problem isn't going away anytime soon. Let's stop pretending AI is magic and start treating it like what it is – a powerful but imperfect tool.

  4. Preserve human creativity. Use AI for the grunt work, not the creative work. Let it handle the tedious bits so we can focus on the interesting problems.

Looking Forward (With Measured Optimism)

Despite my grousing, I'm not a complete pessimist. The Gartner Hype Cycle exists for a reason – we're currently sliding into the "trough of disillusionment," but that's actually healthy. It means we're moving past the breathless hype towards more realistic applications.

I'm looking forward to the day when AI is just another tool in our toolkit, not the answer to every question nobody asked. When we can have a sensible conversation about its strengths and weaknesses without being labeled either a Luddite or a techno-utopian.

Until then, I'll keep fighting the good fight. Pushing back against unnecessary AI integrations. Writing code that's actually needed. And yes, using AI daily when it genuinely helps – just trying to be more mindful about when I'm using it because it's useful versus when I'm using it because it's there.

Now if you'll excuse me, I need to go remember how to use grep without asking an AI first. Muscle memory, don't fail me now.


What's your take on AI fatigue? Are you finding yourself overwhelmed by the constant barrage of AI news and products? Drop me a line – I'd love to hear from fellow travelers in this brave new world of artificial everything.

P.S. - Did I use AI while writing this post? Only to double-check that find command syntax. Old habits die hard.