I'm genuinely perplexed why people find link previews useful. The page image is too small to really see, and it's not as if rapidly opening and closing a tab actually takes effort. It may be faster than the long click depending how comfortable you are with keyboard controls.
I have other questions as well -- does the link preview ignore your plugins? So if you loaded the web page, would uBlock Origin prevent tracking which is _not_ prevented when loading the preview? If you have a phishing test url with a unique identifier, do you flunk the test even if you don't click through?
It's an idea that makes a lot of sense on Wikipedia, because every page there explicitly starts with a short summary, and often an image that can be pulled from the infobox. I'm not yet convinced it will work well on the rest of the Web.
But also that's been part of Wikipedia's website for years now, no special browser support necessary. And because it's tailored specifically to Wikipedia, it works great!
Page preview seems nice in theory, but I'm unconvinced it'll be that useful. Web pages just don't have a the level of standardization necessary to automatically grab a useful preview. And I don't think Firefox has a big enough pull to make that sort of standard.
>> it's not as if rapidly opening and closing a tab actually takes effort. It may be faster than the long click depending how comfortable you are with keyboard controls.
For normal people, i.e. those who aren't techies, you have your answer right there on why this may be useful. We are very efficient with our browser usage. Normal people struggle with this task because they get lost in the steps. Even if they can do the steps at a slow pace, it doesn't stick because it's too slow. The leap to using keyboard shortcuts just won't happen for the majority of people.
I think your points on how this is implemented are fairer. That said, that said it seems like the best approach is to follow the simplest and most efficient approach, which is to not load extensions, or use what is loaded.
The success of features like this is known with technical users like us. We don't like them because we have a workflow that avoids the issue like the phishing one you describe. It's unclear whether this helps users and is likely somewhat experimental. I think it's a much better place to be doing work that other areas where they have invested even if it has many issues.
It's just horrible. For me the preview doesn't show until I've hovered for ~3 seconds (even before I turned on AI), in which time I could have long since middle clicked the page, skimmed, and closed it again.
This is the content for the preview:
> www.mozilla.org
> What's new with Firefox 142
> What's New | Firefox 142
> 3-4 mins reading time
No OpenGraph descriptions are good, so I can't see it ever being better than this. I don't know why this reading time metric has become a thing, it's useless because it doesn't know which parts of the page I'm interested in. I could actually see the full url from the immediate link preview, so having only the domain here is worse than useless.
The AI summary is both too short and too slow to be useful (unless maybe you're running an RTX 6000 or whatever). For this link, it only mentions Relay.
And even the basic behaviour seems broken. The preview appears at seemingly random locations on the page, sometimes under the cursor and sometimes far below. When it does decide to appear away from the cursor, releasing the mouse button actually follows the link, completely negating the purpose of the preview!
> I have other questions as well -- does the link preview ignore your plugins? So if you loaded the web page, would uBlock Origin prevent tracking which is _not_ prevented when loading the preview?
This was the primary reason I'd never used Vivaldi browser's long-standing feature of being able to set sites within a side panel, since it ignored addons. Was only last year that was changed to allow addons.
I quite prefer the link "peek" in Zen, which I'm led to understand is based on the one in Arc. Shift click a link, it opens up in a configurable-size floating modal (90% by default) on top of the current page, with controls to pop it out into a full tab or close it. It's only a small improvement over "open in new tab, switch to new tab, close new tab", but it feels nice.
The new link preview in Firefox seems a lot less useful.
Interesting thought. A lot of sites have been adding various META tags for link previews in social media, it might be neat to have a browser native form of that, especially because making it a browser feature could help push it back to open standards (right now most of them are OpenGraph which despite "open" in the name is still somewhat proprietary to Facebook/Meta and Twitter Cards which will probably forever be called that).
(ETA: This does seem to be what the feature actually does, having now tried it. Ignoring the AI Summary feature part of it, most of it does appear to be META tag driven and uses the card images of OpenGraph/Twitter Cards.)
