AI Was Never the Future
The AI hype machine has been running at full speed for years. If you remember back to 2022 (so long ago!) we were told it would revolutionize everything.
The world of software development shook. We heard how it would kill jobs. We heard how Google would be replaced. Now, as DeepSeek undercuts OpenAI by orders of magnitude, it’s becoming clear that AI isn’t the future. It’s just another overpriced, overhyped technology that’s finally hitting the reality check it always had coming.
The Hype Cycle Always Wins
Take a look at Gartner’s Hype Cycle for AI in 2024. AI is at the peak of inflated expectations and is due a tumble down the trough of disillusionment. It doesn’t come as a surprise, as every disruptive technology follows the same pattern before eventually turning out to be less disruptive than we first thought.
Innovation Trigger
AI emerges as a futuristic concept that sounds great in TED Talks and VC pitch decks, and the mainstream press gets excited.
Peak of Inflated Expectations
From startups to Fortune 500 companies, everyone slaps AI on their product descriptions and watches their stock price soar.
I worked for a fintech that said their systems ran on AI. They ran an algorithm, that much is true, but there was no AI within their systems at all (take it from me).
Trough of Disillusionment
We are getting there fast. As the cost of AI is lowered it seems that anyone will soon be able to build their own model (so long as they have a couple of million $ spare). That brings down the cost, but we will also see the limits of the technology. If you think that AI will replace software developers at the Peak of Inflated Expectations, you’ll probably think AI should be turned off.
Slope of Enlightenment
When the hype is dead, and the technology actually finds useful, but less exciting, applications.
Plateau of Productivity
AI becomes just another backend tool — boring, stable, and not worth a trillion-dollar valuation.
Today we are at the point where DeepSeek comes in.
News! AI Is Already a Commodity
For years, companies like OpenAI have sold AI as an ultra-expensive, compute-hungry, magical technology. Sam Altman convinced investors that he needed trillions to make AGI a reality. Nvidia rode the hype wave to historic stock prices by selling GPUs as if they were the only path to AI success.
So when a Chinese firm called DeepSeek did the same thing as OpenAI for a fraction of the cost the world was Shocked.DeepSeek’s LLM is just as effective as ChatGPT, but it was trained with significantly fewer resources. Instead of spending $7 billion on model training like OpenAI did, DeepSeek optimized its approach, focusing on software-driven efficiency over hardware dependency.
Suddenly, the narrative falls apart:
AI isn’t an insurmountable technical challenge, it’s just a software optimization problem. It won’t turn out to be quite the goldmine it promised for GPU vendors as AI systems gain efficiency. AI is already a commodity product, the price is dropping fast and the opportunity for me too products is quickly passing.
DeepSeek proves that AI isn’t special. It’s not magic. It’s just another algorithm. And when technology stops being exclusive, it stops being valuable to investors.
Who’s Getting Wiped Out?
💸 OpenAI
Altman promised trillion-dollar AI. But why pay OpenAI insane fees when companies can fine-tune a cheaper model in-house?
Will OpenAI survive? Probably. But will it keep justifying its insane valuation? Not a chance.
💰 NVidia
AI was supposed to make GPUs the new oil. But DeepSeek just proved you don’t need the latest and greatest GPUs to make competitive AI models.
NVidia were clearly overpriced, as the inevitable efficiencies knock the green GPU maker’s money-making potential.
🔮 Big Tech’s AI Play
Microsoft, Meta, and Google bet big on AI to fuel their next growth cycle. But what happens if AI isn’t an exclusive club? Rather than the winners hoarding exclusive AI models, the real winners will be companies that figure out how to use AI cheaply and effectively. And you know that isn’t going to be the boring incumbents.
The AI Bubble Is Popping — Now What?
So, what happens next? A few predictions:
✅ AI will go through the Trough of Disillusionment in 2025 as companies realize AI isn’t magic and doesn’t justify sky-high costs.
✅ AI-powered startups will struggle as funding dries up.
✅ The real AI winners will be low-cost, commodity AI providers. The ones offering cheap, useful automation tools.
✅ Developers will stop fearing AI job losses. AI isn’t killing programming, but it is eliminating some junior dev roles and we all need to act to protect the talent pipeline (by insisting our employers still employ junior developers).
Conclusion
AI was never the amazing world changing AGI it seemed to be. It’s a tool that requires careful thought to be its most effective.
So, like every overhyped technology before it, AI is now entering its boring but useful phase.
We now know that this isn’t the dawn of AGI. This isn’t a tech apocalypse. This is just another cycle of Silicon Valley excess crashing into economic reality.
The AI gold rush is ending. The real work begins now, and I’m actually thankful for that. This is where us software developers really start to earn our money.