🐈‍⬛David Sommerseth

F/OSS hacker, mostly working on #OpenVPN
- speaks only for himself.

"Don't aim to be someone. DO something."

#nobridge - because I believe in the real #fediverse, and I don't want my own views/data to be abused by yet another "closed-service which can do whatever it wants for profit".

**BEWARE:** Someone has created a Twitter profile in my name:
https://twitter.com/DavidSommerseth - this is ***not*** me

**If you want to follow me**, you now **MUST** have some content on your profile where we have some common ground on interests. I will no longer accept random profiles wanting to follow with no toots or no other follows or followers in the same interest sphere.

Contact: https://david.sommerseth.email/
PGP Fingerprint: 690D B606 E838 182F A8F9 018C 755A 3AB9 4530 7622

Public Key

npub16undjuczsmkhvdzy3rvxglyvmc9vkjh5p64me8efzv4knetkhvgqds2f6s

NIP-05 Address

[email protected]

Profile Code

nprofile1qqsdwfkewvpgdmtkx3zg3kry0jxduzktft6qa2aunu53x2mfu4mtkyqpz3mhxue69uhhyetvv9ujuerpd46hxtnfduqs6amnwvaz7tmwdaejumr0ds578hdj

Publishing to

relay.damus.io

nos.lol

Author Public Key

npub16undjuczsmkhvdzy3rvxglyvmc9vkjh5p64me8efzv4knetkhvgqds2f6s

Show more details

Published at

2026-07-24T11:11:48Z

Kind type

0 Metadata

Event JSON

{ "id": "a084f8052f3167b6b4b336d2f3743f3e8706a9e443d005ce76e4256e91fc008a", "pubkey": "d726d9730286ed76344488d8647c8cde0acb4af40eabbc9f29132b69e576bb10", "created_at": 1784891508, "kind": 0, "tags": [ [ "t", "fediverse" ], [ "t", "openvpn" ], [ "t", "nobridge" ], [ "proxy", "https://infosec.exchange/users/dazo", "activitypub" ], [ "L", "pink.momostr" ], [ "l", "pink.momostr.activitypub:https://infosec.exchange/users/dazo", "pink.momostr" ], [ "-" ] ], "content": "{\"name\":\"🐈‍⬛David Sommerseth\",\"about\":\"F/OSS hacker, mostly working on #OpenVPN\\n- speaks only for himself.\\n\\n\\\"Don't aim to be someone. DO something.\\\"\\n\\n#nobridge - because I believe in the real #fediverse, and I don't want my own views/data to be abused by yet another \\\"closed-service which can do whatever it wants for profit\\\".\\n\\n**BEWARE:** Someone has created a Twitter profile in my name:\\nhttps://twitter.com/DavidSommerseth - this is ***not*** me\\n\\n**If you want to follow me**, you now **MUST** have some content on your profile where we have some common ground on interests. I will no longer accept random profiles wanting to follow with no toots or no other follows or followers in the same interest sphere.\\n\\nContact: https://david.sommerseth.email/\\nPGP Fingerprint: 690D B606 E838 182F A8F9 018C 755A 3AB9 4530 7622\\n\",\"website\":\"https://infosec.exchange/@dazo\",\"picture\":\"https://media.infosec.exchange/infosec.exchange/accounts/avatars/109/593/468/354/365/379/original/b6da78132916c2e4.png\",\"banner\":\"https://media.infosec.exchange/infosec.exchange/accounts/headers/109/593/468/354/365/379/original/6f7de23acf6f498d.jpg\",\"nip05\":\"[email protected]\"}", "sig": "60dca6e3d40c84051af4480b947a63a3c3cd7a83fd3d055a60eb0c7bec724aea0ec56d2a5bc673029b289d4e1d022555c61b9afaef9ec2537f6e91b0d27cf0d9" }

Last Notes

2026-06-07 16:59:49 UTC

- reply

Wow! I didn't know it was so strict on receiving donations. You anyway need to do what's needed to stay legal. I wasn't aware of this, so it sounded like a lot more work and overhead for a hobby project.

2026-06-07 16:30:38 UTC

- reply

I believe you need to figure out what solves the short time situation and what makes your instance sustainable.
If that is best done through creating an organisation or foundation handling all the costs and receiving contributions, that is a hard question to answer for anyone not being directly involved in it. Many people will hopefully join in, but I suspect most people would care less if it is arranged "this way" or "that way" as long as you have credibility amongst your users.
My point is ... Do what your 917 users would trust most **and** which gives you least administrative management overhead. You and your team is first of all serving the people behind those 917 accounts.

2026-05-25 21:24:13 UTC

I always remap my sshd daemon to listen to a non-standard port, to reduce a lot of noise. Which has worked fine for years. But every now and then there are attempts. All the #Linux kernel flaws found lately has made remote login attempts more interesting for attackers. And they scan much more broadly now than just port 22.
And that's why my second line of defence is to disallow remote root login - and also make use of the AllowGroups feature in sshd_config. Users granted remote access **must** be member of a specific group. And root is also excluded from this group.
That pays off these days. And this is a nice filter match for #fail2ban and similar tools
https://termbin.com/0cf6
I have 293 login attempts on "random users" since May 21. And 259 attempts as root.
#infosec #ssh #sshd #systemhardening #kernel

2026-05-21 09:07:38 UTC

- reply

Hah! That's an interesting project. I have been looking for TB replacements for many years - but not found anything really convincing me so far. I might give this one a spin.
GNOME Geary wasn't that bad at first glance - but way too simple for my taste. Claws was just too much 1995-look and way too many tweakble knobs everywhere. Claws was closest to make me switch; but that was years ago now.
To convince me, it need to do PGP seamless, and support for more IMAP/SMTP accounts + CalDAV and CardDAV support (using EteSync). It need to be able to handle lots of mail folders (I use server-side mail filtering heavily), handle huge mail folders and have threading view (I heavily dislike "conversation view" with passion). Also need to have support for multiple identities (have lots of them as well, with different SMTP servers per identity). And of course, keyboard navigation (I use Nostalgy++ for more useful keyboard tweaks).

