Chegg stock sinks after CEO says ChatGPT hurt growth
finance.yahoo.com
finance.yahoo.com
Selling is likely to be the right move, but blaming the CEO?! There is literally nothing he can do in the near term to avoid losing users to ChatGPT-like services. Their business is likely doomed if they don't already have a rabbit in their hat.
It'd be a convenient yarn to spin to concerned shareholders, anyways.
I found many vulnerabilities in one of their recent acquisitions and although I was paid, they indicated that they won't fix any of them.
That’s interesting to hear they don’t plan to fix them. I wonder if the bug bounty exists as some sort of insurance discount for having a program at all, even if they fix nothing?
What % of the payment would you like to disclose them to me?
Ask ChatGPT for a syllabus foe an introductory calculus course and ask for a study plan. Then ask for a study plan for each section. Than start asking it to explain things.
People will build UIs to automate this process where all you have to enter is what you want to learn, in this case “introductory calculus”
Do schools not consist of this more and more due to increasing reliance on metrics based on standardised testing, where the best way to juke the stats is to teach the test which means teaching kids to memorise facts? Ones which explicitly limit access to the internet or LLMs during test explicitly to make the only route to passing rote memorization of a great deal of facts?
Does GPT and LLMs, which can easily crush any test based on fact memorisation, not present an opportunity to offload the task of mere fact memorisation to the machine and allow one to learn how to do things which would have previously been unviable without first memorising a great deal of facts?
I don't consider knowing WHAT to ask to be all that much of an obstacle either, you just ask it about something you don't know very much about, but that you want to do. If anything I see LLMs as presenting an existential threat towards models of measuring merit through memorisation, as it's exposing said memorisation of being testing for testings sake, disconnected from real world utility or the ability to meaningfully measure merit.
Is there a single standard for "school?" I was using it as a device to separate out the difference between the two styles. I do recognize that the majority of American schools are simply memorization factories but to focus on that is to miss the argument entirely..
> If anything I see LLMs as presenting an existential threat towards models of measuring merit through memorisation
As you've essentially just repeated that same argument with this point. The types of "learning" LLMs are good at are not the types of "learning" people would _expect from a good school_.
I do believe that this will occur, but it simultaneously undercuts your previous arguments in favor of GPTs utility, as they only make sense in a world where education is _purely_ about memorization.
And sorry to reorder this, but...
> which can easily crush any test based on fact memorisation,
What is your source for this assertion?
Back in the day when the internet was becoming popular there were loads of similar stories (likely openai’s marketing team is familiar with these) about people using the internet to learn.
Recent times are proof that that was a mistake for many.
You are creating your own conspiracy about how internet created conspirationist. Social Medias are the fault, not learning stuff on internet.
I try to think about this like wikipedia, which was very unreliable in its early days (and which people said would never be trustable). It still has some untrue information, or biased information. I don't trust it as a definitive source of knowledge. But I do turn to it frequently because most of the time it is either correct, or points me in the direction of correct information.
Personally, imo, there is already immense and easily realizable value in LLMs, in their current truthy-falsy state. For the sake of the people who categorically deny they could ever be of any value to them, I am perplexed.
The risk is that people will assume that this software is right all the time and that means they will uncritically follow it.
For example, I'll watch a youtube video or read a blog post that talks about a topic I'm not an expert in. That type of content often uses jargon or acronyms that I'm unfamiliar with. Asking ChatGPT with some context usually explains the jargon to me, but also can help explain why it's used in this specific context, which is way more helpful than just reading a definition.
Could the fact that you are not expert at things be the reason you think you are learning from the chatbot? If i ask it anything i am an expert in i can quickly tell it’s a waste of time, but if i ask it in fields i am not, i get tricked into thinking it’s right.
Meaning that while it does appear to be right it does so due to my own ignorance about the topic.
But I disagree about time wasting. I use it to generate boilerplate code in languages I’m less familiar with, but can fill in the gaps. It’s stellar for that use case, scaffolding entire CRUD apps with DB connectors and specific libraries.
So, if you keep asking really superficial questions, you will probably get most answers either correct or wrong by some obvious inversion or single change that can be verified easily.
don't get me wrong, a chatbox is absolutely the worst interface for that kind of interaction. Search would be much better. But the big search companies seem to use all of their resources to make search worse, while investing heavily on good chatboxes, so we got here.
The idea that everything the chat bot is saying is wrong is just... well, wrong. You should take it with the same skepticism you would anything else you read on the internet, but I've found ChatGPT4 is still a very valuable tool to have.
So you assume if you are an expert at something you would be so well versed that you couldn't learn new things from a chat bot?
If this is the case, then I am not and will probably never be an "expert" by this standard.
I'm amazed the public was willing to invest in a site that, as far as I can tell, solely exists to help students c̶h̶e̶a̶t̶ get access to homework answers.
At least ChatGPT does things besides cheating :)
So while it may not be the case that all users of Chegg are cheating, at some schools essentially all cheaters are using Chegg, including to cheat on exams (students post course bibles, test banks, etc.).
It doesn't help that Chegg is slow to take down materials even when explicitly told that the material was obtained illegitimately (which makes sense---empirically, illicit materials are the draw for at least some segment of their customers). I know several folks who have had to involve their university's legal team to get their course materials removed from the site.
Maybe you didn't have such an experience, but that doesn't mean it's a paradise for cheaters. I used it for help solving textbook problems and occasionally for homework. For how much it costs, it was a supreme waste of money.
Shouldn't the fact that your university has to put effort into thwarting cheating on Chegg be evidence that cheating is a problem at your school?
People are talking about whether deep learning and AI will affect companies and job, and when that will happen, and in the case of Chegg we already have the answer (barring them making some large pivot).