#dev 2023-04-18
2023-04-18 UTC
geoffo, gRegor, lockywolf and IWSlackGateway joined the channel
# 
[KevinMarks] [snarfed] I'll have a look
# [KevinMarks] The first first S in RSS was "Site" wasn't it?
[marksuth] joined the channel
# 
[pfefferle] [snarfed] do you have signature validation in brid.gy fed?
# 
IWDiscordRelay <capjamesg#4492> What is RSS?
# IWDiscordRelay <capjamesg#4492> https://indieweb.org/RSS
[Jeremy_Keith] joined the channel
# 
@LukasRosenstock ↩️ I think the #IndieWeb community has some interesting ideas of automating links between sites, for example through #WebMention. However, at the end, it is about people writing blogs reading each other, linking each other and building on each other's ideas. (twitter.com/_/status/1648274713439883268)
[dave]1, [chrisbergr]1, [jeremycherfas], [Jeremy_Keith], [KevinMarks], [tantek], IWSlackGateway and [KevinMarks]1 joined the channel
# [KevinMarks] Bugger, Google didn't renew unmung.com for me
# [KevinMarks] I should have moved it off their registrar when Will left
# 
[schmarty] o no
geoffo joined the channel
# IWDiscordRelay <capjamesg#4492> 😦
# [KevinMarks] I can have a look. I'm trying to reclaim it but it's a yak shaving password reset as I haven't used GAFYD in a decade
# IWDiscordRelay <capjamesg#4492> GAFYD?
# [schmarty] Google expletive expecting your domain
# [schmarty] *expletive expletive
# [schmarty] (Google apps for your domain)
[chrisaldrich] joined the channel
# [KevinMarks] unmung is pretty much oneshot cloud functions with a little bit of memcached caching. What's the best thing to host that for Python? AWS Lambda? Google Cloud Functions? Azure Functions?
# 
[aciccarello] capjamesg, I noticed your RSS feed isn't autodiscoverable. Might be something to add to make it easier to subscribe. https://www.rssboard.org/rss-autodiscovery
# [KevinMarks] I'm familiar enough with AWS Lambda to know it's really annoying. I suppose it comes down to how easy it is to do a URL fetch in the environment
# [KevinMarks] it's in AppEngine classic, on Python 2.7, so If I am going to redeploy it I need ot change most of the code anyway
# [aciccarello] Oh, we have autodiscovery in the wiki already https://indieweb.org/RSS#Autodiscovery
# [KevinMarks] aren't the modern ones mostly node flavoured though?
# [KevinMarks] feedparser is pretty key to unmung
# [KevinMarks] they don't seem to let me renew the domain without signing up for Google Workspace, so I am feeling rather 'burn it all down' at the moment
# [KevinMarks] I am probably 3 layers of edge case for them, given that I signed up fro this in 2009 with a.google.com
# [snarfed] interesting, it lets me use https://domains.google.com/registrar/ ok with a non-workspace account, but maybe yours was on an old personal GAFYD domain?
# [KevinMarks] yes, it was old enough to be registered via godaddy, but not using a godaddy account I have access too
# [KevinMarks] that looks like this
# [KevinMarks] do I call hover and ask them to transfer it out?
# [KevinMarks] nope hover want me to unlock it at godaddy
# drmr Steganography might be what I was looking for thanks snarfed++! I've come across the term it in those CTFs (or more like puzzles really) where people hide stuff in images, but didn't make the connection to text
# drmr Oops, forgot the square brachets, [snarfed]++, and thanks for the link [KevinMarks]++!
# [schmarty] what is Karma?
# 
Loqi Karma is a community recognition tool used in the IndieWeb community and powered by the friendly bot Loqi https://indieweb.org/Karma
# 
capjamesg [snarfed] Does the contents of this look right to you? https://jamesg.blog/feeds/posts.xml
# [schmarty] drmr: it's very minimal but there are "instructions" there for using Loqi's karma system ^^
# drmr Thanks! Was just gonna try to look it up 😅
# [schmarty] it did work 😄
# drmr Also capjamesg, saw this the other day, I guess the arms race is on lol https://twitter.com/sohamgovande/status/1641828463584657408
# drmr Thanks for the video link capjamesg++
# drmr [snarfed]++ [KevinMarks]++
# drmr Oh you're right, seems to be deleted since I bookmarked it. Can't find now, but basically, there was GPT-Zero, which is like a turnitin, but for detecting if text was generated by ChatGPT. That link was to a tool for bypassing that, so that your text becomes "undetectable"
# drmr Reminds me a bit of ad-block-blocker-blocker-...
nertzy_ joined the channel
# drmr I really like that video though, because one thing I wanted to look at in addition to the steganography part is having "plausible deniability" that I wrote something. And I've seen the stuff where it's possible to deanonymise someone based on their writing style etc
geoffo joined the channel
# bkil This sounds like a topic for #chat I think.
# capjamesg drmr You may find this repo funny: https://www.deanza.edu/faculty/ramireztono/phil04/weekfiveoverhead.pdf
# capjamesg This is to what I meant to link: https://github.com/jdagdelen/hyperDB/tree/main
# drmr Very nice! Is it fair to say: HyperDB is to SQLite as Pinecone is to MySQL?
# drmr This will come in quite handy! I had a project idea around clustering embeddings, to find relationships between words (and their "categories"), ever since I saw this on HN and wanted to try https://news.ycombinator.com/item?id=35445312
# drmr My first use case is around "core values". There's this exercise where you try to narrow down your core values, but you start with a list of 200-something of them. When I looked at the list, I couldn't help but feel like that some were very similar, some were like secondary to others, etc
# drmr I thought that if you turned them all into embeddings, then clustered them, it might give you some sort of taxonomy for free
# drmr That way, narrowing down core values isn't so tedious. If you wanted to rank those values, tournament-style, you'd have to compare each one to every other one, so like quadratic time. But if there was a "tree" of values, based on this clustering, it could maybe be log time
# drmr Ah oops, you're right, I'll go there 😅
jonnybarnes, mouse[d] and IWDiscordRelay joined the channel
# [chrisbergr]1 I'm searching for quite some time for a solution to automatically take and save screenshots for my bookmark posts. Just playing around with chrome php library. https://github.com/chrome-php/chrome It works very well, so in case anybody else is in need of a php solution... ;)
# bkil Did you consider taking a snapshot of the above the fold DOM instead as an alternative? It would weigh orders of magnitudes less in most cases.
# [chrisbergr]1 Do you mean taking the snapshot manually?
bterry joined the channel
# bkil I would believe tooling should already exist for this. You can find similar tools that optimize & minimize above the fold HTML and above the fold CSS already.
# [chrisbergr]1 Until now I have not found anything satisfying in php that I can run on my server
# bkil Why do you have to run it on your server? If you are bookmarking a site that you are browsing, a bookmarklet would be a reasonable choice.
# [chrisbergr]1 I'd like to use it on mobile, too. Maybe this can be solved also, but I love the simplicity in use of the mentioned library.
# bkil I see it can also be put to good use: https://github.com/chrome-php/chrome#user-content-evaluate-script-on-the-page
gRegor joined the channel