2026-05-12 21:49:26 UTC

- reply

Reminds if of an advice I got ages ago ... "Sometimes it's better to fight the bugs you know, rather than to upgrade and fight the unknown".
The argument was not intended to discourage upgrades, but to encourage testing and auditing upgrades first - so you hopefully can have a better understanding of the unknown.
But testing and auditing hasn't exactly become easier.

2026-05-02 07:58:12 UTC

- reply

And during those 6 years ... Which alternatives has arrived which has been able to get as much traction as PGP and be a fully viable alternative to it?
Also, I frankly think it is a bit disingenuous to call out GPG instead of OpenPGP. The latter is the protocol/standard. GPG is a short form of GnuPG, being an implementation of OpenPGP. And PGP was both referred to as a prior name for OpenPGP and the reference implementation of it.
GPG surely is a user unfriendly implementation in so many ways, and doesn't do much good to make OpenPGP more understandable. Sequoia-PGP and Tumpa both tries to improve that.
But it is currently impossible to ignore OpenPGP. It is still needed in many places, and is still understood good enough on many conceptual levels - which is why it is still widely used. Plus OpenPGP still evolves and improves as well. Where GnuPG has also been seen as dragging its feet to keep up with improvements, which certainly doesn't really help.
GnuPG is finally getting some real competition. And there are now viable alternatives which can "fill its shoes" better.

2026-04-30 10:05:49 UTC

All who use GnuPG gpg, and especially for signing git work ... give Tumpa-cli a real run and report back to @npub18he…8d3x ! This project works very well already, and is just a few weeks old.
https://github.com/tumpaproject/tumpa-cli
I did my install using cargo build --release which gets you started super easy.
Now we just need to convince Kushal to move to a better hosting place! What do you think, @npub1fh7…ee83 😁 😉
#PGP #OpenPGP #GnuPG #gpg #git #Rust

2026-04-30 09:57:09 UTC

- reply

I must say it's a true blessing to use the tcli agent!
This one does indeed cache the password/PIN as it should.
Thanks a lot for great work! Looking forward to see how this moves forward further 🙂

2026-04-29 20:42:20 UTC

- reply

Now it works nicely! 😉

2026-04-27 08:43:48 UTC

- reply

This one took me by surprise 😁 <code>Error: signing failed: inline (cleartext) signing requires a software secret key [...] card-only keys are not supported — use detached signing instead </code>
But it is truly refreshing to see a CLI command with a UX possible to understand without digging into man pages and searching through a million disorganised options. Which is also why I've been drawn towards oct and sq. tcli looks equally convincing currently.
What those lacks are a drop-in replacement such as the tclig ... so I believe you're unto something here.
Will test out the agent as well, to see if that can ease my current pain with git, gpg and the not-cached card PIN.

2026-04-27 07:56:13 UTC

- reply

Did the less lazy approach now .... built it fine from source 😉

2026-04-26 21:52:58 UTC

- reply

I use PGP for git commit signing, using GnuPG and subkey on a Yubikey.
I tried to test your AppImage distribution, but that failed to start on RHEL-10 due to dependency issues. Not had time to explore building from source.
My RHEL-10 setup might not be optimal (upgraded via an unsupported path from RHEL-10 Beta), so there might be some issues related to that too. Ny main gripe now with GnuPG is the complete lack of password caching for Yubikey operations requiring the private key.
I've tried oct too, same issue there.
That said, I prefer the Sequoia-PGP cli over gpg; Sequoia is the sanest to option use between the two of them.

2026-04-20 15:43:40 UTC

- reply

How is this different from Sequoia PGP (except simpler to spell) and OpenPGP Card Tools (oct)
https://sequoia-pgp.org/
https://lib.rs/crates/openpgp-card-tools

2026-03-12 16:17:51 UTC

- reply

I don't know exactly where you got Proton's claim from ... but this is what they say in their support section:
> A kill switch is a security feature that protects your IP address in case you unexpectedly lose the connection to a Proton VPN server. In case the connection is interrupted, a kill switch blocks all external network traffic to and from your device *> *until the connection is automatically re-established to the same VPN server. **
(my highlight)
source: https://protonvpn.com/support/what-is-kill-switch
This is generally how kill switch works basically everywhere. As it doesn't tear down the virtual network interface (including the redirect routes, routing your internet traffic via the VPN) until it has established a new connection to the remote server.
What is being demonstrated here is that the user implicitly *disconnects* the connection before starting to connect to the new chosen server.
A more accurate test for kill switch is to block the network traffic for the amount of time it takes for the VPN client to start "recovering" the connection by establishing a new connection *to the same* server. That new server may have a different IP address.
The reason this isn't possible to achieve when switching servers completely is that the encryption certificates would result in a mismatch, thus tearing down the connection as the client wouldn't be able to identify if it's a man-in-the-middle (MITM) attack happening, with a host trying to impersonate the real VPN server it should be connected to.

