AI: From Bulls to Bears and Back
What a German proverb can teach us about AI's potential and perils.
These days, with the AI doom-and-gloom train running in overdrive, we are well served to take a step back and remind ourselves of the lovely (and admittedly strange) German proverb “Nichts wird so heiß gegessen, wie es gekocht wird”: “Nothing is eaten as hot as it is cooked.”
GenAI is [oftentimes/sometimes] awe-inspiring. From solving complex physics problems to planning intricate tasks that stymie human experts, the potential is undeniable. Systems like OpenAI’s o1 model showcase advancements that let AI “think through” issues, as seen when o1 beat human PhDs at their own game in areas like hard science (and re-created code a PhD student took 10 months to develop in a mere hour (watch here ↗)). This points to a future, where AI doesn’t just assist (or entertains us with party tricks such as creating a very buff version of Harry Potter (you are welcome… ↗))—it actually innovates.
Yet, despite the massive leaps in capability, it’s clear that this revolution is still just getting started. Corporations have been slow to adopt AI at scale. Microsoft’s AI-powered Copilot, for instance, sees adoption rates hovering between 0.1% and 1% of its 440 million users (and many of the people who got access to Copilot don’t use it (link ↗)). GenAI, though powerful, hasn’t delivered enough tangible value to justify mass deployment. And while AI can create complex results, from detailed programming to handling challenging puzzles, its integration into business processes remains minimal.
And then, there might also be reason to argue that “AI has a subprime moment” (link ↗). GenAI, while innovative, has glaring flaws. Hallucinations remain a persistent issue, making AI unreliable for critical business applications (AI in healthcare, anyone?). Worse, the infrastructure costs behind AI are staggering. OpenAI, for instance, will lose a whopping $5 billion this year, with no clear path to profitability (alas, at least they seem to be able to fundraise eye-watering amounts of money—but surely, at one point, investors want to see returns). In fact, major firms are struggling to justify the AI investments they’ve made.
Where does this all leave us? With the Germans… The technology is evolving, but the hype is (as usual) ahead of the current capabilities. It’s an exciting time, but one that demands patience. Personally, I like to regularly read both the doom and gloom case for AI—it keeps being a moving target and it’s too important to not blindly run into either direction. As so often in life—the truth is somewhere in the middle.
@Pascal