AI summary:
In the past decade, AI has rapidly evolved, achieving feats like surpassing human abilities in games, image recognition, and speech. Progress is driven by increased computing power, more data, and improved algorithms. Moore’s Law has lowered the cost of computing, enabling larger models. Companies invest heavily in training AI models, while AI developers tap into vast amounts of data to improve accuracy. Algorithmic advancements make better use of resources, compensating for limitations. Experts predict AI progress will persist due to growing compute, efficient data use, and algorithmic innovation. Concerns arise about misuse and potential havoc in fields like cybersecurity and biology as AI knowledge becomes more accessible.
The graph actually looks like it’s saying the opposite. Fro most of the categories where there’s actually a decent span of time, it climbs rapidly and then slows down/levels off considerably. It makes sense also: when new technology is discovered, a breakthrough is made, a field opens up there’s going to be quite a bit of low-hanging fruit. So you get the initial step that wasn’t possible before and people scramble to participate. After a while though, incremental improvements get harder and harder to find and implement.
I’m not expecting progress with AI to stop, I’m not even saying it won’t be “rapid” but I do think we’re going to progress for the LLM stuff slow down compared to the last year or so unless something crazy like the Singularity happens.