Is AI a digital money pit? Has the sunk-cost fallacy trapped the champions of AI in a death spiral? Should there be a course correction toward more attainable goals?
Week after week, news surfaces of funding deals and partnerships among the major players in the AI space. This activity has fueled talk of circular financing driving an AI bubble. There is also speculation that OpenAI may run out of cash in 2027, barring some sort of sea change.
Perhaps the wrong questions have been asked. The barometer to define industry success has become so warped, it may be impossible for AI to measure up.
Expectations changed fundamentally in the froth of the dot-com bubble. Demand for tech success skyrocketed. Suddenly, every innovation needed to deliver hockey stick growth that reached Mars and just kept going. Surely, the phenomenon that saw FAANG companies take over could be replicated, right? And if those circumstances couldn’t be reproduced, perhaps the next tech revolution could at least push these same companies to even greater heights with perpetual growth.
We expected it with augmented reality, virtual reality and the metaverse, despite examples and warnings that suggested otherwise.
But now AI must deliver on everyone’s wishes with the power of a digital djinni, or there will be a lot of explaining to do — to the tune of billions of dollars. The clock is ticking loudly.
Microsoft CEO Satya Nadella said recently at the World Economic Forum that AI could lose “social permission” to consume energy and have other environmental impacts if real benefits did not manifest soon. That sentiment followed Nadella’s previous request that the public stop calling GenAI content “AI slop .”
When the AI hype cycle kicked off, it was quickly followed by concerns about potential risks, including opposing views on the possibility of AI-driven extinction. Some of those alarms quieted down into simple irritation with slop GenAI content, while valid worries about misinformation spiked amid already strained social tensions.
Discussions on the future of work in the AI era have become something of a running gag. Champions of AI promise the technology will enhance jobs, as long as you embrace it like a long-lost sibling — and then sometimes in the same breath, warn that entire careers are doomed regardless.
InformationWeek’s own coverage has looked at the growing demand for ROI from AI. One year ago, InformationWeek called out the many costs AI could impose on the environment, careers and industry. The pressure on the AI industry to deliver may now be even worse than during prior tech revolutions.
During the dot-com bubble, darlings such as Pets.com rose quickly and crashed hard — with their assets and cute spokespuppet sold off for parts. Instead of tempering expectations from those days, the tech sector seems eager to chase the next big, frothy innovation in hopes that it will devour the market.
So where are the 3D televisions? What about mobile phones with built-in video projectors? What happened to native mobile apps making the internet go extinct? Did every house become a fully automated smart home? Have NFTs replaced gold?
QR codes may finally replace UPC barcodes in 2027 — but that’s also a topic I covered back in my newspaper days more than 15 years ago.
I could list the ambitious ideas and companies I’ve watched stall or burn out, but instead of enshrining their names in a graveyard, it would be more like filling a litter box.
What AI, in its current form and approach, is being asked to deliver may simply be impossible. And that raises the question of just how much more the industry can pour into AI before reality pushes back. I’ve seen online discussions questioning whether tripling down on GPUs to power LLMs can actually deliver artificial general intelligence. Bolting extra gas turbines onto data centers might not get the industry across the finish line.
After all the investments and expended resources, what happens if AI does not deliver the promised exponential leap forward? Who will use future iterations of AI if the world is exhausted from trying to sustain it?
This would all be easier if cold fusion and other next-generation renewable energy sources were readily available. Pity AI cannot solve that riddle — and that is the paradox.
Still, there may be hope to build an AI powerful enough to tell us how to turn a profit with AI.
Or is there?
Get more commentary and insights three times a week with the InformationWeek newsletter.
