Hot Take Firefox shares roadmap as it loses users at an alarming rate

[Victor M]

Yes, I have read good things about Gemini. It is supposedly great at green field coding but so so at modifying existing code. Haven't tried it, don't know for sure.

 

[Victor M]

Anyways, i selected Brave snap as the default browser in Ubuntu over Firefox. It has an adblocker, and can turn off javascript optimization which I heard is exploit friendly, And I unchecked keep system running after shutdown. Any second opinions?

 

[Zero Knowledge]

Even when they don't, Ghostery and Adguard MV3 extensions work just fine on all chromium browsers. I don't understand why people keep preaching Firefox + Ublock Origin as the only solution these days.

No they don't, try accessing YouTube or Facebook/Instagram with ghostery or Adguard. They won't work properly. Regular non security/it just want to access FB/IG/X.

They don't want to be bothered with buggy extensions or technical crap, they just want their social media to work.

I think people do care, they simply don't know how to set it up correctly and are afraid they'll mess something up. I had numerous people asking me to get rid of the ads on their phones at work. Not just for YouTube but also for browsing the web.

Because while they may remove simple ads, they can't remove complex ads which are injected through JavaScript and need advanced scriptlets. They also struggle a lot with those anti-adblock messages. There's no element picker so if there are ads or placeholders that aren't blocked, you can't block them manually. Then there's control argument as uBlock Origin in Firefox gives you absolute control what will load and what won't.

Simply said, if you want to block ads more efficiently and proactively, you have no other options than to use Firefox + uBlock Origin (or Brave).

Oh, they will change their minds once ads become intolerable. My parents were never bothered by ads, until YouTube started implementing them more aggressively as well as news portals they visit. They aren't using PCs though, just phones.

Sorry but bullshit, they are annoyed but they don't care. They just want to read memes and FB/IG posts and watch cat movies on YT.

The moment you ask them or explain adblockers or DNS filtering they zone out. VPNs anyone can get hence why popular but more technical stuff they don't care.

@Marko :) Nah, Athropic cites Madaidan, privacyguides and Profincognito, not Reddit. I asked it to produce citations. It even shows which arguments are dated. AI is improving all the time.

I didn't ask Anthropic which browser is most secure. I asked it to compare security of Chromium and Firefox. And I gave it context: I am using the info for a dmz pwn test.

Then I pick Brave which is Chrome based, and asked ChatGPT the similar question.

Yeah whatever mate, A.I steals regularly from reddit including Claude and ChatGPT and from forums like these since it's models are trained on them.

I guarantee you every A.I model is scrapping every reddit post daily if not hourly, they are also scraping every security forum post like this forum and others daily.

I am not just flooding the forum with nonsense, I am giving you what the paid versions produces. And Anthropic gives more detail. It didn't claim Brave is better, it gave me some caveat about Brave on Ubuntu. The Claim is My understanding and conclusion, I wasn't clear about that.

Come on, do we want to go back to people posting A.I slop as original content. Leave that crap to bot. No one wants to read A.I garbage.

My coding years are behind me. Nowadays I use ChatGPT mostly for coding. And I am not asking just for 3 liner's. And ChatGPT is reasonably good at it. May current project runs at almost 1600 lines of code. It sometimes produces mistakes, and I have to give it the error messages and feedback, but it gets the job done. Without it I would have to learn new languages which would take months. And then learn Linux specifics which would require god knows how long.

Hence is the problem with A.I and the influx of vibe coded garbage apps. There are just so many numbnuts out there making slop apps and posting them online it's gotten unbearable. But I guess with the death of coding/dev jobs people are trying to eeek out a living somehow. You know full well A.I is trained on random posts on StackEXchange, StakcOverflow, ServerFault and other random Linux/coding mailing lists. If your A.I model is trained on garbage amateur posts what will the end result be? Code full of bugs, vulnerabilities and non working code.

 

[Victor M]

@Zero Knowledge I know we all like to think that AI are dumb, and trained with all the garbage on the internet. And that all it can do is regurgitate. But I would like to point out one important thing, it is a neural net, with it's working modeled after our brain. The only difference is that it learns with words, whereas I don't think we use words for all our thinking. But it can reason. So the garbage in garbage out theory doesn't quite fit, because it can filter out garbage just like we can. And it can know to recognize a person is forum posting a bug needing an answer, from the forum post that posts a solution to that problem. The fact that it can reason also means it can learn abstract concepts. That is why it can create programs, when to use an "if then else' and when to use a 'while loop' so as to produce a solution to the programming problem we give it. It can also think via association. When you want it to create a logo for a cybersecurity company it would choose to use shields and padlocks icons.

So when asking AI a question, it answers by filtering and reasoning all the things it has read to produce an answer. It could be called regurgitation, similar to how a doctor would regurgitate the medical text he learned in university, but it can also apply reasoning to fit precise question just like how a doctor would apply our medical history to produce a diagnosis. The more questions you ask of it, the more 'medical history' it has. And thus when I ask it a question like, "what do you think of CyberLock' it would tailor the answer to fit my particular strict stop-the-adversary-at-all-costs hard line security threat model. And point out things that my threat model would like to hear, and features that doesn't fit my threat model and caveats that I should be aware of.

As to your vibe coded apps, I would just like to say beauty is in the eye of the beholder. If a program fits what we want to do, then who cares if it is vibe coded or not?

So if you don't have much history of questions in a particular domain, then you must give it context along with the question. Such as 'Answer like a pharmacist, what does drug XXX do, what illnesses does it treat' . Or "Answer like a security analyst, what do you think of CyberLock'. Then you won't get random quotes from Reddit.

 

[Zero Knowledge]

As to your vibe coded apps, I would just like to say beauty is in the eye of the beholder. If a program fits what we want to do, then who cares if it is vibe coded or not?

I'm just going to reply to this, the problem with vibe coded apps is people with zero skill or coding education or history are creating apps with zero thought about security. Zero accountability, zero audits, zero security, zero design, just ZERO KNOWLEDGE. What on earth can go wrong with LLM trained on garbage StackEXCHANGE posts?

I'm not against A.I, in fact it's absolutely superb. What I'm against is the flood of poo apps/extensions everywhere with people trying to make a quick buck while harvesting user details. I know I know coders/devs have been replaced on mass by A.I and what can they do expect create a crappy fuel price check app and hope for the best.

Anyway this will sort itself out eventually when all these A.I coders find out there is zero money in poo and that security is hard even with A.I. The flood of user data is about to begin.

 

[Gangelo]

No they don't, try accessing YouTube or Facebook/Instagram with ghostery or Adguard. They won't work properly

Perhaps your experience is different than mine. Adguard extension especially works fine for me on Edge on my work laptop, even on YT. My personal machines all have adguard for windows installed which is of course on another level but we are talking about the extensions here.

The way I see it is that MV3 extensions get better and better over time. At least good enough for me not caring about MV2 to MV3 transition anymore.

The argument circulating the internet to use Firefox just because of proper adblocking is a non-issue these days IMHO.

 

[Berny]

1) FF 153.0.4 : Privacy and Security → Enhanced Tracking Protection → Standard (default)
2) Kaspersky Protection → Block data collection
3) uBlock Origin

Works fine for me ...