And this post shows exactly why cluely is such a damp squib destined to fail. Zero substance just grandiose claims that hold no water. They’re spending tens of millions of dollars of investor money “going viral” instead of spending investor money actually building a great product.

Cluely is just some kids larping as founders, the perfect example of the worst of the AI boom, it’s as if the entire thing is AI generated, a simulacrum of a startup. There are a million more talented and capable teams of young people out there.

If someone from a16z reads roy’s post they might finally realize they should ask for their money back.

i'm a cluely hater but the fundamentals of the article are pretty good. simply, "viral content" is and ad. and ads need eyeballs and conversions.

whether they end up actually building a good product that isn't just cheating will determine if they will be seen as clowns or geniuses.

I've been hearing "cluely" everywhere, or at least on a bunch of infosec podcasts about various insane bugs and exploits in their software. And when I heard the description of it, "an AI assistant that sits in meetings with you," and the emphasis on "nobody knows it's there," with the subtle wink-wink in the marketing... I was like, "hang on... this sounds a lot like the guy who got expelled from Columbia for making an app to do this in his Amazon interview..."

And now I realize it's the same guy! I don't know if he intended this from the beginning — he probably didn't, considering nobody forced Amazon to email Columbia and advocate for his expulsion. And maybe he wasn't even expelled. But damn. Love him or hate him, this is a masterclass in viral marketing and playing the hand you're dealt. Good for this dude. I wish him the best but personally the whole concept of a silent meeting participant gives me the creeps, even before considering the ethics of the "interview cheating" use case.

Roy, if you're reading this... if you succeed it will not be due to your engineering skill. Focus on marketing and put some funding towards engineering brains who can make a robust product that passes enterprise muster. Build some corporate growth hacks into the core product (e.g. SSO + calendar invite pathogen + Figma trick of inviting anyone with same corporate email and charging per seat in a post-facto "true up" at end of billing period). Then go hunting for sales people who have no ethical qualms with silent meeting participants taking notes and helping them look less clueless to their prospects.

A Cluely Angel's Thesis

Here are 6 notes + personal takes I wrote down when reading Cluely

@openmartai and myself:

1. Everyone wants to go viral, but virality has its prerequisites: the content itself has to actually be worthy of more attention.

- Why can shitposting go viral? Because it sparks emotional resonance — it’s polarizing, or it’s so absurd it doesn’t make sense.

- Novel ideas can go viral because they use creative methods (a unique celebration post, for example) to present something ordinary (Pylon's bell ringing act)

- Genuine, high-quality videos

2. generally need 10K+ followers base, and that post itself must've reached a decent number of ppl. If it’s truly worthy of virality, then it will spread widely.

3. (My personal take) Virality rewards for those who have been preparing for a long time but haven’t yet been seen. Virality is not something easy to copy.

4. Virality is an acquisition channel — it sits at the top of the funnel, and it does not solve retention. That’s why you need to make sure you’re either in the stage of collecting feedback, or you’ve just post-PMF and need users to grow. These two scenarios are the most ideal conditions for virality. I totally agree with this

5. Even if Cluely’s social media hit several million views, those didn't translate into real downloads. You don't ask for everyone to know you — only the focus group. So maybe, you doesn’t necessarily need to become another pylon/ artisan/ fluently. If our DAU and retention are more important (which are just closer to revenue), then making knowledge content — maybe even just posting a ZoomInfo vs. Apollo comparison on reddit — would likely be more effective than trying some flashy stunts on X.

6. Virality as top-of-funnel acquisition works better for consumer products > B2B. For B2B, going viral = just branding.