Martino Mosna
banner
martinomosna.com
Martino Mosna
@martinomosna.com
94 followers 76 following 330 posts
Non si smette di giocare perché si invecchia, si invecchia perché si smette di giocare. / You don't stop playing because you get old, you grow old because you stop playing. Linkedin: https://linkedin.com/in/martinomosna Web: https://www.martinomosna.com
Posts Media Videos Starter Packs
The bubble bursting is long overdue...
Remove spammed pages from training data and most importantly from the LLM is... an actual nightmare and probably not possible at all.

We'll see how it turns out.
I was actually fearing a prompt injection string.
LMAO Google scrapers are at it again:
- impression sharply up, again
- avg position sharply down, again
- clicks still going their own way as if nothing happened

#dataisuseless #seo #google #searchconsole
All of the above, but with number 3. slightly amended:

3. AI & answer engines pay mainstream media to maintain a functioning content ecosystem.

It's already happening under our noses.
With all the machine learning involved... Vibe SEO is basically just SEO. Again.
I'm still to wrap my head around the fact that the guy named his company as a fantasy artifact that was made for good and ultimately perverted into evil... And looks totally oblivious about it. Like "we're not evil, we're the good guys". Yeah sure, Mr. Saruman. How's the Dark Lord doin'?
Reposted by Martino Mosna
TIL that LinkedIn will start using our personal data and content to train their AI models.

This setting is set to “on” by default, unless you opt out.

➜ You can turn it off here: www.linkedin.com/mypreference...

Or go to Profile > Settings > Data Privacy > Data for Generative AI Improvement
Reposted by Martino Mosna
Ah, yes, the infamous "Cat, I farted" 🤣
ChatGPT notoriously uses a 3rd party scraper as a search results provider (it's been named recently, can't remember the name).

That basically means that Google sees those searches as legitimate user queries.

To see them in GSC is the only logical consequence.
I think I can hear the intro. Actual goosebumps.
Reposted by Martino Mosna
Ilias Ism @il.ly · Sep 5
The best SEO strategy is to create the content that people are actually searching for. It's that simple
Oh, it's gonna be SO MESSY
Reposted by Martino Mosna
This is what happened when someone took a site that used <table> to show data, and had lots of featured snippets, to a website that used <div> layouts to show the same data. (Pre AIOs etc)
For everyone struggling with site names in Google Search
Oh, lovely! I wasn't aware of that page in the docs. Gonna bookmark it 😁 thanks.

(also nice to see that my working solution is actually the officially supported one)
My current working theory is: geographically based websites, on the same IP... a machine learning classifier could easily make a mistake and give a false positive.

Current possible solution: add schema markup for Website & Organization.

Here's hoping that @johnmu.com can help me troubleshoot this!
They also share a somehow loose thematic tie, in the fact that both websites have a "local" directory architecture and share some pages for residential solutions (Phosphoro is core business, Atlantic Station just a sub-section).
Now, for the issue that I'm hoping to have some feedback or some ideas: the site's name for Google is apparently now "Atlantic Station"

It's the name of a neighborhood in Atlanta, USA.

There's a website with that name, but it's a WordPress site.

Only thing they have in common: Cloudflare.