The AI Rollercoaster: Why We Keep Going from 'It's Over' to 'We're So Back'
Been scrolling through AI discussions lately and stumbled across this fascinating chart showing the emotional rollercoaster we’ve all been on with AI development over the past few years. The graph perfectly captures what someone described as the “it’s so over” to “we’re so back” vibes that seem to define our relationship with artificial intelligence progress.
Looking at those peaks and valleys, it really does feel like we’re all passengers on some sort of collective emotional pendulum. One minute everyone’s convinced we’ve hit the dreaded “AI wall” and progress has stagnated, the next minute there’s a breakthrough that has us all believing the singularity is just around the corner.
What struck me most about the discussion was how different people are interpreting the same data. Some see a sigmoid curve levelling off, others spot a sine wave pattern, and a few optimists still believe we’re on an exponential trajectory with temporary dips. The truth is probably somewhere in between, but it highlights how our perception of AI progress is as much about psychology as it is about actual technical advancement.
The investor perspective really resonated with my IT background. One user made a brilliant point about how markets can outpace actual development - essentially pricing in expectations that reality simply can’t match. It reminded me of the dot-com bubble, where companies with barely functional websites were valued like they’d already revolutionised commerce. The Dyson sphere analogy was particularly apt: even if we captured all the sun’s energy, investors would probably price in five suns and then act shocked when we fell short.
Working in DevOps, I’ve seen this pattern play out countless times with new technologies. There’s always that initial honeymoon period where everyone thinks the new tool will solve all their problems, followed by the inevitable disappointment when reality sets in, and then finally a more mature understanding of what the technology can actually deliver. Kubernetes, Docker, even cloud computing went through similar cycles.
The GPT-4o discussion was interesting too. Whether it was actually better than GPT-4 seems to depend on how you were using it and what you were measuring. This speaks to something I’ve noticed in my day job - the gap between benchmarks and real-world performance can be enormous. A model might score brilliantly on standardised tests but struggle with the specific quirks of your particular use case.
What worries me, though, is how this emotional volatility might be masking more serious concerns about AI development. While we’re all caught up in whether the latest model is incrementally better than the last one, we’re not spending enough time thinking about the environmental impact of training these massive models, or the social implications of potentially displacing entire categories of work. The energy consumption alone is staggering - and that’s something that should concern anyone who cares about climate change.
But here’s the thing that gives me hope: the discussion itself shows we’re becoming more sophisticated in our thinking about AI progress. People are starting to look beyond just raw capability improvements and considering different paradigms - neuromorphic chips, spiking neural networks, entirely new approaches to reasoning. That suggests we’re moving past the hype cycle into more nuanced territory.
Maybe the real progress isn’t in the models themselves, but in how we’re learning to think about artificial intelligence as a technology that’s going to develop in fits and starts, with plateaus and breakthroughs, rather than the smooth exponential curve that Silicon Valley marketing departments would have us believe.
The person who set a reminder for one year from now has the right idea. Sometimes the only way to really judge progress is to step back and look at the bigger picture. Whatever happens next in AI development, at least we’re getting better at managing our expectations - and that might be the most important advancement of all.