🔇💼 LinkedIn Is Throttling Your Posts and Won't Tell You

Hey Explorers!

Time magazine built a second version of its website. You can't load it. But ClaudeBot can. That's the week in one sentence… and it repeats across LinkedIn's feed, ChatGPT's citations, and the eight billion simulated people now available to test your product before a real customer touches it.

Here's the signal that matters 👇🏼

🕵️ Time Built a Version of Its Website You're Not Allowed to See
A developer poking at crawler behavior found that Time serves sponsored pages to ClaudeBot, OAI-SearchBot, and PerplexityBot that ordinary readers never load. Ally Bank and the Project Management Institute bought the first slots. Your competitors can now pay to shape what an AI says about your category, and nothing in the answer gets labeled as an ad. (🔗 Link)

🔇 LinkedIn Is Throttling Your Posts and Won't Tell You
Posts flagged as generic AI output don't get removed. They just stop reaching anyone outside your existing followers, which is the part of LinkedIn that was worth anything. The platform claims 94 percent detection accuracy and has published no false positive rate, so the only audit that helps is reading your last ten posts and asking what's in each one that only your company could have written. (🔗 Link)

📍 Depth Beats Breadth When ChatGPT Decides Who to Name
Semrush and Kevin Indig tracked 1,094 categories in ChatGPT over six months. In categories close to a brand's core expertise, ChatGPT cited the domain in 74 percent of appearances and named the brand in 44 percent. Wander too far from what you're known for and named mentions collapse, and appearing in just one of a category's five prompts actively drags your mention share down. (🔗 Link)

👥 8.3 Billion Fake People Are Now Available to Test Your Product
A Harvard and MIT led team released MatrAIx, a population of persona agents built on 1,290 attributes each, with a million of them free to download. You can run a price increase, a checkout flow, or a support bot past a sampled cohort and get back stated reasons, not just scores. They ran 18,189 trials and still tell you to validate against real humans before betting anything on it. (🔗 Link)

✍️ ChatGPT Won't Write Like Stephen King Anymore
Ask for a passage in a named author's voice and you get a polite decline plus an offer to deliver similar atmosphere in its own voice. The block now covers dead writers too, which it didn't a month ago. If any part of your creative workflow leans on "in the style of" prompts, assume that lever is closing across every model you use. (🔗 Link)

🎯 Terence Tao Gets a Better ChatGPT Than You Do
Sean Goedecke compared his own ChatGPT sessions to a transcript of the mathematician working through a hard proof and found the difference wasn't prompting. Tao knows the subject well enough to spot the one useful line in a wall of output and shove the rest aside. Same model, same tokens, completely different results. (🔗 Link)

🛑 OpenAI Hit Pause on Its Own Model
OpenAI told Axios it cannot rule out critical cyber capabilities in Astra, so it's slowing development and hardening internal testing until safeguards catch up. It notified the White House voluntarily, which lands right as the administration drafts its process for reviewing models before release. That process will eventually decide when the tools your team runs on are cleared to ship. (🔗 Link)

🦠 An AI Designed 16 Viruses That Never Existed
Stanford and Arc Institute researchers fed 9 trillion nucleotides to a model called Evo, got back 700,000 candidate genomes, tested 285, and 16 of them came alive. They only infect E. coli, and engineered viruses are already how doctors go after drug-resistant bacteria. The researchers also asked, in writing, that nobody point this at human pathogens. (🔗 Link)

There's a page on your site nobody has opened since 2024. A crawler read the whole thing this morning and took notes.

Your website has an owner. Your deck has an owner. The version of your brand that only machines see is nobody's job yet.

Machines are the audience now. Give them something worth quoting.

Purple Horizons helps leadership teams find the specific thing only you can say, then build the workflows that produce it every week instead of once a quarter. Strategy sprints, executive briefings, buildathons, hands-on deployment.

👉🏼 Work with Purple Horizons

Signal over noise. Always.

Ralph & Gianni

Tired of overpaying GoDaddy or Namecheap? There's a smarter option with Porkbun!

Porkbun is the domain registrar trusted by creators, developers, entrepreneurs, and anyone who wants low prices without the nonsense or the forced upsells.

Here's why so many people are making the move to the Bun:
• Most domains sold at cost
• Low, transparent registration and renewal pricing
• Free features like WHOIS privacy and SSL certificates
• Powerful web and email hosting options
• Real human support 24/7, 365 days a year
• Named the #1 domain registrar by Forbes Advisor and USA Today

Whether you're launching a business, building a personal brand, starting a side project, or creating your very first website, Porkbun makes the whole thing easy.

Get $1 off your next domain registration with Porkbun now.

Get Your Domain At Porkbun

📍 9/1 South Miami Tech Tuesdays @ Bougies in South Miami

📍 9/8 Cafecito & Pastelitos @ Chug’s in Coconut Grove

Want to sponsor our newsletter or events? Drop us an email at [email protected]