If the jump between ChatGPT '5' and 4 is as big as between 3.5 and 4 then we're not in an echo chamber, we're in trouble. If the gap is smaller then they are running out of steam and we'll be fine.
I don’t think it will be if history is any precedent. Look at self driving in the mid late 2010s, the rate of advancement lead some otherwise smart people to make big bets that autonomous cars were right around the corner and we were going to be driven around by robots any by 2022.
We’re at the early phase of AI, it demos well but seems to break down in ways that seem obvious in the real world. We’re at the limit of how big the models we can train now. GPT4 has been described as 8 GPTs in a trench coat by George Hotz which was later confirmed by the founder of PyTorch. I’m not saying we have nothing to worry about, just that the hype seems to overtake reality early in the adoption cycle.
GPT4 is like an unreliable but brilliant employee.
