Ask HN: My CEO wants to go hard on AI. What do I do?

> At a smart business, you work for your customers.

Well, then you're bound to the amount of money customers can give you.

It's usually less than a VC partner with a god complex can give you.

I came here to write this exact response.

Folks work in mission driven organizations. In organizations that are optimizing for dollars you see quality fall by the wayside.

Mission driven organizations that understand the levers of capital are the most enjoyable workplaces of my career.

You want a CEO who is customer obsessed, not a CEO who is maximizing for dollars in the bank at all costs.

It has to be mission first for a startup to be great, really for any critical endeavor.

Call your existing ML models across the stack "AI," and wow him by reporting that you already are "AI"-first!

It seems that another round is necessary, and VCs want AI. You and the CEO are competent and transparent. This is the premises.

You just need to have the conversation with the CEO, but in the right order.

Come to him/her with objective data, but without your conclusions or concerns, just your observations about where this is headed and if he sees at what rate this AI focus is eating ressources. Ask the CEO what it means for the company. Let him/her do the talking.

From this point, the CEO have the data and you have the company policy. Know you can talk on equal grounds on the matter with him or her. Good luck.

This approach looks to me like a classic mistake of assuming all your potential paying users both are and want to be (these are different things!) power users. Similarly it might be suffering from the problem of living on the cutting edge and assuming everyone else is too.

I like the post and what it goes over. I agree with huge swaths of it, especially the misuse of ai tools right now. I don't agree that anyone doing it that way (like the gmail team called out over and over here) is either stupid or naive. I don't agree every consumer tool should become an agent building playground. I don't agree building said playgrounds are easy (I think it's much much harder to design a good agent building version of a product than a good product). I don't agree consumers want everything they use to work this way. I also think it ignores the very real problem of ai bullshitting and the handcuffs that puts on people using it for mission critical things like paying the mortgage.

Ironically I think this post falls into its own trap of not thinking about the next step. Yeah a really good email agent product as demod sounds great in a world where nothing works this way yet. However, a world where every product I use has to be re-engineered from scratch, with various unknown and non-customizable (and enshittifying) LLMs under it, with various training and fine tuning, unknown access to data, and no interop is the wood-frame horseless carriage the author is mocking. That would be a terrible situation worse than the current one.

Rethinking for a world of ai agents, it would be better if products empowered consumer agents instead of tried to supplant them. In THAT world products stay simple and just expose a bunch of tools and patterns that each consumer's custom agents can effectively use. Making that work well for your own product is actually viable AND helpful AND doesn't force your users to change their behavior if they don't want to. Anything I, the consumer, do to handle or prompt my own agent pays off across my entire ecosystem and investment can be focused on the right things by the right people (ie gmail should make email tools not agent tools).