Oslo acquires ChatGPT for 110k students and teachers
digi.no
digi.no
Technologies like GPT-4, much like Google search before it, vastly increase the number of things we can do acceptably well without needing to study it in depth beforehand - letting us save those extra brain cells for where they really count.
I make no statement about AIs influence on intelligence yet.
Unless you have an abundance of "free" brain power eg. you work physical labor or otherwise don't have a job that requires mental work... how isn't brain cells (figuratively speaking) a scarce resource?
A small paper with a linited study on 300 students from Pakistan and China and the result is “Accepting AI without addressing the major human concerns would be like summoning the devils” .
Who knew Nature could sound like a bible thumping southern baptist.
Personal prediciton. AI fears on human IQ is about as valid as fears of the television and computer games have historically been.
Aka incredibly powerful teaching tools wjhen used right, damaging propaganda when used wrong.
Autonomois warfare is what scares me, not this hysterical crap everythime humans get a new useful technology.
Specifically, masses relying on unintelligent responses are exposed to unintelligence (e.g. a number of people may absorb the il-logic), so as an aggregate there are risks.
btw it's called a garden-path sentence. https://en.wikipedia.org/wiki/Garden-path_sentence
Never knew there was a name for these sentences.
When I forget to start Slack and someone sends me a message, Slack sends me an email saying "You have a new direct message in <channel>".
Well, it so happens that when I get this, the notification dialog that pops up to tell me about the email is just the right width that it gets cut off as "You have a new dire.."
The first several times this happened I went straight to Slack eager to find out what was so important.
is visible in the overview.
> [that this and that Meetup needs]"
Is the full sentence visible after clicking.
About the only real counterargument is that you could also interpret the translation as trading the business for 110,000 students and teachers. That would not be garden path as it would be (accidentally) misleading wording in its entirety!
Edit: typo
My FEIDE expires so I cannot try anymore :)
Not really relevant when they will be using GPT-4, is it?
Plus the powerful features from code interpreter is a night and day difference
Not everything is worth saving a bit of money to get worse performance from something that has a worse network effect
I guess organisations move slow.
But atleast now they will have a culture and practices of working around it.
However I think it's a typo when the article says GPT-3.5. It doesn't make sense to "buy" GPT-3.5. They probably meant ChatGPT Plus which includes GPT-4 access (50 messages in a rolling 3 hour window).
Who builds the UI? Who manages accounts?
Mistral , open source etc do not have a sales force. Shame really
Is a school supposed to set up infrastructure for all of that themselves?
The Mistral cloud offers are perhaps a bit too young to have been considered. And perhaps a bit more risky too.
First people have to be taught to use it, for that 3.5 is decent.
I guess its like your grandma using Google, vs as a working professional using Google.
Your grandma using google is still a plus.
at least for professionals, maybe kids don't care too much
Norway: buy 110k licenses
Oslo to Kirkenes (which is further east than Istanbul).
"European" as an identity is a rough diamond still.
Too bad some people want to abuse EU to destroy this and move towards monolithic slob with little variance.
The 'critical thinking' part at this point in time leads you to only one conclusion: that the tool is unreliable and dangerous and shouldn't be used. Or, you're teaching kids that they need to cross-reference all 'authoritative sources' which leads to a complete failure of the system - if you give kids a textbook will they have been 'trained' by GPT to not believe a word of any of the claims in it without cross-referencing? How does that enable better education.
If you provide a tool from a position of authority it needs to be a reliable and believable tool; standing in a position of authority and handing out access (to children!) to a lie box and asking those children to 'rely on it' is frankly abusive. That's beyond mixed messaging, that's just some form of weird mental torture.
I just think it's a real issue to give kids free access to a tool that so confidently makes stuff up. I know that textbooks aren't perfect but they're at least audited by a committee of experts, mistakes slip in but there's oversight, there's a procedure to ensure that they're reliable representations of (currently understood) facts, and that when politics enters the textbook production process it (correctly) becomes a bit of a national scandal.
And its enough when such a technology is 80% right. Nobody will verify every answer if it is correct most of the time.
