The Intel ones are quite a bit easier, but still not as easy as a PC. You need to disable some FW security settings to allow for a non Apple kernel to boot.

Supernote is the alternative I went with. They have a pretty responsive dev team and the cloud integration is optional, you can push stuff over the local WiFi network.

I returned my Remarkable 2 after a couple of days for a Supernote. Can do local network file transfer, and wifi screen sharing.

This is pretty misleading due to its brevity, an attacker on the same network can determine what website you’re going to but not the content being exchanged. A VPN moves the threat of having your browsing destination determined to the VPN provider from the local network.

That said, modern WiFi encryption does prevent other devices on the network from eavesdropping, so the attacker would have to employ a more involved attack (e.g. ARP spoofing) in order to even see the destinations.

It runs well enough in Windows Subsystem for Android!

I’m surprised not to see https://cryptpad.fr/ here, a FOSS, self-hostable E2EE web based office suite. Not as feature rich as GDocs but offers the basics in a more secure manner.

He has been stepping back from Signal over time.

Do you listen to risky.biz?

I installed INCH on all my browsers, it’s obviously not 100% accurate, but it is nice to get a visual cue that the article you’re reading may very well be AI generated.

While Chromium itself is a very solid platform, and correspondingly Chrome is a hard exploitation target, it’s quite easy to screw up a fork of it. Comodo Secure Browser was a chromium fork that was fixed to an old version of the renderer with known security issues and was built to disable the sandbox. It also added libraries that were compiled without ASLR that worsened security for every application that loaded them.

Chrome has an enormous security team behind it in addition to P0, so bounties on Chrome exploits are around $500k. FF bounties are a fifth of that, which is probably a portion of less security, and a portion of lower target market. Brave could be doing terrible things that without an audit would be unknown. Web3 code is pretty terrible on the whole, so adding that to a secure base may not be great…

While Chromium itself is a very solid platform, and correspondingly Chrome is a hard exploitation target, it’s quite easy to screw up a fork of it. Comodo Secure Browser was a chromium fork that was fixed to an old version of the renderer with known security issues and was built to disable the sandbox. It also added libraries that were compiled without ASLR that worsened security for every application that loaded them.

Chrome has an enormous security team behind it in addition to P0, so bounties on Chrome exploits are around $500k. FF bounties are a fifth of that, which is probably a portion of less security, and a portion of lower target market. Brave could be doing terrible things that without an audit would be unknown. Web3 code is pretty terrible on the whole, so adding that to a secure base may not be great…

I pay for Kagi.com for search, I use NextDNS over my personal Tailscale network that blocks all the commercial social networking sites and their CDNs, as well as a ton of ad networks. I use uBlock Origin in firefox to further remove content that may be served 1st party. Opted out of as many analytics services as I can and frozen my credit with all four US credit agencies. I alternate between using a VPS as a Tailscale exit node, or ProtonVPN for country-specific location egress.

I think it comes down to the threat model that you implicitly or explicitly operate under. Most people don’t think about it, and so they equate “more” with better, and VPNs are easily marketed as more, turn it on and rather what whatismyip.com showing a map near your house, now you’re magically somewhere else!

If you are paranoid about everything, then again there is the “defense in depth” mindset, which in theory couldn’t hurt. That said, having a clear mental model for what you are aiming to be protected from is the best way to find a suitable suite of protections. To agree with a number of others in this thread, ad-blockers (I recommend NextDNS personally) are a great step to stop organizations with a financial incentive to learn all they can about you to sell you stuff, or sell your data. There have been large US ISPs that have experimented with injecting ads or other content either into default DNS responses (e.g., if you mistype something in the search bar it will bring you the ISP’s terribad search portal), or even HTTP responses. If you are stuck with one of those ISPs (I’m sorry, and the US monopolies on ISPs are terrible), then a VPN will help you against your threat (the ISP).

If you are an EU resident, and protected by GDPR (or some of the US states that are enacting similar protections), then moving to a more centralized service can be a good thing, since you have a single place to request data deletion, etc., whereas for a non-EU resident, “smearing” your data over multiple non-coordinating entities is a good move to limit the view of you from any single organization.

If you are worried about government surveillance, you have bigger issues. Most people who want to think they are uber valuable to the government are not, and act in counter-productive ways, but co-mingling their data with that of actual baddies, so it all gets revealed in a warrant search. The Lavabit hosting service was used by extreme privacy wonks, and some actual criminals, and when the government went after Snowden, they got all of Lavabit’s data, so being on that platform may have been counter-productive for people hiding from the G-men. The OPSEC needed for countering government-level is beyond what you’ll learn on a public post, and must be incredibly well-curated and maintained; it will cost you, but if someone will outspend you to get you, then it’s table stakes.

I have a ReoLink camera that streams to my Synology NAS that’s joined to my Tailscale network that I can access from anywhere!

ZipPy is much less robust to defeat attempts than larger model-based detectors. Earlier I asked ChatGPT to write in the voice of a highschool student and it fooled the detectors. The web-UI let’s you add LLM-generated text in the style that you’re looking at to improve the accuracy of those types of content.

I don’t think we’ll ever be able to detect it reliably enough to fail students, if they co-write with a LLM.