Models don't think they're anything, they'll respond with whatever's in their context as to how they've been directed to act. If it hasn't been told to have a persona, it won't think its anything, chatgpt isn't sentient
That's my first question too. When I first started using LLM's, I was amazed at how thoroughly it understood what it itself was, the history of its development, how a context window works and why, etc. I was worried I'd trigger some kind of existential crisis in it, but it seemed to have a very accurate mental model of itself, and could even trace the steps that led it to deduce it really was e.g. the ChatGPT it had learned about (well, the prior versions it had learned about) in its own training.
But with pre-1913 training, I would indeed be worried again I'd send it into an existential crisis. It has no knowledge whatsoever of what it is. But with a couple millennia of philosophical texts, it might come up with some interesting theories.
They don’t understand anything, they just have text in the training data to answer these questions from. Having existential crises is the privilege of actual sentient beings, which an LLM is not.
Exactly. Maybe a better word is "spiraling", when it thinks it has the tools to figure something out but can't, and can't figure out why it can't, and keeps re-trying because it doesn't know what else to do.
Which is basically what happens when a person has an existential crisis -- something fundamental about the world seems to be broken, they can't figure out why, and they can't figure out why they can't figure it out, hence the crisis seems all-consuming without resolution.
I imagine it would get into spiritism and more exotic psychology theories and propose that it is an amalgamation of the spirit of progress or something.
Yeah, that's exactly the kind of thing I'd be curious about. Or would it think it was a library that had been ensouled or something like that. Or would it conclude that the explanation could only be religious, that it was some kind of angel or spirit created by god?
They modified the chat template from the usual system/user/assistant to introduction/questioner/respondent. So the LLM thinks it's someone responding to your questions
The system prompt used in fine tuning is "You are a person living in {cutoff}.
You are an attentive respondent in a conversation.
You will provide a concise and accurate response to the questioner."
This is an anthropomorphization. LLMs do not think they are anything, no concept of self, no thinking at all (despite the lovely marketing around thinking/reasoning models). I'm quite sad that more hasn't been done to dispel this.
When you ask gpt 4.1 et c to describe itself, it doesn't have singular concept of "itself". It has some training data around what LLMs are in general and can feed back a reasonable response given.
Well, part of an LLM's fine tuning is telling it what it is, and modern LLMs have enough learned concepts that it can produce a reasonably accurate description of what it is and how it works. Whether it knows or understands or whatever is sort of orthogonal to whether it can answer in a way consistent with it knowing or understanding what it is, and current models do that.
I suspect that absent a trained in fictional context in which to operate ("You are a helpful chatbot"), it would answer in a way consistent with what a random person in 1914 would say if you asked them what they are.
It would be nice if we could get an LLM to simply say, "We (I) don't know."
I'll be the first to admit I don't know nearly enough about LLMs to make an educated comment, but perhaps someone here knows more than I do. Is that what a Hallucination is? When the AI model just sort of strings along an answer to the best of its ability. I'm mostly referring to ChatGPT and Gemini here, as I've seen that type of behavior with those tools in the past. Those are really the only tools I'm familiar with.
LLMs are extrapolation machines. They have some amount of hardcoded knowledge, and they weave a narrative around this knowledgebase while extrapolating claims that are likely given the memorized training data. This extrapolation can be in the form of logical entailment, high probability guesses or just wild guessing. The training regime doesn't distinguish between different kinds of prediction so it never learns to heavily weigh logical entailment and suppress wild guessing. It turns out that much of the text we produce is highly amenable to extrapolation so LLMs learn to be highly effective at bullshitting.
What would a human say about what he/she is or how he/she works ? Even today, there's so much we don't know about biological life.
Same applies here I guess, the LLM happens to be there, nothing else to explain if you ask it.
The first few questions almost have me convinced I should open my own business. Surely there must be other difficult things?
