SVP, SEO & Stable Genius @ Razorfish. Founder SERPrecon.com Works from early in the morning until late in the evening. Makes many calls and has many meetings.

Joined March 2007
2,485 Photos and videos
Pinned Tweet
Made a Chrome extension to bring semantic SEO to everybody for FREE!
Excited to launch our FREE Chrome extension that uses Chrome's built in gemini model to do semantic semantic analysis of websites. Extract keywords, score content, find the most relevant "chunk" and more! serprecon.com/chrome
8
16
97
21,552
we've reached the point where I don't care about the difference in AI models. it's like talking about processor speed or hard drive size in a computer. Most people don't care or notice. What can it now do, or what can't it now do? that's what matters.
5
16
837
Ryan Jones retweeted
Trump's White House UFC fight is reportedly being sponsored by Bud Light, despite MAGA's absurd boycott of the beer.
102
549
2,766
73,820
Ryan Jones retweeted
The same people who said masks were "literally 1984" think this is fine.
🚨 BREAKING: Palantir has received a contract from the Trump Administration to accumulate and centralize data on Americans.
176
9,883
59,471
1,065,271
Ryan Jones retweeted
I found a study that tested the GEO tactics everyone sells. On three AI engines, they lost to doing nothing at all. The tactics are the famous ones: Add statistics. Add quotations. Cite sources. Use an authoritative tone. A new paper ran a page through these edits across three different AI engines and measured how often it got cited. On GPT-4o-mini, the untouched page got cited 13.34% of the time. The same page after the full GEO checklist landed between 10.92 and 12.21%. The AutoGEO method I shared earlier this year scored 12.12%. Every optimized version came in below just leaving the page alone. Before you burn your playbook, two caveats belong right here. This was measured on simulated engines, not live ChatGPT or Google AI Overviews. And it is a single preprint, so not peer reviewed. Other research, including the AutoGEO paper itself, reports the opposite, with citation gains up to 51 percent. So the honest read is not that GEO is bullshit: it is that nobody has settled this yet, and anyone selling you a fixed checklist as proven science is overclaiming. Here is what I think the data is actually pointing at: bolting tricks onto thin content does not move the needle. The same study found that strong credibility signals, meaning real statistics, citations and quotations, pushed one page from 8.4% visibility to 23.7%. The lever was never the tactic: it was whether the content earned the citation in the first place. It is just the fundament, not the strategy. So stop using GEO checklists like they are settled rules (except for Glippy, Haha!). Test your own pages instead. Track which ones AI engines actually cite, and build from what works, not from what a vendor pitched you. Sources: 1. Liu, Z. and Xu, P. "Think Before Writing: Feature-Level Multi-Objective Optimization for Generative Citation Visibility" (April 2026), arXiv. arxiv.org/abs/2604.19113 2. Wu, Y. and Zhong, S. et al. "What Generative Search Engines Like and How to Optimize Web Content Cooperatively" (2025), arXiv. arxiv.org/abs/2510.11438 3. Aggarwal, P. et al. "GEO: Generative Engine Optimization" (2023), arXiv. arxiv.org/abs/2311.09735
1
7
36
8,684
This is interesting. A German court ruled that Google is liable for misinformation in their AI overviews - and noted that 8% of AI overviews contain wrong information. arstechnica.com/tech-policy/…
1
2
96
Ryan Jones retweeted
A developer automated his entire life with scripts. Then he quit. His coworkers found them. The repo is called Hacker Scripts. 49,500 stars on GitHub. Based on a true story. The original story comes from a Russian forum around 2015. A build engineer left his company. His coworkers went through his old machine. Here is what they found: Script 1: "hangover" If it is 8:45 AM and he has not logged in, the script emails his boss "not feeling well, working from home." Excuse pulled from a random list he pre-wrote. Script 2: "kumar-asshole" If a specific client emails with words like "help" or "trouble," the script SSH's into the client's server, rolls back the database to the last good backup, and replies "no worries mate, be careful next time." The client never spoke to a human. Script 3: "smack-my-bitch-up" If it is after 9 PM and he is still logged in, the script texts his wife "working late" with a random excuse from a pre-written list. Script 4: "fucking-coffee" The office coffee machine ran Linux. It had a TCP port open. Every morning, 17 seconds after he logged in, the script opened a telnet session to the machine and sent the brew command. A mid-sized half-caf latte. 24 seconds to brew. His coffee was ready the moment he walked to the kitchen. Every time. For years. His boss thought he was the most responsive developer on the team. His wife thought he was always keeping her updated. His client thought he had the fastest support in the industry. All four were talking to bash scripts. Nobody noticed. For years. His coworkers posted the scripts to GitHub when they found them. Developers around the world reimplemented them in 17 languages. Ruby, Python, Go, Java, Kotlin, PHP, PowerShell, Node.js, and more. One legend. Seventeen languages. 150 commits. The license is WTFPL. A real open source license. It stands for "Do What The Fuck You Want To Public License." He did not automate his job. He automated his life. And nobody knew until he left. 49,500 stars. WTFPL licensed. Legendary.
