Please consider removing the Explain feature entirely

[davidculley]

@Bassario Do you mean my text? Or @BIG_JEFF’s reply?

My text was most certainly not written by/with ChatGPT.

In case you meant my text, let me say this:

You probably just wanted to be funny and make a joke, but I find it tone-deaf and frankly a bit insulting that you allege that I used ChatGPT after just having read a massive wall of text about how much I disdain ChatGPT.

Most people—to pick an arbitrary number, let’s just say 90%—are so bad at writing and following a structured approach that when they get to see someone who is good at writing, they naturally assume that this person must have in some way cheated or used some AI tool. Is it out of envy because they themselves can’t write that good? I don’t know. Talking about what I referred to as “societal costs” … In less than two years, ChatGPT has trained people to assume that good prose must have involved a LLM, thereby disparaging skillful human writers.

You just proved my point that ChatGPT has costs besides the 20 bucks a month, which most people unfortunately just can’t see due to a lack of knowledge.

Why do you think my writing looks as though it was written by ChatGPT? It’s because ChatGPT steals from authors, copying their style, and thereby making the investors of OpenAI rich, without giving anything back to the authors it ripped off. Don’t make the mistake of treating ChatGPT as a person. It’s not a person who learned to write, and developed its own style as humans do over the course of years or even decades of effort and honing their craft, learning to present information in a digestible way. My writing might look like it was written by ChatGPT because ChatGPT rips writers off and presents their work as its own, annulling all the time and effort the writers invested in their craft—just so that I can listen to tone-deaf jokes—and that makes me angry.

[Peter1949]

No, it was written by a descendent of Ned Lud.(zum Scherz) Anyway here is ChatGPT’s response to davidculley’s post:-
"This critique touches on a very real concern: the environmental impact of artificial intelligence and data-intensive technologies. The increasing demand for data centers, which power AI models like ChatGPT, does come with a carbon footprint due to the energy required for both training AI models and handling everyday usage. Here’s a more nuanced look at the argument:

1. Energy Consumption and Carbon Footprint

• Training large models like GPT indeed consumes vast amounts of computational power, translating to significant energy use. Some estimates suggest that training a model like GPT-3 could emit as much CO2 as several transatlantic flights, depending on how energy is sourced.
• The deployment of AI also adds to this footprint. Every time users interact with a chatbot or a similar service, servers must process those requests, increasing energy demands over time, especially as usage scales up.

2. Data Centers and Energy Efficiency

• Data centers, where these models are hosted, are massive consumers of electricity. Historically, these centers relied heavily on fossil fuels, but there’s a growing shift toward using renewable energy sources. Major tech companies, including those behind AI, are pledging to make their data centers carbon neutral. For instance, Microsoft, Google, and Amazon are heavily investing in renewables and are working on making their infrastructure more sustainable.
• Innovations in cooling technology and energy-efficient hardware are also reducing the overall environmental impact. However, these advancements may not completely offset the growing demand for data centers driven by AI and cloud computing services.

3. Comparative Impact of AI

• AI technologies like ChatGPT can contribute to solutions that mitigate climate change. AI is being used for optimizing energy grids, improving logistics and transportation to reduce emissions, and aiding scientific research in areas like renewable energy development and climate modeling. So while it contributes to the problem, it can also be part of the solution.
• It’s also important to compare AI’s environmental cost with the benefits it offers. If AI tools can improve efficiency across industries, reduce human travel through remote working tools, or help discover cleaner technologies, the net effect could potentially be positive.

4. Wasteful vs. Purposeful Use

• The concern raised about “wasteful technology” is valid in some cases, especially when AI is used frivolously or inefficiently. However, like most technologies, its environmental footprint depends heavily on how it’s used. If AI is primarily used for entertainment, advertising, or low-impact tasks, the environmental cost may outweigh the benefits. But when AI is leveraged for high-impact, socially beneficial purposes, it can justify its energy consumption.

5. Broader Context of Climate Crisis

• The statement that AI could accelerate climate catastrophe is somewhat dramatic, but the core concern is legitimate: as we expand our digital infrastructure, we need to carefully manage its energy use. However, AI alone isn’t likely to “kill humanity”—it’s just one piece of a broader system of unsustainable practices. Global industries like agriculture, energy, transportation, and deforestation are far larger contributors to climate change than AI technologies.

