Ai content is why nobody trusts founders anymore
Strong claim, hear me out.
Every founder update reads the same now. "excited to share," emoji bullets, fake vulnerability about the rollercoaster. doesn't matter if it's real or not, it all sounds identical.
Noticed this hard while hiring. read every dm myself for a growth role. most were obviously ai-generated, same tone, zero thought put in. the ones that stood out were rough, specific, clearly written by someone who actually looked at the problem first.
Think the same thing's eating founder credibility. people used to trust a post because it sounded like someone actually figured something out.
Now half of linkedin sounds like the same prompt wearing different names, so nobody believes the "authentic journey" post anymore, they just scroll past it as content.
Not anti-ai, use it daily. just think we've optimized founder voice into something nobody actually trusts and most people haven't clocked it yet.
Curious if that's just me being cynical from too many dms this week or if others feel it too.
Replies
I’ve noticed this too. When every post has the same polised structure, even genuine stories start feeling scripted.
my trust goes up when someone shares a specific mistake, decision, or observation that could not have been copied from a generic startup template.
@bradely_charlesbois i think founders are going to have to work harder to sound like themselves. My LinkedIn feed already has enough polished content that says very little.
Founder updates used to feel like actual conversations. Now you can often guess the prompt behind the post before reaching the second paragraph.
Maybe the better approach is using AI for editing, but keeping the original thoughts and wording human. That balance seems much more believable.
The fake vulnerability part is especially noticeable lately 😅 There’s a big difference between sharing something personal and packaging vulnerability like a content format.
I was saying something similar on Linkedin.
I've come across several websites where the business owner has been asking why they're not making sales, and it's because they're using AI to build the website and write the copy.
They all follow the same design scheme and the copy has the same generic voice. There's no real personality or anything to distinguish them. There's nothing there to connect with their customers.
It's all just causing them to click away instantly, instead of making them read or engage further.
WIT
I think It's the same everywhere on social media now. Feels like I am reading one and the same post/dm all over again. And people who write without use of AI, and do it well, have a great competitive advantage.
yeah, honestly feel this. not against AI either, we use it too, but I think part of it is also that people just don't really read anymore, not just short video, even posts, people mostly skim now. so writing gets built for skimming instead of actually saying something, and AI is really good at that, not so good at the part underneath it.
for us specifically (we're building a messenger, Protocall) this shows up in a really literal way: AI still can't write a good push notification. it'll get the words technically right and still completely miss why anyone would bother tapping it. that gap is basically the whole job, and it's exactly the part AI skips.
i build an ai marketing tool so im exactly who u mean. imo its what u feed it, ask for a founder update and u get every founder update mixed together
When someone calls a post AI written, it is worth looking at what they are actually testing, because it is almost never style. Cadence, tidy paragraphs, punctuation habits, those are proxies, and every one of them comes out of a single prompt, so a founder who tightens their prose gets accused more often.
What a prompt cannot produce is a claim that could turn out to be wrong. Generated text is unfalsifiable because nothing behind it happened, so the only signal that survives is detail specific enough to check and awkward enough that nobody would invent it. One of ours: the scraper reported availability as InStock, the reading code looked for a different field name, and a price that failed to parse fell through to a default meaning out of stock, which pulled live products off a storefront. We fixed the field mismatch and kept the rule that came out of it, that a parse failure is never allowed to express itself as a product state. Trust comes from specifics somebody can hold you to.