The issue arises when the belief you have is tied to your ego so you only cherry-pick the evidence that will satisfy your ego and ignore/question everything else, but not this.
There is no information that they compared open source models and considered whether to build their own data servers.
You should at least do some experiments to show that gpt can improve quality in education, right? Although I know that data manipulation is common in scientific research nowadays by picking the data that better serve the wanted assumption, but this time they even skip this step.
Disclaimer: I live in Oslo; I'm an educator; I usually see similar poor decisions in education due to ignorance.
But the logic in Oslo is that our municipal government has a budget, and buying it at least shows what I have done as a civil servant.
I don’t know if you can get their “must do something” mentality.
That sounds reasonably similar to me. People keep buying iPads for classrooms because “it puts technology in the classroom” or something like that. Then they break and get sold for scrap.
Yes, figuring out how to teach and what to teach using computers was difficult, but students needs familiarity with computers and how to use them.
Saying that "they'll learn on their own in their spare time" is ridiculous and tantamount to saying schools shouldn't be teaching reading, because students will want to read on their own outside school and they already have access to books at home.
BTW a lot of corporations that have yearly IT budgets replace large portions of PC's and server goods and sell them to schools for a dollar (or Euro). I wish that my school growing up knew about this.
Someone coming in here and %s/computers/ChatGPT/g is going to make an argument that sounds equally reasonable—so if ChatGPT is truly, materially different and worse, then the argument against it is a different argument.
My complaint here is that schools are making educational decisions based on technological fads, without good plans for how to use these technologies in education. Yes, a stopped clock is right twice a day—sometimes, it will be the right decision. Sometimes it will be the wrong decision. The trick is knowing the difference.
anyone with more knowledge of scandi languages able to understand why? is the "swedish" (or "norweigan") part implied?
"Utdanningsetaten" translates to "Education Agency". But I guess instead of "The Education Agency" it tries to translate it into context, which should be "The Norwegian Education Agency".
However it seems confused about which country is being talked about so it just think it's Sweden maybe.
You just need one teacher who acts as a supervisor. And another group that sets the curriculum. No need to actually teach for the most part.
How does that matter if kids can access from home?
You could go back even further and ~25 years ago, I had an entire class on using search engines and all manner of advanced operators and search techniques.
With today's computing power it would be awesome to see what the search engines of past decades could have become if not for the monetary incentives to obscure information and increase engagement.
I.e. good google
I spent 5 minutes searching Google for what an email error code meant, it kept showing me threads on Microsoft's forums that never gave me a clear answer. Took about 10 seconds to get the correct answer on ChatGPT.
If you treat it like it's a bumbling assistant that has remarkably encyclopedic knowledge, it's a very helpful tool. If you treat it like it's truly intelligent and let it do your thinking for you, then you're going to have some problems.
But, generally speaking, if you don’t have some way of verifying what the correct answer is, how can you be sure that the answers you get in 10 seconds from the LLM are actually correct?
In many cases, programming included, you don't actually care if it's 100% correct, it is enough to just give you a nudge, show you the way forward, suggest some ideas etc.
If whatever it is you are doing can be verified for correctness, then of course you will check the answer. But for creative pursuits there often is no correct answer, and generative AI (LLMs included) excel at that.
I don't understand how people feel they can use it as a search engine replacement. If you have to go separately check the results then it doesn't seem like it's saving any time. I suspect people are just trusting the output a lot more than they should.
Before ChatGPT I had to manually investigate and read through multiple pages of potentially irrelevant stuff to find an answer I could try to verify. If you find an answer on StackOverflow, you still need to integrate and verify it.
Now I get an answer straight away just by asking a question, and can skip manual knowledge extraction step to get right on answer verification step (i.e. compiling, testing, extending).
Even if ChatGPT is nothing else except a way to fuzzy search and summarize found results, even just that is already hugely helpful and has almost completely replaced Google for me.
But on a serious note I have only gotten good content out of ChatGPT when it's used as a smart Thesaurus, or a code checker (even that was 50:50). If I have a paragraph that I feel needs rewriting or condensing I can throw it in and resubmit until I get a good result - which I still then need to edit to make it less... ChatGPTish. If I open up old scripts (data extraction/analysis) that I forgot what they do I chuck it in and find out what will happen.
Like any large technological advance, the hype is much larger than the eventual reality (see VR goggles).
[1] technically it doesn't even make "mistakes", it's just an LLM doing its thing, but we humans like to anthropomorphise
just like a child, chatgpt can not tell the difference between reality and fiction.
chatgpt lies because it presents fiction as fact. there is no intent to deceive necessary.
I'm done anthopomorphising now.
To me, it is like a junior developer. It does the bulk of the work and whenever there is an issue, you can tell it like in a code review to fix a particular piece. Look past the confidence level and focus on the work product.
Remember, it doesn’t need to always be accurate. It doesn’t need to be perfect. It just needs to outperform the alternative (passable job here) for the price paid (amazing value here).
But indeed it's still quite useful. It's best used in situations where it doesn't matter much if the answer is right, or when the answer can be easily checked if it's correct, yet more annoying to come up by oneself in the first place.
90% of the time I know the code I want to write. ChatGPT just writes it for me, when I read it I can confirm it is what I would have written, or fix it if not (or more likely: tell ChatGPT to fix it)
This means I don’t have to google for syntax, canonical ways of doing things, etc.
If you are asking ChatGPT for facts, triple check everything it tells you. Or better still, paste the Wikipedia entry to it first, then ask for the style or type of writing you are interested in about those facts
> People are enthralled with this tech when they should still be very very skeptical.
Yes, that's the real issue. Everyone should be very skeptical about ChatGPT responses. But instead it's often trusted blindly, because "it's often enough correct".
But I get the idea behind this particular move and at least it can be a chance to teach the students to distrust it.
Isn’t denying these facts just as dogmatic as accepting them blindly?
Instead, think of it as a computer that can perform transformations on text, described in plain language and based on text's tone and meaning.
There's its ability to write small one-off scripts for me. I've given it an example of a structured log file and told it to "write a pyparser script to parse this string" and it will do that. What would take me a hour to learn pyparser and get somewhere now takes a few minutes. You can literally see it iterate on different solutions until it comes up with something acceptable. (ChatGPT 3.5 doesn't do that.)
Another great use case is as an assistant when I'm learning some new topic.
I was reading A Mathematical Theory of Communications, the breakthrough thesis by Claude Shannon. It's quite accessible, but there are always points where you don't quite understand a point that is being made and you want to get more examples. I can ask ChatGPT to create a concrete example, use actual numbers, plot a graph that illustrates the concept. And it will do that, along with the Python script that it used to create the graph so I can double check what it did. "Create an Eb/No graph for the following parameters." And then ask to modify the graph to illustrate a particular point. Google can deliver similar information, but it will push me a document that contains the answer, along with a truckload of info that I wasn't looking for. ChatGPT pulls just the info that I'm looking for.
Or lately I was looking into Laplace transformations, low pass filters, pole/zero plots etc. ChatGPT 4.0 uses the SymPy symbolic math package to derive the math, I can once again examine the scripts, ask for concrete examples with graphs and so forth. It's way more efficient than doing the same thing with Google, which will usually send you StackExchange with an answer that is not quite what you're looking for.
In these kind of cases, I know enough math and subject knowledge to know when it's bullshitting, but it doesn't have to be 100% accurate to a huge help in learning.
Here's an example of some thing I did off the cuff. I have a bunch of hand written notes that are mind maps and graphs written on old notebooks from 10 years ago. I snapped a picture of all of the different graphs, threw them into chatGPT and asked it to convert them to mermaid UML syntax.
Every single one of them converted flawlessly when I brought them into my markdown editor.
The amazing thing here is that that the ChatGPT large language model wasn't even explicitly trained for this at all, but with the massive quantity of data that it's trained on its able to just connect the dots.
