Authenticity Gap: The Frustrating Side of AI {{slop}}
Why 80% of what you see online sounds like it came from the same boring neural net—and how to fight back
Listen to a 5-minute pod on this article if you are on the go or just want a fun way to digest!
We've reached an impasse with AI-generated content. For the first time in the post-ChatGPT era, real people operating under their own personal brands have transformed into "content farms" by outsourcing their daily social media presence to Large Language Models (LLMs).
The numbers are staggering. Late 2024 data showed that over 50% of LinkedIn text was AI slop. Today, this number is probably higher, and we would go so far as to say 70-80% of all posts we encounter on LinkedIn are either generated entirely by AI or heavily polished with it.
Spotting AI-created content isn't difficult, especially when posts land in that uncanny valley where something feels slightly off. Usually, you can tell within the first two sentences. The real frustration kicks in when we see our colleagues—people we've followed on social media for years—start doing this, sounding as inauthentic as inauthenticity can possibly be.
Don't get us wrong—we at One Ten Hundred Ventures are AI maximalists, or what some call e/acc (effective accelerationists). Our core thesis centers on investing in AI-native founders, driven by a deep conviction that existing tech stack incumbents will be decimated by this new generation of entrepreneurs.
But—and this is a significant but—we believe AI should be deployed from a first principles approach, focusing on functions that truly move the needle while preserving originality. The problem lies in execution. Most people today rely on similar, commoditized prompts like "create a viral LinkedIn post on this topic..." because the majority of ChatGPT users are not AI-native.
We call this chasm, particularly in the context of online content, the Authenticity Gap.
Now let’s look at the good, bad and the ugly on how folks are not keeping things OG:
The Ugly 💀
The worst cold reach-out we’ve ever received:
A couple of things are blatantly wrong with this cold reach-out I received last week: They got my name right in the subject line (though why put my name there at all?), but somehow my name becomes ((firstname)) in the actual intro. They assume we're running agile (what gave them that idea??) and think we're a product team when we're actually investors.
How unskilled—or lacking in vision—do you need to be to build AI pipelines that send such sloppy cold outreach emails? This isn't the AI's fault but pure human laziness.
The Bad 💣
Solid content buried under generic formatting and poor tone:
What you see above is the majority of AI users in action today. The tone, aesthetics, and formulaic structure—hook → setup → climax → takeaway → call-to-action—are generic and painfully predictable. We'd wager this individual used ChatGPT to create this content with weak prompting.
Here's what we consistently see across posts like this:
The formulaic opener: First line invariably starts with "here's what everyone misses..." or "here's what people get wrong..."
The false dichotomy: The predictable "It's not about X, it's about Y..." two-liner cadence that follows.
Em dash overload: Excessive long dashes (—) scattered throughout. This post uses them five times! (Also notice me using them in this newsletter intentionally?)
Zero personal voice: Complete absence of personal narration. Instead of saying "when we learned that it wasn't about splashy design," it remains sterile and detached.
No authentic insights: Missing any anecdotes or observations that could only come from real-world interaction or genuine human engagement online.
The sad thing is that such a post has 30k likes, 1000 comments, and 1.4k reposts (!)
The Good 🌞
Genuine insights, unique perspective backed with tactical proof points:
This LinkedIn post—created by Amos Bar-Joseph, co-founder of Swan AI which is one of the handful startups out there that are truly AI-native—follows a similar structure to The Bad getting some things right and others wrong.
The inauthentic, blatantly AI elements are still present:
False dichotomy formula: "AI agents aren't digital labor; they're leverage" follows the predictable It's not about X, it's about Y template.
Variation on the theme: "No team meetings. No approval chains. Just intelligence networks..." delivers another spin on the same It's not about X, it's about Y pattern.
Textbook structure: The entire flow—"I generate $1.5M..." → "Most GTM leaders see AI agents as..." → "100x version of each person... one-person GTM org..." → "I'm sharing the exact blueprint..." → "comment if you want the link..."—follows the rigid hook → setup → climax → takeaway → call-to-action formula.
We categorize this as The Good because it executes three core elements effectively:
Personal storytelling with substance: Amos provides a concrete anecdote with tangible examples of how he's achieving a "$1.5M+ pipeline with a one-person GTM org." He isn’t just giving theory but backing up with specific results.
Human touches throughout: Authentic moments are sprinkled like salt and pepper around the post, such as "p.s. thnx Julia for having me!!" These small details feel genuinely human.
