@gemlog @lupyuen that only blocks the links, no? BCMA will auto redirect to archive.org: https://git.nogafam.es/deCloudflare/deCloudflare/src/branch/master/subfiles/about.bcma.md -- These still only help those who receive a bad link who also happen to have the special plugin. We still need bad links to be replaced at the source so everyone can benefit.
@gemlog @lupyuen the projectile net is more interesting to see: https://invidious.fdn.fr/embed/5DkEwVHI9zk
@lupyuen @gemlog invidious link to demo video => https://invidious.fdn.fr/embed/ADW63thj-Pg
@gemlog @lupyuen That's a #Cloudflare site. This article is open to all: https://www.thedrive.com/the-war-zone/41093/check-out-the-anti-drone-weapons-carried-by-security-at-biden-meeting-in-brussels
@krock another useful one is ISMM: https://git.nogafam.es/deCloudflare/deCloudflare/src/branch/master/subfiles/about.ismm.md That one will mark up pages you visit to flag CF links, so you know before you click if it's a CF link.
@krock the most useful one is #BCMA: https://git.nogafam.es/deCloudflare/deCloudflare/src/branch/master/addons/releases The bcma.xpi/bcma.crx addon will automatically redirect attempts to visit #Cloudflare sites to #InternetArchive, documented here: https://git.nogafam.es/deCloudflare/deCloudflare/src/branch/master/subfiles/about.bcma.md
@krock i've actually come to /like/ the Cloudflare CAPTCHAs, because if i'm in a browser+profile that doesn't have a CF detector, the CAPTCHA serves as a crutch for profiles where I don't have my shit together, ultimately helping me avoid CF. I just hit control-w when I see the captcha.
@krock Tor Browser has anti-fingerprinting features that no other browsers do (AFAIK), but some browser plugins ruin that. There are also shortcomings with TB, so i do as you are, and use a dozen or so profiles with different browsers and never use it in fullscreen. When anonymity is important, TB without plugins is still king.
@krock My comment about Tor users refusing to connect to CF sites is not driven by the threat of CF de-anonymising them, but rather b/c Tor users become intolerant of the CAPTCHAs, and/or they are simply more aware that CF is seeing everyone's traffic. Normies/non-Tor users are mostly blind to Cloudflare, so they don't even know the compromising position they are in.
@krock Cloudflare can't generally ID Tor users (this is why they treat all Tor users with equal hostility). Exceptionally, if you supply data to a CF site either by logging into hCAPTCHA or logging into the target website, then CF can of course distinguish you from others; as well as by browser printing and the like.
@polarhive @simon if you want to have some fun with them, run xroach without telling them. Cockroaches will start walking around and when they move the mouse the roaches scatter under windows. Then when the move a window, it exposes a few which then bolt under another window. Really authentic and creapy. It's an old app though.. not sure if it's still maintained.
@iron_bug @anniemo71 @gomez @davidoclubb @iron_bug #DDG has /proven/ to violate their privacy policy: http://techrights.org/2021/03/15/duckduckgo-in-2021/ An unknown is of course more worthy of trust than a known cheater. Garage operations just don't have the capital to finance a large scale AI engine.. Only very big money interests behind GAFAM have that kind of cash to invest.
@iron_bug @anniemo71 @gomez @davidoclubb i sense some sympathy for tech giants (as you equated scraping Google/MS results with "stealing"). IMO scraping their results is the best scenario, because it uses the resources of tech giants without compensating them, which means not only not feeding privacy abusers, but costing them. That's a good thing.
@davidoclubb @gomez @anniemo71 @iron_bug Facebook is even figuring out which of their users are gay with very high accuracy. You don't have to worry about searx (FOSS s/w) doing that.
@iron_bug @anniemo71 @gomez @davidoclubb Ss is not searx; it's custom. But wrt searx, no kid in moms basement is financially equipped to exploit data to the extent of #GAFAM. We're talking very costly AI.
I've just worked out how to OCR a color doc: use pdf2djvu twice (once to create a mono doc, once for a color doc); run ocrodjvu on the mono doc w/the --save-script option. Use djvused to run the script on the color djvu file (you'll probably have to hack the script to change the "select" page refs). That's it.. throw away the mono version.
@gemlog @mplammers @democracynow yeah, it's a question that would have to be on a case by case basis. Sometimes the info alone is damning enough and evidence can come from something else.
@mplammers @democracynow @gemlog in the case of Reality Winner, it would have been enough to bilevel the doc.
@gemlog @democracynow @mplammers i'm sure wikileaks won't repeat their very sloppy mishap. If I were wikileaks I would OCR and regenerate from scratch assuming a huge font selection, and perhaps a way to add noise to graphical elements - but drop graphics entirely if nonessential.