Conclusion

While the criticism is not baseless, it frames the issue in extreme terms. AI technologies like ChatGPT do have a carbon footprint, but they are also tools that can be used to fight the very climate problems they might contribute to. The real challenge lies in ensuring the development and deployment of AI becomes more energy-efficient and that the benefits AI brings are significant enough to outweigh its environmental costs."

[davidculley]

First of all, respectfully, I’m flabbergasted that someone would respond to my human-written(!) post that took me over an hour to formulate even though I already had a pretty good idea in mind of what I wanted to say, with a cheap reply generated in less than 5 seconds with the very tool that I criticize. The effort spent is quite asymmetric: I spend a lot of effort to raise some valid concerns. Someone dumps some cheap garbage onto me that took no effort at all to write. I, again, have to spend lots of effort to debunk that trash. Talk about societal costs … The spread of misinformation is becoming too big a problem because it’s so much easier to spread misinformation than to fight it.

But I won’t let that bullshit stand for itself, lest someone gets the impression that this AI-generated text has any legitimacy.

Just as a heads-up: When I say, “this is garbage/bullshit”, I’m saying that what ChatGPT produced is garbage. I’m not saying that your opinions or what you say is garbage. That would be disrespectful, and you’re of course allowed to have your own opinions, even if they differ from mine. Please keep that in mind: I’m talking trash about ChatGPT, not you.

but there’s a growing shift toward using renewable energy sources. Major tech companies, including those behind AI, are pledging to make their data centers carbon-neutral. For instance, Microsoft, Google, and Amazon are heavily investing in renewables and are working on making their infrastructure more sustainable.

That is complete bullshit and exactly what these companies want you to believe. It’s the exact opposite of what is actually the case. In other words, this is propaganda. Just look at what Eric Schmidt (former CEO of Google) recently said: “We won’t reach our climate goals anyway, so why even bother trying? Let’s abandon our climate goals and go all-in on ChatGPT instead.” Do I need to search the sources for you to believe me?

Innovations in cooling technology and energy-efficient hardware are also reducing the overall environmental impact.

That’s also complete bullshit. It’s true that companies tried some innovations in cooling technology, like Microsoft building its data centers underwater at the ground of the ocean, but they quickly abandoned these projects because they weren’t sustainable. So that’s bullshit again.

AI technologies like ChatGPT can contribute to solutions that mitigate climate change.

No, they cannot. This confuses AI with ChatGPT, despite me having explicitly said, “Don’t confuse AI with ChatGPT”. So much for intelligence or actually understanding the meaning of the text these LLMs produce. There are indeed some AI tools that can contribute to such solutions, but ChatGPT is not one of them. ChatGPT absolutely does not mitigate climate change. This is again plausible-looking but ultimately wrong horseshit.

reduce human travel through remote working tools

What the hell has that to do with ChatGPT? I myself am working 100% from home as a software engineer through remote working tools, but that has absolutely nothing to do with ChatGPT and is completely unrelated. This is again utter bullshit.

the net effect could potentially be positive.

The problem with “net positive” is the word “net”. Weighing everything positive and negative he did, Elon Musk might also be a “net positive”. His companies advanced electric cars, space travel, etc. but he’s still a fascist abuser and Trump supporter. Should we let him sterilize transgender people just because he’s a “net positive”? I don’t think so. Same with AI or ChatGPT: I don’t care if they “could” “potentially” be net positives, because even if they turn out to be, we should still stop using ChatGPT.

The concern raised about “wasteful technology” is valid in some cases, especially when AI is used frivolously or inefficiently.

This again confuses LLMs with AI, despite me having explicitly cautioned against doing that. And ChatGPT can/will never be efficient. What does it not understand about “scanning the entire internet and then some, and still not knowing how many times the letter ‘r’ occurs in the word ‘strawberry’”? ChatGPT already ingested all data there currently is, what more data does it want to ingest to become more efficient? And my concern raised is not just valid in “some cases” but, regarding ChatGPT, in all cases. So there’s again some utter piece of garbage that undermines valid concerns and spreads doubt and propaganda.

