Hacker Newsnew | past | comments | ask | show | jobs | submit | homebrewer's commentslogin

And this explanation is very likely to be entirely hallucinated, or worse, subtly wrong in ways that's not obvious if you're not already well versed in the subject. So if you care about the truth even a little bit, you then have to go and recheck everything it has "said".

Why waste time and energy on the lying machine in the first place? Just yesterday I asked "PhD-level intelligence" for a well known quote from a famous person because I wasn't able to find it quickly in wikiquotes.

It fabricated three different quotes in a row, none of them right. One of them was supposedly from a book that doesn't really exist.

So I resorted to a google search and found what I needed in less time it took to fight that thing.


And this explanation is very likely to be entirely hallucinated, or worse, subtly wrong in ways that's not obvious if you're not already well versed in the subject. So if you care about the truth even a little bit, you then have to go and recheck everything it has "said".

It cited its sources, which is certainly more than you've done.

Just yesterday I asked "PhD-level intelligence" for a well known quote from a famous person because I wasn't able to find it quickly in wikiquotes.

In my experience this means that you typed a poorly-formed question into the free instant version of ChatGPT, got an answer worthy of the effort you put into it, and drew a sweeping conclusion that you will now stand by for the next 2-3 years until cognitive dissonance finally catches up with you. But now I'm the one who's making stuff up, I guess.


Unless you've then read through those sources — and not asked the machine to summarize them again — I don't see how that changes anything.

Judging by your tone and several assumptions based on nothing I see that you're fully converted. No reason to keep talking past each other.


No, I'm not "fully converted." I reject the notion that you have to join one cult or the other when it comes to this stuff.

I think we've all seen plenty of hallucinated sources, no argument there. Source hallucination wasn't a problem 2-3 years ago simply because LLMs couldn't cite their sources at all. It was a massive problem 1-2 years ago because it happened all the freaking time. It is a much smaller problem today. It still happens too often, especially with the weaker models.

I'm personally pretty annoyed that no local model (at least that I can run on my own hardware) is anywhere near as hallucination-resistant as the major non-free, non-local frontier models.

In my example, no, I didn't bother confirming the Russell sources in detail, other than to check that they (a) existed and (b) weren't completely irrelevant. I had other stuff to do and don't actually care that much. The comment just struck me as weird, and now I'm better informed thanks to Firefox's AI feature. My takeaway wasn't "Russell wanted to nuke the Russians," but rather "Russell's positions on pacifism and aggression were more nuanced than I thought. Remember to look into this further when/if it comes up again." Where's the harm in that?

Can you share what you asked, and what model you were using? I like to collect benchmark questions that show where progress is and is not happening. If your question actually elicited such a crappy response from a leading-edge reasoning model, it sounds like a good one. But if you really did just issue a throwaway prompt to a free/instant model, then trust me, you got a very wrong impression of where the state of the art really is. The free ChatGPT is inexcusably bad. It was still miscounting the r's in "Strawberry" as late as 5.1.


> I'm personally pretty annoyed that no local model (at least that I can run on my own hardware) is anywhere near as hallucination-resistant as the major non-free, non-local frontier models.

And here you get back to my original point: to get good (or at least better) AI, you need complex and huge models, that can't realistically run locally.


No idea about most Linux users, but here's what little we know for sure:

Arch pkgstats (opt-in): ~64% FF, ~41% Chromium, ~17% Chrome

https://pkgstats.archlinux.de/fun/Browsers/current

Debian popcon (opt-in): 2.2% Firefox, ~10.3% Chromium

https://qa.debian.org/popcon.php?package=firefox

https://qa.debian.org/popcon.php?package=chromium

Flathub installs: 10kk Firefox, 10kk Chrome, 1.8kk Chromium

https://flathub.org/en/apps/org.mozilla.firefox

https://flathub.org/en/apps/com.google.Chrome

https://flathub.org/en/apps/org.chromium.Chromium

snapcraft statistics isn't public, afaik.


The usual package for Firefox on Debian is `firefox-esr`. 44.18%.

https://qa.debian.org/popcon.php?package=firefox-esr

The `firefox` package is only in `unstable` and thus installing it requires jumping through hoops.


According to Mozilla's own stats, most Firefox users do not have any extensions at all:

> Has Add-on shows the percentage of Firefox Desktop clients with user-installed add-ons.

> December 8, 2025

> 45.4%

https://data.firefox.com/dashboard/usage-behavior

Note that language packs are counted as extensions.

Some have disabled telemetry, of course, but how many? Here we can only rely on our own observations, and of all Firefox users I know, it's zero.

(I keep it enabled because I want my voice to be counted — people who have never lived in an autocracy tend to have peculiar views on this.)


