Skip to content

Duolingo CEO tries to walk back AI-first comments, fails

Technology
134 90 549
  • I hear good things about Pimsleur as an alternative.

    It can go either way, some people like the method, others hâte it because it's not gamified. Pro tip, get pimsleur courses from your library if you want to try them for a real trial rather than what they give you

    • Following backlash to statements that Duolingo will be AI-first, threatening jobs in the process, CEO Luis von Ahn has tried to walk back his statement.
    • Unfortunately, the CEO doesn’t walk back any of the key points he originally outlined, choosing instead to try, and fail to placate the maddening crowd.
    • Unfortunately the PR team may soon be replaced by AI as this latest statement has done anything but instil confidence in the firm’s users.

    How do these people become CEOs they're as thick as several short planks nailed together.

    Firstly every single company that has tried to replace its employees with AI has always ended up having issues. Secondly even if that wasn't the case, people are not going to be happy about it so it's not something you should brag publicly about.

    If you're going to replace all of your employees with AI just do it quietly, that way if it fails it's not a public failure, and if it succeeds (it won't) then you talk about it.

  • Check out "Language Drops" and "Rosetta Stone" if you're looking for replacements. They both have very different approaches to language learning (both from each other and from Duolingo), but their content is at the very least much better curated than Duolingo's.

    I haven't gone out of my way to check but AFAIK neither of them is jumping on the AI-before-anything-else train.

    Drops had an ai feature where it would show you a “fact” at the end of each session, which was often completely wrong etymology.

  • AI is social cancer

    It's a lie told by marketing companies that have gaslit artists into automating their creativity and gaslit governments into automating fascism

    Automated fascism completely defeats the purpose of fascism. The whole point is to lord power over people, if a computer is going to do it automatically then it's no fun.

  • How do these people become CEOs they're as thick as several short planks nailed together.

    Firstly every single company that has tried to replace its employees with AI has always ended up having issues. Secondly even if that wasn't the case, people are not going to be happy about it so it's not something you should brag publicly about.

    If you're going to replace all of your employees with AI just do it quietly, that way if it fails it's not a public failure, and if it succeeds (it won't) then you talk about it.

    If they do it quietly, they won't get the stock price bump every company gets from saying they're going to replace (costly) employees with AI.

  • I'm a long time user of Duolingo and you earn plenty to give yourself the occasional streak freeze if you can't go two days without doing a lesson. It's not really as predatory as it sounds. It's nothing like pay to win type games.

    Fuck Duolingo for the AI shit though, don't mistake me for a Duolingo simp thinking their blameless. It's just that the monetization is not as predatory as it sounds.

    "Freezing" your streak is just silly, even if they offer it for free. Is this just for online clout, so you can brag (falsely) to others how long you haven't broken a streak?

    If an alcoholic goes 10 years without drinking, then has a beer, the streak is broken. Doesn't mean you can't recover and improve, but it is what it is. It's dishonest to pretend it didn't happen, especially if you're comparing yourself to others...

  • In another thread someone told me you can buy gems or something to keep your streak going.

    That would've made me uninstall long before his comments.

    It's worse than that.

    Yes, you can pay for a streak freeze. If you don't, you'll probably find that you were given one for free anyhow. You'd have wasted your gems.

    Yes, you can pay to undo a streak loss. It's more than paying for a freeze.

    It'll give you multiple chances to pay for all that, too. If you're out for days and then come back, you can pay to fix your streak.

    What is the point of a streak if you can just buy your way back to it?

    Also, I had paid for the last couple years, which (IIRC) includes free streak freezes. It still asked if I want to pay for them. I'd say no, and find I had one anyhow, or a friend had miraculously given me one.

    But during the last year (365 days) my streak was actually only at 190 or so because I'd used so many streak freezes that I got for free. I wasn't even trying to keep my streak.

    When I finally let my streak die, the icon started trying to guilt trip me into coming back with horrible icons of Duo being sad, heartbroken, or even dead.

    The constant mental manipulation that was well beyond what gamification should ever be was what finally drove me to just quit playing altogether. I had already canceled my sub long ago, but I'm not even going to use the remainder of this year I've already paid for.

  • The old tree system was much better, it allowed you to mix exercises from different topics. The new path system locks you into one topic until you know all the sentences by rote.

    They changed that a couple years ago, right? I think that was when I quit.

  • How do these people become CEOs they're as thick as several short planks nailed together.

    Firstly every single company that has tried to replace its employees with AI has always ended up having issues. Secondly even if that wasn't the case, people are not going to be happy about it so it's not something you should brag publicly about.

    If you're going to replace all of your employees with AI just do it quietly, that way if it fails it's not a public failure, and if it succeeds (it won't) then you talk about it.

    Intelligence has nothing to do with success. These people are born into wealth, are wealth-adjacent, or are expert ass-kissers.