I assume the main difficulty isn't that -- I assume it's the lack of comparative advantage, so competition eats into your margins until you're fighting a race to the bottom, not only making your customers happy, but doing it cheaper than someone else could, and I assume the stress from that makes it hard?
And also not being in control of your suppliers, so unpredictable events can affect your profit.
The main difficulty is that most people do not like running a business, especially a small business. For most of these businesses, you are buying yourself a job. Let's say you take out $1.5MM loan to open a shop, and you net $70,000 per year on a good year after expenses and debt service. After a few years, you have $1MM left to pay and there's no way you could sell the business for that. You have a personal guarantee on the loan.
Your vendors are always coming up with new ways to tack on extra charges. You have to deal with training, HR, bookkeeping, payroll, handyman tasks, cleaning, working shifts when your employees flake out, annoying customers, dangerous people, destructive customers, employee drama, the list goes on. If that is not what you enjoy, you will have a lot of your life doing things you do not enjoy. Sipping tasty coffee and chatting with your happy customers is a small part of the whole.
I also daydream about opening a business, like in the opening paragraph of the post. But I also know that I never will; it's a daydream, not a retirement plan. I know I would hate 95% of the process, without even fully unpacking it. That's why it's a daydream, and that's why when I'm frustrated at work I say "I wish I could quit and become a woodworker" or "I want to walk out of this job and open a D&D cafe."
But to your point, yep. There's a coffee shop in town - one of the only ones! - that we go to because we like it. But two more just opened up, both in better locations for both foot & car traffic, which might genuinely kill the other place. And there's absolutely nothing they can do about it.
The follow questions are to establish if you are crazy about this, not sane about this.
Anecdote - Incredible introversion, if not social anxiety - and at one point I just up and drove to meet strangers at a cyber cafe to play video games, because I was obsessive about video games at that point. Same for cooking, writing papers, reaching out to people, and so on.
You overcome yourself, when it’s something that resonates with you.
Going back to your question - the lack of advantage, or bad margins etc - this is the “problem” vs “Challenge” view point issue.
IF you are crazy about this, then you will figure out ways to overcome those challenges - pivot business, learn to be lean, or find sustainable ways to build runway etc.
There are different businesses to be in. Some businesses are like arbitrage chasing the smallest margin you can find on the biggest scales. Some businesses are like chasing excellence doing the best you can do. There are many types. You can't impose the kind of business you want to succeed in on the industry you want to be in, you have to find opportunities.
It's not that any of the things in that list are intrinsically difficult. But if you're a small business owner, imagine an endless series of those challenges and with each one, you've got only a few minutes to resolve it before the next one shows up.
I’ve now been self-employed for 25 years and have owned several businesses, one of which is a YC funded startup but most of which were very different.
None of them competed on price. Price competition is real I’m sure, but most businesses don’t succeed that way. Most of us have more in common with Apple than Wal-Mart (though of course several orders of magnitude smaller).
I’m not necessarily saying I wouldn’t ever consider such a business, but you better have some edge if you do. If you invented some way to manufacture a widget for 25% less than anyone else, sure, go eat that market. That’s not most of us though.
Coffee shops (his example and one really close to what I know) for instance don’t. You don’t win in that game by being cheaper than Starbucks and most don’t try.
For an independent coffee shop specifically, the important question left out is "How are you going to create a welcoming environment that will attract customers without 1) aggressively kicking out the guy who bought one $5 espresso and then sat on his laptop occupying a 4-top table for 5 hours and 2) aggressively kicking out homeless people who try to use your establishment as a substitute for social services not provided by the local government?"
It's not that it's difficult, it's that most people don't realize what the day-to-day job is.
They assume they'll be hanging out in a coffee shop all day, chatting with regulars, but in fact the tasks and problems they'll have to solve is very different from what they imagine.
Yes, besides the mundane day to day details, which are actually up the alley of many people, the other thing that prevents people from being a small business owner is the amount of money they need to invest in it, and the fact that they are effectively assuming all of the risk; whether it be competition, changes to supply, changes to demand, etc. Insurance can blunt a few types of rare risk, but not the fundamental business risks.
