Are you also wondering why ChatGPT is not getting much smarter? OpenAI is just releasing new features for their platform, not better models. You can take examples like the Ghibli trend, voice-to-voice, memory feature — all of these are cool, but what about AGI (the thing people promised)? What about smarter AIs?
Let me tell you, it's not just OpenAI. This is happening with every other AI company you can name. They all started great, but now their progress is slower than expected.
Let me give you a one-word reason: "DATA."
Yes, the lack of data is the thing behind the slow progress — maybe even what's stopping us from reaching AGI.
The short reason is: they've already trained their AI models on most of the publicly available data. (Yes, your private data might be included too — through partnerships or tricky terms and conditions.)
There's just not enough fresh data to train these models.
I'm attaching a graph where you can see the crazy growth that happened before — and how it's slowing down now. This kind of growth is called logarithmic growth.
If you're hoping this graph will suddenly go up again, you might be disappointed. This slow progress will keep happening until we find something that gives us a huge amount of new data to train on.