This edition is brought to you by our friends at Venngage.
Turn long reports into infographics people actually read
Nobody read the report.
Solid work, no doubt - but it was 20 pages, and half the people you sent it to opened it on a phone between meetings.
That's the problem Venngage AI solves: turning a report nobody finishes into a one-pager everyone gets in seconds. Paste in your findings or the raw stats, and it pulls the numbers that matter into a layout that already looks like your brand. Then you edit it - colors, fonts, the layout - make it completely original…and yours.
Superpath readers get 30% off any plan with SUPERPATH30. Build yours today at venngage.com
Hey folks, it's your content friend Eric here.
As long-time readers know, I do my best not to be timely or topical in this newsletter. The less I touch The Discourse, the better, in my opinion.
This week, I’m going against my better judgment and joining the hordes of folks reacting to the news that Claude will now mark AI-generated content.
The TL;DR of what’s going on:
The EU recently put the AI Act in place to regulate AI products and protect normies like you and me from potential risks of nefarious AI use. They’re mostly trying to prevent stuff like manipulation (e.g., deepfakes) and classification (e.g., facial recognition databases).
Part of the AI Act forces LLMs like Claude to clearly label when content has been processed by AI. Starting…last week…Claude will be watermarking any text produced by its various systems (vanilla Claude, Claude Code, Claude Design, etc.).
What’s fun and spooky is that the watermark doesn’t lie in the formatting—it’s hidden in the words themselves (oOoooOoOoooooO 👻 ), undetectable by humans.
My understanding of how it works: LLMs build sentences by choosing words one at a time based on the statistical likelihood of the next word—like a series of coin flips between all the words. If an LLM wants to hide a secret watermark in text, it just has to choose the occasional (slightly) unlikely word (e.g., a synonym) based on a secret key. Hide enough unlikely words in one piece of text, and the odds are this could have only been written by the holder of the secret key.
Humans can’t spot this. We’re way more likely to quietly spot all the other ridiculous AI writing tells. But if you are an AI model that knows the likelihood of all the words (and the secret key), congratulations, you can now detect AI writing. (If you care more about the technical details, this is one such example of how it might be done.)
TL;DR: Claude is gonna hide secret messages in your AI-written text. (Be sure to drink your Ovaltine!)
The reason I’m writing about this today is to combat the predominant narrative on LinkedIn that this will once again change everything.
I am willing to put my stellar reputation on the line and confidently predict this will have little-to-no impact for B2B content folks who use (or don’t use) AI.
But people are going to be able to tell you’re using AI to write now!
Whaaaaat? Are you saying that Series B company’s CEO didn’t quietly write 365 thought leadership posts and 500 SEO articles this year completely by hand? Thank the AI heavens we have watermarks now 🙏
Snarkiness aside, I do have many B2B content thoughts on this. In the interest of brevity, I’m just going to rocket out a stream of thoughts in bullet point form:
Let’s be clear: This is just Anthropic abiding by EU laws. They don’t have any incentive to discourage you from using AI.
Anthropic specified that any processing of text by its AI products has the potential to introduce a watermark—even a quick round of copy edits. So will the watermark be useful in any way to call out truly “AI-written” text? Doubtful.
How much AI is needed to call something “made with AI”? I was reminded of The Puzzle of the All-American BBQ Scrubber, an episode of the Search Engine podcast in which an American businessman tries to make a completely “Made in America” product, but discovers it’s basically impossible to build every single piece of the thing you’re making in America. There’s always some component that’s only made offshore (e.g., a screw, or the cast that makes the screw). That’s true of much of our content workflows now. How much of our content workflows never goes through at least one touchpoint with AI (e.g., spellcheck)?
I really don’t think B2B consumers care if something is made by AI if it’s helpful/quality/educational/not obviously slop. Even pre-AI, all the B2B content we created was a means to an end (business/career growth). I don’t think that changes with an AI label.
That’s clearly not true for “real-life” content—I categorically reject all AI content IRL when I can detect it (looking at your Focus for Work playlists, Spotify)—but I don’t think B2B consumers care as long as something isn’t overly laced with quietly-s and that matters-es.
Even if B2B readers did care a little bit, I don’t expect the average reader to go out of their way to copy-paste your blog into an AI detector, or download some Chrome extension that does so. There has always been a not-so-hidden incentive to B2B content. Just pass the vibe test of usefulness, and you should be fine. (Although it’s possible that browsers and social platforms are forced to label AI-written content in the future. Will people care more?)
What about our secondary audience, Google? Google has so much business incentive to gain from AI usage (e.g., Gemini, AI overviews). They’re investing billions here. So no, I don’t think Google will ever devalue AI-marked content in its SERPs or elsewhere. So much of the internet is made up of AI-written content now that I think we’re already past the point of no return for Google filtering it out.
As flagged by Rachel Bicha in the Superpath Slack community, could this open up a can of worms re: intellectual ownership? Quick Google lawyering tells me content made purely by AI falls in the public domain, but content produced by a human with AI falls under their personal ownership (not the ownership of OpenAI, Anthropic, etc.). But plagiarism lawsuits with/by AI are already rampant, so theoretically, nothing much should change here for now.
This will no doubt revive the “AI detector” conversation. Freelancers and agencies have already been subject to these detectors for a few years. The smart writers already disclose their use of AI to their clients. In theory, this should create fewer false positives from AI detectors, so maybe that’s a net positive?
Of course, the more you edit the text, the less likely it is to be detected. I’m sure many folks will take ridiculous steps to avoid being labeled as AI, making all of this moot. Lots of “scrambling” skills and products will emerge soon so that we can all prove how human-written our AI content is.
