I looked through Common Crawl and found over 300,000 parseable RSS/Atom feeds, confirming that Web feeds are still a major part of the Open Web. But most aren’t high quality, and autodiscovery often points users at stale or abandoned feeds.
Remote
Mark Nottingham
@mnot@techpolicy.social
mastodon 4.7.3Co-chair IETF HTTP & AIPREF working groups, member Internet Architecture Board (again), standards lead at Cloudflare. Former W3C TAG, W3C BoD. Interested in the intersection of legal regulation and technical standards.
Social media is mostly-write for me - if you want to make sure I see something, e-mail.
0 Followers
0 Following
30 Posts
Joined November 18, 2022
In 2026, kids skating in public spaces isn’t annoying - it’s FUCKING WHOLESOME, carrying on a 40+ year legacy, and certainly should not be considered any kind of crime.
More belated introspection from the adtech industry.
To help find relevant RFCs no matter what your network connectivity, https://rfc.fyi now implements a service worker that caches the data locally for offline use, and can be installed as a Progressive Web app (PWA) -- i.e., you can install it to your iOS or Android home screen.
Note that it still points to online-only RFCs at the RFC Editor site.
In the last few weeks, Trump has done more for the worldwide adoption of electric vehicles than any previous US president. Congratulations.
Guidelines for aspiring Nobel Prize winners:
1. Do not threaten other countries with wholesale destruction and possible genocide.
"A stable competitive market never has more than three significant competitors, the largest of which has no more than four times the market share of the smallest."
https://www.bcg.com/publications/1976/business-unit-strategy-growth-rule-three-four
Every online interaction is a lopsided negotiation. For AI to truly work for us, we need more than just safety -- we need to start building true agency as a form of collective bargaining.
https://www.mnot.net/blog/2026/04/24/agents_as_collective_bargains
Standards work is notoriously hard to track. Let’s explore if grounding AI in working group records can make that history more accessible.
AI seems to have the same pathology that leads humans to name files like "report_final_done_final2_lastone_finanimeanit_revision3"
This needs to be spread wide and far.
Replying to
@edavies@functional.cafe aha - I know what happened there (old version got pushed). Will fix thanks!
Replying to
If this is true, AI creates even more pressure towards centralisation than is obviously apparent (and there's a lot that's apparent).
For previous thoughts, see:
https://www.mnot.net/blog/2024/11/29/platforms
Replying to
@huitema that's the cheese cart at a secret supper club about five minutes from where I live... unfortunately they have since closed :(
Replying to
Replying to
@liaizon the thinking there is to use HTTP content negotiation, rather than repeat content in the document in different languages. Generally sites that are sophisticated enough to have translations can do this.
Replying to
@mro @liaizon
Yes - the current draft tries to accommodate mixed language (but may not express itself well). Question is whether it's technically viable for *choice* of language to be done with proactive in enough cases, vs also accommodating reactive negotiation (i.e., a data structure with multiple language tags in the format).
Replying to
Replying to
Replying to
@KasTasMykolas@river.group.lt Yeah, I realised that the sniffing code wasn't working properly - doing another run now to see how that changes things.
Replying to
@KasTasMykolas@river.group.lt it's a good question. You can see stats for extensions that give an indication.. raise an issue?