I think the correlation of people using extensions and people disabling telemetry is pretty high. I do both myself. Even a decent password manager requires one (though not on android because it has an API for that). On android I do use others obviously.

Always appreciate people citing real data! I honestly would not have been able to guess one way or the other but unfortunately most comments are kind of hip firing in random directions that are impossible to keep track of, so it helps to keep these discussions grounded.

But what if you weigh this by usage time? The firefoxes without extensions might be hardly ever used

> About a month before expiration it somehow got renewed for 10 years, which is weird because it was not available

I've seen some domain registrars auctioning off domains during the last 2-4 weeks before they expire. If nobody buys it, then it actually expires and is then released.


Which registrars? I would want to avoid those.

At the end of the day, no matter your domain, ICANN can just take it for their VC bros. Happened to a friend of mine that owned a pretty novel domain name that a certain social media company wanted. He refused to sell. ICANN and his registrar just transferred it out from under him. Gone. See ya.

Wow. In light of this it's amazing that Mr. Nissan (RIP) and later his heirs managed to not only retain control of nissan.com, but regain it after it was stolen years after his passing.

Money talks

There's a difference between trademark issues and your registrar auctioning off the name

Unless you prove you had the trademark first and they did it anyway.

Out of curiosity, what was the domain?

You gotta name the domain!

I'd rather not face the ire of ICANN, sorry.

I know better. They read this site. They know that all it takes is some company to issue some trademark litigation and they fold. No basis, no question, just here you go.


what would they do to you...?

Oh you sweet summer child, you haven’t met their lawyers…

well, condescension aside, literally what would they do? there's nothing remotely illegal about posting the name of a site in a forum. and here you are trying to get me to be as scared as you are about posting a basic fact in a forum and why would I be?

You might want to look up Thunderbird crowd funding over the past couple of years. Spoiler: it's been very successful.

Check Firefox's annual budget compared to Thunderbird's annual budget and get back to me.

Selfhosting gitea is trivial, I'm saying this as someone who has been doing it at work for almost 6 years. Our experience has recently prompted another org (run by people we know) to move off GitHub, they also seem to be happy.

There are plenty of alternatives to Github, from Gitlab and Gittea to Forgejo, but Codeberg is not one of them, which is what I wanted to stress.

Brave is 100% FOSS. At least the client side, I've not looked into their server applications.

https://github.com/brave


Fair enough. I'd still be very hesitant to use it on account of it being a chrome fork. Moreover, I don't really understand how Brave expects to be a viable business without deeply betraying their userbase at some point.

It admittedly is a gut feeling, but Brave started out with a browser and some handwavy crypto magic beans and seemed like it careened from idea to idea looking for a business model, occasionally stepping on toes along the way. They have products like AI integration, a VPN and a firewall, but those aren't particularly stand-out products in a very crowded market.

As a point of comparison, Kagi started out with a product that people were willing to pay for, and grew other services from there. I feel comfortable giving them money, and I'd be willing to at least try their browser - if it ever releases for Windows.


Your points are valid. But what made me finally switch was that it is open source, that it has been out for roughly a decade now, and that Brendan Eich's opinions from 2014 are mostly based on his Catholic faith at the time (which obviously is likely to have changed/evolved now that we're a decade later).

> Moreover, I don't really understand how Brave expects to be a viable business without deeply betraying their userbase at some point.

They have a way better merch store than Mozilla. They should expand that.

"MERCHANDISING! Where the real money from the movie is made!"


> which obviously is likely to have changed/evolved now that we're a decade later

I refuse to make any assumptions there. Either he says he changed, or I treat him like he hasn't changed.


> Either he says he changed, or I treat him like he hasn't changed.

That kind of attitude is why things like right to be forgotten are being pushed. The world isn't set in stone.


Right to be forgotten is fine for people who aren't celebrities. If you want to be forgotten then don't have a CEO job at a huge company.

Also it generally applies to situations where people are associated with a crime and either there wasn't enough evidence or they already served their time. Neither of those is true here. He has faced nothing except people being mad at him for something he definitely did.

It would be so easy for him to say he changed. Why should I pretend it never happened if he won't do that little thing?



Not sure who or how, but someone somewhere confused Gnome for Mozilla/Firefox. The claim was that Mozilla has had an "literal witch as CEO" but that article is about Gnome.

You've confused them with GNOME. The witch is out, she did not last long.

Oopsie. Yeah that was GNOME. My bad.

This may very well be caused by my incompetence, but Sentry's docker-compose setup has never survived for more than a few months under my control. Something always destroys itself without an obvious reason sooner or later, and either refuses to start, or starts and doesn't really work. I tried updating it regularly, tried never updating it, getting the same treatment either way.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: