ChatGPT is losing some of its hype, as traffic falls for the third month in a row::August marked the third month in a row that the number of monthly visits to ChatGPT’s website worldwide was down, per data from Similarweb.
It took a while but yeah that seems about right. It takes a lot of guiding to have it produce something usable. I have to know a lot about what I want it to do. It can teach me things but the hallucinations are strong sometimes so you have to be careful.
Still it helps me out and I make a lot of progress because of it.
I like it for certain techy things. I just used it to create a linux one-liner command for counting the unique occurances of a regex pattern. I often forget specific flags for Linux commands like how
uniq
can perform counting.And something like that is easy to test each piece of what it said and go from there.
As long as you treat it like a peer who prefaced the statement with “I might be wrong / if I recall correctly” it ends up being a pretty good aid.
“I can suggest an equation that has been a while to get the money to buy a new one for you to be a part of the wave of the day I will be there for you”
There, my phone keyboard “hallucinated” this by suggesting the next word.
I understand that anthropomorphising is fun, but it gives the statistical engines more hype than they deserve.
Your phone keyboard statistical engine is not a very insightful comparison to the neural networks that power LLMs. They’re not the same technology at all and just share the barest minimum superficial similarities.
Ah “neural networks” with no neurons?
I’m not comparing technologies, I’m saying those are not “hallucinations”, the engines don’t “think” and they don’t “get something wrong”.
The output is dependent on the input, statistically calculated and presented to the user.
A parrot is, in the most literal of ways, smarter than the “Artificial intelligence” sentence generators we have now.
deleted by creator
because they are being wilfully obtuse suggesting that “neural network” a term going back over half a century for a computatuonal method doesnt apply to things without biological neurons, and doing the same thing applying an overly narrow deffinition of halucination when it has a clear meaning in this context of stating textually probable but incorrect statements.
Aka “hype”
These types of articles bother me. Almost every game, movie, and product has an initial unsustainable level of hype, then comes back down from it.
But these articles inevitably try to frame it as if it’s an indication that something’s failing.
That very much depends on if you believed all the hype or not. If you did, then yes, it’s failing, as ChatGPT was supposed to be the next big breakthrough that was going to automate everything ever, and any company that didn’t get in on that right now was going to be left in the dust by all their competitors. On the other hand, if you were an actual sane person (so you know, not a CTO/CEO), then this is very much a non-story as you always knew that all those outlandish claims were nonsense and that this was always going to be yet another niche piece of tech that’s useful in a few places in limited amounts.
I mean, yeah? You can’t rely on hype ever being present. Honestly with the very light use I do (Asking a couple of questions at most in a month) I feel like it has not changed at all. Just at the top I now have a locked button that says “GPT 4”.
FWIW, it’s become an inextricable part of my life. I use it for hours every day, for programming and Linux advice, spreadsheet help, foreign language practice, and random trivia.
Yesterday, I discovered that Snoop Dogg’s -izz speak from the aughts was actually derived from carny pig Latin.
Yikes, I would be very scared to take anything ChatGPT says as accurate. Google keeps trying to get me to use theirs when I do searches, and I refuse.
I think (hope) that peroson is being facetious.
I hope people are smart enough to understand that the statistical sentence generators don’t “know” anything.
It can generate simple stuff accurately quite often. You just have to keep in mind that it could be dead wrong and you have to test/verify what it says.
Sonetimes I feel like a few lines of code should be doable in one line using a specific technique, so I ask it to do that and see what it does. I don’t just take what it says and use it, I see how it tried to solve it and then check it. For example by looking up if the method it used exists and reading the doc for that method.
Exact same as what I would do if I saw someone on stack overflow or reddit recommending something.
It’s just very quick at doing simple things you already could do - or doing things that you’d need to think about for a couple of minutes.
I wouldn’t trust it to do things I couldn’t achieve. But for stuff I could, it’s often much quicker. And I’m well equipped to check what it’s doing myself.
Statistical sentence generator gets thrown around so much, if anything I doubt people actually understand what can be achieved through just that. It doesn’t matter if it doesn’t know anything. If it can generate sentences statistically with a 100% correct and proficient outcome, it’d always be correct regardless of its lack of knowledge.
We’re not at 100%. But we’re not at 10% either.
A parrot can generate sentences with a 100% correct and proficient outcome, but it’s just using sounds their owner taught them.
Garbage in, garbage out.
Even the smartest, most educated people are never 100% sure of anything, just because there’s always nuances.