2026-03-02 09:35:37 UTC

This is a follow-up post on the [sad state of Mozilla](https://infosec.exchange/@dazo/116133378756919378 )
First, notice the date of the commit identified (as highlighted in a few posts below this one).
Secondly, Mozilla has done further changes to their Privacy policy since this initial change. I am not fully convinced about them - since the Privacy FAQ at the same time is not aligned. The **reason** for my continued mistrust to Mozilla is that they have gradually, over many years, moved in a direction I do find privacy unfriendly. And they have ties/agreements/contracts/partnerships to companies who does not have a good track record on privacy topics. I generally trust people and organisations *actions* more than their *words* of what they want to do.
Thirdly, it should be fairly clear to most that AI/LLM is not preserving privacy well when data is sent to a remote server to be processed there. And even running parts of the LLM engines locally does not fully disentangle the privacy aspects fully - data is still being exchanged with a remote server (otherwise there would not need to be "AI service provider URLs" in about:cofig). Mozilla did force AI/LLM unto users, enabled by default with the only way to disable that in the beginning via about:config. And it took several releases before more user friendly approaches to disable it arrived. Due to this delay, I really wonder "does these new knobs really fully disable AI/LLM?". I have that doubt, because of how Mozilla has behaved over many years.
On top of this, the Mozilla leadership is extremely well paid while they have reduced their engineering teams working on Firefox and other products. That is a too strong indication for me to ignore, that profit and leadership compensation seem to be way more important than the core mission of making Internet a better place.
I have little trust in Mozilla for the time being. And I doubt I'm alone, due to the traction this toot thread triggered. Currently, I believe trust can be built up again. But it will take a lot of efforts now to repair what has been broken. For that to improve for me, I will need to see a lot of actions from Mozilla, where they clearly does changes in the whole organisation and communicates them clearly and that the communication is aligned across all aspescts - including policy documents, FAQs, source code. Until that happens, I will use some of the Firefox forks. And leadership compensation need to be completely transparent and come down to a level which is not in an astronomic level comparable to large for-profit enterprise companies who generally cares little for anything than their own egoistic wealth.
If a person taking a leadership role in an organisation claiming for a better Internet and fighting for its users is getting uninteresting unless there is a million dollar yearly compensation when the people doing the grunt work, delivering code resulting in a real product, has a 5th or 10th of that compensation, then I **do** question the values this person holds. And I will especially highly question the leadership when they need to reduce cost and choses to cut among the engineers doing the grunt work while the leadership not considering their own compensation.
So basically, I find the Mozilla organisation fairly rotten currently. It preaches the nice words but ends up doing something completely different.
#mozilla #firefox #privacy #foss #opensource

2026-02-25 21:39:17 UTC

- reply

#Mozilla has lost their ground and is now in a free fall into a sinkhole. I doubt they'll ever get out if this again unless they do a 180-turn within the coming days. Mozilla has lost a lot of trust and credibility over the last couple of years. This accelerates that distrust even more.
https://blog.mozilla.org/en/firefox/update-on-terms-of-use/
It looks promising, until you hit the last paragraph (my highlight)
> In order to > **make Firefox commercially viable**> , there are > **a number of places where we collect and share some data with our partners**> , including our optional ads on New Tab and providing sponsored suggestions in the search bar. We set all of this out in our privacy notice. > **Whenever we share data with our partners,**> we put a lot of work into making sure that the data that we share is stripped of potentially identifying information, or shared only in the aggregate, or is put through our privacy preserving technologies (like OHTTP).
In my book, that's indirectly **selling data**.
Goodbye, #Firefox.
#privacy #ads #foss #opensource #web

2026-02-25 21:33:20 UTC

- reply

The discussions are heated ...
https://connect.mozilla.org/t5/discussions/information-about-the-new-terms-of-use-and-updated-privacy/m-p/87735
https://connect.mozilla.org/t5/discussions/an-update-on-our-terms-of-use/m-p/88320

2026-02-25 21:25:29 UTC

This is sad 😢
https://github.com/mozilla/bedrock/commit/d459addab846d8144b61939b7f4310eb80c5470e#diff-a24e74e4595fa85440a2f4e7e5dcfe68aba6e1e593aef05a2d35581a91423847
#firefox #privacy #mozilla #foss #opensource #web
https://media.infosec.exchange/infosec.exchange/media_attachments/files/116/133/372/441/577/988/original/5005aed288333386.png

2026-02-25 18:48:46 UTC

Just read @npub1p5z…krfz's experience over a longer time span with how the government in #USA, through #FBI are attacking press freedom deliberately. It makes me angry how #Trump reverted protections the Biden administration put in place after the the first Trump administration.
The press freedom must be fought for by all means everywhere. There are no chance to have an honest government without a free and independent press. And what happens in the #US now is just terrifying.
#press #pressfreedom #politics #uspol
https://this.weekinsecurity.com/fbi-agents-visited-my-home-about-an-article-i-wrote-and-now-i-cannot-go-to-mexico/