17
85
510
36,765
so let me get this straight. California democrats have the ability to rig elections but they ONLY rigged the LA mayor election and not the Gov election where the other candidate only lost by 3%? they just chose ONE race to rig?
94
based on their understanding of counting votes, does MAGA complain that football games are rigged when the team leading at halftime ends up losing?
1
126
convo I have too often: Why aren't I ranking or showing up in AI? Well, your site is just a list of products by category with no other copy. Everything that ranks and is in AI has use cases and more upper/mid funnel context.
8
229
Ryan Jones retweeted
I just open sourced my "Is this slop?" simple test
115
1,044
19,162
449,510
Ryan Jones retweeted
My god, its incredible.
375
516
22,331
2,941,154
Logic classes should be required in middle AND high school. literally every MAGA argument on here is a textbook logical fallacy.
1
5
184
I'm not a fan of this whole "edge cloaking for AI" GEO strategy. Just make your actual website work for visitors AND machines. Your visitors really don't want all the fancy javascript and giant hero images that your creative team thinks they do. Ask them, you'll see.
1
1
20
979
it's always the same story - and the reason none of the automated SEO case studies you see run more than a couple months.
These are 3 well-known brands that are also listed as case studies of AI content scaling tools They scaled a lot of page templates that can manipulate AI answers, like excessive "best X" content, self-promotional listicles, scaled comparison/alternative pages, etc. There's also some spammy structured data in the mix. Each was hit hard by the Jan 20 update this year, and the crash has continued ever since. This last core update only accelerated the declines. These are views of the US organic Google visibility across their entire domains (via @sistrix); even though the AI content generation appears to have been isolated to one subfolder on the site (from what I can tell), the impact is happening across the full domain Now all companies are at the lowest visibility they've seen in 5 years Google is not playing around.
1
1
19
2,990
how does a search engine measure experience (if they wanted to) They're not looking at your author bio, I promise you. It's all math. When you vectorize the entire internet, the "experience" sites will cluster together. Reddit, Forums, reviews, etc. There will also be a cluster of non-experience: wikipedia, dictionaries, news, etc. The engines just see what cluster's centroid is closer to your website. They don't need you to spell out experience - because your site is an entity, and it fits on a knowledge graph - and they can compare it to other "experts" on the same knowledge graph. Hope that helps.
6
1
18
1,053
I made my own free share of voice tool. It's open to anybody to bring their own API keys to use (if you have no API keys it does nothing) but the cool part is, the more users I get, the more I can analyze macro level data: freesov.com/pulse.php

2
1
10
810
At the Catholic Church festival and then and on stage is playing zombie by the cranberries. And I feel like if you know the meaning of the song that’s a very odd choice
1
2
141
i gave claude code the ability to commit and push to git, and now it's like an eager toddler asking me if it should push/commit on EVERY little thing I ask it to do.
3
2
352
Ryan Jones retweeted
Turns out black mirror was a very good source for startup ideas.
88
117
2,139
91,604
Ryan Jones retweeted
This is so insanely corrupt, I can’t even believe it. More than half the donors to Trump’s $400 million White House ballroom just won over $50 billion in new federal contracts in six months. And here’s the part that should make your blood boil. Sixteen of these 27 donors were facing federal enforcement actions, antitrust reviews, labor cases, securities charges. Many of those cases have been quietly dropped or scaled back since Trump took office. You write a check, your legal problems disappear. That’s not a coincidence. The White House won’t even release the full donor list. They’re hiding it on purpose, because daylight is the one thing pay-to-play can’t survive. A federal judge already ruled ballroom construction has to stop until Congress authorizes it. Government is supposed to serve the people, not auction itself off to the highest bidder. When access goes to whoever pays the most, working families always end up paying the price. We either end the corruption, or the corruption will end us. wapo.st/3QmJjSz
1,575
23,849
49,786
1,353,679