In the oft-repeated words of Marc Andreessen, "software is eating the world." I agree with that, but that doesn't mean you can eat software. When it comes to how we sustain ourselves, the art of food can't be lost in pursuit of the science and technology of food.

I ordered a number of prepared meals recently from a delivery service that shall remain nameless. Among the dishes was an omelette. Carefully packed in ice packs inside a cardboard box, shipped from almost 800 miles away. It was clearly cooked a couple of days prior, and emerged from its wonderfully intricate packaging a shapeless, colorless, blob. That sad meal was economies of scale, logistics and software wizardry gone horribly wrong.

Here are 4 lessons from that experience.

1. Taste rules

Just because this company could prepare that meal in a factory, and then put that poor omelette on a 800-mile road trip, doesn't mean it should have. Similarly, just because you can strap wireless glasses to your face and record everything that parades in front of you, doesn't mean anyone will want to join you in your tech-enabled isolation.

Most startups don't make it. That's the nature of startups. Recently we've seen a handful in the food tech business shut down as competition for customers and funding has ratcheted up. There is more where that came from.

The question for those that remain is how to win? A key part of the answer is very simple: taste rules. And that very straightforward maxim is something that every startup can learn from, especially as tech entrepreneurs make their way into every industry -- from farming to finance, healthcare to transportation.

2. Don't get blinded by blind spots

The blind spot many of these food tech companies have is that they are treating food like so many lines of code or units of hardware. They are companies run by software engineers who got into food with the technologist's mantra of better, faster, cheaper. But few seem to care enough about what the food tastes like, what it looks like -- the total sensory experience. That is a deadly recipe in the food business.

It doesn't matter how cool your packaging is, or how great your app is. If the food doesn't deliver on quality, integrity, and taste, it's never going to work.

3. The customer should always come out on top

The same warning applies to whatever product or service your company offers. Does it deliver what customers want or need, above all else? Does your business model improve or detract from the experience? Are the tradeoffs you are making in the people you are hiring, the quality of materials, or the source of manufacturing destroying your shot at a spectacular future with your customers? Does your unboxing experience trump what lies inside?

We make trade-offs on a daily basis in business. But if you don't know the consequences of those choices you can easily kill the thing you are devoting your time, energy, and creativity to build: a strong, lasting connection to customers and a market; a company that is adding to the world rather than subtracting from it.

4. Software as an (elevated) service

This doesn't mean there isn't a place for technology in food companies, and across every industry. There is a huge role for technology to play in the way it can digitize a business and bring it scale. Technology has brought a "chassis" of convenience to the things we do on a daily basis whether that is ordering food on demand from an app, or watching a car magically appear with a single tap. But the next layer of all these tech-enabled products and services should be about the people they serve. That means more personalization, more customization, and more ways to delight riding on top of that technology-led chassis.

If software is better, faster, cheaper, that should manifest as more convenience, greater access, fewer inefficiencies and cheaper, higher quality products as a result. What it doesn't mean is that a truck (or a drone) should travel almost 1,000 miles to drop off my breakfast.