Conversational call-to-action: Instead of generic corporate speak, the CTA feels personal and approachable: "comment if you want the link to join me and I'll send it over!" It's clear, friendly, and easy to act on.
The process for building a post like this with AI is straightforward. You feed ChatGPT all the insights you've gathered authentically through real-world experiences—this is what preserves the content's authenticity. Then you prompt the AI to create a "viral LinkedIn post for an audience of founders, venture capitalists," and so on. Finally, you manually review and refine as you deem fit.
The missing component here, however, is the tone of voice that should be authentic to Amos, but unfortunately is the distinct voice of OpenAI's transformers. Since the content itself is original enough, Amos is able to get impressive engagement, sometimes thousands of comments per post.
The Great 🦅
You'll virtually never spot The Great using AI because they craft content so skillfully that it delivers not only unique, personal insights but also maintains the author's authentic tone of voice. The AI becomes invisible.
Below, you can see how I build content—transforming this Substack article into an X post while working with AI to preserve my own true voice:
Check out the result here posted on X, which is super close to my own voice. And here are the prompts I used to achieve this—feel free to use them yourself :)
As you can see from this relatively simple LLM workflow, creating top-tier social media posts without sacrificing authenticity isn't that difficult. You can preserve both your original ideas and your authentic tone of voice.
How to be original: A human-guided, AI-native playbook 👇🏼
If your goal is genuine thought leadership on social media—measured not by virality or follower count, but by the quality of engagement—you need to hit three notes with perfect pitch: pathos, logos, and ethos. Just like the Ancient Greeks taught us.
Pathos creates emotional connection. Place this strategically at the beginning of your post, leveraging one of the seven deadly sins, and your audience will be hooked.
Logos provides the intellectual foundation. Statistics, functionality, or methodology satisfy the neocortex (the brain's higher-level cognitive center) while the emotional energy from pathos provides you the energy to stay focused.
Ethos builds the trust and reputation needed for your audience to take you seriously and keep returning. This often merges seamlessly with your call-to-action, driving traffic to other marketing content you want to amplify.
The secret sauce is not really rocket science folks.
Just sprinkle in your strategic insights, friendly narration (as if talking to a colleague), and personal anecdotes while avoiding overengineered grammar or random mistakes. And you're well on your way to authentic content! Oh, and don’t forget to build a custom GPT on this so that you can use it all day, every day;)
Unfortunately, most people going viral online these days with AI do a great job on pathos, sometimes a good job on logos, but fail big time to establish solid ethos once their generic, boring, frustrating ChatGPT tone of voice is exposed.
Are we being naïve for criticizing inauthentic content? Is there actually market demand for AI slop—or has authenticity become irrelevant at the point of consumption?
The short answer is, yes. A market (potentially a few large ones even) exists. And all depend on the target audience. Let’s break it down:
Low-information audiences and casual consumers are passive content users who prioritize clarity, familiarity, and simplicity over depth or authorship. They're highly receptive to AI-generated material. As long as content appears credible and answers their questions quickly, they rarely question its origin—and often can't tell the difference anyway. Example: your uncle who watches YouTube videos to fix things around the house.
Non-native English speakers and emerging markets include users in developing regions, education-restricted contexts, or non-English-speaking areas who prioritize clarity, accessibility, and practical value over authorship or stylistic nuance. AI-generated content—especially when simple, neutral, and informative—fills critical gaps and gets embraced without concern for human versus machine creation. Example: a college student in rural Indonesia surfing the internet learn Python for an upcoming job.
Time-constrained professionals encompass executives, managers, and knowledge workers who care far more about clarity and efficiency than authorship. They welcome AI-generated content if it saves time or accelerates decision-making. For this audience, authenticity takes a backseat to utility: a well-structured summary or actionable insight has value regardless of whether it came from a human analyst or machine (as long as it's correct—or sounds like it’s correct (!)). Example: a product manager at a startup who has back-to-back meetings.
Short-form platform users consume content where attention spans rule everything. On TikTok, Instagram Reels, YouTube Shorts, or X, users care more about instant entertainment or novelty than authenticity of creation. Because this content gets consumed rapidly and judged in seconds, AI-generated material easily blends in and often outperforms human content, provided it's engaging. On the other hand, long-form content gives audiences more time to detect and reject shallowness, making human authorship more consequential. Unfortunately, most people consume primarily short-form content. Think all the fans of WWE wrestling matches who actually know it’s all staged but couldn’t care less.
