AI is overhyped. You should use it anyway.
Pandora's box has been opened.
Over the past few years, OpenAI has raised eye-popping amounts of money. The latest funding round was a $6.6 billion investment in October 2024, the largest VC investment ever, putting the company's valuation at $157 billion. The company is reportedly in talks to raise another $40 billion at a valuation of $300 billion. That's on top of the announcement of the Stargate project, which will invest $500 billion over the next four years to build AI infrastructure in the U.S.
Meanwhile, OpenAI isn't profitable. The company saw $5 billion in losses in 2024. When DeepSeek announced a few weeks ago that it had built a comparable model at a reported cost of $6 million, everyone wondered: is AI worth all the money that's been sunk into research and development?
AI optimists claim that our entire world will be changed by AI. It will fundamentally shift how we work (for money) and how we get stuff done (at home). The average user is rightly skeptical. For a $20/month subscription to one of these models, generative AI can do a few things. But not life-changing things.
Is that coming? Or do companies overhype AI's capabilities simply to justify the astronomical investments?
AI's capabilities are exaggerated
If you're unfamiliar with venture capital, here's how investors make money: the biggest return happens with an exit event. The company that took the investment is either acquired or goes public. Sometimes, early-stage investors can sell their shares during later funding rounds.
Because of this, VC-backed companies feel pressure to perform. Investors will often give the startup more money, even if the company is underperforming. This is a sunk cost fallacy: if investors stop funding the company, they lose their investment (tens, hundreds, or — in OpenAI's case — billions of dollars). If they invest more, there's a chance that the company could recover and the VCs get a return.
Founders at AI companies everywhere are desperate to assure investors that their investments will pay off, hence the AI hype. Every company is an AI company these days, with even well-established, legacy products rushing to add features so they can slap "AI" on their websites. The alternative is that the company will be perceived as outdated, or, worse, a dying company. Almost every product I use these days has some type of AI capabilities. Very few are actually useful.
When cryptocurrency became mainstream with the launch of Bitcoin in 2009, investments flooded the market. Today, the total cryptocurrency market cap is around $3.2 trillion, despite limited practical uses for most people. We were told — repeatedly — that crypto would revolutionize financial systems. More than 25 years later, that hasn't exactly panned out.
On top of that, crypto has gone through several bubbles with the worst being the crypto winter and the collapse of several crypto exchanges, mired in scandals. As it turns out, this mostly unregulated industry was a hotbed for scammers like Sam Bankman-Fried (now serving 25 years in Federal prison).
The AI hype feels similar for many reasons. We're expected to believe that AI will revolutionize our daily lives. Yet AI, as it exists today, is notoriously unreliable. It lies and makes stuff up. I asked ChatGPT to provide me with links from my own blog around a specific topic, using the web search feature. I got a few links, and some references to articles that don't exist.
OpenAI is pushing the use of agents. With some prompts, an agent could research and book a hotel for you. But is that a real-world use case? Or will a human always need to double-check the agent's work, so there's no actual time savings?
But AI is a useful work sidekick
Much as I think AI is overhyped, I do think it's a useful tool. It has some very practical daily use cases, such as speeding up research or acting as a sparring partner while I'm writing. Do these micro use cases justify the cost invested in AI? Maybe not. The investments are still hinging on a return that hasn't panned out yet and it's unclear if it ever will. But that's really not my problem. That's the investors' problem.
I'm a proponent of anything that allows you to get your work done in less time. Back when I worked for a tech company, I had to do a lot of manual data entry, moving information around between tools. Then I taught myself to use automation (Zapier). Suddenly, that data entry was gone. Did I volunteer to take on more work to fill this now-freed-up time? Of course not. Because my work was based on results, not hours. If I could get more work done in fewer hours, yay for me.
The same is true of AI. Whether it saves you time or simply lets you work in different ways, it's worth a shot. Yet most people haven't explored AI use cases. According to National University, 88% of non-users are unclear about how generative AI will impact their lives.
(If you're concerned about the environmental impact, this article has a good perspective of the energy/water usage of AI compared to other activities.)
Admittedly, I didn't see the use cases immediately. I couldn't trust the output and didn't get anything that actually helped me during my day. Then I tried using AI to reassign a bunch of data in an app I use called Airtable. AI could do in minutes what would have otherwise taken me hours. Were the results perfect? Nope. But with a task that didn't require precision, it was "good enough" with the tradeoff of giving me time back.
I'd probably find some dark humor if AI only amounts to helping people with small tasks, rather than having a transformative impact on individuals — simply because the current hype often feels like smoke in mirrors. I listened to a webinar recently with a host that talked about analyzing large sets of data with AI. And while AI can do that, the risk of a bad business decision based on faulty data was high (in this scenario).
As it exists today, AI is a shortcut, not a replacement. But, by all means, go for it. Pay OpenAI $20 per month and use the product in a way that costs OpenAI far more. Figure out how to save yourself a little bit of time and mental energy. Pandora's box has been opened, and AI isn't going away (much like crypto isn't going away). Might as well figure out how to use it to your benefit.
Most issues of this publication are free because I love sharing ideas and connecting with others about the future of work. If you want to support me as a writer, you can buy me a coffee.
If you love this newsletter and look forward to reading it every week, please consider forwarding it to a friend or becoming a subscriber.
Have a work story you’d like to share? Please reach out using this form. I can retell your story while protecting your identity, share a guest post, or conduct an interview.