Remarkably, none of these are particularly cool, nor profitable. Democratization is about doing banal things that make AI banal.
Things that do democratize AI: - Open-source software - Blog posts and tutorials - MOOCs and books - Arxiv, arxiv-sanityhttps://twitter.com/math_rachel/status/856601368420679680 …
-
-
- End of conversation
New conversation -
-
-
What about data? With today's MOOCs, blogs, and insane arxiv, data is probably the most scarce resource for democratisation of ML.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Ok, let me add: - New open datasets (lots of these recently) - Cheaper hardware, especially in the cloud (happening exponentially fast)
-
The open datasets should also be easy to use, same for cloud processing. I use exclusively
@docker, on local and cloud GPUs, to share easily
End of conversation
New conversation -
-
-
Diversifying hardware. Most of deep learning is done on Nvidia cards. Competition can only bring benefits for the consumer.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Yeah but you're speaking as a specialist who can use that stuff. The public needs a product like an iPhone or Facebook to get utility.
-
You're context is misaligned with what these folks are talking about.
End of conversation
New conversation -
-
-
Would love to see more: https://medium.com/@mslavescu/learn-by-example-f539ad814117 … and easy to use demos like https://twitter.com/GTARobotics/status/855870333097836544 … so everyone can try them on their data.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
- Publicly available research papers (no paywall)
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.