Hacker News Reader

Hacker News

Comment by meremortals | Hacker News Reader

Parent Full thread

meremortals

Which criticisms? I've just started using it after Firefox's blog post

View on HN

Perhaps this?

The Brave web browser is hijacking links, and inserting affiliate codes

https://davidgerard.co.uk/blockchain/2020/06/06/the-brave-we...

https://news.ycombinator.com/item?id=23442027

The phrase "hijacking links" is, in a word, a lie. Its standard meaning implies that Brave changes the URL loaded when you click on a link (any or many links) in a web page. That never happened, and would be fatally scandalous if tried by any reputable browser.

What happened was a bug in a keyword and domain autocomplete tier we added. Reminder for those not aware: all browsers add reference codes identifying the browser make (not the user) to keyword queries when you type words into the address bar. This is industry standard. See https://twitter.com/BrendanEich/status/1273327455105773568?s....

The bug in Brave was a flag set wrong for two domain names, which are not keywords: binance.us and binance.com. These should not have default-completed when typed into the address bar with a Binance affiliate code, they should have been suggested completions the user would have had to pick by arrow down and <return> or mouse equivalent. We fixed the defaulting bug and turned off the entire suggested completions feature. As noted by someone else in a comment nearby, we instructed Binance not to pay us for any referred new users who traded (Binance does not pay on the referral, they pay only after if the user trades, sharing a fixed proportion of trading fees).

So this was not "link hijacking" in any case. I'm sorry for the blunder, and we added a process step around any changes to address bar to audit harder.

Brendan Eich mentioned it was a default completion bug they got no revenue from, which was fixed:

https://twitter.com/BrendanEich/status/1270128401760743424

Please see https://brave.com/referral-codes-in-suggested-sites/ regarding this claim. It's important to note that Brave never hijacked links, modified pages, our injected codes into content. The browser offered a pre-search list of suggestions for a small set of keywords (see blog post for screenshots). Happy to answer any questions you may have beyond the contents of that blog post. Nothing malicious here; no data or privacy impact either. We were able to fix the behavior within 48 hours (IIRC), and burned the associated affiliate code.

> “Show Brave suggested sites in autocomplete suggestions” setting’s default to “off”

This may be the change I needed to know about to try Brave at some point.

Are there any plans to get into any Linux distro's repositories? Brave wasn't in Ubuntu's last time I checked, although that was a little while ago.

E: Formatting doesn't like asterisks.

I just added it to an Ubuntu (actually Kubuntu), very easy instructions, and familiarly default too - https://brave.com/linux/.

One slight gotcha, if you view the link from Tor it offers an .onion site for the apt repos string, but I wanted the regular repos as I don't use OS-level Tor but happened to be using tor-browser.

Peter Thiel (Palantir) is one of their main investors. that should raise all alarm bellas for a product that advertises itself with "privacy"

Peter Thiel is not personally an investor in Brave. The firm he's a partner in (Founders Fund) has a small seed series investment, which was led by Cyan Banister. Some first floor engineers at Brave have more equity than that small investment constitutes. Repeating falsehoods once you know they are false is lying, so please don't.

> Some first floor engineers at Brave have more equity than that small investment constitutes.

Is "first floor engineers" an SV-ism (or Brave-ism) that indicates something like "early employees"?

It's wider than SV, startup folks anywhere know it. Yes, early employees.

Palantir exists to improve privacy of citizen data within the government. Without it, government workers have massive, untraceable and unaccountable access to private data. (See Snowden for more info.)

I don't like that there's even a need for Palantir, but given the need, I'm glad it exists and I'm glad someone like Thiel is behind it.

I very, very much do not buy this at all.

I don't honestly care all that much about Palantir specifically (pros and cons, though lots of cons), but one thing they do not do by simply existing is help citizens protect their data. They do a lot of things, but not that.

That's just marketing/PR nonsense.

Instead we give that access to Palantir workers? How is that any better?

> Palantir exists to improve privacy of citizen data within the government.

Huh. So the argument to overcome objections to a

Well, that's annoying. Here is what I meant to post, more or less:

Huh. So the argument to overcome objections to a system designed for ingesting, canonicalizing, normalizing, and correlating private data acquired via dubious means from dubious sources is that at least the access to and use of the data is controlled monitored and auditable, reducing the incidence of LOVEINT and similar abuses as compared to ad-hoc systems constructed on the sly?

That's really damning with faint praise, especially since Jevon's Paradox ensures that both the use AND abuse of such data will increase when you reduce the 'costs' of doing so, and that's assuming there is no unsanctioned/off-book/'black' use or egress of the data (which is far from assured, IMO, given what we know about the history of these systems and projects, eg. Total Information Awareness and Trailblazer vs. ThinThread).

Which blog post?

Probably the one by the Mozilla CEO, "We need more than deplatforming"[1]

1: https://blog.mozilla.org/blog/2021/01/08/we-need-more-than-d...

That headline is certainly inflammatory, but I don't think the actual content of the post is nearly as bad - they're not suggesting web browser start flagging/blocking sites or anything similar like it initially made me think.

They're simply suggesting further transparency in algorithmic suggestions and research be conducted. You could even argue that they're suggesting that deplatforming simply isn't the ultimate solution.

The conclusion doesn't follow from the premise. They take the banning of Trump and conclude that there needs to be transparency in advertising? I mean, I agree, but it's certainly a weird article. I'm not sure if they're just piggybacking on a hot issue and inflammatory title to push this message. Whatever it is, I don't like it very much.

You're right that the content itself doesn't suggest Firefox doing something troubling, but the headline certainly does. It bothers me much more from a company known for privacy than it would from someone else.

How is "more transparency from platforms" not in line with what you'd expect from Mozilla? It's not like they just suddenly started to talk about the topic.

The GP specifically said the headline.