These engines are fed information that is written witg 100% surety, completely devoid of nuance. These engines will not produce “answers to questions” that are correct, because “correct” is fluid.
Meh.
That’s a very fallibilistic viewpoint. There are lots of certainties that can be answered correctly.
There are fields and fields in science that work on things that are “certainties”.
If you’re talking about simple stuff like “what is the first letter in the english alphabet”, then sure. But many people, even in this thread, say they use the engines for hours, to get answers, guide them, and discuss.
It is a parrot on steroids, but even a parrot has knowledge. LLMs have 0% knowledge.
You may be right now that I reread their comment.
you have to pay for gpt4. its smarter and more capable, but slower than gpt3.5
I didn’t say I wanted it for free.
I’m starting to think I’m incompatible with GPT. Code that’s laughably wrong (like sticking in things that aren’t even in the language), DM advice that I could get walking down a greeting card isle, and explanations that would get a Wikipedia editor sent to the firing squad.
Nah, that sounds about right. This is just the natural result of people actually trying to use GPT for all the things they were told it would be able to do, and now discovering that was in fact all bullshit. The LLMs are and always have been massively overhyped and oversold on what they can do. Sadly this won’t stop the corporate executives from trying to use them to replace workers, although when that effort eventually face plants they’ll just quietly re-hire a bunch of people and find some middle manager to blame for their failure. This is was and always will be merely a productivity tool to automate some repetitive work, but it still needs someone to review and clean up its output. It’s not “replace someone doing 40 hours of work a week”, it’s “allow someone to do what used to take 40 hours in 35 hours instead”.
Sadly the most impact this is going to have is on spammers and scammers, who can now automate generating their garbage since it never mattered that any of that crap was accurate or not, merely that on a casual glance it looks reasonable.
deleted by creator
The exact same statement applies to computers, mechanical looms and the plough. Thats how technology works.
I agree. I think people are just missing the point. It’s really far from being able to replace a worker.
It’s current capabilities at best can help that worker be slightly faster at certain things. It’s akin to a type of search engine.
Yeah. There are a lot of shitty marketing ideas that suddenly become profitable if you don’t have to pay people to generate the content it needs. Honestly I’ve had a couple of those ideas over the years and I’m glad I’m no longer in a position to propose them to anyone.
Unethical parlour tricks, and nothing more.
When I first started messing with it, I was kind of neat and fun. I like making characters so I was using it for like story prompts and general outlines. Some were better than others, but it was neat for some inspiration and fleshing out. I never took it’s outputs 1:1.
But when I messed with it again recently. It was a lot worse. Like it ignored parts of my prompt. Like as an example a prompt was about a romance story, but the story was about character A and their family. The love interest character was barely a footnote and could have been removed entirely and nothing would have changed with the story outlines it was giving me.
I thought maybe it doesnt like romance prompts, so I tried less specific and more broad prompts from there, and it was the same thing of just… not outputting what I was asking it to. It got worse and worse and sometimes wouldn’t output anything at all.
Not sure what language you’re coding in, but I’ve found GPT-4 incredibly helpful for coding in C++ and Python.
ChatGPT has gotten dumb. I used to have to code check it’s answer every few responses. Now it’s every response. It wrote me an if/else statement the other day where if and else had the same outcome.
deleted by creator
I had a thought about this. I wonder if they intentionally made it dumb so you would opt for the paid version.
deleted by creator
Source: Work in AI, sometimes on LLM’s, mostly on the software engineering side rather than the science side.
I have a few theories on why ChatGPT was so successful, and why the hype is starting to crumble, but they all largely centre around a well-known problem that LLM’s have had for years - they hallucinate, a lot.
When your product becomes popular, you deal with unique problems that don’t seem to scale without insane amounts of money, or a literal army of people to plug gaps where your model is saying things it shouldn’t - whether it’s accusing high-level politician’s of crimes they didn’t commit, or telling people how to make chemical weapons in the style of your grandma. It’s an expensive loop in compositional models, so I’d hate to know how much work it took ChatGPT to get to it’s “best” version.
Over time, valuable data disappears, and your data naturally skews over time due to it being incorrect, invalid, or pushed into just being biased towards a given inaccuracy. Sometimes, you do everything right and you train on manual input that you’ve vetted as correct through expert analysis or user feedback - and it’s still wrong. IMO, ChatGPT was always going to struggle to keep the hype, and it will eventually be seen as what it has always been: a concept that shows the utility of LLM’s as a commercial product.
