the limits of shipping early

3 minute read

Rabbit R1 and the Humane AI Pin launched to non-critical acclaim. For many, they were unimpressive or just simply not good. Marques Brownlee, of MKBHD fame, drove the point home. The discourse remained hot for weeks, with many piling on and others coming to the defence of said companies. Regardless of your position, either defending the pioneers and supposed earth shakers or joining in laughing at the new main character, there’s a lesson in it all. A couple brain cycles later, it dawned on me: there’s a limit to shipping early.

My Twitter thread captures the sentiment. We’ve had the slogan banged into our heads for a decade:

Ship early, get feedback, iterate

Let’s not be fooled, this is unequivocally good advice. In pure software plays, especially those delivered through a SaaS model, this is easy to do. Customers pay their subscription and are continuously shipped improvements. A win for both parties; the company validates their product, the customer receives a product that is forever nearing perfection1. But this advice is the same as all advice – it only works in the correct petri dish of assumptions.

Change the assumptions, it breaks down. Rabbit R1 and Humane’s Pin are hardware products (though funnily enough, both come with a subscription to maintain functionality but who’s keeping score). Hardware products have a different set of constraints. The most glaring is that the hardware device itself cannot be updated with ease. I need to upgrade my hardware each time for improvements. That comes at the cost of heartache from a separation of my sweet, hard-earned dollars. The cost-benefit is wildly different. Purchasing a bad device leaves me with a perpetually bad device. If I am to purchase one, it must be worth the money. Contrast to software where I can come onboard for $10/ mo, try it out and have the option to quit when I feel like2. If I like it enough, I know (or hope) it’ll improve over time. The risk is substantially lower, I’ll be willing to separate from my dollars. The hardware constraint adds a dimension that’s not easily ignored. Ship too early, you’ll be found lacking. In cases, you’ll be fighting reputational damage, hurting your future products. In others, customers will wait for the next version of the device, leaving you starved of the almighty dollar.

Both the devices are attempting to displace an incredibly troublesome adversary: smartphones. The bar is ridiculously high. It matters for several reasons. The singular inconvenience of having to carry a second device is enough for customers to switch off the idea entirely. A smartphone can do most, if not all, of what these devices can do, either natively or with the addition of apps. And in the case of the Humane Pin, with a better interaction model: screens. A product that rivals a smartphone has a comically high activation function – the point at which a product is good enough a customer is willing to suffer its flaws. To succeed in that arena, the device has to be equally as good, if not better, otherwise it’s over before the starting pistol fires.

Just with those two assumptions gone, that you’re playing in a market with existing low(er) quality products and that you can rapidly update your product3, the effectiveness of shipping early is diminished. The product has to be good out of the gate just to survive, let alone succeed. This is all to say that there are circumstances where shipping early may come with a side order of negativity. Sometimes the bar for entry is just sky high, it can’t be ignored. There are limits to shipping early.

  1. In theory, not in practice ↩︎

  2. Yes this is not universally true. There are products, like DAWs (e.g Ableton), that have substantial upfront cost. They’ll face similar constraints as hardware vendors ↩︎

  3. For both products, the assumption that they can rapidly update their product is only partially false. They can still update the software, which both have been doing. ↩︎