︵✿мαиєкι иєкσ‿✿ 😼

[Charles Floate 📈 @Charles_SEO]

REAL ENTITY SEO LESSON #2 - SEMANTIC SEO 📈

The sentence and formatting work that decides whether ANY of your entity text gets extracted...

Because you can have perfect entity strategy and still lose every single triple at the writing stage.

So ALWAYS Subject FIRST!

"Charles Floate founded Charles Floate Training." - NOT "Charles Floate Training was founded by Charles Floate."

Passive voice inverts the order and makes the machine work harder to reconstruct the relationship.

Active, subject-first, verb in the middle.

And REMEMBER, pronouns BREAK the chain.

"PressWhizz has 50,000 publishers. It operates in 90+ countries. They deliver in 24 hours."

You've just handed over two orphaned facts.

Coreference resolution is not free and it is not reliable. Every "it", "they", "this", "the company", "we" is a chance for the entity to drop out of the sentence.

Repeat YOUR BRAND - It reads slightly worse to a human but you get 100x better results from a machine 😅

And Hedging DESTROYS fact confidence and trust signals.

"Charles Floate is one of the best SEO consultants and you should hire him for..."

That's not a claim, there's nothing in there to assert.

"Charles Floate is a black hat SEO expert with 18 years experience, has been personally targeted by Google and Charles Floate founded several multi-million dollar SEO companies"

THAT can become a fact.

Every "one of", "arguably", "may", "helps to", "can be considered" is you voluntarily downgrading your own triple - Be confident, but realistic and try to derisk your direct promotion issues.

Negation has to be LOCAL - "X happened. This is false." does NOT reliably register as a negation.

The claim and the correction get separated, the claim survives on its own, and the machine keeps the association.

"X did NOT happen" works. Negation inside the sentence, attached to the verb.

Which matters enormously for reputation work - if you're rebutting something about your brand across your site, and your rebuttals are structured as "claim, then correction", you might be reinforcing the exact thing you're trying to remove 🙃

ALWAYS keep the entities close.

A subject and object (subject: Charles Floate > Object: SEO Expert) separated by 40 words of subclauses and qualifiers is a weaker relationship than the same two entities sat next to each other.

Distance dilutes, always try to tighten the sentence.

AT A PAGE LEVEL - Chunk boundaries are the silent killer... This is the big one for AI SEO (retrieval layer) right now.

Your content gets split into chunks. If your claim is in one chunk and the evidence supporting it is in the next, the retrieval layer can (and often does) pull one WITHOUT the other.

Half a fact, ends up unusable and not actually cited.

Do not rely on the machine inferring for you...

Claim and support have to live in the same block. Don't let your best triple sit orphaned at the bottom of a section.

Blocks need to be split up by headings, and headings should be treated as query strings, not creative writing.

"What Is Entity SEO" gets matched.

"The Hidden Architecture of Modern Search in 2026 by Charles Floate" gets nothing...

Every H2 is a chance to declare exactly what the block underneath answers. Use them...

Tables and lists are extraction gold.

As long as they match the consensus, add unique value or/and are formatted specifically for SERPs then a table is a stack of pre-parsed triples - Row header, column header, value with subject, predicate, object, already separated for you.

Comparison content that lives in prose is doing it the hard way for no reason - Clarity and confidence win everytime.

Just make sure to name yourself the same way in every paragraph.

Not "Charles Floate" then "Charlie" then "God of SEO" then "Charles SEO" then "CF".

Copywriters call that variation, but machines call it five different unmatched strings...

The sequencing of ALL this across a whole campaign (Which claims to push, in what order, across which sources) that's the hardest part and what takes serious reverse engineering of your specific niche, competitors and SERPs+AI outputs.

But the writing layer is FREE and it costs you nothing but discipline and time to put into practice... Plus you can train the AI to do 95% of the work for you now! 🎩

[Charles Floate 📈 @Charles_SEO]

HUGE AI SEO TIP THAT COSTS $0 AND TAKES UNDER 30 MINUTES ⏲️🔥🔥

STEP 1 - Pull the data

Search Console → Performance → last 90 days → filter to positions 9-30 → export > CSV.

Those are pages Google ALREADY trusts - It's testing them, its just hasn't committed.

And you don't need to build anything NEW, just update what is already there with the next few (easy) steps:

STEP 2 - Find the queries you rank for but never actually ANSWER

This is the bit almost nobody does.

Filter each page by its queries - You'll find terms it's picking up impressions for where the page never directly answers the question.

Google is telling you exactly what it thinks the page is about, and exactly where the page is failing to deliver it.

Free, sat there, but inored.

STEP 3 - Feed it to the AI properly

Drop the CSV into Claude (Cowork if you've got it) with the page content alongside it.

Do NOT ask it to "optimise the page" - You'll just get 400 words of air.

