Mine isn't from academia specifically. It was along the same vein though. I wanted to figure out how different components worked and ended up making my own full stack NLP platform exploring every topic (question answering, named entity recognition, summarization,...). Within about nine months or so, I got it done enough (fully implemented, maybe not perfect like a full team of engineers could do) to build things on top of it. However, it's accurate enough relative to competing or similar products.

I decided to think of ways of making money on it, and just built something on top of it, now I have customers and all that fun stuff.

A research project and a startup (software obviously) both have code parts to them.

It really comes down to the problems you're attempting to solve.

In my case, I love the position I'm in now since I have the ability to make a lot of different kinds of products by just adding a web app in front of a rest API.

As long as it solves a particular need, customers don't care how or even when you built it as long as it does what's advertised.

Research projects are how startups can start though. That's why I think they say get out there and just build something interesting to you (in this case a research project) and see where it takes you.

At risk of simplifying too much: they fulfill very different goals, acquiring public knowledge vs getting rich.

You might be forgiven :-) for thinking that an academic research project can be identified with its publications and declared outcomes, because that's what is usually presented.

At its best, an academic research project is an exploration of some narrow corner of the space of ideas, perhaps with no immediate application. Depending on the area, its publications represent both improvements and dead ends. (At its worst it's just a load of exaggerations and distortions of the truth designed to fool an already self-deceived funding body into providing salaries, a travel budget and equipment. Most projects strike a balance).

On the other hand, startup=growth [1].

[1] http://www.paulgraham.com/growth.html

I think it has more to do with the intent behind the project.

Startups are businesses, and the intent behind them is typically to make money.

Research projects don't intend to make money. Or, they do, but as a fairly low priority.

It seems to me that research projects typically come across as "cooler" or more interesting because they don't exist to earn money, they exist to see if something cool could be done. To me, it seems that the reason a large number of startups are uninteresting is because they're attempting to piggyback on existing success (hence the X for Y stuff).

A useful idea from academia can be used to create a startup, but a startup is more than idea. The key thing that is there is that it must have growth. From Steve Blank: "A startup is a temporary organization in search of a scalable, repeatable, profitable business model." Paul Graham also has the essay http://www.paulgraham.com/growth.html where growth is again emphasized.

In 'Academia', researchers may not actually own the intellectual property resulting from their research.

> Ask HN: What's the difference between a cool research project and a startup?

The outcome. If the outcome is a success, that was the plan all along. If the startup fizzles out, it was really an interesting research project.