What I'm Reading, Saving, and Rethinking - June 13th, 2025

Happy Friday! Every week, I save dozens of posts, articles, and newsletters that challenge my thinking. Here's what stood out this week and why I think it's worth your time too.

Was this email forwarded to you?

🔥 This Week's Saves

🔗 LLMS.txt isn’t robots.txt: It’s a treasure map for AI

Summary: The article introduces LLMS.txt, a new file website owners can place at their domain root to curate the best, AI-friendly content for large language models at inference time (i.e. when LLMs generate responses). Unlike robots.txt which tells crawlers where not to go, LLMS.txt acts as a “treasure map,” guiding AI systems to your most valuable, well-structured pages.

Why It Matters: Traditional SEO signals (e.g. sitemaps, metadata, crawling patterns) aren’t enough to guarantee your content is even seen by LLMs during response generation. LLMS.txt gives site owners a more-or-less direct line to AI systems, increasing the likelihood that their content is prioritized and cited in AI-powered outputs (which is the goal, obviously).

My Take: This is a sharp take on AI and SEO thinking. It signals a move from passive optimization to active, curated visibility, which is the first I’ve seen. But like the initial days of robots.txt and sitemap.xml, it’s early, non-standardized, and unevenly supported.

Bottom Line: LLMS.txt is not a Wonka-esque golden ticket. It’s a precursor. It’s a piece of the puzzle. In order to capitalize on LLMS.txt you have to already be creating “AI-ready” content that is: short, structured, context-rich, and citation-friendly. LLMS.txt works best with clean, easy-to-digest pages with clear headings, short paragraphs, bullet points, semantic cues.

🔗 When Growth Plateaus: How and When to Layer on New Acquisition Channels

Summary: Every product eventually exhausts its initial growth loop. The real question isn’t if growth will slow but what you do when it does. Most startups achieve product/market fit by mastering a single acquisition channel (e.g. viral growth, paid advertising, SEO, or direct sales). However, they often hit a ceiling when they optimize that channel to its limits and assume the solution is adding more channels on top.

Why It Matters: Adding new channels before the fundamentals have changed usually delivers only 5-10% growth bumps, which is not the step-function increases founders need. Meanwhile, it divides attention and resources across multiple mediocre efforts instead of focusing on what could actually unlock the next phase.

My Take: Your first successful channel worked for fundamental reasons, i.e. it was the most optimal path given your constraints at the time. New channels don’t become viable just because you want them to; they become viable when something structural has changed: your unit economics, your content base, your product breadth, or your funding position.

Bottom Line: Rather than asking “what new channels can we try?” the better question is “what has fundamentally changed that would make previously unviable channels now work?” This forces teams to focus on the leverage points that actually matter: better economics, richer content, stronger engagement. This, then, will naturally unlock new growth opportunities.

🔗 The Who Cares Era

Summary: The advent of AI has produced a worrying trend: sloppy, AI-generated content proliferating because nobody cares. There’s so much “disposable content” that is shoddily produced for people who are double or triple screening. Adding on top of this, AI is “a mediocrity machine by default, attempting to bend everything it touches toward a mathematical average. Using extraordinary amounts of resources, it has the ability to create something good enough, a squint-and-it-looks-right simulacrum of normality.“

Why It Matters: Culture is grinding towards a lowest common denominator as AI dumbs down everything. Even Netflix showrunners admit to writing scripts and scenes for viewers on second screens. Netflix shows become elevator music or background noise, never challenging the viewer to think about the show they’re watching. Everything becomes shallow, and cultural entropy is accelerating.

My Take: I use AI…for a lot of things. To assist in my writing, to prototype an idea, to give me ways to say “no” to an invite for drinks over text. But the more AI is used, the more average and similar our work becomes. Everyone is so quick to “put something out there” they never stop to think about what they’re actually saying (or if it’s just noise). But it’s inspiring that the author, Dan Sinker, doesn't reject AI wholesale: he simply insists AI shouldn't replace the thought and heart we bring to our craft.

Bottom Line: In a world of disposable content, giving a damn is a form of resistance. Don’t wait for perfection or polish, just inject some humanity and sincerity into your work.

This newsletter is for you. What marketing challenges are you facing in your marketing journey? Reply directly to this email with your questions or topics you'd like to see covered in future issues.

Until next week,

Adam

P.S. Found this helpful? Forward it to another founder who might benefit—we're all in this together.