Ask it to do this instead:

"For each query this page ranks 8-20 for but does not directly answer, write a self-contained block: an header (H2 or H3/H4 if it can be under an existing block) matching the query as literally as possible, then a 1-2 sentence direct answer, then 2-3 supporting lines or bullet points or claims depending on the query.

Claim and evidence must sit in the same block - No intro, no padding, just give the machine what its looking for.

Add in 2-3 of the top surfacing competitors for even better consensus+uniqueness matching, and that's it! You've got a page that should skyrocket in visibility, and a system you can do over and over again.

STEP 4 - Understand why THAT structure

Because retrieval doesn't work like ranking.

Your page gets split into chunks (or "passages" for AI Overviews - If the answer is in one chunk and the proof is in the next, the model pulls one WITHOUT the other.

Self-contained blocks aren't a formatting preference.

They're the difference between being extracted and being skipped.

STEP 5 - Internal links, done PROPERLY

Get the AI to map every semantically related page on the site and plan the links both ways - Use your sitemap/s to support this + a Screaming From (or SiteBulb, or Ahrefs, or whatever you prefer) crawl of the actual site.

Rule of thumb still holds: Roughly 1 internal per 50 words, 1 external per 150.

But anchors matter more than they used to - "Click here" and "read more" pass link value and ZERO association - Use the entity, and use the relationship.

STEP 6 - Date it and ship it

Visible last-updated in the HTML, not just schema.

Freshness is becoming a live retrieval signal now, not a vanity stamp. An unmaintained page from 2023 loses to a mediocre one updated last week.

I ran this exact process on ONE old post earlier this year - it went from 17 visits/month to 4,000+ uniques, and $16,700/mo in organic traffic value (according to Ahrefs).

One post - One 30 minute update, no new linksn nothing built from scratch and AI did 90% of the work...

And every one of those pages is already inside Google's trusted set for that query. You're not fighting for entry, you're just fighting for the answer.

That's a completely different battle and it's a much easier one to actually achieve...

Stop building.

Start finishing what you've already got 🎩

[︵✿мαиєкι иєкσ‿✿ 😼 @__manekineko]

@pierreeliottlal @romanbuildsaas TRIO

[Chris Long @chris_nectiv]

Everyone is talking about how G2 is important for AEO and they're right.

BUT nobody is talking about Capterra...and it's arguably a more important data source for ChatGPT.

Recent extraction of B2B fan-outs found Capterra is searched +50% more frequently than G2.

[Corey Haines @coreyhainesco]

A Conversion Factory brand optimization:

Sensible. They had great raw assets but a weak arrangement, so our team tightened the layouts, used their patterns and color more intentionally, and brought well defined hierarchy to the page.

Combined with our other product marketing work, their committed ARR doubled from $500k to $1M in a year.

[🐧 @Pentosh1]

There are so many positives to cover from $BB earnings that it’s just not possible to do.

Massive amount of US gov contracts from military to the white. I’ll try to condense everything into this thread. The growth trajectory of BB is off the charts over the next few years

[Jake Castillo @jakecastilloooo]

https://t.co/M1FkRTGivW

[︵✿мαиєкι иєкσ‿✿ 😼 @__manekineko]

@IMJustinBrooke Add me

[︵✿мαиєкι иєкσ‿✿ 😼 @__manekineko]

@IMJustinBrooke Zero

[Tim Soulo 🇺🇦 @timsoulo]

In the last 6 months at @Ahrefs, we analyzed over 1 billion data points across 14 studies. Here's what we learned about AI search optimization:

1) "Best X" blog listicles are the single most prominent content format cited by AI chatbots. They make up 43.8% of all page types cited by ChatGPT specifically.

2) 67% of ChatGPT's top 1,000 citations come from sources marketers can't influence: Wikipedia (29.7%), homepages (23.8%), app stores (6.6%). Only 32.3% are influenceable content like educational pages, reviews, news, and blog posts.

3) 28.3% of ChatGPT's most-cited pages have zero Google organic visibility. These pages get cited repeatedly by ChatGPT despite not ranking in Google at all. A completely separate discovery layer.

4) ChatGPT only cites about 50% of the URLs it retrieves. It fetches dozens of pages per query but uses half as background context without attribution. This means that being retrieved and being cited are very different things.

5) Adding schema markup had zero meaningful impact on AI citations. AI Overviews actually dipped −4.6%, while AI Mode (+2.4%) and ChatGPT (+2.2%) showed changes indistinguishable from zero.

6) YouTube mentions have the highest correlation (0.737) with AI brand visibility out of all the factors we studied (including all the conventional SEO metrics like backlinks, page count, DR, etc). This held true for both Google-owned and OpenAI products.

7) AI Overviews reduce clicks to the #1 result by 58%. That’s up from 34.5% just 10 months earlier. The trend is accelerating.

8) 99.9% of AI Overviews appear on informational intent queries. Transactional, navigational, and local searches are almost entirely AIO-free. Shopping triggers AIOs just 3.2% of the time.