If you're using chatGPT as nothing more than a glorified fact checker and not taking advantage of the multimodal capabilities such as vision, OCR, Python VM, generative imagery, you're really missing the point.
I don't think gpt-4 is worth its 10fold price premium for all but niche usecases (coding may be one but I don't do that too much). And the much longer response delays are annoying.
I think 3.5 turbo is a really nice general use model. And the price difference makes it worry free (I use it a lot and most months I used less than $1) whereas gpt-4 really tends to trigger disturbing invoices after using it a bit :)
I still have 4 on tap for a second opinion if I'm dissatisfied with gpt-3.5 but in general it doesn't really yield much better results, whereas better prompt engineering helps much more with both models.
Of course YMMV but I wouldn't write gpt-3.5 off just because it's no longer the coolest kid on the block.
OpenAI and Microsoft aren't a lot better but it is somewhat better at least.
So Gemini isn't an option for me. Curiously enough, LLaMa 2 is, even though it's from meta, but the fact that you can run it at home is of course a real winner for privacy.
when i changed schools and realized that students scored more because they have access to past exams and they would rather practice them then study the material. in a competitive environment, it nudged me to also adapt to this strategy rather than learning the concepts the intended way.
regarding hallucinations, i wonder how much critical thinking would matter. back when students referred to wikipedia or random search results for an essay, a fair bit of them just take the material for its word.
Even still, they use quite a lot of money too.
From the ChatGPT terms:
> Minimum Age. You must be at least 13 years old or the minimum age required in your country to consent to use the Services. If you are under 18 you must have your parent or legal guardian’s permission to use the Services.
Copilot is completely free when using Edge with a Bing account. Why spend a dime of taxpayer money on this??
In a future where every person has an all-knowing assistance in their pocket --a nearly perfect AI who knows it all-- what should we teach in schools, from early childhood to university?
The corollary to that question is:
When that happens, what will be important in people? How do people add value to an organization?
The answers are, as might be expected, varied.
Some are dismissive. They simply do not believe that moment will come or that it will be as described. That it cannot be as described.
Others are more apocalyptic and come very close to being fearful of this future.
Yet another group is glad they are older and will get to watch this happen while retired.
In some cases there's a struggle understanding how to teach certain subjects, mathematics being a typical example. If you have a device that can solve any and all mathematical problems --and even explain them-- what do you need to know about math?
My take so far --I say "so far" because this doesn't even approach a supportable conclusion-- is that the differentiating factor will be experience and the ability to effectively connect, or use, multidisciplinary knowledge.
Maybe I look at this from an entrepreneurial perspective. A small experienced team will be able to rent or buy thousands of experts and workers across every possible discipline and, using their experience, put them to work towards a stated objective. A person who is "mono-skilled" (I'm sure that's not a term), will not be able to utilize these tools to the same level of advantage.
And, yes, I think you will have to know a lot about a lot of subjects. That's the only way to understand if the AI tools and workers are rowing in the right direction. How we impart this knowledge is a different matter.
One thing is certain, education will have to change radically to address the devaluation of raw knowledge in favor of being able to take different types of knowledge lego blocks to make things happen.
It is possible that Norway's leaders just made a genius move by making ChatGPT available to everyone in this way. Brilliant.
Those abacus civilisations are still laughing at our stupidity.
So too here, one could argue the positive/negative impacts of things like the calculator on our mental capabilities, but chatbots are taking this automation of mental processes up an order of a magnitude both in terms of the breadth of things that can be "input", and the variable quality of what is output.
The unfortunate thing is really the only way to test this is to basically do what we did with social media - just let a generation have at it, and see what happens.
As opposed to straight up government propaganda in textbooks? At least chatgpt averages out everyone’s propaganda.
edit for clarity: this comment applies to history books, literature, stuff like that. If you look carefully you’ll notice they’re full of stuff designed to instill a sense of national identity and patriotism. Especially obvious when you look at different nation’s textbooks about the same historical events described from opposing sides.
ChatGPT in general has been pretty decent about saying “Well different people describe this differently …”. Much better than any textbook I remember from school.