2026-02-24 07:20:53 UTC

I'm checking out @npub1f8e…63tv on RHEL-10 ....
This isn't fine. This is high-level obscurity <code>$ rpm -q --scripts jotta-cli preinstall scriptlet (using /bin/sh): #!/bin/sh set -e mkdir /tmp/alien.$$ perl -pe '$_=unpack("u",$_)' << '__EOF__' > /tmp/alien.$$/script M(R$@+W5S<B]B:6XO96YV(&)A<V@*"G-E="`M=0H*=F5R8V]M<"`H*2!["B`@ M:68@6UL@)#$@/3T@)#(@75T*("!T:&5N"B`@("!R971U<FX@,`H@(&9I"B`@ M;&]C86P@2493/2X*("!L;V-A;"!I('9E<C$]*"0Q*2!V97(R/2@D,BD*("!F M;W(@*"AI/21[(W9E<C%;0%U].R!I/"1[(W9E<C);0%U].R!I*RLI*0H@(&1O M"B`@("!V97(Q6VE=/3`*("!D;VYE"B`@9F]R("@h:3TP.R!I/"1[(W9E<C%; M0%U].R!I*RLI*0H@(&1O"B`@("!I9B!;6R`M>B`D>W9E<C);:5U](%U="B`@ M("!T:&5N"B`@("`@('9E<C);:5T],`H@("`@9FD*("`@(&EF("@h,3`C)'MV M97(Q6VE=?2`^(#$P(R1[=F5R,EMI77TI*0H@("`@=&AE;@h@("`@("!R971U M<FX@,0H@("`@9FD*("`@(&EF("@h,3`C)'MV97(Q6VE=?2`\(#$P(R1[=F5R M,EMI77TI*0H@("`@=&AE;@h@("`@("!R971U<FX@,@h@("`@9FD*("!D;VYE M"B`@<F5T=7)N(#`*?0H*86-T:6]N/21[,3HM?0II9B!;("UZ("1[86-T:6]N M?2!=.R!T:&5N"B`@97AI="`P"F9I"FEF(%M;("(D>V%C=&EO;GTB(#T](")I M;G-T86QL(B!=73L@=&AE;@h@(&5X:70@,`IF:0H*=F5R<VEO;CTD>S(Z+7T* M"FEF(%L@+7H@)'MV97)S:6]N?2!=.R!T:&5N"B`@97AI="`P"F9I"@IV97)C M;VUP("1[=F5R<VEO;GT@(C`N,3$N-#DT.#<B"@II9B!;6R`B)#\B(#T]("(R M(B!=73L@=&AE;@h@('!R:6YT9B`B270@;&]O:W,@;&EK92!Y;W4@87)E('5P M9W)A9&EN9R!F<F]M(&$@=F5R<VEO;EQN(@h@('!R:6YT9B`B;V8@:F]T=&%D M('1H870@<G5N<R!I;B!S>7-T96T@<VQI8V4N(%1H:7,@=F5R<VEO;EQN(@h@ M('!R:6YT9B`B<G5N<R!I;B!U<V5R('-L:6-E(&%N9"!T:&5R96)Y(&1A=&%D M:7(@:&%S(&-H86YG960N7&XB"B`@<')I;G1F(")296%D(&UO<F4@=&AE<F4Z M(&AT='!S.B\O9&]C<RYJ;W1T86-L;W5D+F-O;2]E;B]A<G1I8VQE<R\Q-#8Q M-38Q+7)E;&5A<V4M;F]T97,M9F]R+6IO='1A8VQO=60M8VQI7&Y<;B(*("!P M<FEN=&8@(E1O(&AE;'`@>6]U(&UI9W)A=&4@>6]U<B!S971T:6YG<RP@=V4@ M<')O=FED92!T:&5<;B(*("!P<FEN=&8@(G-C<FEP="`O=7-R+V)I;B]M:6=R M871E7VIO='1A9%]S971T:6YG<R!T:&%T('EO=5QN(@h@('!R:6YT9B`B8V%N M(')U;B!A9G1E<B!Y;W4@:&%V92!U<&=R861E9"!T:&ES('!A8VMA9V4N7&XB ,"F9I"@IE>&ET(#`* __EOF__ chmod 755 /tmp/alien.$$/script /tmp/alien.$$/script "$@" rm -f /tmp/alien.$$/script rmdir /tmp/alien.$$ postinstall scriptlet (using /bin/sh): #!/bin/sh set -e mkdir /tmp/alien.$$ perl -pe '$_=unpack("u",$_)' << '__EOF__' > /tmp/alien.$$/script M(R$@+W5S<B]B:6XO96YV(&)A<V@*"G-E="`M=0H*0D%32#TD*&-O;6UA;F0@ M+78@8F%S:"D*"G-Y<W1E;61?97AE8U]A<R@I('L*("`@('5S97(])#$*("`@ M('-H:69T"B`@("!U<V5R7WAD9U]R=6YT:6UE7V1I<CTB+W)U;B]U<V5R+R0H M:60@+74@(B1U<V5R(BDB('Q\(')E='5R;B`Q"B`@("!S=2`M+6QO9VEN("(D M=7-E<B(@+7,@(B1"05-((B`M8R`B6$1'7U)53E1)345?1$E2/21U<V5R7WAD M9U]R=6YT:6UE7V1I<B`D*B`R/B`O9&5V+VYU;&PB"GT*"G-Y<W1E;61?=6YI M=%]A8W1I=F5?9F]R*"D@>PH@("`@=7-E<CTD,0H@("`@<V5R=FEC93TD,@h@ M("`@8V]M;6%N9"`M=B!S>7-T96UC=&P@)CX@+V1E=B]N=6QL("8F('-Y<W1E M;61?97AE8U]A<R`B)'5S97(B(")S>7-T96UC=&P@+2UU<V5R("UQ(&ES+6%C M=&EV92`D<V5R=FEC92(*?0H*4UE35$5-1%]215-405)4141?4D5$25)%0U1/ M4CT*<WES=&5M9%]R97-T87)T7VEF7V%C=&EV92@I('L*("`@('5S97(])#$* M("`@('-E<G9I8V4])#(*("`@(&EF('-Y<W1E;61?=6YI=%]A8W1I=F5?9F]R M("(D=7-E<B(@(B1S97)V:6-E(CL@=&AE;@h@("`@("`@('-Y<W1E;61?97AE M8U]A<R`B)'5S97(B(")S>7-T96UC=&P@+2UU<V5R(')E<W1A<G0@)'-E<G9I M8V4B"B`@("!F:0I]"@IS869E7W)E<W1A<G1?<WES=&5M9%]S97)V:6-E<R@I M('L*("`@('=H:6QE(')E860@+7(@<&ED.R!D;PH@("`@("`@(&EF(%L@+7H@ M(B1P:60B(%T[('1H96X*("`@("`@("`@("`@8V]N=&EN=64*("`@("`@("!F M:0H*("`@("`@("!U<V5R/2(D*'!S("UO('5S97(]("UP("(D<&ED(BDB"@h@ M("`@("`@(&EF(%L@+7H@(B1U<V5R(B!=.R!T:&5N"B`@("`@("`@("`@(&-O M;G1I;G5E"B`@("`@("`@9FD*"B`@("`@("`@<F5S=&%R=%]I;G-T<G5C=&EO M;G,](E)E<W1A<G0@:F]T=&%D(&UA;G5A;&QY(&)Y(')U;FYI;F<@)W)U;E]J M;W1T860G(&%S("1U<V5R+B(*"B`@("`@("`@:68@(2!S>7-T96UD7W5N:71? M86-T:79E7V9O<B`B)'5S97(B(")J;W1T860N<V5R=FEC92([('1H96X*("`@ M("`@("`@("`@96-H;R`B:F]T=&%D(&YO="!R=6YN:6YG('9I82!S>7-T96UD M(&9O<B`D=7-E<BXB"B`@("`@("`@("`@(&5C:&\@(B1R97-T87)T7VEN<W1R M=6-T:6]N<R(*("`@("`@("`@("`@8V]N=&EN=64*("`@("`@("!F:0H*("`@ M("`@("!E8VAO(")!=71O<F5S=&%R=&EN9R!J;W1T860@=FEA('-Y<W1E;60@ M9F]R("1U<V5R+B(*("`@("`@("!S>7-T96UD7V5X96-?87,@(B1U<V5R(B`B M<WES=&5M8W1L("TM=7-E<B!D865M;VXM<F5L;V%D(@h@("`@("`@('-Y<W1E M;61?<F5S=&%R=%]I9E]A8W1I=F4@(B1U<V5R(B`B:F]T=&%D+G-E<G9I8V4B M"B`@("!D;VYE(#P\/"`B)"AP:61O9B`O=7-R+V)I;B]J;W1T860@?"!T<B`G L("<@)UQN)RDB"GT*"G-A9F5?<F5S=&%R=%]S>7-T96UD7W-E<G9I8V5S"@h` __EOF__ chmod 755 /tmp/alien.$$/script /tmp/alien.$$/script "$@" rm -f /tmp/alien.$$/script rmdir /tmp/alien.$$ </code>
#cloud #cloudstorage #linux

