top 39 comments

sorted by: hot top controversial new old
[–] 53 points 1 year ago
[–] 65 points 1 year ago (4 children)

CEOs are SO INTELLIGENT! I would NEVER have Thought to invest BILLIONS OF DOLLARS on Chatbots and Summarizers which ALREADY existed!

  • source
  • hideshow 4 child comments
  • [+] -53 points 1 year ago (3 children)

    Trying not to be too douchey here, but ironically, your message is actually a very good example of where this technology could be beneficial.

    IT is ACTUALLY not EASY to read a MESSAGE when THE CASE randomly SWITCHES back AND forth.

  • source
  • parent
  • hideshow 3 child comments
  • load more comments (1 reply)
  • [–] 22 points 1 year ago

    "Caden, it looks like Airlynn just said you're a hopeless loser, and she's been banging your personal trainer Chad. Is there anything else I can help you with?"

  • source
  • [–] 22 points 1 year ago (2 children)
  • [–] 20 points 1 year ago

    They're interjecting themselves between us and our contacts. They'll have the power to "summarize", which is also the power to subtly re-interpret meaning. The AI will be given a broad goal, and it will chip away at that goal bit by bit, across millions of summaries to mold public opinion.

  • source
  • [–] 6 points 1 year ago

    I get it, big stupid money hungry tech wants to put itself between you and every other person on earth for the almighty dollar.

    How about we use tech to fix problems not create new ones!?

  • source
  • [–] 8 points 1 year ago (4 children)

    To be fair, my father tends to make messages quite incomprehensible by adding irrelevant information all over the place. Sometimes going on for multiple screens while it could easily have been a 2-3 sentence message.

    Sadly I think AI would even be worse at picking up what information is important from that. But I understand why people want it.

    As for very active groupchats, I am not gonna read +25 messages a day, but being able to glance the gist of it would be awesome.

  • source
  • hideshow 4 child comments
  • [–] 8 points 1 year ago (3 children)

    The exact point at which the gist of it can be manipulated, leaving out context and nudging you toward a different opinion than you might have formed if you'd read the whole thread.

  • source
  • parent
  • hideshow 3 child comments
  • [–] -2 points 1 year ago (2 children)

    Friend, I think you need to reconsider your world perspective a bit. Not everyone is out to get you all the time.

  • source
  • parent
  • hideshow 2 child comments
  • [–] 7 points 1 year ago

    Tinder needs this function right now!

  • source
  • [–] 1 point 1 year ago (14 children)

    I don't use WhatsApp, but this immediately made me think of my dad who doesn't use any punctuation and frequently skips and misspells words. His messages are often very difficult to interpret, through no fault of his own (dyslexia).

    Having an LLM do this for me would help both him and me.

    He won't feel self conscious when I send a, "What you talkin' about Willis?" message, and I won't have to waste a ridiculous amount of time trying to figure out what he was trying to say.

  • source
  • hideshow 14 child comments
  • [–] 40 points 1 year ago (12 children)

    If he’s not communicating in an explicit and clear way the AI can’t help you magically gain context. It will happily make up bullshit that sounds plausible though.

  • source
  • parent
  • hideshow 12 child comments
  • [+] -10 points 1 year ago (11 children)

    A poorly designed tool will do that, yes. An effective tool would do the same thing a person could do, except much quicker, and with greater success.

    An LLM could be trained on the way a specific person communicates over time, and can be designed to complete a forensic breakdown of misspelt words e.g. reviewing the positioning of words with nearby letters in the keyboard, or identifying words that have different spellings but may be similar phonetically.

  • source
  • parent
  • hideshow 11 child comments
  • [–] 12 points 1 year ago (1 child)

    the same thing a person could do

    asking for clarification seems like a reasonable thing to do in a conversation.

    A tool is not about to do that because it would feel weird and creepy for it to just take over the conversation.

  • source
  • parent
  • hideshow 1 child comment
  • [–] 1 point 1 year ago*

    The intent isn’t for the LLM to respond for you, it’s just to interpret a message and offer suggestions on what a message means or rewrite it to be clear (while still displaying the original).

