That's correct. LLMs are plausible sentence generators, they don't "understand"* their runtime environment (or any of their other input) and they're not qualified to answer your questions. The companies providing these LLMs to users will typically provide a qualification along these lines, because LLMs tend to make up ("hallucinate", in the industry vernacular) outputs that are plausibly similar to the input text, even if they are wildly and obviously wrong and complete nonsense to boot.
Obviously, people find some value in some output of some LLMs. I've enjoyed the coding autocomplete stuff we have at work, it's helpful and fun. But "it's not qualified to answer my questions" is still true, even if it occasionally does something interesting or useful anyway.
*- this is a complicated term with a lot of baggage, but fortunately for the length of this comment, I don't think that any sense of it applies here. An LLM doesn't understand its training set any more than the mnemonic "ETA ONIS"** understands the English language.
**- a vaguely name-shaped presentation of the most common letters in the English language, in descending order. Useful if you need to remember those for some reason like guessing a substitution cypher.
If you can watch the video demo of this release, or for that matter the Attenborough video, and still claim that these things lack any form of "understanding," then your imagination is either a lot weaker than mine, or a lot stronger.
Behavior indistinguishable from understanding is understanding. Sorry, but that's how it's going to turn out to work.
Why are people so eager to believe that people can? When it comes to the definitions of concepts like sentience, consciousness, thinking and understanding, we literally don't know what we're talking about.
It's premature in the extreme to point at something that behaves so much like we do ourselves and claim that whatever it's doing, it's not "understanding" anything.
We've studied human behavior enough to understand that there are differences between animals in the level of cognition and awareness they (outwardly) exhibit.
Are we not generally good at detecting when someone understands us? Perhaps it's because understanding has actual meaning. If you communicate to me that you hit your head and feel like shit, I not only understand that you experienced an unsatisfactory situation, I'm capable of empathy -- understanding not only WHAT happened, but HOW it feels -- and offering consolation or high fives or whatever.
A LLM has an understanding of what common responses were in the past, and repeats them. Statistical models may mimic a process we use in our thinking, but it is not the entirety of our thinking. Just like computers are limited to the programmers that code their behavior, LLMs are limited to the quality of the data corpus fed to them.
A human, you can correct in real time and they'll (try to) internalize that information in future interactions. Not so with LLMs.
By all means, tell us how statistically weighted answers to "what's the next word" correlates to understanding.
By all means, tell us how statistically weighted answers to "what's the next word" correlates to understanding.
By all means, tell me what makes you so certain you're not arguing with an LLM right now. And if you were, what would you do about it, except type a series of words that depend on the previous ones you typed, and the ones that you read just prior to that?
A human, you can correct in real time and they'll (try to) internalize that information in future interactions. Not so with LLMs.
Not so with version 1.0, anyway. This is like whining that your Commodore 64 doesn't run Crysis.
We've put an awfully lot of effort into figuring that out, and have some answers. Much of the problems in exploring the brain are ethical because people tend to die or suffer greatly if we experiment on them.
Unlike LLMs, which are built by humans and have literal source code and manuals and SOPs and shit. Their very "body" is a well-documented digital machine. An LLM trying to figure itself out has MUCH less trouble than a human figuring itself out.
Just wanted to say that I think your rendition of Zellige is really nice, thank you for sharing! This is a very cool drawing language, and I think your art here is a really solid example of how creative you can be even with very limited tools.
I made a pinwheel, which is maybe not that exciting, but was a lot of fun for me. I like the bracket syntax a lot, and figuring out how to make this work actually recaptures some of the old sense of exploration and fun that I felt before programming was my day job.
Thanks, please consider sharing your pinwheel, I'm sure it was awesome to make too!
You're touching on an important point. I tried to figure out CFR[] by trial and error and it was a lot of fun, the kind of which I didn't feel for a long time. I think this boils down to two things:
- Bounded complexity, which encourages you to explore it to its fullest
- Immediate feedback, the result is not disconnected from the act of editing the code, so your iteration cycle is super short.