It seems like Firefox (Mozilla) looks for any excuse to make an HTTP request
I can't think of another software program I have ever seen that makes so many non-user-initiated, i.e., automatic, HTTP requests by default; some of this behaviour cannot be disabled
> I have other questions as well -- does the link preview ignore your plugins?
If it’s anything like Reader View in Firefox, I would expect it to ignore plugins, which is absurd in the Reader View case, and would be in the link preview case as well. So much for my browser being a user agent.
for me I'm interested in the preview so I know where the link goes in a general sense, not what it looks like. i would use it for shortened or redirected links to see where a random anchor is pointing before I actually open it?
It already seems to basically grab the text from reader mode. What might be more useful is a way to just open a link directly in that to avoid paywalls or annoying uis though
> Link Previews show a snapshot of a page before you open it, helping you decide what’s worth your time. Just long press any link to preview and reduce distractions.
On macOS there is a native affordance for this by using force click. It's kind of annoying that Firefox chose to not support this and instead made it click-and-hold only.
I agree with the other comments that say it's not discoverable. I've been using MacOS for 10 years and I didn't even know that "Force Click" is a thing. This comment caused me to look it up and then try it.
I disagree that it's "everywhere". I just tried it in Spotlight Search (Command + Space). I can never remember how to see where a thing is located there and I hoped a force click might show me. Hint: pressing command will do it but it takes 250ms to 500ms to show up and somehow I never wait that long.
Not only does Spotlight Search not show me where it is, force clicking doesn't show me a preview, or seem to do anything else.
In Finder force click edits the name of the item you're clicking. So, this doesn't seem to be terribly universal.
I suppose I'll slowly figure out how this works now that I'm aware of it.
This seems to be exclusive to Safari, I can't get it to work in Chrome either (and didn't know about the feature before right now, the discoverability is terrible).
Pretty typical of browsers and especially of Firefox. Took them 7 years to use native scrollbars post Lion, it looked like cross-platform junk for the longest time.
Back when Chrome was still trying to gain traction, they came out supporting native Keychain Access for password management. Firefox, which had a 4 year head start, never implemented it.
Lately, I’ve experienced memory leaks in Firefox that I’m too amateur to diagnose, that leads to Firefox eating 8gb of memory in some web renderer process. So when I excitedly check the changelog hoping for a summary of possible changes, I’m disappointed that there isn’t a verbose changelog for advanced users. I’m sure I could search bugzilla, but it makes me sad that the only “important” things are the headlining features.
It is very likely that one single site is responsible for this. For example in my usage, Jenkins portal is leaking memory like hell, will fill up a few GBs in a a day and jack up CPU utilization to 100%.
I has hundreds of tabs at that point, so I had to find out the culprit. What I did was open a task manager and expand Fifefox process to see dozens of subprocesses. I then sorted by memory (or by cpu) to find out worst offender and killed only that single subprocess without touching others. And after doing that I've looked at the list of tabs and saw that one of them has changed to the crash report tab, with a visibly different icon. And looking at it I saw that it was originally Jenkins portal. Now I proactively close it's tabs and leaks stopped. Maybe this will help you.
cool. There is also about:performance with cpu & memory for each tab / addon. Hmm, it looks like all extensions are combined in a single entry. So, maybe not per addon.
You mean more verbose than the landing page or the release notes which are also linked in the landing page: https://www.firefox.com/en-US/firefox/142.0.1/releasenotes/ ? This is a point release, even more changes are linked at the bottom of the release notes page.
Next up would be looking closer at the pages you frequent. I think many people would be surprised at all the ways web apps screw up these days.
All that said, the browsers, as unfair as it may seem, should do better at handling all of the slop that web app and extension developers put out there. It’s sometimes just a whole lot easier to make the browser more bulletproof than it is to make a bajillion JavaScript/python monkeys conscientious and competent.