So, in conclusion, I conclude, conclusively, that a whole lot of time and energy will be spent worrying about this, and entire markets might pop up around this, but you, as a content person, can safely ignore all of it.
There are many more productive ways to think about AI and when to use it or not. I don’t think this should factor in.
If I were to give you some fortune-cookie advice on things you can do: whether you write with AI or not, continue to add your own secret key. That little flair or quirk that makes your writing unquestionably yours.
Cheers,
- Eric
PS. Thanks to everyone who took our tool survey last week! Congrats to Annabel and Casey for winning the raffle.
🗽 A note from Alex: Superpath in Times Square!
Alex here! A few weeks ago, Superpath was in Times Square. It was a fun culmination of our partnership with beehiiv and our move to the platform over the last few months.
But if you're a long-time Superpath subscriber, you know that my face is the least interesting thing to feature. It's Jimmy who started this thing and Eric who writes this "email worth reading" every week for all of you.
Unlike many of the other creators featured in this campaign, Superpath is a community built not on one person, but a village of folks who make it happen.
I'd like to celebrate this moment along with the many folks like Chloe, Rachel, Ruth, Ronnie, Carol, Anna, Brinda, Tanaaz, Michele, Lisa, Cady, Kisha, Alethea, Amy (A.K.), Anton, and the many, many other Superpath contributors who have made 2026 so special so far.
And thanks to Beehiiv for the shout-out! You can still claim our partnership offer of 30% off Beehiiv with the code SUPERPATH30.
📝 AMA Recap: Jake Block, Head of Clay Studios at Clay
Jake Block heads up Clay Studios, Clay’s social, narrative, and media team. Previously a Supervising Producer at HubSpot, Jake has a simple ethos: make something people want to watch.
Jake was kind enough to do a Slack AMA with the Superpath Pro community a few weeks ago. Here’s the quick summary.
TL;DR: Jake’s answers mostly covered three topics:
Inside Clay Studios: Clay Studios is an in-house media company staffed by two full-time producers and a network of director-level ICs. There is no budget process and no content calendar, only conviction and a publishing cadence. The team is goaled on business value, including workspaces created.
Creative philosophy & craft: Hire for charisma, don’t script, and keep the moments most teams would cut. Clay is deliberately tool-agnostic and product-agnostic, and ideas come from whoever has conviction rather than from consensus.
Advice for other teams and creators: Replace the word “algorithm” with “audience.” Find the niche thing only you can talk about, borrow a format that already works, and make as much as possible. Reps are the differentiator.
Definitely worth a read if you’re video-curious!
📆 Upcoming Superpath Community Events
To get the invitations to our member events, join Superpath Pro. You get 30 days free, so you can attend all these events!
AI Show & Tell (Aug 19): Three community members will show off what they’re building with AI. (Past replays are also available for members.)
Slack AMA: Sabreen Haziq (Aug 26): Buffer’s Senior Brand and Community Manager, Sabreen Haziq, will answer all our questions about the wildly successful Team of Creators program she started at Buffer.
Superpath Social (Sep 2): Meet a bunch of content marketers and chat about content in our monthly breakout chats.
Change My Mind (Sep 16): We’ll debate 3-4 hot content takes as a group—with more nuance than you get on LinkedIn.
🎙 New on Content, Briefly: Think Like a Broadcaster with Ronnie Higgins
Ronnie Higgins, legendary content marketer and longtime Superpath Pro member, joined Alex to unpack his recent piece on thinking like a broadcaster instead of a publisher.
His argument is that content marketers should study their audience's actual lives, not just their interests. From there, you can build repeatable formats for different moments in the buyer journey and program them on a rhythm until they become a habit that drives real commercial value.
Ronnie spent years in TV and film production (CBS, Comedy Central, HBO) before moving into content marketing, and it shows in how he thinks about this stuff.
Alex and Ronnie discussed:
The 95/5 rule and what it takes to land on a buyer’s "day one list"
Why an editorial calendar and a programming schedule aren't the same thing
Borrowing Disney's "synergy map" to build a real brand identity across formats
I also loved Ronnie's trick for turning event guests into podcast guests without hunting for new ideas.
💬 Great Slack Threads This Week
Here were some great discussions in the Superpath Slack community this week:
The AI talking point du jour: Claude is watermarking AI-written content to comply with EU's AI Act. Thoughts everyone?
How are people finding productive ways to work with leadership who are obsessed with “AI tells”?
How do you present content marketing assets to Sales in a way that encourages use? Would love to hear what worked for others: format, cadence, where you surface them, etc.
Does anyone do brand partnerships on LinkedIn as a creator? I want to get a sense of the pricing models
📙 The Reading List
Here are some links that got the Superpath Slack community talking this week:
Clay’s AI writing policy | Varun Anand
Black Hat USA 2026: The 'Breaking' News: The OpenAI–Hugging Face Incident | Black Hat (YouTube)
How Claude marks AI-generated content | Claude Help Center
🆓 Get a free 30-day trial of Superpath Pro
Superpath Pro is our paid community membership. On your free trial, you'll get access to:
A private Slack community with 400+ in-house and freelance content marketers
Monthly 1:1 peer networking calls
Monthly group events with breakout sessions
Monthly AI Show & Tell workshops
Graduate-level content courses
Niche channels and events for freelancers, content leaders, and more
"Most professional groups boast about their member numbers. Superpath intrigued me for the opposite reason: they deliberately shifted from a massive free community to a much smaller, paid one - so they could focus on it, nurture it, and make it genuinely excellent. And they absolutely deliver."
— Farah Cormack, Senior Marketing Leader