@firefox STOP. UPDATING.
The browser was done TEN YEARS AGO. There is literally ZERO FUCKING JUSTIFICATION for you pushing updates EVERY GODDAMN DAY.
Fucking KNOCK IT OFF. WE DON’T WANT YOUR AI FUCKING GARBAGE.
This is why you’re losing to a fucking fork.
You were supposed to be defenders of the faith, not fucking corporate whore sellouts like Google.
Wow, shut up. I’ll take security over your weird delusions any fucking day.
@ReptilianCleric moron
@ReptilianCleric loh look, another useless twat to block
Did you see the article lately about the AI agents that found over 100 credible exploits in Firefox? We should leave those?
@badgermurphy Also, using the words “AI” and “credible” in a sentence only proves you are completely fucking misreading the room.
They are credible because they were independently verified by software developers after the AI identified and called them out. “The room” can think whatever it wants, but it cannot define objective reality.
I am not a proponent of the vast majority of AI’s uses, but to claim it has no valid use cases is intellectually dishonest. This is a case in point.
@badgermurphy it’s not reality. If you throw a million gigawatts of compute at a problem, you’d DAMN WELL better come up with an answer. But your answer has very little value, because those vulns aren’t realistic exploits - it’s an extortion scheme disguised as a security exercise.
You are still talking exclusively about LLMs. I watched an AI tool that was not an LLM running on local hardware devise a new space shuttle engine mount with a highly unintuitive design that a human would almost certainly never make, for example. Surely advancing an entire field of study is worth a couple servers, I would think.
To be clear, burning down The Amazon to code better is a horrific waste, which seems to be what youre describing, but I am not talking about that.
@badgermurphy I dont think it is actually, considering the damage that the compute build out is doing to the planet. That is literally my original point. The browser was done 10 years ago. We have efficient chemical rocket engines already, I’m a fluids engineer, I know all about them. We don’t need a LLM to regurgitate a marginally better design from existing constructs, we need a human to think of a better way to get into space than burning gigatons of hydrocarbons through a nozzle.
The build out of the data centers appears to be almost entirely for LLM AIs. So, I think it is fair to say that locally-running, non-IP-stealing, non-scraping, non-LLM machine learning tools are exempt from these ecological criticisms. The only thing they share with the technology doing that damage is their broad name, AI.
I think you may be falling into the same trap of famed Charles Deull: https://en.wikipedia.org/wiki/Charles_Holland_Duell
Innovation is never “done”. There are new high-quality ideas even today. If a new tool can help you test new ideas faster, that can still be a good to the world if it is a responsible technology like some of these are (but nearly all LLMs are not).
Regarding the browser, unfortunately it will always be an arms race. Even if Firefox never implemented a single new feature or changed any user-facing anything, vulnerabilities are frequently discovered and exploited. It is not hypothetical. I assure you that if you go get the oldest version of Firefox you consider “done” and run it today, it will be exploited either immediately or soon.
I agree that many browser technologies have no business in the browser, and too much recent browser work has essentially been a corporate dick-around that we may have been better off without, but the security and operational components underneath really do need to change often. The Internet may look pretty calm on the surface, but I assure you, it is an absolute warzone.
@badgermurphy You’re LIMITING your innovation by relying on a machine that can only rearrange the data it has, not think outside the fucking box.
That is only true if I use a tool instead of doing my job. Surely you dont think Folding at Home was a waste and an innovation limiter? It used tons of distributed CPU around the globe for years, but it also discovered novel proteins and advanced the fields of chemistry and biology.
It is true that scientific advancements can cause some of our previous skills to atrophy, like, for example, most humans can’t walk 25 miles per day in the modern era, but essentially all primitive humans could and did. I think that the human skills that LLMs are causing to atrophy are communication and creative thought, which are about the only things that separate us from the other apes, though. For that reason alone, I think they’re ominous and potentially disastrous. I believe that they could be worse for your brain than binging hard drug use, and if true, should be killed with fire.
However, other ML tools are not used in anything like the same manner as LLMs. I find that an endocrinolologist can use a tool to help her/him find incredibly early tumor development like a fighter pilot can use a targeting computer, for example. That doesn’t burn through resources like an LLM does, and objectively can help millions. LLMs promise the world, but almost nobody can point to a tangible good they’ve done to even slightly offset their ecological devastation.
I am by no means proposing that its a good idea to hand over the keys to the kingdom of innovation to any AI things. What I am saying is that they can offer a new way to allow us to leverage the thing computers are best at: doing practically the same thing repeatedly a zillion times. These are things that humans could do if there were enough humans, but there aren’t. Computers let an expert teach a computer to do one of his tedious tasks once, then having it do that for him way more than he ever could. He will get rusty at that tedious task, but he’s also doing potentially hundreds of times more work than before. Its a good trade.
