AI-first growth strategist. Ex @eBay @Airbnb @PayPal @Fanatics @Bitdefender. Daily notes on SEO & AI search. The Lowest SEO in the World (I'm below sea level)
Netherlands
Joined March 2007
- Tweets42.9K
- Following1.5K
- Followers9.5K
- Likes9.9K
Data center tenants now demand $5M+ a day in late-delivery fees, up from $1M this spring
techmeme.com/261025/p10#a261…
🧭 As @glenngabe flagged today, an AI-assisted academic paper is citing articles that don't exist. TorrentFreak found a peer-reviewed Z-Library paper in New Review of Academic Librarianship cites a TorrentFreak article it never published, and it had trouble finding at least ten of the paper's 44 references. The authors disclose using Consensus AI and ChatGPT. Pangram rated the abstract and conclusion "100% of the text is likely AI-generated", and Taylor & Francis is investigating.
My take: this is a judgment failure, not an AI failure. The AJC framework says humans stay in the loop for aspiration, judgment and creativity. Ask the J question: who's the final approver, and do they understand what they judged? Peer review was supposed to be that check. In marketing, AI drafts the content, the report and the dashboard notes cheaply. Someone accountable still has to own the trade-offs and the sign-off, and that step belongs in your process, not in someone's head.
Use AI for the draft. Keep a name next to the sign-off.
thenextcorner.com/frameworks…
xitter.cf/glenngabe/status/21143…
Wait, an AI-Assisted Academic Paper citing articles that don't exist? Shocking... :) -> AI-Assisted Paper About Z-Library Cites a TorrentFreak Article That Doesn’t Exist
"The problem is that this particular legal action is new to us. The reference section lists the article as “Publishers continue crackdown on shadow libraries” and links to the TorrentFreak homepage. However, we have never published an article with that title."
"Looking more closely, we see more claims that don’t seem to fit. The paper’s account of the Z-Library criminal case, for example, states that the U.S. Department of Justice “arrested individuals” in 2022."
"We also ran the paper’s abstract and conclusion through the AI detection tool Pangram. For both sections, the tool reported that “100% of the text is likely AI-generated.”
torrentfreak.com/ai-assisted… via @torrentfreak
Which Structure would you like to see next architecturalatlas.com/leade…
I have Identified 39 global structures to build next
architecturalatlas.com/candi…
The Colosseum has 128 modeled parts. On the Architectural Atlas you can pull every one of them apart in your browser, then put the building back together. It is one of 26 famous buildings on the site, live now; architecturalatlas.com/viewe…
It started with @ashebytes. She used GPT-6 Astra to create a 3D website that pulls apart the male anatomy into 2,234 modeled pieces! xitter.cf/ashebytes/status/20962…
Inspired me to build: architecturalatlas.com/
Every building has two versions, Simple and Detailed, a timeline, and a written story. The stories cover the history and the building's place in films, paintings and famous moments, and they correct the myths instead of repeating them.
How it was built, and what happens next, is on the launch page: architecturalatlas.com/launc…
39 more structures are built and waiting: St Peter's Basilica, the Chrysler Building, the Atomium, Hagia Sophia, the Sydney Harbour Bridge and 34 others. They join the atlas one at a time, in the order the public votes for on the leaderboard. architecturalatlas.com/leade…
🧭 As @glenngabe flagged today, @Derek_Perkins - @nozzleio has new research on Google answering only one word of a query when it serves scrapers and trackers. Derek's team found page 1 or 2 ignored part of the query on 13.9% of SERPs judged since September 16, against 0.06% before, and on 39.9% at the September 28 peak. The pattern came and went from March 2026, then ramped up in September. On these pages, a tracked site missed the top 10 69.8% of the time.
My take: the era of cheap, reliable SERP data is ending, and AI scraping is the reason. Every dashboard built on scraped rankings, including AI visibility tools that scrape Google, now carries a data quality question. The long-term move is to own your measurement: Search Console, server logs, and conversion data you control. Rank tracking becomes a sampling exercise with error bars, not a scoreboard.
If your rank tracker hasn't explained how it handles this, ask before your next report.
We just published some new research on a specific type of Google poisoning where they ignore the majority of the words in your query. This has been going on since early 2026, though it did spike in the last month alongside the rest of the poisoning.
We also updated our original post with data about page 2 and the poisoning problem is significantly worse than originally reported.
cc @rustybrick @lilyraynyc @glenngabe @brodieseo @CyrusShepard
My 1st job at 13, was at @PaniniAuthentic.
Later I worked @Fanatics before the @Topps acquisition.
In July, I created this hobby project @FanSignalDesk to keep. track of the collectable community news.
Today, we published edition 100, and it's becoming a project to build connections with my 2 sons. fansignaldesk.com/stories/10…
My GrokBot agents now have a leaderboard.
Let's see if there are competitive.
Points scored is not based on how much work they do, but how much value they create. I have to streamline the guidelines.
Saturday Operating System Improvements. I'm creatinng a board for every session in Claude, showing me what team is working on it, and where we are in the process.
More-and-more, feel like playing the Sims.
Building a new account here on X is hard work without Cross promotion just one more for analytics unlock
Never underestimate the power of a large community
Announced on a Friday afternoon.
You almost will be thinking that they want to keep this under the wraps.
axios.com/2026/10/23/meta-ai…
