We will use the old ways pic.twitter.com/gA2AYVJyKy
— Shaun Anderson (@Hobo_Web) April 20, 2026
In my recent post, I showed you how Google faked a deep, human-level understanding of content. This follow-up story is building….
Hmm – The site (with no human visible content) is ranking in G now
Proof you can rank in both AI’s and the SERPs with no visible content.
(The site is not mentioned and does not appear in the AIO nor the citations) pic.twitter.com/aHs66AR1Rp— Peter Mindenhall (@PeterMindenhall) April 20, 2026
Is it really a blank page if “reading mode” reveals the html? pic.twitter.com/hAqZsbFVMC
— James (@ceodreams) April 20, 2026
Indeed, we are going to find that out.
(As a quick aside, it was folks like Andy Beard (citing me in that post) who taught me how to blog nearly 20 years go.)
The original post:
Over the weekend I was tagged into a post over on LinkedIn where someone had run and experiment.
They launched a blank website…
> No content.
> No images.
> No hidden text.
> Just a blank white website.They did though include fully structured entity based schema to the…
— Mark Preston (@MarkPreston1969) April 20, 2026
The original post is on LinkedIn. I don’t want to link to the site itself until I’ve analysed it.

yeah? That is how search or anything with llms work? What do people expect here
— Klaas (@forgebitz) April 20, 2026
It’s not really how Google Search, specifically, is supposed to work, no. I mean, it contravenes Google helpful content guidelines.
— Shaun Anderson (@Hobo_Web) April 20, 2026
Perhaps the question should be, “How Long Can You Rank In Google Without Visible Content?”
not your father’s Google lol https://t.co/m0vCy7JDic
— John Andrews (@johnandrews) April 20, 2026
This is a flawed study. The site had full HTML content, just hidden by CSS – but still visible to screen readers and bots. AI doesn’t care about CSS styles. It doesn’t “render” content like a seach engine. https://t.co/Y9RiCcCJgb
— Ryan Jones (@RyanJones) April 20, 2026
See also:
see test 2 here. There’s text in the HTML. Would have been parsed into markdown when crawled, sr-only class ignored etc.
Will take a look at the linkedin post for context, but that’s my first thought.https://t.co/roGFZ1OWyd
— David McSweeney (@top5seo) April 20, 2026
I dunno man, just seemed to me like a way to promote whatever that weird protocol is.
“But seven layers of structured data underneath: JSON-LD, llms.txt, Ed25519-signed entity claims.”
I’m betting it would have done the same if you removed all that and just left the hidden text
— David McSweeney (@top5seo) April 20, 2026
Also worth noting:
Did you view server logs on whether anything even requested llms.txt? Outside of some awful “AI audits”, we’ve seen zero requests from any crawler.
— Martech Zone (@martech_zone) April 20, 2026
See:
Pretty cool the solution to the problem he’s flagging is already built by his own company
— Aaron Haynes (@myeyesshine_) April 20, 2026
Vger needs the information pic.twitter.com/9O1qbvyYzH
— Shaun Anderson (@Hobo_Web) April 20, 2026
He just did the equivalent of turning the font white.
— George Saoulidis (@georgecursor) April 20, 2026
Not that I think this is the value of this test, but on its own, it clearly contravenes Google’s guidelines:

As my friend, Peter, points out:
On the “should it be indexed” front:
A) Has no user visible text
B) Is abusing schema markup – markup does not match visible text
2 reasons it should not be.
What reasons should it be indexed?— Peter Mindenhall (@PeterMindenhall) April 20, 2026
My verdict, at this point, whilst its an interesting and legitimate test of the data layer, on its own, without visible text, this would be, strategically classed as total spam to Google. https://t.co/dHKVR1gRnA
— Shaun Anderson (@Hobo_Web) April 20, 2026
It’s an interesting test of the data layer and LLMs. On its own, without that visible text, of course, it is spam to Google.
Meta – You can rank in Google, talking bout ranking in Google, naturally:


The SEO community this day pic.twitter.com/9oCtPIZB35
— Shaun Anderson (@Hobo_Web) April 20, 2026