@badgermurphy And we already HAVE bug bounties and ways to find vulnerabilities that don’t risk exploit, so destroying the planet (which, YES, data centers are definitely contributing due in excess percentage) ISN’T FUCKING WORTH IT.
@badgermurphy lol I’ve seen those “credible” reports. I should invite some people to come laugh in your face. I bet @jonny could rebut.
you guys are a joke on yourself every time you respond, you get that, right?
And STILL missing the fucking point.
For smart guys, you sure are fucking stupid. You think you’re the only ones out here who know stuff. Some of us are ACTUAL engineers buddy, not “computer science” majors.
What a joke.
Go back to ruining the internet and stop bothering me.
I have this job that calls me that. They could be wrong, but they’re paying me to do it, and by all accounts so far from anyone that knows me, I am pretty competent.
It seems that the “point I am missing” is that many people dislike AI so much that they are willing to gleefully rush to throwing the baby out with the bathwater.
I think anyone speaking in good faith about AI must acknowledge that it has led to many advancements in many industries. There are all kinds of ML algorithms, not just the barely useful “mechanical Turk” LLMs.
I’m opposed to using AI to autocorrect my code or tell me the name of the 5th king of England, but I’m not against it doing things like identifying cancer in patients practically before they even have it, or devising unintuitive physical structures with uncanny structural properties that advance space technologies, for example. Just because the technology is 99% waste doesn’t mean its not 1% useful. That waste is the result of a different problem: corporatocracy and regulatory capture.
@badgermurphy Stop conflating machine learning with LLMs. WE KNOW THE DIFFERENCE.
When you do that, we hear “lie, lie, bullshit, lie” and stop listening.
Where/how am I doing that?
I understand that ML is to LLM what a motor is to a car. They are not the same, but one is a piece of the other. I dont believe I’ve ever conflated those terms. I am trying to specifically draw that distinction, in fact.
I can’t tell if you’re trying to persuade me, convince others,or are simply trying to shout so I can’t be heard. Youre telling me to stop doing something I dont believe I ever started, but are not saying what that is in a way that I can decipher.
@badgermurphy you’re in a thread about AI in Firefox and uneccesary updates to push that garbage. I think YOU are the one who is trying to change the discussion. We’re talking about LLMs and consent and security. No one gives a fuck that your argument is “there are special use cases where LLMs are really cool!”
No. There are not.
I dont understand. I didn’t say you were changing the discussion and I didn’t say LLMs are good. I noted that they were used to find exploits in Firefox, many of which were confirmed to be true. I didn’t make a value judgment on whether using that tool was good. If I had, I would have said that it is not worth burning down the planet to write more secure code.
That said, they already did it. It would be even more of a waste if we disregarded the one small good that came out of it. I’m saying that we now know for a fact things that are not done about Firefox, that’s it. If you look at my words, I didn’t make the claims youre saying, and you are reading them into my words, presumably due to assumptions you’ve made about my opinion, which I assure you are also untrue.
@badgermurphy There is NO BABY IN THE BATHWATER. The LLMs lie, cheat, and steal, and are destroying the planet. We don’t want AI, we don’t want the data centers, and we don’t want the gas turbines.
LLM are a strategic, business, and personal liability, and not only did you STEAL the bathwater, you stole the fucking house.
Machine learning is not AI, we know the difference. What we don’t want is your fucking autochat autocomplete copyright theft machine anywhere near us.
You are talking about LLMs, and I’m talking about all of AI. LLMs are the ones that lie, cheat, and steal. Other ML algorithms use scientific data licensed for such uses. They share some technology, but are not at all the same in assembly, construction, or use. Much like a gas generator is nothing like an earthmover, but they operate on the same core technology.
@badgermurphy You can’t talk about ALL OF AI as one thing. It is not one thing, and we are well aware that the version you’re pimping is a copyright thieving LLM, not a machine learning algo running inside a closed system on proprietary data. Stop being disingenious. We’re not fucking stupid, and you’re just making yourself look bad. Worse, I should say.
I think you may misunderstand me. I am doing the opposite. LLMs are nothing like the other ML technologies other than that they use machine learning algorithms. They are as different as a leafblower and a space shuttle.
@badgermurphy In fact, let me tell you this: I have worked on projects worth $5 billion per unit with over a million parts each that are actually vastly more complicated than your little runtime routines. And do you know how many times in a 30 year service life we asked the customer to update the product? A product with sensors, network connects, comms devices, safety interlocks etc etc?
