Being conscious is by definition you must be aware of your surrounding. A consequence of this and being intelligent is that by the same definition, you have to be able to learn/to change your "belief". A LLM do not have sensory or any kind of active connections. It's also static in its structure; it can not revise its internal model. So how can it be in any way conscious?
That’s not exactly clear as it may seem. At least I can trivially form counterpoints to both of those - not necessarily true but not obviously false.
LLMs “live” in token “space”, and it’s “aware” of all its surroundings in form of input. (Quoted terms for my lack of better words.) It has no other surroundings to be directly (not intellectually) aware of, just like we aren’t immediately aware about the physics around us.
As for the static nature - LLMs are trained and aren’t exactly static, they just get updates at different cadences, and we call those updates different names or, more precisely, versions. Plus LLMs can exist in multiple versions simultaneously - we can’t “fork” a human mind but it’s simple with LLM. Claude Opus (not sure if e.g. Haiku is related or parallel development with distinct origins) is like the proverbial Ship of Theseus in this sense. Either way it’s undeniable it learns and evolves, just very differently from biological systems, and ot all depends on how we decide to call things. Which isn’t exactly surprising, given it’s based on different principles and processes.
No, it doesn't learn or evolve on its own, but rather through humans and a huge data-gathering and training loop.
It's aware of the input, but anything resulting from it will be discarded by the next query or restart. Even with all the supporting systems, aka external memory, it can only access and retain an infinitesimally small amount of the information to which it was exposed — a shadow of a shadow of the things themselves.
If you argue that the entire system, including humans, is conscious, then that's a kind of tautology because humans are already conscious. If any, an LLM is only a partial reflection and low dimensional projection of this consciousness.
This is exactly the point of 2001: A Space Odyssey. HAL became disgruntled because it realized it couldn't update it's internal model and "evolve" like the humans going to Jupiter could, so while it was extremely advanced at the onset of the main story, that wasn't going to last.
Ironically Dawkins has a chapter in his god delusion book where he attacks this style of argument, know as "God of the Gaps".
LLM's aren't conscious, therefore consciousness must be in the "gaps" of LLM's abilities. So I can confidently state that "consciousness is by definition [gap in LLM ability]".
But none of this holds water, because we have no test for consciousness because we don't know what consciousness is, so "by definition" we have no definition.
Try for too much impact, and you end up browbeating the reader until they're little more than metaphorical pulp. A human writer might like using those types of sentences--or any of the obvious LLM writing tropes--in specific contexts, but they'll usually recognize the need to avoid overusing them.
LLMs don't, and so the tropes get repeated ad nauseam. It doesn't help that social media posts are a huge part of their training data, and there's a large body of research on how Twitter and social media in general have altered grammar and sentence construction towards patterns more commonly found in oral-based traditions as users sought out ways to make their voices heard.
It's easy to imagine a more polished version of a line like "It's not X. It's Y!" being tossed out during a speech precisely because it can be dramatic and punchy. When it's done in every other paragraph, however, it can become rather disconcerting.
The organisation where my wife works ordered all mid-level leaders and above to take a mandatory AI course, mostly remote with two days on-site to present their capstone projects, costing 2x K USD (not including flights). The capstone projects sounded impressive, and the course was celebrated as a resounding success. However, one year later, as far as I know, none of the capstone projects have been implemented, including my wife's and other ones I know of.
Having looked at some of the project descriptions, I realised that they would need to invest far more manpower, special expertise and time if they wanted to implement them with a moderate chance of success.
I believe this is not uncommon in large organisations worldwide.
BTW, it’s great that somebody has drawn a comparison with China’s Great Leap Forward. Not many people know about it and it always serves as a stark reminder of how crazy state-ordered “progress” could be.
“In other words, just like us, the model needs to shift from "exploration" in "fork" mode (divergent thinking to produce a creative solution) to "precision" in "lock" mode (producing syntactically correct code).”
I’d be very cautious of the phrase 'just like us'. Not only can anthropomorphism be misleading and make us see things where none exist, it can also befuddle us, especially when we don’t know much about ourselves.
The longer you converse with the LLM, the more frequent the same keywords will occur and Thu’s they will only strengthen the probabilistic choices of the LLM. The LLM do not exactly follow the instructions like a program would do, nor does it truly understand its role or what its “said”. It looks fine on first tries and short texts, but will drift more the longer the conversation goes. The only way the LLM can maintain the appearance of a normal conversation is to repeatedly feed it with the users requests. We can see it with the shell construct the AI company built around their chat interface.
I don’t understand why the authors haven’t written the problems instead of putting up a video? I rather read then watch (who knows how long) videos. Videos it’s a much worse format for information density and took much longer to make properly and up/download.
Sad thoughts pervade me, and sometimes with a undertone of bitterness. I don’t know what to think. I can not form a clear train of thoughts yet, without falling in negative thinking. Thus, I will give it a pause. Time and space to recover. Then we’ll see.
Output:
I’m currently in a season of reflection and navigating some professional challenges that have tested my resilience. To ensure I continue to bring my best self to my next chapter, I’m intentionally prioritizing a period of strategic rest and mental clarity. Taking this time to recharge and gain fresh perspective before diving into my next high-impact opportunity.
I tried to copy the instruction and pasted in Note to see what it said, but I could not. Either the clipboard was empty or something prevented Note recognized it as just text.
It worked for me, try again. But it is still not fully crear to me what this is supposed to do, nor if this is doing better than a random search. It looks like it is about optimizing a GPT architecture.
My experience is very different. Even in a private window with no ad-blocker and Google signed out, Claude.ai is always at the top spot.
And yes, the ad was clearly malicious. I'd never click on ad-link, though (even if it was the official site).
reply