The statement that AI could accelerate climate catastrophe is somewhat dramatic

“Somewhat”? We’re all going to die. Well, not all of us. Millionaires and billionares are already building bunkers for the apocalypse because the very people who push this technology onto us are very keen on surviving when the starving people dying of thirst will storm their villas where they’ll hoard the remnants of drinking water on Earth. For example, in Germany, private investment funds are already buying concentration camps from the Third Reich Nazi period to turn these bunkers where Hitler hid while Germany was bombarded into luxurious bunkers for the ruling class once the shit hits the fan.

However, AI alone isn’t likely to “kill humanity”—it’s just one piece of a broader system of unsustainable practices.

What an utter piece of garbage. So you expect you can stack unsustainable practice onto unsustainable practice onto unsustainable practice onto … and expect nothing will come of it? This is basically denying climate change. You can’t expect to stack unsustainable practices indefinitely without serious consequences to come of it.

they are also tools that can be used to fight the very climate problems they might contribute to.

Again, completely wrong, despite looking like it was written by a human.

The real challenge lies in ensuring the development and deployment of AI becomes more energy-efficient and that the benefits AI brings are significant enough to outweigh its environmental costs."

This makes me furious. We already know how to stop climate change. We know it since at least the 1970s, when oil companies predicted with astonishing accuracy what would happen. Floods, more powerful hurricanes, wildfires, etc. Has anyone read the news these past weeks and months? Does anything sound familiar? We know it since the late 18xx when it was first discovered that climate would change due to the industrial revolution. We already know what we need to do. We don’t need AI or ChatGPT for that.

I will not deign to pull apart the rest of the cheap garbage written by ChatGPT as I think I sufficiently argued what garbage that chatbot generated. I’d appreciate it if future commentators could refrain from replying to a thoughtfully written text with carelessly spewed out propaganda generated by GenAI. If you do want to take part in this discussion, I’d love to hear your thoughts, but please at the very least spend the effort to write your reply yourself.

[davidculley]

@Peter1949

No, it was written by a descendent of Ned Lud [sic].

Also @Peter1949, this is the second time you called me a Luddite, which I consider an insult. (A Luddite was a worker in the 18xx’s who feared technology, therefore destroyed it, and wanted to go back to a time before technology; basically a terrorist and irrational ultraconservative.) Ironic to call a software engineer and computer scientist a Luddite. I don’t want to go back to a time before technology and sabotage every machine I see. I just point out the problems with this one particular technology and ask for using technology in a reasonable manner.

So after being insulted twice, I think I’m done helping @Peter1949 learn German or answering his questions about German sentences.

[Judith42]

I like the Explain feature. I find it useful, and accurate enough for my needs. I’ve not knowingly found errors there, and it has definitely helped me understand sentence structure, cases and grammar better. It’s the feature that made me choose to pay for Clozemaster.

If you really want to reduce your climate impact, do what I do:

1. Never get on an aircraft (all flights, even local ones, have a very high climate impact)
2. Never take a cruise holiday. High pollution and high carbon.
3. Give up eating meat (unless it would otherwise go to waster eg. Un eaten sausages at the end of a friend’s BBQ)
4. Don’t buy new clothes - wear what you have for a long time, repair if possible. Buy from second-hand shops whenever you can. If you really have to buy new (may be the case for decent shoes, but I’ve had some great ones from charity shops when I’ve been lucky), then buy good quality ones that will last.
5. Buy everything second-hand if you can. eg phones.
6. IF you have things you no longer need, then donate to charity shops, or give away via local Facebook community groups (this latter option, is probably the most efficient as there is less risk of it being thrown away)
7. Insulate your home to the best of your ability - especially the roof -people have very out-dated ideas on how effective their loft insulation is.
8. Install a heat pump (had mine several years now, and am very pleased with it)
9. Make your next car electric (preferably second-hand)
10. Ask for donations to environmental charities for Xmas and birthdays. (I have that arrangement with several relatives, and I donate to their favourite charities in turn)
11. Maintain any garden you have, as having actual plants in it, and don’t cover it in tarmac/gravel/etc.

Once you’ve done all that, you can shout about Chat GPT - the explanations on Clozemaster are saved once they are generated, so the emissions are relatively low.