Let's examine these audience segments from a purely economic perspective. Let’s imagine that a human creator produces 10 top-quality posts per week with 100% engagement, generating 1,000 total engagement points. Meanwhile, an AI-assisted creator produces 100 posts per week at 70% engagement, generating 7,000 engagement points—a 7x advantage despite the authenticity penalty. Given the audience segments that either don't care or can't discern AI from human content, simple production-scale economics will beat human creators every. single. time.
So who remains? Who will continue caring about authenticity long-term?
We believe it’s the Highly educated, tech-native users—especially those fluent in AI themselves, understand media trends deeply, and are able to pattern match faster than others. They can often detect the subtle cues of AI-generated content and tend to penalize content that feels formulaic, impersonal, or derivative. For this group, originality, voice, and lived experience matter, and they’re quick to disengage from content that feels like a low-effort prompt output rather than a genuine human perspective. We put ourselves in this group. (Big Q: is this a big enough market?)
(We are not including Highly-educated, non-tech-native users here [e.g. the older generations] as they, in all likelihood, will not be able to categorize AI content vs non.)
Where does the golden nugget lie? Can we have best of both worlds?
Yes. We see two paths forward:
1. Nuanced, segment-by-segment approach Individual and organizational content creators should tailor their backend creation processes to specific audience segments. While great creators already customize output for their target audience, doing this well requires choosing the right AI-human collaboration behind the scenes.
For Highly-educated, non-tech-native users, more manual curation of final output is essential. For Low-information audiences and casual consumers, AI can take over significantly with minimal curation, as long as it drives engagement with AI slop (!).
2. AI-native full-stack, undeniably human frontend Someone will soon crack the code (if you're building in this space, get in touch!) on handling the majority of content creation—from ideation to distribution at superscale—in a highly autonomous, hands-off fashion.
For this to succeed, especially for brands, the output must "sound" authentic, original, and fully aligned with previous content published by the individual or brand. It needs to be HQ content that appeals to the general population without feeling like "AI slop."
We emphasize the word "sound" here because as long as the audience doesn't realize content is AI-generated and it's exceptionally failproof, even "deeply human" AI-generated storytelling can penetrate consumer hearts. We believe AI will advance to this state-of-the-art level, but it will likely require integration into the physical world—always with humans wherever they go, whomever they meet, whatever they do. Think AI-first smart glasses.
Let's put on our sci-fi hats for a moment and think about the far-future. Our investigation here raises profound questions that demand deep contemplation as we navigate our relationship with information in the age of AI:
What happens when the consuming audience is also AI? Does content creation pivot to serve free-roaming AI agents rather than human consumers? Given enough time, do we simply trust AI agents to "read" what needs to be read for us, making decisions on our behalf in the background? Y Combinator has an intriguing request for startups in their Spring 2025 batch: B2A: Software Where the Customers Will All Be Agents.
What if AI never truly generates content that deeply connects with humans across short-form, long-form, multimedia, and real-world physical experiences? Or perhaps humans will reject AI-generated content altogether? This would mean AI will always need human consciousness as a partner, because conscious beings will only trust other conscious beings before consuming content.
Or will we abandon our professional ambitions because AI consistently makes the best decisions for us—including consuming content at supernatural speed and creating individualized, day-to-day roadmaps until death that ensure maximal happiness? E/acc advocates would call such a future indistinguishable from utopia, surpassing anything Thomas More imagined—more akin to our notions of what a day-in-the-life of a solarpunk civilization would look like.
Whatever will be, one thing is for certain, we will crave authenticity & originality even more when everything around us is artificially generated.
One potential side effect we're optimistic about is an unexpected sailing ship effect, where the flood of AI-generated content could force us to: 1) supercharge our creativity, 2) establish impenetrable systems of record for authentic human content, and 3) explore our true unique differentiation as humans, with AI as half-competitor, half-ally.
Never forget: when AGI has taken over and we meet the singularity head-on, the neural net you have sitting in your human head only exists in one entity at a time, across all space-time! Your brain and consciousness is the most non-fungible entity of intelligence out there, which AI will never be able to replicate with all your memories, needs, emotions, intelligence, and most importantly, your consciousness.
How do you supercharge yourself with AI while being authentic? We’d love to hear from you—email us or comment below. 👇🏼
Stay autonomous folks, until next time. 🌵|110100
If you are building semi-autonomous companies, we want to talk to you.
Get in touch here: hello@onetenhundred.vc