Today is the day I start my own #AITeaParty. I am throwing that shit in the harbor and I invite you to do the same.

For reasons, I am documenting multiple generations of birth certificates. I have scanned some and been emailed some. I want them all together in a good place and I started to upload them to my #Paperless install in my docker. Then I thought "Is this fucking thing hooked up to any #AI anything? Will my family's birth certificates become part of a training set if I collect them in this place?"

The fact that I had to go and check before I put sensitive information in a tool under my full control was a wake-up call. I am disconnecting everything from everything because I don't want to ever think that thought again.

I have my #Karakeep hooked up to #OpenAI but I am disconnecting it today. I'd rather not have automatically generated tags than the history of what I find interesting in a training set. They are cute but not worth it.

Generalizing from there: is there any benefit I receive from any generative AI process worth the risks of touching it at all, or of it having any aspect of my data? Fuck no. Today's the day. I've been an indifferent but unmotivated critic. From this second on, I vow to remove it from anything under my control. If it is a third or fourth party risk because the product I use, I will work to extract myself and my data from those products. If your work product was generated in whole or in part with generative AI tools, I don't want it. If you think they are a good idea, I don't want you and your judgement affecting me.

Clearly this is an uphill climb because the percentage of commercial products and freelancers that use #OpenAI or #Gemini or #Grok or #HellscapeAI rounds to everything. It won't be perfect and it won't be easy but I have the time and energy and I'm fired up.

#FuckAI

尝试部署了Jumble和Nostr-Relay，接下来再托管blossom-server作为媒体服务器，全部为不稳定测试使用

相关链接：✅完成 ❌未完成
✅Nostr笔记：https://gist.github.com/KipJayChou/475bb46a87b9b976407e044829c4e8d0
✅Jumble：https://github.com/CodyTseng/jumble
✅Nak：https://github.com/fiatjaf/nak
❌Blossom-Server媒体服务器：https://github.com/hzrd149/blossom-server
❌SoapBox：https://docs.soapbox.pub/
❌strfry中继：https://github.com/hoytech/strfry
❌khatru-pyramid中继：https://github.com/github-tijlxyz/khatru-pyramid
❌Nostr!moe：https://join.nostr.moe/
#jumble #nostr #blossom

接下来再迁移部署些服务
将VSCode软件迁移到code-server了，方便但有的时候还是需要软件
✅code-server：https://github.com/coder/code-server
Ntfy和Gotify也部署好了，链接Gatus工作了快24小时了
✅Ntfy：
https://docs.ntfy.sh/
1. 和Gatus联动：https://github.com/TwiN/gatus?tab=readme-ov-file#configuring-ntfy-alerts
2. 和Grafana联动（测试成功，还需要在grafana配置报警规则）：https://github.com/academo/grafana-alerting-ntfy-webhook-integration
3. 和Beszel联动：https://beszel.dev/zh/guide/notifications/
4. 其他intergrations：https://docs.ntfy.sh/integrations/
✅Gotify：
Github：https://github.com/greenhandatsjtu/gotify
1. 和Gatus联动：https://github.com/TwiN/gatus?tab=readme-ov-file#configuring-gotify-alerts
2. 和Beszel联动：https://beszel.dev/zh/guide/notifications/

#codeserver #ntfy #grafana #beszel #gotify

❌然后是更新Loops到最新版本，不知道能不能解决无法看到Loops.Video主站视频
https://loops.feddit.social
❌从HostDZire迁移Loops-Server（转码CPU需求）到NetCup
❌迁移Karakeep，然后把媒体文件（PDF、HTML）迁移到Cloudflare R2
✅Down掉Readeck
❌迁移Sharkey到NetCup
✅Down掉LinkStack，❌寻找替代品迁移到ClawCloud
对Down掉Funkwhale
✅Down掉Albyhub
❌迁移RSSHub到ClawCloud
❌迁移eddrit并部署其他社交媒体前端https://ghost.feddit.social/protecting-privacy-using-open-source-alternatives-to-social-media-front-ends-part-1 到ClawCloud，并且加上SSO
❌迁移GoToSocial-Main到NetCup

#loops #docker #karakeep #r2 #readeck #sharkey #linkstack #funkwhale #albyhub #rsshub #eddrit #gotosocial #frontend

Finally got #Wallabag working through Docker! For some reason that was completely opaque to me, it was a complete ball-ache to get working with my proxy manager, but finally it works!

I've been using #Karakeep up until now as my article bookmark/read-it-later app, which works really well, but I really wanted to make use of the #KOReader Wallabag integration to read saved articles on my eBook reader.

Was it worth it? Almost certainly not.

Anyway, next on my list to figure out - FitTrackee.

Các bạn dùng dịch vụ tự host nào thường xuyên nhất? Một thành viên chia sẻ top 5 của mình: 1️⃣ SearXNG (search), 2️⃣ Vaultwarden/Bitwarden (password), 3️⃣ Jellyfin (media), 4️⃣ Karakeep (note), 5️⃣ Mealie (recipe). #selfhosted #tựhost #SearXNG #Vaultwarden #Jellyfin #Karakeep #Mealie

https://www.reddit.com/r/selfhosted/comments/1qgcaxi/what_selfhosted_services_do_you_use_the_most/

在iOS/iPadOS上使用karakeep最舒适的方法：删掉官方丑爆&无用的app，使用karakeeper

https://apps.apple.com/us/app/karakeeper/id6746722790?l=zh-Hans-CN

KaraKeeper同时集成了karakeep客户端和插件：
客户端：使用实例网址、API Key登陆karakeeper，页面比官方更加顺眼，信息密度更高。
插件：在Safari使用实例网址、API Key链接扩展即可，点击插件，save to karakeep，AI Tag Summary会自动生成好。

符合刻板印象：软件都是第三方的更好（

@carstenraddatz 😃

#karakeep #karakeeper #bookmarks #selfhosting #ios #ipados #safari #apple

RE: https://mstdn.feddit.social/@admin/115915116983323489

基本完成了从Raindrop和Chrome收藏栏迁移到KaraKeep

还有两处没做好：
1. karakeep支持S3，但是我把各种存档都存到了本地volume，虽然不是很占空间
2. 没有开始就使用CRAWLER_STORE_PDF=true
希望大家不要踩我的坑了（）

是否开启 INFERENCE_ENABLE_AUTO_SUMMARIZATION=true 倒是不是很要紧，不过强迫症肯定是需要在刚刚托管时就开启

让我再看看karakeep的文档，还有什么遗漏和功能
对了，开启每日备份，一次不过几十KB

#karakeep #raindrop #bookmark #chrome #docker