2026-02-12 14:13:20 UTC

- reply

such a policy (which I don't mind at all!) is kinda ironic on masto.ai .... 😉

2026-02-12 08:33:23 UTC

- reply

I'm very new to testing out Radicle - so I probably don't have the full overview yet.
I would expect any thing using git to work fine. What is different is that rad init adds a new rad remote where you push and pull from.
Interacting with the rad remote triggers the pushing of data to/from Radicle nodes in the Radicle network.
An important detail though, anything you push to the Radicle network marked as a **public** repository, will be visible across all the seed nodes interacting with your repository. And **there are no way** (currently?) **to delete or remove a public repository**. Making it private will only stopp updating the public "copies" of your data, it will never completely disappear.
If you create a private repo in Radicle, you will then need to grant access to specific Radicle nodes which will seed your repository, otherwise the data will not ben pushed anywhere outside your local machine. And due to the nature of Radicle, those selected nodes need to be nodes you fully trust.
In addition, IIUC, these selected nodes with access to your repository need to have direct access to each other (think p2p / peer-to-peer). This is easily achieved installing tor and configuring a "Hidden Service" for your nodes, which simplifies the firewalling and network routing across nodes on different networks.
If I've misunderstood anything, please correct me!

2026-02-05 21:01:41 UTC

- reply

Ahh, alright. It's not a bad idea. Might even be possible to (ab)use DNS records for such lookups too. But then you'd need to know a domain as well, unless there is a 1:1 relation between git author/committer e-mail address and the Radicle alias name.

2026-02-05 20:33:22 UTC

- reply

A question about your radicle URL with ~defelo ... How does that work? Something you configured yourself on your web instance, or something I can do with my Radicle identity?

2026-02-05 18:13:39 UTC

I've been doing some initial testing with @npub1h39…qfkm ... https://radicle.xyz/
First thoughts ... <li>it's an interesting idea, and I really like the concept.</li><li>It's a lot of new concepts I need to get used to, learn a lot of new terminology</li><li>I need to learn stop comparing more directly against how things works in Codeberg/GitLab/GitHub</li><li>it's pretty fast after all</li><li>Searching for existing projects on radicle seems to be an impossible task or based on pure luck. The "random" <code>rad:z3g.....</code> strings identifying a repository (needed for <code>rad clone</code>) are even worse to remember than IPv6 addresses. I'm missing a "DNS" equivalent for looking up those RID values.</li><li>There seems to be more URLs out there pointing at documentation for various versions. Several of them are completely outdated and examples no longer work. In other words: Confusing documentation for newcomers.</li>
I have so much to learn, and need to sit down and read the docs way more carefully. Just need to find the right ones.
#programming #source_code #version_control #vcs #opensource #foss #git #radicle

2026-01-22 13:20:42 UTC

- reply

I'm probably moving back to @npub1vpu…608s. I usually close all my tabs by default, so didn't fall into your trap.
My complaint is that @npub1c7u…lvlq is resource hungry and makes my machine sluggish.
Maybe I should give Floorp a test run as well. Not being Chromium based, privacy focused and not pushing AI into my face are my base requirements.

2025-10-02 13:22:53 UTC

- reply

*sukk* Ja.
Norgespris er den klassiske X-Y problemløsningen. Vi har problem X og løser den med en annen løsning Y. Som til syvende og sist ikke løser noen ting. Bare skyver problem X foran seg.
Norgespris vil bli dyrt for Norge. Og løser ikke utfordringen med kraftfordelingen som er for svak i Norge. Og den regninga som kommer etterhvert blir enda dyrere.
Hadde politikerne med bidratt til å bedre strømfordeling i landet - ja, så vil også prisforskjellene også fordele seg. Baksiden er at man får høyere priser der de er lave i dag (fordi det er for mye kraft tilgjengelig i regionen) - og derfor blir dette en død plan for politikerne, spesielt i et valgår.
Men samtidig vil prisen vil gå ned i regionene som får tilgang til mer kraft (fordi underskuddet av kraft blir bedre dekket opp fra regioner med overskudd).
Hva får vi for denne "dyrere" strømmen ... Jevnere priser mellom regionene og et mer robust strømnett. Dette er åpenbart for dyrt for politikerne, som i stedet løser dette med å bruke penger på betale forbrukerne for den dyre strømmen de ellers må betale for. Og det er alt vi får igjen for den statlige pengebruken.
Norgespris er mer som om politikerne roper ut ut: «Oj, SEEEE! En rosa elefant!"» når landet skriker etter bedre fordeling av kraften i landet.
Man kaster penger ut av vinduet uten få noen ting igjen for det. Bokstavelig talt pisser i buksa for å holde seg varm en vinternatt.
Det er så idiotisk.
Men nå har vi Norgespris - og for mange så er det like greit å ta den avtalen og få litt penger igjen fra staten. De kaster jo bokstavelig talt de pengene mot oss forbrukere, så hvorfor ikke ta dem imot?