The other answer to this is correct, but a little more detail in case you're interested:
^W is a control code. This specific one represents the keys ctrl+w, a keyboard command in Vi and Bash among other things. It deletes the previous word. You often see something similar with ^H as well, which is a single-character backspace. https://en.wikipedia.org/wiki/Backspace#^W_and_^U has some more information about these.
Some people use them more to make visual jokes in written text, more or less the same way you'd use strikethrough formatting in text.
And in my experience ties in nicely with the fact that when a dumb terminal glitched or lost sync with the server, control codes would start to litter the screen instead of being interpretted. You'd see things like:
…thsi typo^Ŵ^H^H…
start to appear, sometimes followed by several random letters hit in frustration, before the final hard-reset of the terminal and resignation to the fact some unsaved work has been lost.
This story resonates with me a lot. I hope you don't mind if I share my own story (not as a counter-argument or anything, but out of solidarity from an addict in recovery). I don't really talk about this with most people, because I feel very weird about it. I used to be debilitatingly addicted to video games. I played World of Warcraft in particular compulsively, to a degree that I joke about but am also deeply ashamed of. There's an addon that you can install that will calculate the amount of real-world time you've played, across all of your characters.
I would really rather not know.
I'm not addicted anymore, and I couldn't tell you what changed. I have a guess though; I think it was mobile games? But maybe not the way you might think. I got really hard in to some of the early mobile gacha games for a little while. Summoner's War was one of them, it's about 10 years old now. I spent some money on them; I have a tech job, and splashing a little cash on pulling the slot machine lever for powerful new monsters felt good for a little while. But those things were barely games, they were basically the absolute shortest, least ornate form of the hedonistic treadmill that you could offer and still call it a game, they were essentially just designed to form habits and get you addicted to a routine with predictable rewards, and then slowly dial back the rewards as you progress but also slowly dial up the number of advertisements for the in-game cash shop.
Regular pc/console games would at least try to hide the ways in which many games are grinds, they'd try to make an actual experience and give you value for your money and make you feel satisfied about your purchase. There's a lot of psychological tricks involved there, please don't get me wrong, pc or console gaming isn't free of that kind of predation at all (especially these days). But with mobile gaming, that's almost all there is. It's like the difference between going to a nice restaurant and having a couple glasses of wine with dinner vs going to a 24 hour liquor store at 3am and buying a case. At the liquor store, there's no music playing, there's no ambiance, nobody cares if you're having a nice time, you're not there for the atmosphere, you're there because there's something in your brain that will hurt you if you're not.
I think that ultimately, spending money on four seconds of mild excitement over a few special monster summon scrolls that ended up being garbage, and then going straight back into the endless pointless grind, was what actually did it for me. I think there was a moment where I saw completely the whole machinery of the game, laid completely bare. I felt like I could finally see that it wasn't really a game it was like a garbage disposal that I could throw money into. It drove home what a sad little addiction I really had. The total merciless clarity that I was functionally just a type of depressed rat slamming a little lever hoping to get a good treat that would help me clear the next level of a dungeon so that I could grind a marginally better set of drop rates that would improve my clear speed on that dungeon slightly pretty much instantly burned out my dopamine receptors for that kind of gameplay.
The funny thing is, I actually can enjoy video games normally now? I buy games that look fun, I play them a normal amount, and then I put them down when I'm done, whether or not I've beaten them. I feel really weird about the way I experienced that specific type of clarity / addiction burnout, it's a very difficult feeling to describe, but I was just like "oh I hate doing this, I don't want to be this kind of animal anymore". I know for a fact that this is not something that everyone gets to do, and that in a way I'm really fortunate that I came out the other side of it a basically normal casual gamer instead of someone who would do the video game equivalent of sucking the tar out of a cigarette butt to get a nicotine fix, that's not usually how that story goes.
Thanks for sharing your story. I had a similar moment with the mobile game skinner boxes where I just got repulsed and lost all urge. But I still get sucked into higher quality games, thus, this post :)
Mogstation is truly incredible, it's like the digial equivalent of those "this is not a place of honor" signs, it feels designed to actively repel use.
