Ask HN: What should we fund at YC Research?

This is so fundamental I doubt that it would raise an eyebrow for more than a few seconds:

Ask yourself what sorts of environments entities like YC depend on for their very existence...?

Freedom to build wealth and disperse that wealth, essentially without constraint...

Please choose something that helps support the formation, or maintenance, of an environment that values human welfare and actively supports it....

In other words make your choice a philanthropic one, in the truest sense of the word...

I think we should find a way to take the best content from Math Circles and get that to far more students in a digestible format.

Drilling down even further, if the Question is "what is the _single_ most useful thing that can be done to increase future scientific literacy and accelerate technological progress ?" .. then I think the answer must be to teach multiplication by drawing rectangles.

vis : https://quantblog.wordpress.com/2013/04/13/distributive-rule...

Its a really good antidote to 'math is magic, a set of rules you apply without knowing why'... its based on physical intuition we already have, and it leads very naturally to deeper more abstract thinking :

You can go quickly from counting out areas of integer sides via grouping, to multiplying fractions and decimals, to expanding (a+b)*c, to (a+b)(c+d), you can introduce 1+3+5+7+...+n = n^2 and 1+2+3+..+n=n(n+1)/2 , you can introduce primes as 'non-rectangle' numbers .. all the while improving facility with basic computations, and developing better math intuition.

Kids who believe they are dumb at math can relate to this kind of physical/visual construction - if you play tennis or basketball or fold paper or saw wood you already have a good intuition of area.

It seems not much math is being taught in school before age 12 .. We don't have to wait or invent anything groundbreaking - we can take concrete steps now, we know what works - ie. we have the vaccine, we just need willpower to get it to everyone.

Software productivity tools

We should look at how we can improve the ROI for education.

Millions across the world, especially in developing countries, drop out of school because they (and/or their guardians) see no benefit from long-term investment in education. Others who somehow manage to stay in formal institutions are exposed to decontextualized education that they cannot realize their full potential.

There will be many different solutions to it. One of them could be a large-scale, technology-immersed learning system that teaches a broad range of topics to students through a vocation. The vocation could be decided based on the learner's interest and the local resources. For example, in northern Nepal, child walk through perilous snow-covered hills and mountains to recover Yarsagumba ("Himalayan viagra"), a fungus with aphrodisiac and medicinal value. Instead, the kids can be educated progressively in details about different aspects surrounding Yarsagumba - mountain climbing, biological systems, business, marketing (where they could sell the collected Yarsagumba), greenhouse and high-tech farming systems, technology, etc. - without disturbing their Yarsamgumba collecting activity.

This is a simple example. Since a diverse topics are being taught and practiced, learners would not be restricted in the same vocation.

Scaling and automation of biological, medical, and biomedical research. There are huge tracks of research areas which essentially are technologically backwards. From poor usage of statistics (see all the the replication issues in cancer research, psychology, and etc) to the lack of high quality programming tools to efficiently solve problems. More startups building cheap, applicable, and open-specification hardware with open source software could be a huge boon. The hardware and equipment situation of many scientific labs is stuck in the 90's at best and often based on equipment developed in the 50's and 60's.

Having just finished a master's in material science in applied bio-mechanical research I am personally aware of the lack of basic equipment and tools. Machine learning techniques for example could be applied to many areas of experimental research in biological settings but most labs do not have the time, money, or expertise to apply it. Another possible option would be a sort of "professional software development co-op" for scientific research. Being able to buy (subsidized) development resources might be incredibly helpful. The cost could possibly be offset by forming partnerships to commercialize valuable research, similar to a YC model where the few successes could fund the other.

The system for basic scientific research also seems broken. Much of the data used in many fields (especially nuclear research) is based on (very excellent) research from the 50's. We're essentially riding on accrued capital which will likely reach a limiting point in the coming decades.