2025-09-11 19:53:36 UTC

- reply

Ikke uenig i målet her. Det ville vært vakkert og flott. Og ja, vi ville ikke angret om 50 år. Tvert i mot.
Men hvor skal vi ta de pengene som trengs fra? Joda, klart vi har masse penger i Norge - med "oljepengene". Men hva skjer med den norske makroøkonomien da? Slike grep har ofte en stor negativ konsekvens som dukker opp litt senere.
Det er jo nettopp derfor økonomer er svært bekymret for "budsjettpulsen" i statsbudsjettene. Hvis vi bruker for mye penger i forhold til inntektene (skattene, i praksis), så påvirker dette alt fra inflasjon til kronekurs - og mere til.
Det sies at "å være rik er en like stor forbannelse som å være fattig, om ikke større". Fordi bruken av midlene vi har gir store ringvirkninger - som oppleves senere.
Siden ingen av oss besitter evner til å spå inn i fremtiden ... så blir slike store og kostbare grep ganske risikable - også i et 50 års perspektiv.

2025-08-30 17:35:57 UTC

- reply

Ehm .... This is from the bottom of that HTML site's source view:
🤪
https://media.infosec.exchange/infosec.exchange/media_attachments/files/115/118/916/307/712/937/original/ed32f80c80517a06.png

2025-08-27 16:01:24 UTC

- reply

> A common misconception is that GrapheneOS is anti-Google, which is not true. Its goal is to provide a secure and private operating system; ideology has nothing to do with it.
Thanks for that detail. I will refrain from mentioning GrapheneOS in context with Android and privacy moving forward.
Hardening and security are great and really important; don't get me wrong. However, *privacy* is not the same as *security*. Security have some overlapping features with privacy - but it is not fully overlapping.
It is possible to quite some degree to have security without privacy and vice versa.
For my own thread model, privacy is far more important. Your threat model might be different.

2025-08-27 15:44:28 UTC

- reply

Yeah, that's unfortunate - and they have been working to resolve this. But this is not an issue in CalyxOS itself.
This issue is an organizational challenge where they lacked control mechanisms to revoke access properly and in due time - which affected the build infrastructure CalyxOS is built on.
From the URL you point at:
> First, we want to assure you that we have no reason to believe the security of CalyxOS and its signing keys have been compromised.
> As you know, we announced a recent leadership transition. When senior personnel have access to signing keys and leave a team, it is security best practice to update signing keys and conduct audits. So in accordance with that, we are using this transition period to update our security protocols, including updating the signing keys and taking other steps to further protect our users.

2025-08-27 12:55:34 UTC

- reply

Unfortunately, there are still leakage to Google with #LineageOS ... hence there are other more privacy aware alternatives. Such as [@e_mydata](https://mastodon.social/@e_mydata )
https://e.foundation/about-e/
Other alternatives are #CalyxOS and #GrapheneOS. They take some other approaches. There are pros and cons of all these various alternatives, though.

2025-08-18 11:16:06 UTC

- reply

Det funker iallefall ikke for meg. Så da er jeg nok ikke i målgruppen til MDG.
@npub1r7t…tegu

2025-08-18 11:11:30 UTC

- reply

Ellers kan det kanskje være bedre om MDG snakket mer om politikk enn at de må komme seg over sperregrensa ... at det mer føles som at en stemme går til politikken og ikke å "redde" partiet.
Skjønner veldig godt at sperregrensa er et stort og stygt monster MDG som parti vil holde seg unna ... men det er mindre motiverende for velgerne.
https://www.nrk.no/ytring/stadig-mer-skamlos-velgerjakt-1.17531044
https://media.infosec.exchange/infosec.exchange/media_attachments/files/115/049/459/285/966/537/original/91125b9c843a6712.png

2025-06-20 13:01:46 UTC

- reply

Vipps hadde en tilsvarende greie for noen år siden, men det løste seg etterhvert. Det var før /e/OS kom opp i versjon 1.0. Nå er de på version 3.0.1. Og dette surret dukket opp i dag. Var inne i Vipps i går uten problemer.
TeliaPlay funker ikke. Ellers funker alt annet helt fint. BankID, førerkort, Ruter og bruker flere bank apper - null problemer der. (HBO) Max, TV2 Play, Netflix, NRK appene funker alle fint.
[@vegard_berget](https://tutoteket.no/@vegard_berget ) [@e_mydata](https://mastodon.social/@e_mydata ) @npub1hn8…g0mt

2025-06-20 07:42:41 UTC

- reply

pre-installert på FP4 av @npub1hn8…g0mt.

2025-06-20 06:09:07 UTC

"Takk!" f***********************, j**** #Vipps.
Å **ikke** bruke Google Android betyr **ikke** at telefonen er "rooted".
#Norge #bank #norsktut #allheimen
https://media.infosec.exchange/infosec.exchange/media_attachments/files/114/714/185/545/307/636/original/d974866f2baf60de.jpg

2025-05-30 20:12:29 UTC

- reply

Is there a similar poll on Bluesky too, so the data can be more correlated? Otherwise you end up with a very biased result.

2025-05-13 12:15:29 UTC

- reply

I agree with what you say. All I'm saying is that knowing - or being given a possibility to learn - the language is of tremendous help to becoming more integrated.
Also, culture cannot and should not be forced upon anyone. But the language is an important puzzle to better understand the culture. A lot of our own culture is embedded into how we express things through our language.