9) For a given search query, Google’s AI Mode and AI Overviews reach the same conclusions 86% of the time — but cite almost entirely different sources (only 13.7% citation overlap).

10) AI Overviews change every 2.15 days on average, with 70% of content differing between consecutive observations. But semantic similarity stays at 0.95. The words, sources, and entities constantly shuffle, but the actual meaning barely moves.

[AYi @AYi_AInotes]

Karpathy这句话，算是点破了2026年用AI的最高境界

现在用AI最蠢的方式，就是一句一句写prompt，

你像个保姆一样，每一步都盯着，改完这版改下一版，

看起来效率很高，但其实你自己才是最大的瓶颈😆😆😆

AI一秒钟能生成几千个token，
你一分钟才能敲出几十个字，
你插手的次数越多，整个系统的速度就越慢😂😂

真正的顶级玩家，已经在把自己彻底踢出循环了，
所以你不用当下一句prompt的提供者 要当整个系统的设计师，提前写好规则，设定好评估标准，建好反馈循环和安全网，
然后放手让AI自己跑，自己迭代，自己纠错，

以前的杠杆是雇人，写代码，用资本，
现在多了一个更加恐怖的杠杆，叫做token杠杆，
就是说你只需要输入几个token的大方向，背后就能发生海量的事情，

一个人就能干过去一个团队的活，
当然这不是说完全不管，
你还在循环里，
但你的循环从每一分钟一次，变成了每一周一次，
你不用再盯着每一步的执行，
你只需要在关键节点做最终的判断，
别再当AI的保姆了，直接当它的老板🚀

[Garry Tan @garrytan]

The secret to an articulate agent like mine isn't one file. It's three:

SOUL.md — Who the agent IS. Voice, values, operating principles, what good output looks like, what bad output looks like. Not a system prompt, a constitution. Mine says things like "brevity is mandatory," "humor is mandatory," "never open with 'Great question,'" "swearing is allowed when it lands." The more specific and opinionated this is, the less your agent sounds like a chatbot. Write it like you're briefing your smartest friend on how to be you, not like you're configuring software.

USER.md — Who YOU are. Not a bio — a deep model. How your mind works, what you're building, your strengths, your blind spots, your family, your temperament, what triggers you, what you care about. The more the agent understands about you, the better it can serve you. Mine is ~4000 words.

AGENTS.md — Operational rules. What to check on every message, what to never do, how to handle failures, lookup chains, path rules, brain-first protocols. This is the playbook for how it works, not who it is.

The articulation comes from SOUL.md being brutally specific about voice. Generic instructions → generic output. If you write "be helpful and concise" you get ChatGPT. If you write "speak like a peer with taste, one sentence when one sentence works, uncomfortable truths welcome if actually true, language with voltage" — you get something alive.

[Leon Lin @LexnLin]

brandkits skills is out

https://t.co/ByHmn0BZb7

always open for feedback or requests :)

[Meng To @MengTo]

I recorded a 43-min video on how to turn a DESIGN.md into landing pages, mobile screens and motion design

[ROFI @bidah]

Releasing https://t.co/NERoDiSPte. More than a hundred design.md to choose from for free with desktop and mobile previews, kitchen sink design system samples, filters by category and search. Also accesible through the CLI `npx designdotmd add gym-kettlebell`.

[︵✿мαиєкι иєкσ‿✿ 😼 @__manekineko]

@Aidanb2b DESIGN

[shadcn @shadcn]

Meet Sera. A new foundational style for shadcn/ui.
Minimal. Editorial. Typographic.
Shaped by Print Design Principles.

Available Today. Try on shadcn/create.

[︵✿мαиєкι иєкσ‿✿ 😼 @__manekineko]

@LoganTGott Prompt

[Romàn @romanbuildsaas]

I’ve hired over 50 LinkedIn influencers to promote GojiberryAI over the past 3 months.

Average ROI is 1.5x on revenue.

I spend $1 to generate $1.5 in MRR, which is honestly insane.

Here’s what I’ve learned to make it work:

1. Follower count doesn’t matter. Look at likes and comments instead

2. We pay $300 for small influencers, $500 for mid-sized and $750 for large ones. Never above that

3. Check who is liking and commenting on their last 10 posts. If it is always the same people, they might be using a POD

4. Look at their lead magnet posts. If they get between 100 and 300 comments I pay $300. Between 300 and 500 I pay $500. Above that I pay $750

5. I write every post and choose every visual. No guesswork, I know what performs

6. I create a high quality lead magnet that the influencer shares in replies to every comment

7. The influencer’s job is simple. Copy paste my post and reply to comments. That is it

8. Each lead magnet has unique tracking links so I know exactly how much each influencer generates

9. If it is profitable, I double down

10. To find influencers I search by keywords or just scroll. Anytime I see strong engagement in my niche, I reach out

Hope this helps 🙂