Hacker News Reader

Comment by alantrrs | Hacker News Reader

Hi! I'm building a platform for scientists to run and share their experiments (anything computable) including their whole research environment. I want to make scientific research easily discoverable and replicable.

First, what are your thoughts on this market?

Also, some advice: I'm currently building the prototype based on my own experience as I'm my own user. That's the only thing I've been focusing on, I haven't looked for external feedback yet and haven't spent much time looking for people to join me. I figured both of those things will be much easier once I have the prototype ready. Am I on the right track or should I be doing everything at once?

View on HN

As a former academic I will share my thoughts here. I agree that there is a great need for this, but not demand from those with the data. In science you are effectively discouraged from sharing data since doing so means you lose the ability to rehash the data and publish a new paper. The second problem is scientists as a general rule won't pay for software or any sort of saas.

Can I suggest an idea for a related product; a platform to make collaboration between different areas easier. Right now if you have an area where you lack a skill or background it can be very time consuming finding the right people to collaborate with, especially internationally. Make it easier to create collaborations that result in publication and you will have a product with paying customers.

Daniel, thanks for sharing your thoughts. My initial focus will be in public datasets (e.g. Imagnet). People are already making their code and results available online through different mediums. However, I'm hoping having a specialized platform where you can share your whole environment will improve discoverability and reproducibility which should have great impact in collaboration as well.

Another suggestion might be to collaborate with an established academic to put in a grant to develop this platform. The funding agencies are worried about this problem and they may fund a good application to build this. You could then concentrate the commercial effort on the front end.

I wish you luck with this project as it certainly is needed.

I think that's a great market. The lack of reproducibility in scientific experiments is an elephant-in-the-room problem. I feel like I read a new story about it every week.

Even though you are your own target user, it's dangerous to build a full prototype without talking to actual users. It's likely that you will find that what you thought people wanted does not match 100% with what they actually want.

Do you have mockups or screenshots of the tool? I would recommend lining up some research scientists who can serve as product advisors. Tell them about your idea and ask them what they would want in this kind of tool. Try to get a commitment from them to use it after you build it - that will force clarity on their side.

Not only will that prevent you from potentially throwing away code, it will get you started building the kind of deep advocates in the science community that you will need to get a movement like this started.

Jared, appreciate the feedback!. Any chance you could introduce me to some people in the Open AI team? I think this would be a great fit and I'm very interested in their thoughts.

Reply on news.ycombinator.com