2025-05-13 10:00:58 UTC

- reply

Try not to make this worse than it really is. I'm not living in the UK, so I'm not too deep into the full political situation there except of what is being reported abroad.
For any individual living in a foreign country, it is really valuable to know the language. It helps understanding the culture of the country you live in. It helps integrating you into the society and makes it easier to get to know more people - like your neighbors. It makes it easier to get a job.
There are a tons of reasons learning the language of the country you want to move to.
For people not doing that, they often end up isolated in the country. If there are many of them, they typically move to live close to each other - which **is** understandable. But this is where the beginning of frustration grows more quickly, resulting in more segregation and social issues across the whole region slowly grows. Combine that with lack of job possibilities and the crime starts to rise
Language is key to reduce these risks. It's not the solution alone. But to get newcomers to integrate better, language is the first of many steps. And when newcomers gets chance to properly integrate, they have a chance to start contributing back to the community and society.
There are no easy nor simple solutions. There are no quick-fixes here. But it's a large puzzle with many pieces which all need to fit together.
However, f the political goal is to use language knowledge to block immigrants to arrive to UK - I can agree that is highly questionable. But if it is one of more steps to ensure better integration of immigrants in the longer run, it might not be as bad.
Integration is, after all, the most important aspect of a well functional society.

2025-05-10 08:04:23 UTC

- reply

After I read «Silmarilion», I got a broader perspective of who elves were and their history. What we see in LOTR is the continuation of their different perspectives and historical disagreements. What fascinate me is how they in the start have this, how to say it, respectful distant disrespect for each other - across both "elf tribes" but also across dwarfs, hobbits and humans. But through the battles they all fight together on the same side - it seems to mend some of these old historical wounds - and the disrespect turns to admiration and respect for each other.
Each "race" and "tribe" in LOTR has different strengths and weaknesses, which becomes very powerful when collaborating and giving others room to excel where they have their strengths. And they all discover they have more strengths within them than they believe, which they discover through their individual challenges.
I believe our world today could have so much to learn from LOTR. United we stand stronger, no matter where we come from.

2025-04-22 12:50:50 UTC

- reply

De vilkårene der burde jo være lett å akseptere ......... 😉

2025-02-23 00:28:55 UTC

- reply

It almost feels like this is a deepfake video ... I struggle to believe this really happened.

2025-02-20 10:49:42 UTC

- reply

Altså noen som lesser ned et piano med "pynteting" og tilbyr det som pinneved fortjener ikke å være i besittelse av et sånt instrument.
Dessverre sies det ingenting om hvor stemt det er eller hva slags vedlikehold det har hatt ("ingen" er det jeg tipper). Det kan være at det har blitt så forsømt at en uopprettelig skade har skjedd.
Hadde det vært en institusjon for vanskjøtsel av instrumenter - alla barnevernet - så ville disse vært på plass sporenstreks.
Hadde vi hatt plass og mulighet til å å hente det, hadde jeg vært i en bil allerede.
Og ja, for de som lurer: Et **instrument** er IKKE et møbel som man putter pynteting på.

2024-12-28 21:26:16 UTC

- reply

og hvorfor trenger man feltet for kjønn i det hele tatt?

2024-12-19 11:41:56 UTC

- reply

I want to highlight a detail about DNS as well.
All information in the video is correct. However, there would be benefits of using external DNS servers providing encryption, such as @npub12us…h86y
If you use such services over a VPN, it will be very hard for the VPN provider to see what you are visiting - and impossible for your local ISP connection.
But as with VPN service providers, you need to be equally careful selecting a DNS service provider.

2024-12-19 11:39:28 UTC

- reply

Finally a more sane consumer VPN walkthrough!
I do wish though that it would be highlighted that most VPN advertisements and services are about **consumer** VPN services.
There is a real VPN use case where you want to access information inside a network not directly accessible from the public Internet - which is where VPN **is** the right tool. This, and only this, is the real challenge VPNs were designed to solve. Avoiding geo-blocking is merely a side-effect of this.

2024-12-17 12:57:14 UTC

- reply

What would you say is the advantage of running a closed Forgejo instance vs just pushing your git trees to home server via SSH?
I'm also using @npub1fh7…ee83 for public stuff (also donated too!) ... which is why I wonder about the advantage of also running a fully private instance.

2024-12-09 14:13:29 UTC

Hmmm så #Vipps hare lansert et nytt domene ... https://vippsmobilepay.com/
Eller har de? Det lukter #scam og #phishing, spesielt med tanke på de tusenvis av mailene #spam filteret har luket ut dette året.
#norge #norsktut #allheimen

2024-12-06 10:40:36 UTC

- reply

Mastodon will indeed require far more efforts if your life goal is to end up as an influencer. You need to actually make real efforts to get people to follow you.
In contrast to "make something viral once and the algorithms might boost you to users not knowing you".
To me, that's an argument why the algorithmic approach has failed. You never really know what the algorithm might boost - and if what it boosts has real value to the SoMe users.

2024-12-06 10:36:28 UTC

- reply

<li>nobody likes to do anything time consuming nowadays</li>
Uhm ..... doomscrolling would like to have a say here .....

2024-12-01 19:59:01 UTC

- reply

My first ThinkPad laptop was Pentium 3 with had the RAM expanded to 64MiB. Megabyte.
I piece of me dies each time I download an Android app exceeding 64MiB.

2024-11-29 19:48:44 UTC

- reply

The USA need to be flooded with this kind of shows, unfortunately. Hopefully more people there can then experience those "wait a minute?!" moments to reflect on their own rejection of systematic racism issues in the USA today.

2024-11-14 18:10:54 UTC

- reply

Yeah, encryption is something conspiracy theories often fail to account for. But many of the CT believers don't understand the important details and ignore them instead of trying to understand it.