[davidculley]

I appreciate your feedback, Judith. Genuinely.

I already do (almost) all items on your list:

• I’m a vegan, so I never eat meat. I also care for chickens and honeybees.
• I don’t own a car, not even an electric one. I travel exclusively by train, public transport, and bicycle.
• I work exclusively from home, therefore never travel senselessly for work when a video call suffices.
• I never fly, and sacrifice more than most people. (I’d like to visit Japan someday, but you can’t feasibly reach Japan from Germany by train.) I don’t ask for recognition for doing that, it’s out of my own moral convictions.
• I donate to various charities.
• I don’t care about fashion, so it’s been years since I last bought clothes. I especially despise fast-fashion, including Shein and Temu. I bring my Merino-wool cloths to a local tailor to fix them when they have holes.
• I buy my computers and phones second-hand / refurbished.
• I’m not a home-owner and can’t convince my landlord to install a heat pump, so I’m guilty of heating with gas. That’s why I said almost.

Have I earned the right to criticize ChatGPT in your opinion? (Regardless of your answer, I think a person should be allowed to criticize ChatGPT even if that person is not perfect herself/himself.)

@Judith42

Once you’ve done all that, you can shout about Chat GPT

You have my highest respect if you do all these items on your list yourself. I think doing the items on your list is very laudable (otherwise I wouldn’t strive to do them myself).

Unfortunately, you underestimate the environmental costs of ChatGPT. You underestimate the sheer scale of it. You also make the mistake of shifting responsibility to individual citizens. I don’t blame you at all for that, please don’t get me wrong. It’s not your fault. Because shifting the responsibility to individuals is exactly the strategy the oil companies employ since the 1970s. It’s all the media talks about, so how could you know better? It’s not your fault. The fossil fuel industry has been feeding us their lies and propaganda that individuals can save the world and stop climate change if individuals only start recycling their plastic bottles or taking the bike to work instead of the car. While they themselves keep burning unimaginable amounts of oil and dumping toxic waste into our rivers.

It definitely doesn’t hurt if individuals do all these actions, but these are merely drops of water on the hot stone that evaporate into nothing.

Individuals can’t stop climate change if we don’t at the same time change the capitalistic foundation of our society and stop the fossil fuel industry from burning oil and gas.

Do you want to know who is one of the largest owners of OpenAI? The oil sheikhs from the United Arab Emirates. They really like it if western countries waste lots of electricity because that makes these countries dependent on oil to produce electricity, and the Arabs make profits from selling us that oil.

I don’t deny ChatGPT’s usefulness. I don’t deny that for Clozemaster’s needs, the accuracy of ChatGPT is high enough. After all, it’s not that hard to identify in a simple, short Spanish/Italian/German/… sentence what the subject or object is. Even if ChatGPT gets it wrong, then what of it? It’s not that a mistake of that sort would have any serious consequences. People wouldn’t start running around, religiously preaching about subjects and objects, or word orders, as they do if ChatGPT spreads misinformation about vaccines or other conspiracy theories.

What I’m talking about are the costs. I’m saying the usefulness of ChatGPT is not worth its costs.

I know you meant well, and from your 11 points I know you’re a good person, but unfortunately you miss the point. Again, it’s not your fault. How could you know if the media deliberately keeps us in the dark because they’re all in the pockets of the fossil fuel industry?

As Clozemaster users, we can’t stop the fossil fuel industry, but we certainly can stop using ChatGPT.

[Judith42]

First off, I’m genuinely impressed that you share my beliefs and act by them. I suspect I’d like you a lot if I met you in the flesh. (I do do all the things on my list, but not all of them are available to people who don’t own their own homes, and I’m also fortunate that I live in any area with a lot of very good charity shops, and can also pick up items (and donate them) via a local Facebook group where people list unwanted stuff)
I miss my friends in other countries that I’ll never see again, as I used to go to a lot of overseas SF conventions. But the last year I flew was 2002.

I hate Chat CPT in most uses. I find it infuriating that Google now always generates an unasked GPT reply to any of my search queries. As my queries are generally very specialist, it is a complete waste of time and I never click on it as I know it won’t have considered the sources I’m seeking. If you know an option to stop it offering them, I will gladly switch it off.

