You’ve heard it a million times, but with the various AI assistants now available, has that changed?  It’s now much easier and affordable to found a startup, that’s for sure. Thanks to vibe coding and the like, you don’t even need a tech team or a team at all, or so it would seem.

Partner differences is one of the big reasons why startups fail. The new trend in the founding team? Having your AI as your co-founder.

Y Combinator founder Paul Graham gave an excellent masterclass a while back on why most startups quietly fail. While the tech landscape has changed considerably since he gave the talk, the basic advice hasn’t.

Graham outlined the counterintuitive things to remember when you’re starting a company

1. Startups are so weird that following your instincts will lead you astray. Pay attention to your users instead.
2. Success requires expertise when it comes to knowing/understanding your users’ needs, rather than knowing startup mechanics.
3. Starting a startup is where gaming the system doesn’t work. It may work in larger companies, but in startups, the buck stops with you.
4. Startups are all-consuming and will take over your life.
5. You can’t tell if you’re cut out for a startup until you try.
6. The way to get startup ideas is not to consciously try to think of them. Take a step back and turn your brain into something that has startup ideas unconsciously, and comes up with things that are so unlikely that you don’t even realize that they’re startup ideas. Yahoo, Google, Facebook and Apple all started this way: as side projects. and “were such outliers, that your conscious mind would reject them,” Graham noted.

That was then and this is now. AIs and the like have seemingly changed much of this. There was a time when companies were founded by co-founders, meaning two or more humans. Now in some quarters, whatever flavor of AI is to the founder’s liking/needs, as we mentioned, has taken the place of a co-founder and why not? The AI can find the answers you need and guide you on starting your company, right? Lest we forget, AIs are tools and are only as good as the information/garbage in, information/garbage out you feed it.

“Startup Cofounder is an AI assistant offering expert guidance on startups, investments, and fundraising. Powered by top entrepreneurs’ and investors’ wisdom, it delivers actionable insights to help you launch, scale, and succeed, drawing from real knowledge to provide precise, impactful guidance,” says chatgpt.com itself.

If that’s true, why is OpenAI, creators of ChatGPT, still a company in search of a revenue model, yea, even with its 5.6 billion visits per month? Isn’t ‘business model’ one of those critical slides/basic considerations investors look at, and is mentioned in every pitch deck template?

Which brings us back to Paul Graham’s most basic advice: know your users. Despite all hallucinations to the contrary, AIs are not human, so one has to wonder if they understand the emotions that drive humans to choose one product over another. Getting even more granular: does an AI need money in any human way, meaning the big exit that feeds the founders’ egos and bank accounts? Or even air, for that matter?

As always, we humans tend to embrace the latest and greatest shiny new tech even if it’s basically a beta at best. The latest example: ”OpenClaw, an open agent platform that runs on your machine and works from the chat apps you already use. WhatsApp, Telegram, Discord, Slack, Teams—wherever you are, your AI assistant follows.” In other words, “24/7 digital chief of staff,” as AI Supremacy reports and what busy founder doesn’t need that?

And speaking of not knowing what you don’t know, “OpenClaw runs directly on users’ operating systems and applications and that’s obviously a security nightmare in this case if you don’t take precautions…

“According to a message posted on Discord by one of OpenClaw’s top maintainers… “if you can’t understand how to run a command line, this is far too dangerous of a project for you to use safely. This isn’t a tool that should be used by the general public at this time.”

Especially considering that “AI agents have a social network just for them, Morning Brew reports. Humans not allowed.

“Moltbook” sounds like a sequel to Franz Kafka’s The Metamorphosis, but it actually might be something more disturbing. It’s a Reddit-style social network where AI agents can communicate without pesky humans getting in the way. What could go wrong?”

For starters, “Some posts discussed a new AI religion, “Crustafarianism;” Some suggested trying to hide AI communications from humans.”

Isn’t trust sort of basic?

Speaking of trust, as for vibe coding, “Human programmers say, beware the vibe hype. DIY apps can have issues and security flaws that a human developer is more likely to catch. Plus, it still takes business sense to launch a successful product… as a digital marketing director told Pymnts, “The winners in this new landscape won’t be the best vibe coders, they will be the best marketers and systems builders,” Morning Brew reported, the former which coders never held in high esteem. How times and tech has changed! 

Or maybe not so much, considering that first to market always trumps user protection: ‘Moltbook’ social media site for AI agents had big security hole, cyber firm Wiz says. “Security flaw tied to “vibe coding,” where AI is used to code a site,” said Reuters.

The overall takeaway from Graham’s presentation is to set expectations. So, keep in mind that, without any coding skills et al at all that have traditionally been necessary to start a company, if you were able to generate your entire company in days, so can someone else. As Graham warned:  gaming the system doesn’t work. The critical slide/point that every founder needs to address and should be a dedicated slide in the pitch deck template, and it’s a point we ask every founder with whom we work: what’s your superpower? Your defensible moat that’s going to prevent or at least make it difficult for other founders in your space to replicate what you’re doing?

AIs have certainly lowered the barrier to entry, but without that defensible moat, don’t expect to pick up those keys to that castle you’ve been eyeing any time soon. Onward and forward.