2024-11-11 13:42:26 UTC

- reply

Den butikken legges ned.
Har hørt at den oppvoksende generasjonen i familien som har drevet denne butikken siden tidenes morgen ønsket å gjøre noe annet.
Veldig synd. Men samtidig forståelig hvis man ikke ikke føler kallet å drive butikk selv.

2024-10-28 21:02:03 UTC

- reply

The only full #E2EE e-mail services I've seen so far are Tuta and Proton. cTemplar was a thing some years ago, but is now shut down. These services has E2EE enabled by default and cannot be disabled. And this is why neither provides standard SMTP/IMAP configuration against their servers. Proton provides the mail Bridge, which runs locally and "bridges" the web service API on the server sise with an SMTP/IMAP service on your local device mail clients can connect to.
With "full E2EE", it means the service will try to encrypt all sent mails opportunistically, depending on the recipient e-mail address. And incoming plaintext mails will get encrypted as soon as possible after the spam/malware checks. All account data (including messages) are stored within zero-knowledge policy; the service provider cannot decrypt the data without your knowledge - only you have access to the unlocked private encryption key.
Then there is mailbox.org and Posteo.de. I'm not fully up-to-date on Posteo, but I believe it is similar to mailbox.org in regards to E2EE. Encryption need to be explicitly enabled through extra steps. These services provides standard server side SMTP/IMAP setups. If encryption is enabled, you need to configure PGP yourself in your mail clients. Neither of these services has much PGP key management tooling and you will need to handle that yourself. Proton handles this mostly fully automatic, also for external recipients if their domain support WKD.
Mailfence is anothet popular European mail service, but does not provide any E2EE support like the services above. You can make use of PGP through IMAP/SMTP clients, but anything received and sent unencrypted will always be unencrypted on the server. Most PGP enabled mail clients should support WKD as well these days.
StartMail is also very polular, which is similar to Mailfence and also in the EU region. FastMail as well, but that's an Australian company, where the Australian government has crippled the privacy legislation so much E2EE isn't really realistic; the government can force companies to implement backdoors to accounts.
PGP is the most widely used and available encryption method among all of these, with Tuta being the exception with their own proprietary encryption. Tuta claims they are the only able to encrypt the subject field, which isn't entirely true.
There is a PGP extension to encrypt the subject (and the human readable name in e-mail addresses). Many IMAP/SMTP clients through their respective PGP add-on supports this. Proton webmail can decrypt them, but will not encrypt it yet - as the PGP extension for this is not formalised as a public standard yet. However, using a IMAP/SMTP client and enable SMTP submission (requires a paid plan and can only be used on your own domains) can support this by PGP encrypting the e-mail in your mail client directly.
Tuta's encryption is implemented in a way that it basically mangles e-mail headers completely and exporting e-mails will break DKIM signatures. That's because the e-mail is recreated on each export. Internally, all communication between Tuta users uses an internal proprietary protocol - which also never reaches the Internet at all. That's why most consider Tuta's e-mail service to be more a chat/messenger service with SMTP support (only) for interaction with external users unencrypted.
Tuta (and Proton) does provide an way to send encrypted messages by using pre-shared passwords. But this is essentially just a simple login to the Tuta/Proton service via unique URL and the shared password - so all communication happens only on that platform.
That's basically the most common options you have when trying to avoid the big tech companies.

2024-10-27 19:57:10 UTC

- reply

First of all, the AI and crypto aspects in Proton is overhyped and more trolled than anything else in media and social media. You don't need to use any of them. And while there are fair questions to be asked - AI (aka Proton Scribe) can be use fully local in your browser if you want to use it (no data is sent to Proton) *and* they don't train the AI on users data regardless of local or server side mode in use.
Proton Wallet also exploded by people which either didn't care to read about how it has been implemented or deliberately just wanted throw a shit show for fun. While there are aspects around BTC which are privacy unfriendly, there are aspects around the Proton approach which makes it potentially a bit harder to trace down everything and correlate different transactions. But there's also been some talks about support for more crypto currencies too.
Proton releases new features and products early, where they provide merely a minimum viable feature set - and spend time improving it over time. This may give an impression of half-baked features or slow development. But in fact, over time, they roll out a lot of features gradually.
This year Proton Docs arrived too (only collaborative text editing currently), as an extension to Proton Drive and both Proton Drive and Pass has gained capabilities of sharing with other Proton users (also outside your own domain) in both read/write and read-only mode. Proton Drive Windows got the sharing support very recently too.
Proton Mail is just superior to Tuta in any aspect, where Tuta export is horrendous (had to manually edit lots of issues in the exported .eml files) and lack of import ... It just makes Tuta a very strong lock-in provider. Proton Mail has decent import/export tools. And then the mail filtering in Tuta is still just a silly joke compared to the capabilities Proton Mail provides. Not to mention that a Bridge capability will not arrive for Tuta, according to Tuta.
For the negative stuff on Proton ... Proton Calendar does not have a Bridge alternative (I use EteSync instead, not using Proton Calendar - with a CalDAV bridge, it would be good enough for me and my users) and Proton Drive sync client for Linux is still missing (yes, there is rclone, which kinda works - but that's cumbersome and is dreadfully slow in "mount mode").
Considering that Proton is self-sufficient economically, not depending on VC at all and has grown the service to a lot more than 100mill users and staffing being 500+ today, with their own server infrastructure ... I'd say Proton is a real solid alternative.
Even though I have no interest in Proton Scribe (AI) or Proton Wallet (Crypto currency), I would still highly recommend Proton first, hands down. They have been rock solid for years with little downtime, with a strong commitment to provide solid privacy services, all to a reasonable price - when considering what you get back for the money.

2024-10-03 13:03:22 UTC

- reply

s/things changed dramatically/it went downhill ever since/

2024-09-03 15:54:36 UTC

- reply

I do wonder how this will fly with companies depending on PCI-DSS, HIPAA and similar certifications where their employees screens might contain sensitive data - even for a brief time.