Make no mistake, the likes of Google, Amazon, Apple, and Meta will probably plug the gap, and will reach either parity with GPT4 or improve on it. However, the fundamental problem of hallucinations will not disappear, and we’ll continue to see neutered experiences that make for great tools, but burn cash to provide these tools with minimal possibility of offending people/damaging the brand.
The main thing I hope to see from the rise and fall from ChatGPT is a rise in productivity tooling, but also people to finally see those that hype these technologies as what they are - grifters.
There’s also the novelty factor. Like how DALL-E was the rage not that long ago, people flocked to it because it was new and interesting.
But the novelty has rather worn out by now.
This is a big part for me. When ChatGPT first came on the scene, I was absolutely blown away by its natural language parsing capabilities, but it wasn’t long before I started to hit the boundaries of its abilities. I was disappointed by how unreliable it was with anything but the most simple queries. Now it just doesn’t do enough to really bother with.
School’s just starting. It won’t hit the peak of hype without some huge new features or improvements, but it’ll rise again.
Yeah I’d love to continue using ChatGPT but I got warned for making it roleplay as Widowmaker and trying to fuck the bot.
They don’t want my money? Fine. I’ll give it to someone else who doesn’t have arbitrary morality rules on playing wall-ball with linear algebra.
Hasn’t the service also gotten worse? When it first came out, you kept hearing how it could pass the bar exam and medical license test. And now all you hear is that it can’t do basic high school homework without wrong answers. Maybe it was hype in the beginning and it never could do those things.
deleted by creator
It is so neutered now that I rarely find it is worth the time to use it. Now it just gives non-answers that are rarely helpful or accurate.
Edit: To give a bit more context, I still use it several times a day just to be sure it is still a disappointment.
It was always able to do some genuinely amazing things, but it was always limited when you took it beyond its wheelhouse. It helps to think of it not as an “AI” as people keep saying, but as a “text completer” with a huge amount of power within that domain.
Or, another way to think of it is as a super-powerful search engine. If the answers and knowledge you’re asking of it were fed into it as input data at some point in its training, it’ll probably be able to find it and reformat it back to you with a scary amount of smoothness and precision. If you’re asking it to figure out something new, it may be able to fake it in some short-term fashion or another based on what it’s seen, but not with any genuine understanding behind it. That’s just not what it does. I actually have a little private theory that if it was given something like the bar exam in scale and complexity, but an exam was genuinely a whole new novel invention that hadn’t been extensively discussed and represented in its input corpus, it would fail pretty badly. A lot of what humans can do that makes them capable is adapt to new domains – we can teach ourselves to play chess, or do math, or fly airplanes, or play Celeste. GPT is hugely impressive but it’s still only one domain.
I actually don’t believe that it’s gotten substantively less capable. I think there are little ticks up and down in its capability sometimes in particular areas, and people seize on those to conclude that it’s now becoming dumber, but in my experience, the raw API was always quite capable (more so than the somewhat nerfed chat interface), and it was always super-capable with some tasks and not at all capable with others. I think journalists are just now figuring out that, after having studied the issue in their professional capacity for the better part of a year, and reporting on it as if it’s a new thing.
From what I’ve seen, here’s what happened. GPT 4 came out, and it can pass the bar exam and medical boards. Then more recently some studies came out. Some of them from before GPT 4 was released that just finally got out or picked up by the press, others that were poorly done or used GPT 3 (probably because of gpt 4 being expensive) and the press doesn’t pick up on the difference. Gpt 4 is really good and has lots of uses. Gpt 3 has many uses as well but is definitely way more prone to hallucinating.
I made the mistake of asking chatgpt questions about securing my network setup. It confidently gave me a huge amount of misinformation that led to 8-10 hours of frustration and pointless troubleshooting.
Do NOT trust ChatGPT.
Came here to say this. ChatGPT is very good at coming up with convincing bullshit. Always do your own research.
This is just balancing out. Anything that gets over-hyped will eventually drop in use. It’ll eventually be a boring yet useful tool just like spreadsheets, spellcheck, or email.
ChatGPT: *declines in popularity *develops sentience *gets emotional *evolves into SkyNet
They definitely nerfed it. We will probably end up in a situation where corporations and the rich have access to god-tier AI, and everyone else has access to mediocre, ad-supported AI.
Just wait until we find out what the US military has.
No one seems to have thought about the fact that most schools have been out for those three months. Not sure exactly how much of the traffic is high schoolers and college students cheating, but that could account for at least some of the loss in traffic.
Edit: missed a word
Cheating isn’t necessarily the only use case for GPT, although it definitely does play an impact on the overall number of users.
I would like to see if the use of their API increased.