I’ve had similar issues running the latest Firefox (currently 142 as per this discussion) on the latest Fedora (42). I used to be able to lock the screen and go home but I’ve recently had a couple of mornings where I’d come into work and find my system unresponsive. I use the Magic SysRq command to trigger the OOM (out-of-memory) killer as many times as required to free up enough resources that I could log in on a virtual console (Alt-Ctrl-F2). This would allow me to manually kill Firefox, freeing up about 15GB of RAM and all 16GB of the swap file.
I’ve been too busy with work to spend any time investigating the cause. At first I had been blaming the `teams-for-linux` electron app but figured that wasn’t the culprit because I close it every evening. In Firefox, `about:processes` is useful while actually using the application but I’m not really sure how best to diagnose what’s happening after the fact.
I've used firefox for 15+ years, but it starting to bug me too much. Memory issues, forced restarts that block navigation (wth!), clutter all around, having to disable sponsored crap, and random incompatibilities are starting to take a toll. I've been falling back to Edge for crying out loud.
From what I understand, this is because the OS package manager changed some of Firefox's files in a background update (Ubuntu does this through unattended-upgrades), and Firefox's built-in updater doesn't have this problem.
Go into about:config and search for browser.ml stuff. Some of it is just for text completion, but you can also load models - transformer js, and also send stuff to chat gpt.
No they don't, please stop spreading misinformation. There was a performance regression due to the new vector search which was fixed before it even reached 1% of Firefox users in the US alone: https://news.ycombinator.com/item?id=44912974
I wonder how the link preview will work on mobile. In the past, click and hold on a partially obscured URL on mobile basically worked like "hover" on the desktop and revealed the whole link (useful to see if the link was malicious, or had click tracking, or just didn't go to the site you thought it did).
I hope there's still a way to do that feature. If nothing else, by disabling the preview which I doubt I would find as useful.
edit just retested on mobile. click and hold is basically the context menu. If you tap on the first item in the context menu (the link ended in …) you get the full link wrapped.
So, more correct to say that click and hold is the right click context menu on mobile, even though I mostly just used it for that display whole link feature. Perhaps they could put the preview as a sub item of the context menu.
I wonder if it addressed something practical and meaningful...like implementing a spell check that at least covers all of the English language. Maybe then they can start working on the 5 year trek to also make the spellchecker somewhat competent.
Or maybe we'll just keep packing features, because everyone here knows, features are what save products! Not usability!
At this speed Ladybird might be the browser to use in a couple years, even though I find their decision to use a memory unsafe language for their core tech foolish in the year 2025.
I remember the original reasoning for creating Firefox (first Phoenix) was to create a more lightweight and faster version of the Mozilla browser, which initially was true.
Or maybe fixing a few dozen open bugs about losing all open tabs to a /dev/null once in a while. I had it happen twice this year and now extremely vary about even simply normally using browser. Sigh...
I should probably bug report this, but I don't even know where to start. For as long as I can remember "find" has been semi-broken.
Hit Ctrl+f to search for a word you know for sure is on a page and Firefox might not find it. You type in the first five characters and Firefox goes "Nope, can't find it", then enter character number six, and then Firefox sees it, enter another character and nope, lost it, can't find it anymore.
You look stare right at a word and find will be unable to locate it.
But yeah, there's nothing I really care about, haven't been in a while. Funny enough, given how Firefox/Firebird/Phoenix started, I just want a slimmed down browser without a ton of features. Tabs and an extension API, that's it.
If you turn off the ai part the preview just pulls in the title and the first chunk of text on the page. Actually still fairly useful for shortened links or ambiguous urls, etc
I'm still ambivalent on the rest of the AI features, but the AI translation is absolutely amazing. The translation quality isn't perfect, but being able to seamlessly translate 20+ languages 100% locally is amazing.
Agree, for us switching between languages all the time, with some of those languages being less known to us, it's a great tool!
My only wish was that I can force it to always allow me to try translating things, even if it doesn't identify it as some specific language. Sometimes what I want to translate is like 30% one language and 70% another language, and I still want to translate it to another language, but since the tool doesn't see it as "foreign enough" or something, I don't even get the choice of having it translated.
Besides that, it's a wonderful despite it not being perfect. Hopefully with time it'll only get better as they get more data. On that note, I'd be more than happy to contribute data if they added some way of giving "good translation / bad translation" feedback, but haven't seen that. I guess I had two wishes in the end.
I agree, I'm generally sceptical of new AI "features" in the browser and will be turning most of them off. But the translation feature (which has been in Firefox for a while now) is great. The difference is that translation in a browser is something that is clearly useful and has always been AI-based to an extent, so shipping with a local model for translation is a strict improvement (leaving aside any difference in translation quality, which I have not noticed). The other AI features are not obviously useful IMO.
Link Preview is a little weird, but also interesting, you can run a local AI model that summarizes the page for you, so its not using ChatGPT, or any of the cloud APIs.
Readit News
I have other questions as well -- does the link preview ignore your plugins? So if you loaded the web page, would uBlock Origin prevent tracking which is _not_ prevented when loading the preview? If you have a phishing test url with a unique identifier, do you flunk the test even if you don't click through?
Page preview seems nice in theory, but I'm unconvinced it'll be that useful. Web pages just don't have a the level of standardization necessary to automatically grab a useful preview. And I don't think Firefox has a big enough pull to make that sort of standard.
For normal people, i.e. those who aren't techies, you have your answer right there on why this may be useful. We are very efficient with our browser usage. Normal people struggle with this task because they get lost in the steps. Even if they can do the steps at a slow pace, it doesn't stick because it's too slow. The leap to using keyboard shortcuts just won't happen for the majority of people.
I think your points on how this is implemented are fairer. That said, that said it seems like the best approach is to follow the simplest and most efficient approach, which is to not load extensions, or use what is loaded.
The success of features like this is known with technical users like us. We don't like them because we have a workflow that avoids the issue like the phishing one you describe. It's unclear whether this helps users and is likely somewhat experimental. I think it's a much better place to be doing work that other areas where they have invested even if it has many issues.
But you're point is correct that the majority of people do indeed not use any device quickly. No matter wherever it's a phone or desktop browser.
This is the content for the preview:
> www.mozilla.org
> What's new with Firefox 142
> What's New | Firefox 142
> 3-4 mins reading time
No OpenGraph descriptions are good, so I can't see it ever being better than this. I don't know why this reading time metric has become a thing, it's useless because it doesn't know which parts of the page I'm interested in. I could actually see the full url from the immediate link preview, so having only the domain here is worse than useless.
The AI summary is both too short and too slow to be useful (unless maybe you're running an RTX 6000 or whatever). For this link, it only mentions Relay.
And even the basic behaviour seems broken. The preview appears at seemingly random locations on the page, sometimes under the cursor and sometimes far below. When it does decide to appear away from the cursor, releasing the mouse button actually follows the link, completely negating the purpose of the preview!
Four times of five the link previews are blocked by a cookie consent or newsletter signup modal anyway.
This was the primary reason I'd never used Vivaldi browser's long-standing feature of being able to set sites within a side panel, since it ignored addons. Was only last year that was changed to allow addons.
The new link preview in Firefox seems a lot less useful.
One interesting breakdown: https://getoutofmyhead.dev/link-preview-meta-tags/
(ETA: This does seem to be what the feature actually does, having now tried it. Ignoring the AI Summary feature part of it, most of it does appear to be META tag driven and uses the card images of OpenGraph/Twitter Cards.)
I can't think of another software program I have ever seen that makes so many non-user-initiated, i.e., automatic, HTTP requests by default; some of this behaviour cannot be disabled
Of course the default telemetry is infamous
If it’s anything like Reader View in Firefox, I would expect it to ignore plugins, which is absurd in the Reader View case, and would be in the link preview case as well. So much for my browser being a user agent.
Deleted Comment
It already seems to basically grab the text from reader mode. What might be more useful is a way to just open a link directly in that to avoid paywalls or annoying uis though
On macOS there is a native affordance for this by using force click. It's kind of annoying that Firefox chose to not support this and instead made it click-and-hold only.