The main different with Clozemaster, is that the query is only done once. Thereafter, the reply is recorded, and everyone else who clicks ‘explain’ for a particular sentence will NOT trigger a new CHAT GPT search.

I’m all too aware of the environmental cost of AI. It’s driving up energy consumption, and I would happy ban it. But the use on Clozemaster really is a drop in the ocean (because of the ‘use once’ factor. If it had to be regenerated every time, I would agree with you).

Like you, I get infuriated by people who want to ban plastic wrap on food, and then take a cruise to Antarctica for a month… Plastic wrap on food is complex, as it increases the shelf life and hence reduces food waste. And reduced food waste saves a lot of emissions - these things are complex and finding the balance…

We need an end to the massive fossil fuel subsidies around the globe.
We need carbon taxes (that are used to support those on the lowest incomes).
We need an end to the tax-free status of aviation fuel.
The list is endless, but a reformation of the global tax system just might give my granddaughter a chance of a liveable planet when she grows up…

One of my favourite charities is Client Earth - they provide legal support to people suing oil companies and other acts of environmental vandalism. My sister has just donated to them for my birthday (she’ll be getting support for her local hospice for hers)

My pet hate is school ski trips. The damn glaciers are melting, and what example are they setting their kids? It’s okay to fly…

[Spielratte]

@davidculley

So after being insulted twice, I think I’m done helping you learn German or answering your questions about German sentences.

Please don’t stop. You’re not going to convince the already-convinced about AI, but you can continue to help build and maintain a user forum where actual people answer questions that other actual people have, a forum I personally have used to great benefit since I abandoned Duolingo for Closemaster (the final straw being Duo abandoning its forum).
Thank you for your detailed look at ChatGPT and the broader context it’s involved in. Maybe you could write something about using the comments button–make it a thread, get people talking about how to use one’s intelligence and curiosity to find answers instead of having them served up by AI. You and all the others who have taken the time to reply to my questions are valued.

[Hyacinthe]

So after being insulted twice, I think I’m done helping you learn German or answering your questions about German sentences.

You are free to do whatever you find appropriate, although this sounds a bit like you’re determined to punish thousands of other users, because one user said something you found insulting. The only result of this will be that those thousands of users will now have no other option than to use the AI Explain function, because your explanations won’t be there anymore. So in my opinion you’re throwing the baby with the bathwater, but again, it is your right to do so. Personally I think you would achieve your goal much more if you kept answering users’ questions, adding a warning about the AI being less than useful in the same post.

[davidculley]

We’re starting to get a bit off-topic. I don’t want the discussion to be about whether I continue answering the questions of users who insult me.

I want to show @mike that enough (paying) users would understand and even appreciate if @mike and team removed the Explain feature.

[Hyacinthe]

Indeed, this is getting a bit off-topic, because now you’re talking about users who insulted you and about not answering their questions, which is not what I understood from your previous posts. Sorry, my bad!

About having enough paying users who would understand and even appreciate your proposal, do you have any statistics to support your claim? In other words, if @mike opened a new topic on the forum and asked everyone if they want this feature to be removed, how confident are you that the majority would happily agree with your suggestion?

[mike]

Thanks for raising these issues as well as the discussion! We will continue to consider removing the explanations with respect to their accuracy and utility as well environmental and ethical concerns. For now, however, we will likely keep them.

Most users seem to like the explanations and benefit from them despite the shortcomings like you described. As far as environmental concerns, our usage of ChatGPT is quite minimal, and we will continue to minimize our use of it like @Judith42 mentioned.

Happy to revisit and reconsider as things change.

[Floria7]

Please please leave Explanation - I use it regularly for quick reference, and it’s been accurate so far re the Italian course. It’s optional anyway.

[seveer]

The Luddites are much maligned and the word’s unfortunate use as an epithet obscures a rather more nuanced history than the modern caricature of their positions suggests. Like most slurs, it is used to invalidate reasonable arguments by ignoring their actual content. This recent book on the subject may interest you:

Blood in the Machine: The Origins of the Rebellion Against Big Tech

[Diverse_Melody]

I find the explanation feature extremely helpful for hebrew, despite its faults. The explain feature is one thing i really love about clozemaster. It would be really disappointing if it was removed.

[davidculley]

@seveer

Like most slurs, it is used to invalidate reasonable arguments by ignoring their actual content.

I agree with you and thanks for the recommendation of Brian Merchant’s book. I heard Brian on the podcast “Tech won’t save us”. It’s unfortunate that Brian is a columnist for the LA Times (correction: he no longer is, which is good), owned by billionaire Patrick Soon-Shiong who forbade the newspaper staff to stand up against fascism in the recent US presidential election because, well, capitalists turn to fascism when they feel they might lose even just the tiniest bit of their wealth.

I know the true meaning of the word Luddite and their history, and consider myself a Luddite in the true sense of the word. Many people, however, have no idea regarding the history of Luddites and capitalism (including the person who called me a Luddite), and simply believe the lies they are told by the ruling class, and therefore adopt the twisted meaning of the word Luddite, which serves, like you say, to discredit them and ignore their actual content.

Maybe, if I find the time when I’m back on my desktop computer, I’ll elaborate. But now I’m on my mobile device where it’s inconvenient to type longer texts.

Since the person who insulted me obviously fell for the lies and meant the word Luddite as an insult, I didn’t see the use to explain how the meaning of the word Luddite changed after the Luddites lost the fight against the capitalist ruling class. History is written by the winners, after all. The person who insulted me simply would’ve ignored the history lesson, same as he ignores the actual content, so I didn’t see a need to waste my time arguing with strangers on the internet.

[seveer]

I assumed as much, no need to defend yourself. I really just wanted to plug that book for any of those reading who wish to inform themselves. Cheers.

[davidculley]

I want to tell a short story to all the people who say they find the Explain feature useful and would be disappointed if it were removed.

Before the Americas were discovered in the Middle Ages, sugar was exorbitantly expensive. No regular person could afford it, only kings and lords and dukes and such. The sugar canes of which sugar is made require a tropical climate, so they were grown only in India and were called the “reeds that produce honey without bees”. It was extremely labour intensive to make sugar out of these canes, and farmers had to break their backs in a hellish heat.

The sugar was imported from India for prohibitively high prices and was used by the ruling class as a “secret ingredient” to improve the taste of disgusting bitter medicine.

When the Americas were discovered, the Europeans learned that there was another place on earth, besides India, where the climate was right to plant sugar canes. Of course, it was still labour intensive and therefore expensive. But by moving to slave labour, the plant owners could drastically lower the price of the sugar that was exported to Europe.

Thanks to slavery, regular people could now finally also enjoy the sweet, sweet sugar. Regular people started to sweeten their tea, to produce bon-bons, et cetera. But working on the sugar plants, in the tropical climate, riddled with Malaria-carrying mosquitos, as a slave, was still hell. Millions of African-Americans died: of exhaustion, of diseases, of whiplashes and other “punishments” by the cruel slave owners. I hope I have to tell nobody that slavery is one of the worst atrocities committed by humanity.

I am very sorry if I offend anyone, but when I read some of the answers in this forum, all I read is, “I like sugar. I really like how my tea tastes after I sweetened it with sugar. I would be very disappointed if the sugar were taken away from me.”

Just to be very clear, before anyone comes after me, I’m not accusing you of slavery or any such thing. The horrors of slavery on these sugar plants didn’t happen because the Europeans hated the slaves slaving away on the sugar plants in some distant land. These horrors happened because the Europeans didn’t care about the slaves. They didn’t even think about them. They only ever cared about how much better their tea tasted with some sugar in it. They only ever cared about their own experience.

So many answers in this community remind me of this. They only stress how ChatGPT helps them with their target language, and completely neglect the other sides of the issue.

Let me ask you all a question: Is the difference between having and not having some sugar in your f***ing tea worth the unimaginatively large suffering that is very purposefully kept away from you becoming aware of it? And then tell me again you’d be disappointed if the sugar exports and exploitation of slaves were stopped after people finally found their moral conscience.

[Floria7]

Sorry everyone but for me, this is far too political for our friendly forum.

[mike]

I appreciate the passion here! We’re going to keep the explanations for now, and will revisit as the technology and our usage of it changes.