That said, FF14 is a ton of fun! And IIRC these days you can play the free trial up through the end of the first expansion, which is a LOT of gameplay.
> The player politicized their neutral platform. The NBA fined players that wouldn't stand for the anthem for the same reason.
Standing for the national anthem is not a politically neutral action. FORCING players to stand for the national anthem is not a politically neutral action.
>Standing for the national anthem is not a politically neutral action.
Oh brother ... but OK.
The NBA went through something like this. Their policies called for standing for the anthem if you're out on the court while the anthem is playing (i.e. and cameras are rolling, and fans are watching). If you happen to have an objection to this, then you are welcome to stay in the locker room for the duration of the anthem and come out after.
I think that's entirely fair. You don't get to politicize a neutral platform, and you also aren't forced to do anything you morally object to.
The "1000 dollar apple phone" thing is such a weird myth.
My wife has been using the iphone 5s for years now, because the later models are too big for her to hold comfortably. It originally came out in 2013. It is still supported by Apple, and hers currently runs iOS 12.4, which came out last month. She is on her second 5s, because she dropped her first one in the lake trying to take a picture of our dog. I use it from time to time to help pull up music or navigation in the car and it runs just fine, it's a perfectly usable phone.
These days, you can get a "renewed" one for less than a hundred dollars, unlocked, straight from Amazon.
Android is the phone ecosystem where you actually need to spend a thousand dollars* on phone hardware, because all of the inexpensive android devices are cheaply made garbage that feel like they should come filled with candy. They are stuffed with manufacturer bloatware (if you're lucky, or malware if you're not), and are obsolete right out of the box. If you spend a hundred dollars on an android phone, I absolutely guarantee you that you will regret it.
I'm not exactly an apple partisan here, I work on android apps for a living, I've used android phones for years, I actually like the Android OS and I want it to thrive, but the current state of the overall android hardware and software ecosystem, outside of the very high end devices that most people frankly can't afford, is just dire.
*- I have a pixel 3 XL, I know a thing or two about spending a thousand dollars to get a usable android device.
There are enough cheap to mid range devices that do the job in Android. I dont think you understand the value of even 100$ in a developing country. Most people don't care about updates in this scenario. They want a phone that works. And far more people in the lower strata have access to cheaper Android phones than Apple.
If you are looking for a brand-new phone in the solidly mid-range for around $250, you can look at pocophone or OnePlus for smaller but still well-known companies making excellent phones. These phones compete in pretty much every way with the more expensive offerings by google or samsung, and have no vendor bloatware.
OnePlus's latest models have become a bit pricier, going on $450, but you can still buy the oneplus 6 straight from them for a good price.
This is, as I understand it, the goal. I mentioned this in a different comment, but getting companies to think of unsecured consumer data as a liability is absolutely key to getting them to take privacy seriously. Companies need to consciously decide if the risk of accruing this data is worth the downside. Pre-gdpr there was functionally no downside at all.
If personal data doesn't have value to the company and it's not an asset, then the company shouldn't be collecting and storing it. Making storing personal data a liability is good if the goal is to prevent companies from storing personal data unless there's a strong business need (and corresponding level of care) involved.
I noticed that too (mtg meme group just showed up as a wall of "image may contain:text" placeholders), but I didn't realize it was an outage and I thought that I'd accidentally turned on some kind of really annoying safe-mode.
Obviously, people find some value in some output of some LLMs. I've enjoyed the coding autocomplete stuff we have at work, it's helpful and fun. But "it's not qualified to answer my questions" is still true, even if it occasionally does something interesting or useful anyway.
*- this is a complicated term with a lot of baggage, but fortunately for the length of this comment, I don't think that any sense of it applies here. An LLM doesn't understand its training set any more than the mnemonic "ETA ONIS"** understands the English language.
**- a vaguely name-shaped presentation of the most common letters in the English language, in descending order. Useful if you need to remember those for some reason like guessing a substitution cypher.