3 ms·

This should not be considered an LLM issue. It's a user experience issue. Mozilla has made changes that happened by default before. Often I have had to find t

by Lerc 1y ago

This should not be considered an LLM issue. It's a user experience issue. Mozilla has made changes that happened by default before. Often I have had to find the setting to put it back to how I wanted. I remember when it moved the URL bar to the bottom. I don't think it is always an easy call to make. Tabs were a significant user experience improvement, but hiding it behind an opt-in would have limited it to people who knew about it. I use Firefox as main on desktop and mobile. I have noticed messages on upgrade pointing to LLM features. I haven't engaged with them an from thereon haven't noticed any change because of them. Saying there are reports of excessive memory or CPU use isn't terribly useful without references to those reports. One such report posted on HN was shown to have been unrelated to the LLM. Are there any reports actually showing degradation because of LLMs rather than post hoc ergo propter hoc?

themafia 1y ago
> This should not be considered an LLM issue. LLMs are garbage and they add nothing to the browsing experience. > Are there any reports actually showing degradation because of LLMs rather than post hoc ergo propter hoc? You can control this option with a setting. It seems like it would be really really easy to just test this. As a result I can't see any reason to doubt this by default or apply a legalistic evidentiary standard when considering it.
- celsoazevedo 1y ago
  .
  - reyqn 1y ago
    And is also not using an LLM. It's neural machine translation.
    - superkuh 1y ago
      NMT is a category containing both transformers and deep RNN. The Mozilla translation models are transformer LLM NMTs trained via Marian https://marian-nmt.github.io/ https://marian-nmt.github.io/ (ref: https://github.com/mozilla/translations/blob/main/docs/README.md#training-pipeline https://github.com/mozilla/translations/blob/main/docs/READM...)
      - reyqn 1y ago
        NMT doesn't "contain" tranformers and deep RNNs, it can use them. LLMs use a transformer architecture, not everything using a transformer architecture is an LLM. NMT can actually use an LLM, but that's not the case according to the documentation you linked, they use a parallel dataset to train their models.
        lostmsu 1y ago
        > they use a parallel dataset to train their models If you want to be pedantic you should look up the LLM definition.
        reyqn 1y ago
        Care to explain why?
- skybrian 1y ago
  It's easy to test, therefore there is no need to test it? I don't see how that follows. If you haven't tested it then you haven't tested it.
  - themafia 1y ago
    > therefore there is no need to test it? What evidence do you have that it wasn't tested? Or are you just making an assumption based on one article? Did you do any follow up searching? It might have led you to this: https://www.neowin.net/news/mozilla-under-fire-for-firefox-ai-bloat-that-blows-up-cpu-and-drains-battery/ https://www.neowin.net/news/mozilla-under-fire-for-firefox-a...
    - Lerc 1y ago
      Which says. Update (August 13, 2025, 03:57 GMT): While the community correctly identified a performance issue, their attribution of the cause was mistaken.
- gucci-on-fleek 1y ago
  > LLMs are garbage and they add nothing to the browsing experience. The builtin translation feature [0] is LLM-based [1], and that adds a ton to my browsing experience, since it's made web pages in other languages accessible to me. [0]: https://support.mozilla.org/en-US/kb/website-translation https://support.mozilla.org/en-US/kb/website-translation [1]: According to Wikipedia, "A large language model (LLM) is a language model trained with self-supervised machine learning on a vast amount of text, designed for natural language processing tasks" [2]. The translation code is transformer/RNN-based and trained on raw texts [3, 4], and translation definitely qualifies as a natural language processing task, meaning that the translation feature is LLM-based. [2]: https://en.wikipedia.org/wiki/Large_language_model https://en.wikipedia.org/wiki/Large_language_model [3]: https://www.youtube.com/watch?v=J06koBcfm5w https://www.youtube.com/watch?v=J06koBcfm5w [4]: https://marian-nmt.github.io/features/ https://marian-nmt.github.io/features/
- smcleod 1y ago
  That's a pretty big opinion made as a statement there. I use LLMs for many incredibly useful things daily, an example in the browser is summarisation and outlining.
- orbital-decay 1y ago
  Hold on... For starters, LLMs made Chinese, Spanish, French websites comprehensible to me. Even the best models like Claude still make some mistakes and miss some nuance, but they largely have genuine understanding of the cultural context and can adapt it. They're qualitatively better than classic machine translation models, and the mistakes will eventually be fixed. I can finally communicate in person's native language and be reasonably sure that most subtlety is preserved. This is something that actually breaks language barriers, defies the Babylon, connects people! If that is nothing to you, I don't know what else is. Deep search is also extremely useful, it saves me actual days and finds things and cool people I wouldn't be able to find in reasonable time otherwise, a lot of that in isolated communities like discord servers normal search has no insight into. Sure, you have to be aware about things left in the dark, but you also have to be attentive to this with normal search as well. Video summarization lets me extract the actual message from the pile of clickbait, marketing, and three-hour long interviews - all while pointing at the timecodes for me to verify. Now, the pressure of bullshit will inevitably shift to LLM training. It's only a matter of time. But that's another question, the tech itself works amazingly well, what the hell are you even talking about?
- laterium 1y ago
  > LLMs are garbage That's just your opinion man.
dathinab 1y ago
> Are there any reports actually showing degradation because of LLMs rather than post hoc ergo propter hoc? AFIK no nearly all AI features only do any compute if you explicitly ask them to do stuff the only exception is that when you create a tab group (and only then!) it will auto recommend a name. And while I don't find a very useful for me (and I use tab groups a ton) I also have observed that non technical users get confused with unnamed tab groups and the UI around them so I really understand why they do that. Also while it's "technically" a LLM it's tiny and fast to run and could also have been implemented with just embeddings at a slightly loss of quality (and again only runs if you create a tab group and not at any other time and not if you computer has less then 3GiB of RAM etc. etc.(yes FF disables AI features on very low RAM devices AFIK)) Put people hear AI think LLM get angry and stop thinking. Or considering that maybe a non negligible amount of people outside of their bubble want exactly that and might leave your browser if you don't supply that.