Never. Never in 10 years. That was our literal criteria. For a product with a 30 year life span.
@badgermurphy Beat that, and then come back and try and convince me it’s about security.
Because the thing I worked on was THE MOST SECURE MACHINE IN THE WORLD. A nuclear fucking submarine.
You don’t know shit about security. ESPECIALLY if you’re using a FUCKING LLM to do your code checks. Jesus, now that you told me that, I’m inclined to just delete the fucking thing.
@badgermurphy You should ALSO, instead of arguing with me when you don’t even understand the prompt, is to see the responses on my OP and how many people agree with my thoughts on your product. You are losing your customer base. You ought to be worried about that.
@badgermurphy Again, you’re being beaten by a FORK OF YOUR OWN PRODUCT.
Why the fuck do you think that is, smart guy?
@badgermurphy I’ll tell you why, since you’re obviously clueless: FUCK. YOUR. LLM. It’s NOT intelligent, it’s NOT alive, and it is NOT better than an actual senior developer. It’s also theft, copyright violation, proprietary data loss, intellectual property theft, environmental damage, and fucking brain rot.
@badgermurphy and I don’t know who is running the @firefox account, but THEY at least seem to get it.
No? Firefox is way behind and even if it wasn’t, the browser is an enormous attack surface so they have to keep improving security.
@muusemuuse nonsense. complete and utter bollocks.
Please explain.
@muusemuuse The security excuse for constant updates is a lie. There are not constant security patches that need to be fixed. The “risk” is extremely low, and is not reduced much by your browser checking a security certificate to see if the fraudulent website has paid google this month or not. THey update to put privacy and security violations on my machine - because that’s all AI is, a security NIGHTMARE. .
@muusemuuse Furthermore, I go to the same three websites every fucking day, not wandering the far depths of the darknet in search of snuff films and infected jpgs. There is NO RISK that requires daily or even weekly updates to a piece of software that is essentially an image viewer. And if there is, they DID IT WRONG. Browsers were done being developed ten years ago, and there are no “new features” worth a damn.
As an amateur web dev I can say stuff like jpegXL which allows for smaller and higher quality images is cool, and regardless browsers let websites run arbitrary javascript code, so it’s naturally insecure.
@muusemuuse Furthermore, if security is pushed all the way down to the user level, and we expect users to do the work to keep the internet secure, then the internet IS BROKEN AND BADLY DESIGNED. Security is the province of OS’s, ISPs, and networks, not individual instance users. If the last line of defense is ME< THEY DID IT WRONG.
You are demonstrably wrong with literally every damn post here so I’m just going to assume you are mentally ill and move on.
Counterpoint: No, YOU are wrong. This rant hits a lot of key points, and hits them well.
Related:
Number of times in my 60++ years I’ve had sensitive personal data compromised because I had a sucky password or didn’t use MFA: ZERO.
ZERO.
Number of times due to some giant corpo server being hacked or them JUST SELLING OR GIVING AWAY MY DATA: DOZENS. That I know of.
Stop blaming the fucking victims.
@muusemuuse go fuck yourself you clueless fuck
Oh, well if you say so then.
@ilinamorato yep, I did, and your sarcasm falls not upon me. go away.
That was an opportunity to explain a comment that clearly no one understands.
I’m sure that whatever platform you’re on lets you not update if you want.
I wouldn’t advise that from a security standpoint, but if you’re set on it…
They’re not pissed at lack of updates really, but at features that feel anti-consumer, such as always-on telemetry and opt-out AI
Then they should say that instead.
@Omodi @gwl all you people in this side thread are completely missing the point. and you’re not contributing anything here. No one wants your “Im sure there’s a way to block update notifications” BS responses There isn’t. And that’s not the point.
I don’t want to do your maintenance. I don’t want to read Firefox update release notes. I DON’T WANT TO FIND WORKAROUNDS FOR MY SOFTWARE.
The POINT is developers don’t understand consent OR security if they think the end user update solves security.
It sounds like he doesn’t want update notifications. I’m sure that that’s possible, but I’ve no idea what sort of update notification he’s upset about. Could be from some kind of package manager, Firefox itself, whatever. He might also have a nightly build version installed or something, when he could benefit from extended support version.
A browser is something that you should keep updated. I’m fully aware of the “Remind me in next time” trap that a lot of tracking and data harvesting is shoved behind, but just update your fucking browser dude. Sometimes, the developers actually know better than the users.
@wyatt_h_knott @firefox
What they said.
Way to many updates and the containers make the updates very disruptive … the total opposite of what *nix updates are supposed to be.If you want security updates, but not necessarily the latest features, there is always Firefox ESR.
@folekaule I actually use libre wolf when I care