  • "Freezing" your streak is just silly, even if they offer it for free. Is this just for online clout, so you can brag (falsely) to others how long you haven't broken a streak?

    If an alcoholic goes 10 years without drinking, then has a beer, the streak is broken. Doesn't mean you can't recover and improve, but it is what it is. It's dishonest to pretend it didn't happen, especially if you're comparing yourself to others...

    Really comparing missing a day of a language learning app to alcoholism recovery?

    Your streak doesn't go up on days you use a freeze.

  • “AI is creating uncertainty for all of us, and we can respond to this with fear or curiosity. I’ve always encouraged our team to embrace new technology (that’s why we originally built for mobile instead of desktop), and we are taking that same approach with AI. By understanding the capabilities and limitations of AI now, we can stay ahead of it and remain in control of our own product and our mission,” writes von Ahn.

    Now please explain in more detail how this advice should be followed, practically, by someone you just fired because AI was cheaper. Give examples of how they can "stay ahead of it" so as to "remain in control of the product and mission" they are no longer employed to work on. How should they "embrace" this transition and "respond with curiosity" to no being newly unable to afford food or rent? "Uncertainty for all of us" my ass.

    The former employees are now curious about how they will pay rent and eat, so there’s that.

  • What I'd wonder is why it's such massive expensive for Duolingo to hire 2 or 3 people to cover a language anyway. Presumably most of the work is contractual - hire somebody competent to produce a course, get somebody to say the lines, refine the course based on feed back and that's mostly it.

    Terminal MBA brain

    • Following backlash to statements that Duolingo will be AI-first, threatening jobs in the process, CEO Luis von Ahn has tried to walk back his statement.
    • Unfortunately, the CEO doesn’t walk back any of the key points he originally outlined, choosing instead to try, and fail to placate the maddening crowd.
    • Unfortunately the PR team may soon be replaced by AI as this latest statement has done anything but instil confidence in the firm’s users.

    There should be a federated system for blocking IP ranges that other server operators within a chain of trust have already identified as belonging to crawlers. A bit like fediseer.com, but possibly more decentralized.

    (Here’s another advantage of Markov chain maze generators like Nepenthes: Even when crawlers recognize that they have been served garbage and they delete it, one still has obtained highly reliable evidence that the requesting IPs are crawlers.)

    Also, whenever one is only partially confident in a classification of an IP range as a crawler, instead of blocking it outright one can serve proof-of-works tasks (à la Anubis) with a complexity proportional to that confidence. This could also be useful in order to keep crawlers somewhat in the dark about whether they’ve been put on a blacklist.

  • Really comparing missing a day of a language learning app to alcoholism recovery?

    Your streak doesn't go up on days you use a freeze.

    No? It was a comparison of the streak, not the subject of the streak. That was just an example. My point remains. Unless you can literally stop time, the streak died. It's okay that it did, but why pretend it didn't?

  • How do these people become CEOs they're as thick as several short planks nailed together.

    Firstly every single company that has tried to replace its employees with AI has always ended up having issues. Secondly even if that wasn't the case, people are not going to be happy about it so it's not something you should brag publicly about.

    If you're going to replace all of your employees with AI just do it quietly, that way if it fails it's not a public failure, and if it succeeds (it won't) then you talk about it.

    People who are smart in one or two domains often overestimate how smart they are in other domains. They develop a mental model, confirm it quickly, and never re-asses it.

    The issue with AI, is we're probably hitting our first real S curve with the current technology's performance but a lot of people who bet big are only see the exponential part and assuming there won't be a level off, or that the level of is far away.

    There is no Moore's law for AI.

  • No? It was a comparison of the streak, not the subject of the streak. That was just an example. My point remains. Unless you can literally stop time, the streak died. It's okay that it did, but why pretend it didn't?

    If you can't see why someone might have a different criteria for a streak in days without alcohol as a recovering addict and days in usage of a learning application I can't help you.

    • Following backlash to statements that Duolingo will be AI-first, threatening jobs in the process, CEO Luis von Ahn has tried to walk back his statement.
    • Unfortunately, the CEO doesn’t walk back any of the key points he originally outlined, choosing instead to try, and fail to placate the maddening crowd.
    • Unfortunately the PR team may soon be replaced by AI as this latest statement has done anything but instil confidence in the firm’s users.

  • How do these people become CEOs they're as thick as several short planks nailed together.

    Firstly every single company that has tried to replace its employees with AI has always ended up having issues. Secondly even if that wasn't the case, people are not going to be happy about it so it's not something you should brag publicly about.

    If you're going to replace all of your employees with AI just do it quietly, that way if it fails it's not a public failure, and if it succeeds (it won't) then you talk about it.

    How do these people become CEOs they’re as thick as several short planks nailed together.

    Being a CEO has absolutely nothing to do with intelligence, I guarantee you that Duolingo has employees who are far more intelligent than the CEO.

  • You're just pretending you are doing something useful with your life instead of just slaving away at your job enriching others.

    /s, obviously. What a wild take they have.

    You had me to the /s, NGL

  • What I'd wonder is why it's such massive expensive for Duolingo to hire 2 or 3 people to cover a language anyway. Presumably most of the work is contractual - hire somebody competent to produce a course, get somebody to say the lines, refine the course based on feed back and that's mostly it.

    I would assume it is more about time than money.
    It is a big investment making a whole functional system with llm (I have a hard time believing it is actually AI), it will cost a lot if done wrong (just like everything else). You can't just prompt "make a course in Spanish" and get anything good out of it and you still need ppl who can quality check the output. I could see them use it to mass produce sentences and stories in different levels (not the actual story) and voice recordings, but not actually anything creative and I would assume that is the goal. But we have seen too many shitty products to believe in anything with llm.

  • 258 Stimmen
    38 Beiträge
    217 Aufrufe
    F
    A whole article about how terrible this is, then towards the end they got clarification from Google and, surprise surprise, it was all an overreaction and they were fear mongering. “This update is good for users: they can now use Gemini to complete daily tasks on their mobile devices like send messages, initiate phone calls, and set timers while Gemini Apps Activity is turned off. With Gemini Apps Activity turned off, their Gemini chats are not being reviewed or used to improve our AI models. It’s just giving Gemini more local assistant abilities.
  • 17 Stimmen
    2 Beiträge
    26 Aufrufe
    T
    Yeah, sure. Like the police need extra help with racial profiling and "probable cause." Fuck this, and fuck the people who think this is a good idea. I'm sure the authoritarians in power right now will get right on those proposed "safeguards," right after they install backdoors into encryption, to which Only They Have The Key, to "protect" everyone from the scary "criminals."
  • Airbnb Hosting Assistants

    Technology technology
    1
    2
    0 Stimmen
    1 Beiträge
    15 Aufrufe
    Niemand hat geantwortet
  • A Tech-Backed Influencer Wants to Replace Teachers With AI

    Technology technology
    32
    1
    114 Stimmen
    32 Beiträge
    128 Aufrufe
    sturgist@lemmy.caS
    Heck yeah! Gotta watch that again, thank you kind stranger!
  • 24 Stimmen
    14 Beiträge
    86 Aufrufe
    S
    I think you're missing some key points. Any file hosting service, no matter what, will have to deal with CSAM as long as people are able to upload to it. No matter what. This is an inescapable fact of hosting and the internet in general. Because CSAM is so ubiquitous and constant, one can only do so much to moderate any services, whether they're a large corporation are someone with a server in their closet. All of the larger platforms like 'meta', google, etc., mostly outsource that moderation to workers in developing countries so they don't have to also provide mental health counselling, but that's another story. The reason they own their own hardware is because the hosting services can and will disable your account and take down your servers if there's even a whiff of CSAM. Since it's a constant threat, it's better to own your own hardware and host everything from your closet so you don't have to eat the downtime and wait for some poor bastard in Nigeria to look through your logs and reinstate your account (not sure how that works exactly though).
  • 219 Stimmen
    119 Beiträge
    384 Aufrufe
    L
    Okay, I'd be interested to hear what you think is wrong with this, because I'm pretty sure it's more or less correct. Some sources for you to help you understand these concepts a bit better: What DLSS is and how it works as a starter: https://en.wikipedia.org/wiki/Deep_Learning_Super_Sampling Issues with modern "optimization", including DLSS: https://www.youtube.com/watch?v=lJu_DgCHfx4 TAA comparisons (yes, biased, but accurate): https://old.reddit.com/r/FuckTAA/comments/1e7ozv0/rfucktaa_resource/
  • 21 Stimmen
    3 Beiträge
    28 Aufrufe
    B
    We have to do this ourselves in the government for every decommissioned server/appliance/end user device. We have to fill out paperwork for every single storage drive we destroy, and we can only destroy them using approved destruction tools (e.g. specific degaussers, drive shredders/crushers, etc). Appliances can be kind of a pain, though. It can be tricky sometimes finding all the writable memory in things like switches and routers. But, nothing is worse than storage arrays... destroying hundreds of drives is incredibly tedious.
  • 588 Stimmen
    77 Beiträge
    273 Aufrufe
    F
    When a Lemmy instance owner gets a legal request from a foreign countries government to take down content, after they’re done shitting themselves they’ll take the content down or they’ll have to implement a country wide block on that country, along with not allowing any citizens of that country to use their instance no matter where they are located. Block me, I don’t care. You’re just proving that you can’t handle the truth and being challenged with it.