by Lerc 1y ago

themafia 1y ago

> This should not be considered an LLM issue. LLMs are garbage and they add nothing to the browsing experience. > Are there any reports actually showing degradation because of LLMs rather than post hoc ergo propter hoc? You can control this option with a setting. It seems like it would be really really easy to just test this. As a result I can't see any reason to doubt this by default or apply a legalistic evidentiary standard when considering it.

celsoazevedo 1y ago
.
- reyqn 1y ago
  And is also not using an LLM. It's neural machine translation.
  - superkuh 1y ago
    NMT is a category containing both transformers and deep RNN. The Mozilla translation models are transformer LLM NMTs trained via Marian https://marian-nmt.github.io/ https://marian-nmt.github.io/ (ref: https://github.com/mozilla/translations/blob/main/docs/README.md#training-pipeline https://github.com/mozilla/translations/blob/main/docs/READM...)
    - reyqn 1y ago
      NMT doesn't "contain" tranformers and deep RNNs, it can use them. LLMs use a transformer architecture, not everything using a transformer architecture is an LLM. NMT can actually use an LLM, but that's not the case according to the documentation you linked, they use a parallel dataset to train their models.
      - lostmsu 1y ago
        > they use a parallel dataset to train their models If you want to be pedantic you should look up the LLM definition.
        reyqn 1y ago
        Care to explain why?
skybrian 1y ago
It's easy to test, therefore there is no need to test it? I don't see how that follows. If you haven't tested it then you haven't tested it.
- themafia 1y ago
  > therefore there is no need to test it? What evidence do you have that it wasn't tested? Or are you just making an assumption based on one article? Did you do any follow up searching? It might have led you to this: https://www.neowin.net/news/mozilla-under-fire-for-firefox-ai-bloat-that-blows-up-cpu-and-drains-battery/ https://www.neowin.net/news/mozilla-under-fire-for-firefox-a...
  - Lerc 1y ago
    Which says. Update (August 13, 2025, 03:57 GMT): While the community correctly identified a performance issue, their attribution of the cause was mistaken.
gucci-on-fleek 1y ago
> LLMs are garbage and they add nothing to the browsing experience. The builtin translation feature [0] is LLM-based [1], and that adds a ton to my browsing experience, since it's made web pages in other languages accessible to me. [0]: https://support.mozilla.org/en-US/kb/website-translation https://support.mozilla.org/en-US/kb/website-translation [1]: According to Wikipedia, "A large language model (LLM) is a language model trained with self-supervised machine learning on a vast amount of text, designed for natural language processing tasks" [2]. The translation code is transformer/RNN-based and trained on raw texts [3, 4], and translation definitely qualifies as a natural language processing task, meaning that the translation feature is LLM-based. [2]: https://en.wikipedia.org/wiki/Large_language_model https://en.wikipedia.org/wiki/Large_language_model [3]: https://www.youtube.com/watch?v=J06koBcfm5w https://www.youtube.com/watch?v=J06koBcfm5w [4]: https://marian-nmt.github.io/features/ https://marian-nmt.github.io/features/
smcleod 1y ago
That's a pretty big opinion made as a statement there. I use LLMs for many incredibly useful things daily, an example in the browser is summarisation and outlining.
orbital-decay 1y ago
Hold on... For starters, LLMs made Chinese, Spanish, French websites comprehensible to me. Even the best models like Claude still make some mistakes and miss some nuance, but they largely have genuine understanding of the cultural context and can adapt it. They're qualitatively better than classic machine translation models, and the mistakes will eventually be fixed. I can finally communicate in person's native language and be reasonably sure that most subtlety is preserved. This is something that actually breaks language barriers, defies the Babylon, connects people! If that is nothing to you, I don't know what else is. Deep search is also extremely useful, it saves me actual days and finds things and cool people I wouldn't be able to find in reasonable time otherwise, a lot of that in isolated communities like discord servers normal search has no insight into. Sure, you have to be aware about things left in the dark, but you also have to be attentive to this with normal search as well. Video summarization lets me extract the actual message from the pile of clickbait, marketing, and three-hour long interviews - all while pointing at the timecodes for me to verify. Now, the pressure of bullshit will inevitably shift to LLM training. It's only a matter of time. But that's another question, the tech itself works amazingly well, what the hell are you even talking about?
laterium 1y ago
> LLMs are garbage That's just your opinion man.

dathinab 1y ago

> Are there any reports actually showing degradation because of LLMs rather than post hoc ergo propter hoc? AFIK no nearly all AI features only do any compute if you explicitly ask them to do stuff the only exception is that when you create a tab group (and only then!) it will auto recommend a name. And while I don't find a very useful for me (and I use tab groups a ton) I also have observed that non technical users get confused with unnamed tab groups and the UI around them so I really understand why they do that. Also while it's "technically" a LLM it's tiny and fast to run and could also have been implemented with just embeddings at a slightly loss of quality (and again only runs if you create a tab group and not at any other time and not if you computer has less then 3GiB of RAM etc. etc.(yes FF disables AI features on very low RAM devices AFIK)) Put people hear AI think LLM get angry and stop thinking. Or considering that maybe a non negligible amount of people outside of their bubble want exactly that and might leave your browser if you don't supply that.