Remember when every startup slapped “AI” on its pitch deck and investors threw money at them? Engineer.ai was the poster child for that era, and it just got caught with its pants down.
The Wall Street Journal dropped a report that confirms what many of us suspected: Engineer.ai isn’t using AI to build apps. It’s using human engineers. Plain old, flesh-and-blood coders. The company, which raised nearly $30 million from SoftBank and others, has been selling the dream of a fully automated app factory. The reality is a lot more mundane.
Engineer.ai’s founder, Sachin Dev Duggal (who apparently also goes by “Chief Wizard” — yeah, that’s a red flag right there), claimed on stage last year that his platform could build 80% of a mobile app from scratch in about an hour. The WSJ found that the company doesn’t use AI to assemble code. Instead, they outsource the work to human engineers in India and elsewhere. The AI? It’s mostly a decision tree for assigning tasks and some basic natural language processing for estimating prices. That’s not cutting-edge machine learning. That’s just a fancy spreadsheet with a chatbot wrapper.
The company was sued earlier this year by its own chief business officer, Robert Holdheim, who alleged that Duggal was telling investors the product was 80% done when it barely existed. Classic startup drama, but with an AI twist.
This isn’t just one bad actor. It’s a symptom of a much bigger problem. According to MMC Ventures, startups that claim to use AI attract up to 50% more funding than those that don’t. And they suspect 40% or more of those companies don’t actually use any real AI at all. The incentive structure is broken: investors want to hear “AI,” so founders say “AI.” Everyone knows the game, but nobody wants to call it out.
I’ve seen this pattern before. It’s the same as the “blockchain” craze a few years back, or the “cloud” hype before that. Slap a buzzword on a traditional business model, and suddenly you’re a tech unicorn. The difference here is that AI is harder to fake convincingly. You can’t just say “we use AI” and expect people to nod along forever. Eventually, someone asks to see the code.
Engineer.ai isn’t alone. The article points out that a lot of “AI” in the real world is just humans behind the curtain. Think about content moderation on Facebook and YouTube — they claim AI does the heavy lifting, but it’s actually armies of contractors in the Philippines or India reviewing the worst of humanity. The AI handles the easy stuff; the humans handle the mess. That’s not AI. That’s outsourcing with a marketing budget.
What gets me is the scale of the deception. Engineer.ai raised $30 million on a promise they couldn’t deliver. That’s not a startup pivot or a failed experiment. That’s deliberate misrepresentation. And they’re not the only ones. The .ai domain (from Anguilla, of all places) has doubled in use over the last few years. Companies are literally buying a domain name just to signal they’re an AI company. It’s absurd.
Look, I’m not anti-AI. I use it every day. But this kind of bullshit hurts the entire field. It makes it harder for legitimate AI companies to get funding, because investors get burned by the fakes. It erodes public trust. And it wastes money that could go toward actual innovation.
The Engineer.ai story is a cautionary tale. If a startup claims to use AI to automate something complex, ask for specifics. What model? What training data? Show me the code. If they can’t answer those questions, they’re probably running on human power and hype. And that’s not AI. That’s just a regular business with a good PR team.
Comments (0)
Login Log in to comment.
Be the first to comment!