I agree with the other comments that say it's not discoverable. I've been using MacOS for 10 years and I didn't even know that "Force Click" is a thing. This comment caused me to look it up and then try it.
I disagree that it's "everywhere". I just tried it in Spotlight Search (Command + Space). I can never remember how to see where a thing is located there and I hoped a force click might show me. Hint: pressing command will do it but it takes 250ms to 500ms to show up and somehow I never wait that long.
Not only does Spotlight Search not show me where it is, force clicking doesn't show me a preview, or seem to do anything else.
In Finder force click edits the name of the item you're clicking. So, this doesn't seem to be terribly universal.
I suppose I'll slowly figure out how this works now that I'm aware of it.
Back when Chrome was still trying to gain traction, they came out supporting native Keychain Access for password management. Firefox, which had a 4 year head start, never implemented it.
I has hundreds of tabs at that point, so I had to find out the culprit. What I did was open a task manager and expand Fifefox process to see dozens of subprocesses. I then sorted by memory (or by cpu) to find out worst offender and killed only that single subprocess without touching others. And after doing that I've looked at the list of tabs and saw that one of them has changed to the crash report tab, with a visibly different icon. And looking at it I saw that it was originally Jenkins portal. Now I proactively close it's tabs and leaks stopped. Maybe this will help you.
Next up would be looking closer at the pages you frequent. I think many people would be surprised at all the ways web apps screw up these days.
All that said, the browsers, as unfair as it may seem, should do better at handling all of the slop that web app and extension developers put out there. It’s sometimes just a whole lot easier to make the browser more bulletproof than it is to make a bajillion JavaScript/python monkeys conscientious and competent.
I’ve been too busy with work to spend any time investigating the cause. At first I had been blaming the `teams-for-linux` electron app but figured that wasn’t the culprit because I close it every evening. In Firefox, `about:processes` is useful while actually using the application but I’m not really sure how best to diagnose what’s happening after the fact.
Is that what you're looking for? If you sort it by updated, I assume everything after the 27th is probably in the current release.
From what I understand, this is because the OS package manager changed some of Firefox's files in a background update (Ubuntu does this through unattended-upgrades), and Firefox's built-in updater doesn't have this problem.
Go into about:config and search for browser.ml stuff. Some of it is just for text completion, but you can also load models - transformer js, and also send stuff to chat gpt.
I hope there's still a way to do that feature. If nothing else, by disabling the preview which I doubt I would find as useful.
edit just retested on mobile. click and hold is basically the context menu. If you tap on the first item in the context menu (the link ended in …) you get the full link wrapped.
So, more correct to say that click and hold is the right click context menu on mobile, even though I mostly just used it for that display whole link feature. Perhaps they could put the preview as a sub item of the context menu.
Or maybe we'll just keep packing features, because everyone here knows, features are what save products! Not usability!
Hit Ctrl+f to search for a word you know for sure is on a page and Firefox might not find it. You type in the first five characters and Firefox goes "Nope, can't find it", then enter character number six, and then Firefox sees it, enter another character and nope, lost it, can't find it anymore.
You look stare right at a word and find will be unable to locate it.
But yeah, there's nothing I really care about, haven't been in a while. Funny enough, given how Firefox/Firebird/Phoenix started, I just want a slimmed down browser without a ton of features. Tabs and an extension API, that's it.
Why am I reminded of the Blackadder episode, "Ink and Incapability"?
[0]: https://github.com/mozilla/firefox-translations-models/
[1]: https://github.com/davidventura/firefox-translator
My only wish was that I can force it to always allow me to try translating things, even if it doesn't identify it as some specific language. Sometimes what I want to translate is like 30% one language and 70% another language, and I still want to translate it to another language, but since the tool doesn't see it as "foreign enough" or something, I don't even get the choice of having it translated.
Besides that, it's a wonderful despite it not being perfect. Hopefully with time it'll only get better as they get more data. On that note, I'd be more than happy to contribute data if they added some way of giving "good translation / bad translation" feedback, but haven't seen that. I guess I had two wishes in the end.