Super strange to see that a site-query lists articles like the "headless stopgap" but a exact search for ["headless stopgap"] or even [site:aifoc.us "headless stopgap"] doesn't. I would love to learn the solution (if it's not a bug on the search side)
@rustybrick.com asking jinxed it 🤣
even AI mode gets it
I remember proposing The Cloisters – crazy times.
The issue isn't that Google is using an excavator but that Google is using an excavator to build their store building (AI mode) on our property (content) and claim everything as their own.
If Google only would have a quick cache ;-)
As for the 204 status code: I believe it's primary use isn't caching but acknowledging that the request was successful and also indicate that there is no body to send for a complete response. 304 status code also has no body but there is a body which already has been sent and is known to the client.
I stumbled a bit about the part on TLS where it somewhat implies that TLS implies authenticating that the server is who he says he is. I did some reading and my conclusion: It's not wrong but it's not completely correct either.
I really enjoyed listing to the newest episode of "search of the record". So I would like you to continue this conversation @divingfor.fun & @methode.bsky.social
Two comments:
Happy birthday! 🥳
Google AI Overviews will give you an answer even if what you’re searching for is the meaning of a made up idiom.
They are killing thousands of websites for this crap.
Pet peeve of mine: mobile is dominating almost always. Yet most SEO slides show desktop serps
I just tried and it “works” the same in German. I also checked ChatGPT search and it can’t be fooled as easily
a screenshot showing that Google handles duplicate url parameters by joining the values – the q parameter to be exact.
I'm guessing Apple Germany is spoiling the April fools? To be honest: I would be it if possible: www.apple.com/de/mac/lumon...
Are those slides going to be shared?
Google Search Console API gains 24-hour hourly data for past 8 days coming soon searchengineland.com/google-searc...
this is what it looks like. Right now Google stores roughly 10 days of data on a hourly basis. I guess in theory a week over week look of yesterdays data would be something Google could "easily" implement.
The Google CoreUpdate might be a good moment to share a new Chrome extension I wrote. chromewebstore.google.com/detail/gsc-u...
Allows you to see more hourly data than just the last 24h. cc @markwilliamscook.com @lilyray.nyc @aleyda.bsky.social @glenngabe.bsky.social …
I love it
Adapt or die, as always. I don't believe there is a one size fits all solution to this. I've seen Slashdot, Facebook, … traffic come and go and maybe it's now the end of getting traffic from other sources. Maybe it's the era of direct traffic and owned networks.
Maybe my wording is off but I am not arguing that free content is eliminating the subscription market. I am saying that AI that is scraping free, ad-driven content is eliminating that content. On the subscription side of things people will not be willing to pay for every site only a few that remain
My predictions: Publishers will try to keep LLMs away from their precious content but most will ultimately perish. This is because ecommerce websites (or other websites with alternative monetizations other than subscriptions and ads) will provide content “for free” and open to LLMs. Very few subscription model publishers (I’m guessing very big ones like the NY Times) will survive. In the end the web will be less diverse and much more expensive.
this is what I wrote last Friday in a slack channel:
Impression of our Berlin office
how to exit vim 😏
I was thinking about flying to the US this year. I guess I won't 😞
I have to say I am really impressed by Claude 3.7 Sonnet. "Create a website in the style of the 'Still Alive' from at the end of Portal during the credits." didn't even take a minute to create a nice page which even could play music. Sure this isn't advanced programming but still very impressive.
Grok output saying not to mention Musk or Trump as disinformation spreaders
Don't mention Trump or Musk as disinformation spreaders -was apparently a Grok system prompt that was exposed in its "thought process" output. A few people on Reddit claim to be able to reproduce this:
Hbd!