  • source
  • parent
  • [–] 9 points 1 year ago* (last edited 1 year ago) (8 children)

    An LLM could be trained on the way a specific person communicates over time

    Are there any companies doing anything similar to this? From what I've seen companies avoid this stuff like the plague, their LLMs are always frozen with no custom training. Training takes a lot of compute, but also has huge risks of the LLM going off the rails and saying bad things that could even get the company into trouble or get bad publicity. Also the disk space per customer, and loading times of individual models.

    The only hope for your use case is that the LLM has a large enough context window to look at previous examples from your chat and use those for each request, but that isn't the same thing as training.

  • source
  • parent
  • hideshow 8 child comments
  • [–] 1 point 1 year ago (1 child)

    My friend works for a startup that does exactly that - trains AIs on conversations and responses from a specific person (some business higher-ups) for purposes of "coaching" and "mentoring". I don't know how well it works.

  • source
  • parent
  • hideshow 1 child comment
  • [–] 1 point 1 year ago (5 children)

    There are plenty of people and organisations doing stuff like this, there are plenty of examples on HuggingFace, though typically it's to get an LLM to communicate in a specific manner (e.g. this one trained on Lovecraft's works). People drastically overestimate the amount of compute time/resources training and running an LLM takes; do you think Microsoft could force their AI on every single Windows computer if it was as challenging as you imply? Also, you do not need to start from scratch. Get a model that's already robust and developed and fine tune it with additional training data, or for a hack job, just merge a LoRA into the base model.

    The intent, by the way, isn't for the LLM to respond for you, it's just to interpret a message and offer suggestions on what a message means or rewrite it to be clear (while still displaying the original).

  • source
  • parent
  • hideshow 5 child comments
  • [–] 2 points 1 year ago* (4 children)

    Huggingface isn't customer-facing, it's developer-facing. Letting customers retrain your LLM sounds like a bad idea for a company like Meta or Microsoft, it's too risky and could make them look bad. Retraining an LLM for Lovecraft is a totally different scale than retraining an LLM for hundreds of millions of individual customers.

    do you think Microsoft could force their AI on every single Windows computer if it was as challenging as you imply?

    It's a cloned image, not unique per computer

  • source
  • parent
  • hideshow 4 child comments
  • [–] 1 point 1 year ago* (3 children)

    Hugging Face being developer-facing is completely irrelevant considering the question you asked was whether I was aware of any companies doing anything like this.

    Your concern that companies like Meta and Microsoft are too scared to let users retrain their models is also irrelevant considering both of these companies have already released models so that anyone can retrain or checkpoint merge them i.e. Llama by Meta and Phi by Microsoft.

    It’s a cloned image, not unique per computer

    Microsoft's Copilot works off a base model, yes, but just an example that LLMs aren't as CPU intensive as made out to be. Further automated finetuning isn't out of the realm of possibility either and I fully expect Microsoft to do this in the future.

  • source
  • parent
  • hideshow 3 child comments
  • [–] 2 points 1 year ago* (2 children)

    Your concern that companies like Meta and Microsoft are too scared to let users retrain their models is also irrelevant considering both of these companies have already released models so that anyone can retrain or checkpoint merge them i.e. Llama by Meta and Phi by Microsoft.

    they release them to developers, not automatically retrain them unsupervised in their actual products and put them in the faces of customers to share screenshots of the AI's failures on social media and give it a bad name

  • source
  • parent
  • hideshow 2 child comments
  • [–] 1 point 1 year ago (1 child)

    They release them under permissive licences so that anyone can do that.

  • source
  • parent
  • hideshow 1 child comment
  • [–] 1 point 1 year ago*

    yea someone could take the model and make their own product with their own PR and public perception

    that's very different from directly spoonfeeding it as a product to the general public consumers inside of WhatsApp or something

    it's like saying someone can mod Skyrim to put nude characters in it, that's very different from Bethesda selling the game with nude characters

  • source
  • parent
  • load more comments (1 reply)
    [–] -3 points 1 year ago

    It does make sense in big groups with tons of irrelevant discussion but also few messages you actually need to read.

