In 2020 - I've found myself caring less about #AI research and more about the infrastructure society needs to use it

I just feel like we've got cars - but no highways

Pick almost any problem you'd actually want to solve with #AI and there's basically no data available to you
Like - I'm sitting on a sofa right now - no clue whether it's likely to cause me back-pain or not.

Presumably millions of people are sitting on couches right now.

... but am I able to use all my fancy #AI tools to study it?

Nope...
I ate some food today - it might make me healthier - it might not....

Presumably millions of people ate similar foods to me today...

And they'll all have different health reactions

Can I study it?

...nope
I could literally go on and on for hours about all the things we don't know about our world around us - despite the fact that we *could* in theory be studying all of it.

We have loads of statistical power-tools at our disposal.

#AI is not blocked by research.
#AI is blocked by the fact that computer scientists don't collaborate with social scientists to build data-sharing networks that people can trust - that are aligned to people's social good.

Btw - this is what #privacy technology is really about.
Anyway - I think this is the most important #AI problem to solve because all this work we've done isn't really that interesting in the real world unless:

1) data is available to solve real problems
2) the solution and the resources to create it (data) are public goods
You can follow @iamtrask.
Tip: mention @twtextapp on a Twitter thread with the keyword “unroll” to get a link to it.

Latest Threads Unrolled:

By continuing to use the site, you are consenting to the use of cookies as explained in our Cookie Policy to improve your experience.