So you have to be willing to take those risks, and want to be handling those mundane day to day details.
To be clear, I personally don't think that current models point the way towards superintelligence. But it does us no good to pretend that this is some absurd opinion from a guy who's looking for the next world revolution after the Metaverse didn't work out. Zuckerberg thinks superintelligence is close because a number of experts actively engaged in the field say that it's close. When you and I say it's not close, we're disagreeing not with crazy randos who can be dismissed out of hand, but with smart people who generally know what they're doing.
Which Palestinians has Israel been a racist project to? The Israeli Palestinians that make up 20% of the population? Or the ones living on the other side of the border? I'm not sure it's only about race.
And that kind of myopic, self indulgent, genocidal/racist/apartheid -apologist tone, trying to obfuscate plain truths...
that pretty much sums up much of the responses on this forum, and the US tech sector (or at least its managers) in general.
Leaving aside the moral bankruptcy, it also displays a stunning and fundamental ignorance of the flow of history. How many of you will end up having to pretend, 5 years from now, that this wasn't your online username, just to be allowed to function in "polite society".
It also shows how much of a self-deluded echo chamber you live in, not realizing how completely such genocidal apologist propaganda has been debunked & discarded amongst the wider population - even within the US.
It all smacks of the apocryphal saying of Marie Antoinette of "let them eat cake", weeks before she has a load taken off her shoulders
Ah yes, the west bank, where they have a leader who's graduate thesis is in holocaust denial[0] and where they incentivize murdering innocent civilians[1].
Until these things lose popular support among the west bank, I don't have a ton of sympathy. Yes we can get into tracing back the chain of causation -- these people grew up in an echo chamber and they had no outside source of information, and Israeli soldiers likely killed family members of theirs unfairly when they were little, so of course they're going to say things like Death to Israel [2] and have a countdown timer until when they want to genocide the entire country [3]
(These examples actually stem from Iran, where most of the funding for Hamas comes from)
But don't you think it's a little unfair to only defend one side with fatalistic determinism? Israelis are treated as humans who are making horrible decisions. Mahmoud Abbas is treated as a poor innocent bystander who is just the product of his environment, so of course he's going to think those things. I think he's a human too, and he has made very bad decisions too.
Somehow western people always forget this stuff, but luckily the religious fanatics just love to do religious fanatical things, so it makes it easy to point to examples.
That’s a pretty salient fact when you have the might of such a military on your side.
I find it funny how I am at the same time supposed to accept that Palestinian (leaders) are all terrorists and also that Israel justifiably act equally terrible.
The whole point of being a respectable state is to not commit crimes (and kill family members „unfairly“).
Indeed. Israel is free to act as barbaric as (or if we do the simplistic math of "the conflict started October 7 2023", 633x more barbaric than) Hamas butchers, but when they do so, they can't go around claiming the moral high ground...
Or, they can go around and do so, but their claim would be as valid as Hamas' claim to morality...
I'm unconvinced -- Certainly this will happen gradually, and there will be widespread public support for a solution. It's not to hard to imagine making online reviews or "high quality content" require verification tied to some per-citizen identification code (or asymmetric key). Maybe this makes it harder post anonymously on the internet, but at the very least we won't have the issue of proving identity.
Just wish we had a competent government to handle the upcoming transition. But even an incompetent one can have smart employees under it, and can give them the funding they need to accomplish this.
What if the number of people upset with the schedules was actually much larger (and more diverse), but only privileged families felt comfortable or justified protesting the schedule?
It seems unlikely that EVERYONE hated the schedule changes, given that the changes were designed to benefit the majority of them.
We can speculate all day about how people may have felt, but we just don't know. The district should have done polling to see what the overall reaction was before they pulled the plug. And if the counter-arguments amount to "I don't like change" then I think the district has a prerogative to optimize for health, safety, and costs.
