4 ms·
I think that well before we get to those kinds of runaway scenarios, we will have AI that is multiple orders of magnitude faster than humans at thinking/output.
by ilaksh 3y ago
I think that well before we get to those kinds of runaway scenarios, we will have AI that is multiple orders of magnitude faster than humans at thinking/output. At that point it doesn't need to become "infinitely powerful", it's already more than powerful enough.
It's a matter of making more efficient hardware and more efficient paradigms for the specific LLM application. The history of computing is engineers doing this over and over again. AMD just released an AI chip that can handle an application which normally requires multiple accelerator boards from Nvidia. And this sort of thing is just scratching the surface with relative minor architectural differences. Radical new approaches with memristors or other compute-in-memory departures are coming in less than a decade.
There is no possible way you can prevent people from connecting AI to the internet or tools.
So we are very close to the point where the people cannot compete with a cluster of agents working on a task. People lose control because in order to compete, they have to give the AI broader and broader goals. Since the AI will be working dozens or hundreds of times faster than humans, pausing for human input means the competition will race ahead.
The real issue impending issue is hyperspeed AI that is above average human intelligence. In some measurements this is already here. Within a few years the models or collections of specialized agents will be so good and so fast that no one will be able to deny it. And anywhere a human is in the loop routinely will fall behind due to the limitations of the human.
This does not require runaway AI or AI that is alive. Its just the consequence of deploying faster and faster hardware and gradually improving the model training. It will be quite obvious within 2-3 years that we need to consider _not_ making the AI hardware any more efficient than what we have at that time.
- TeMPOraL 3y agoThe difference between what you describe as not runaway AI, and runaway AI, is in whether or not the AI or AIs are able to optimize their own performance. That is, if they're doing any single thing of the following non-exhaustive list: - Helping design next-gen hardware on which they'll run, or write its firmware; - Procuring more compute, or making money for or otherwise improving ability of a company/group to procure more compute resources; - Improving their own algorithms; - Improving anything upstream in the dependency chain of their own algorithms, for example fundamental mathematical libraries; - Directing or convincing people to increase the number of AIs in the AI cluster; - Making profit in general, or otherwise incentivizing humans to work on next-gen AIs; ... then they're already capable of runaway self-improvement. The term "self-improving" carries connotations of an AI working on it own sources at runtime, as if it was self-evolving Lisp image. However, being a compute artifact and therefore trivial to copy and modify, the AI needs to be treated as a category. With that in mind, GPT-4 helping people to improve code that leads to LLaMA ${whatever-B} that then helps create next gen of Stable Diffusion, which prompts more resources to be put into merging the next gen SD with upcoming GPT-6 into a multimodal model, which then helps make 3D integrated circuits viable, which allows to ... - this is also the case of AI self-improving.
- rainonmoon 3y agoAn LLM doesn't "think", it just puts one word in front of another based on the likelihood of it having been placed in that sequence before. On top of that function is an application layer that contributes additional rules. At no point are any of these pieces "thinking" in any sense. As for speed of output as a metric, I can write you a Python script in 60 seconds that can generate thousands of HTTP requests, print thousands of words, and even find and display for you a decent recipe for chilli con carne. Despite this incredible capability for output, it's about as frightening as a satsuma. Just because something can do something a lot isn't a reason to abandon sense. While these misguided cinematic fantasies captivate everyone's attention, automated decision making is already deciding whether refugees are settled or sent back to be killed in their originating countries, whether people are given housing or not, or "deserve" welfare. The emerging research illustrates quite clearly the problematic biases entrenched in these models. Why this doesn't dominate AI discourse instead of hallucinatory speculating about some theoretical future is beyond me.
- TeMPOraL 3y ago> An LLM doesn't "think", it just puts one word in front of another based on the likelihood of it having been placed in that sequence before. You are confusing LLMs with Markov chains.
- rainonmoon 3y agoLuckily for me, I'm not.
- ilaksh 3y agoIt will make more sense to call it thinking if you realize that thinking does not need to connote all of the other human or animal cognitive characteristics that people tend to bundle with it. ChatGPT can think and reason but it doesn't have a high bandwidth sensory experience or subjective stream of consciousness. It doesn't think in the same way as humans. Etc. Etc. But thinking is still the best way to describe what it's doing. It has a model of the world and what you are asking it to do. It absolutely has to synthesize information to be able to fulfill requests.