Screenshot shows how going from chat to voice and back again alters what model ChatGPT identifies as.
I’ve been wondering why gpt-5 didn’t seem like it was any better than previous models when I had longer conversations with it.
Turns out it out I was talking to gpt-4o most of the time, as voice isn’t supported yet.
A bit of deceptive design I have to say, as the nothing indicates the switch
Iterations*
Fix is on its way.
Uh, I had some many interactions I missed that.
Robots.txt - Can it run doom? Turns out it can:
doom.wunsch.dk/robots.txt (html)
doomxml.wunsch.dk/robots.txt (xml - Firefox only, due to inline xslt not supported in webkit based browsers)
More:
www.wunsch.dk/blog/robots-...
Inspired by @johnmu.com: bsky.app/profile/john...
Almost sounded like a challenge:
100% selfcontained robots.txt file that's also an xml file:
doomxml.wunsch.dk/robots.txt (Firefox only, due to inline xslt not supported in webkit based browsers)
html'ish version:
doom.wunsch.dk/robots.txt
Details:
www.wunsch.dk/blog/robots-...
And its live:
www.wunsch.dk/rate-robots/
Post your domain, and get a review of your robots.txt file :)
MST3K cinema style view of a robots.txt with the commentary: This robots.txt is like a participation trophy for SEO experts: “Good job showing up, now go away”.
Playing around with an interactive version of “rate my robots.txt” and got this gem from testing on Reddit:
—
This robots.txt is like a participation trophy for SEO experts: “Good job showing up, now go away”.
I saw the notice, didn't realize the deadline was this soon though.
I'll have to come up with another technique - polyfill would never trigger since this is sent as an XML content type which wouldn't trigger the JS.
A shame that www.wunsch.dk/sitemap.xml will deprecate though.
Games broke atm, my first version put the game data in the actual sitemap, and GSC didn't like that, so I made the sitemap more compliant, but need to fix the game.
Felt inspired by @johnmu.com an his robots.txt file that was also a wave file, and did a robots.txt that is also an xml document and Mystery Science Theatre 3K (via xslt).
Https://www.wunsch.dk/robots.txt
Huh. We don't expose underlying BGP stuff in our platform but we do have all the PCAPs from our global observation grid. Some really odd things started happening back on the 22nd
But at what pace will they ban the next round of abusers? They need to prove it’s properly and quickly enforced - otherwise these parasites will just keep signing deals with new sites.
Right now this feels like they’re trying to empty the ocean with a teaspoon.
🤯
Anything particular in this that is wave only, or would this work with any RIFF container based files?
Thanks :)
This particular project is on static hosting, so 429s based on redirects is what we’ll try then.
We all know that soft404s are a thing. But how does a SPA signal to Google that it's backend api can't keep up with crawl speeds?
Are soft 429s a thing?
@searchliaison.bsky.social @johnmu.com
I love how Microsoft’s billing strings on your account matches the invoice. This makes it easier to match up receipts.
The Wordpress saga takes another turn:
old.reddit.com/r/WPDrama/co...
Hiding in plain site is a somewhat different version of how WP governance is supposed to happen.
Long overdue from Googles side - one wonders if they could have gotten away with the cloaking if they hadn’t been called out:
www.linkedin.com/feed/update/...
It’s crazy that all they had to do to get out of the penalty was to hide the pages from manual reviewers.
Same time, chrome browsers are reporting the pages as still up into the CrUX reporting tools.
The offending pages are still indexed even.
It can’t be this easy @searchliaison.bsky.social ?
My outmost respect and support for @joost.blog on this.
Even though I think they’ll find widespread support, it think this journey to reform will take time.
Whichever course has to be taken to secure the future of the project, I only hope it can be accomplished without fracturing the community. But I agree something needs to be done.
They did provide Extended Validation Certificates though.
Today there’s no easy way to know if you’re on the site of a legit business, or chinese dropshipper.
Couldn't agree more:
"While domain authority outranks all, efforts to combat parasite SEO will just be pushing water uphill. Only the restoration of algorithmic respect for topical authority, eroded since the ‘helpful content’ update, will actually change the landscape."
Impressive deep-dive into blackhat seo:
recleudo.com/under-google...
Looks like this has gone under the radar. Seems like a significant change.
I can write a message and have it cryptographically signed to prove that I wrote it.
But why on earth do browsers not have support for this? Feels like a missed opportunity that we cannot prove and check authorships across platforms via signing.
I wish the performance team could somehow address that so much cpu is wasted by doing full page builds on every single request.
It would be the biggest environmental win for WP ever.
I can’t help but think about how Drupal 6 was the newcomer, and people leaving Wordpress for it.
D7 was even better, but then Drupal 8 effectively managed to stop the progress: They redid the technical platform, while sacrificing user experience in the process.
www.drupal.org/forum/suppor...