  • source
  • [+] -10 points 1 year ago* (5 children)

    Ahh, intellect of the average American on display. So... -óóóh, is that a Donut?

  • source
  • hideshow 5 child comments
  • [–] 10 points 1 year ago* (4 children)

    People in the US don't use WhatsApp, for the most part.

  • source
  • parent
  • hideshow 4 child comments
  • [–] -2 points 1 year ago

    I have one specifically that is mostly bursts of 2 or 3 hours of chat between whoever is online in the group and with some worthwhile coordination messages mixed in at random. I don't want to read 80 messages about mortgage rates and VTI stocks to find a couple of lines I'm actually interested in about kid plans for the evening or something I'd actually care to talk about.

  • source
  • 14 comments in c/technology@lemmy.zip view that post
    [–] 1 point 1 year ago

    Uh hunh. "Friends & family", surrre. 🤣😶

  • source
  • [–] 15 points 1 year ago (4 children)

    Here's a list of reasons why you should consider moving to Signal , if you haven't already:

    • End-to-End Encryption (E2EE) by Default : Signal uses the Signal Protocol, which is considered the gold standard for E2EE. This means that all your messages, calls (voice and video), and file sharing are encrypted on your device before they leave and can only be decrypted by the intended recipient's device. Not even Signal can read your communications. This is a crucial differentiator from many other apps (like Telegram, which doesn't encrypt all chats by default, or WhatsApp, which collects metadata).
    • Zero Data Collection (No Metadata) : Signal is designed to collect as little user data as possible. They don't store information about who you talk to, when you talk, your contacts, group memberships, or even your profile ID. This means even if subpoenaed, Signal has almost no user data to provide, in fact all they can provide is the date + time you registered with the service and the last day (not time) one of your clients pinged their servers. Other apps, including WhatsApp, collect significant metadata, which can reveal patterns about your communication even if the content is encrypted.
    • Open Source and Audited : Signal's code is completely open-source, allowing security experts and the public to inspect it for vulnerabilities and ensure its integrity. This transparency fosters trust and makes it highly resistant to hidden backdoors or malicious features.
    • Non-Profit Organization : Signal is run by the Signal Foundation, a non-profit organization funded by donations. This means there are no commercial interests, no ads, and no pressure to sell user data for revenue, unlike many commercially driven messaging apps.
    • Phone Number Privacy (Usernames) : While Signal historically required a phone number for registration, it now offers the option to communicate via usernames, allowing you to connect with people without revealing your phone number.
    • Relay Calls : For voice and video calls, you can opt to "relay" your calls through Signal's servers, which helps hide your IP address from the other party.
    • Standard Messaging Features: Signal offers all the core features you'd expect from a modern messaging app, including text messaging, group chats, voice and video calls (both individual and group), file sharing, and voice messages.
    • Cross-Platform Availability : Signal is available on Android, iOS, Windows, macOS, and Linux.
    • User-Friendly Interface : Despite its strong focus on security, Signal maintains an intuitive experience, removing the barrier to entry for anyone who isn't technically inclined.
  • source
  • hideshow 4 child comments
  • [–] 3 points 1 year ago (2 children)

    From Signal's terms and conditions

    Other instances where Signal may need to share your data

    To meet any applicable law, regulation, legal process or enforceable governmental request.

    If i understand correctly, their servers are in the usa. So the usa government has the same level of access as compared to whatsapp? It's non profit now, but so was openai...

    WhatsApp is definitely taking a step in the wrong direction. However, switching to another app is difficult, it's hard to get people ingrained in an ecosystem switch once let alone twice....

  • source
  • parent
  • hideshow 2 child comments
  • [–] 8 points 1 year ago* (last edited 1 year ago)

    Any company operating in any country has to comply with their laws. The difference is that Signal has almost no data to share to comply. Message content and metadata is encrypted so they have no access to it. Your phone number is the only identifier they have and would be obligated by law to share.

  • source
  • parent
  • [–] 3 points 1 year ago

    their servers are in the usa.

    Their "home office" is in the US. That doesn't necessarily mean they don't have servers distributed globally.

    So the usa government has the same level of access as compared to whatsapp?

    No, the US government does not have the same level of access to Signal as they do with Whatsapp. The only reason the US has so much access to Whatsapp is because Whatsapp only bothered to implement End-to-end encryption (E2EE). Unfortunately, in 2205, E2EE is the bare minimum. E2EE via the Signal protocol has been a "solved issue" since 2013 and Whatsapp implemented it 3 years later (great!) but they have not improved privacy since. Whatsapp still collects a metric-fuck-ton of metadata like:

    • Who you communicate with
    • When you communicate
    • How long your calls last
    • The frequency of communication
    • When you're "active" on the platform
    • Group memberships and group titles
    • Your profile information (this is E2EE on Signal)
    • Your contacts get uploaded in a way that's visible to Whatsapp. Signal does contact discovery in a privacy preserving way.

    Then they correlate this data with everything else they have about you to "fill in the gaps". Signal doesn't collect any metadata.

    It's non profit now, but so was openai...

    The difference here is there's nothing of value for Signal to "sell" since they don't collect metadata and have engineered it to work without being able to see anything. The Signal server and client are already open source, there's no "secret sauce". Lastly, because they collect zero data they can't even sell it for ad-serving purposes. Who would buy Signal?

    switching to another app is difficult, it's hard to get people ingrained in an ecosystem switch once let alone twice

    100% agree. The best way I've found is to drop the offending platform (whatsapp) and move to Signal. Let others know you accept text/SMS or Signal messages. Over the years the people on Signal (at least in my group) has steadily grown.

    I would like to close by saying that Signal is not shy about complying with the law, they will not go to prison for anyone's potential crimes. That said, they publish the data they provide when compelled by law and the only data they collect is the day + time you signed up with their service and the last day (not time) one of your clients pinged their servers, source: https://signal.org/bigbrother/

  • source
  • parent
  • [–] 6 points 1 year ago (4 children)

    If the AI could help me decipher what the emoji mean that would be legit helpful

  • source
  • hideshow 4 child comments
  • [–] 3 points 1 year ago* (last edited 1 year ago) (3 children)

    ✌️

    ✝️

    4️⃣

    🧑‍🦲

    ——

    Edit: Too Catholic For Head

  • source
  • parent
  • hideshow 3 child comments
  • [–] 5 points 1 year ago*

    I think there is an entire meme genre about Apple AI notification summaries and the sometimes hilarious results.

    https://www.theverge.com/2024/11/12/24289939/apple-intelligence-ai-notification-summaries-awkward-funny-bad

    This will be about the same.

  • source
  • [–] 4 points 1 year ago

    Most messages are so short I don't see why you'd need a summary. Are people that poor at reading that like 10 sentences is a struggle? If so, education seems like what we should spend millions of dollars on.

  • source
  • So is this the reason that I see massive full page ads saying that WhatsApp is secure? They are worried that introducing ai will make everyone worried it can read all the messages? Or is it that just that they now have a way to read your messages?

  • source
  • 2 comments in c/theonion@midwest.social view that post
    [–] 13 points 1 year ago
    [–] 2 points 1 year ago*

    I personally wouldn't mind AI on my local device that would work with apps that are compatible. I don't understand why every company has to shove their ai into their software. It doesn't make sense how it's profitable for them. Good AI use cases I can think of rn are translation, transliteration, maybe summarise the topic and intelligent operations like parsing times or days into easily bookmarkable dates(like add to calendar button) etc.

    Idk why they moved away from the google assistant type features to this slop. It makes no sense at all. At least we saw two or three good features in the recent apple presentation.

    Edit: shit this is not real :O should i delete my msg :-\

  • source