A
1
ABOUT THIS DOCUMENT
2
This document is a living repository of high-leverage, underserved opportunities where a new organization could radically enhance America's AI deployment readiness and competitive position.
These infrastructure, coordination, and capability gaps either slow down deployment of transformative AI, create unnecessary bottlenecks, or leave critical vulnerabilities unaddressed.
It is therefore a WORK IN PROGRESS.
3
We will be prioritizing making progress on addressing many of these, including attempting to help start teams/organizations for a small number of them at a time.
4
This is more To-Do list than wish list.
However, "done" will not look like we've solved the problem; rather it will look like a person or organization who is focused on incremental progress toward addressing a critical need, and could effectively leverage 10-100x more resources allocated to this problem.
5
6
DISCLAIMERS & STATUS
7
Draft: This artifact is currently in draft form. We are sourcing early feedback and collaboration from various individuals.
8
9
Confidentiality, Ownership, & Licensing: This list is maintained by our team, but we do not claim it as our IP.
We're standing on the shoulders of giants and neither ask for nor deserve credit for all of these ideas; analogously, we accept blame for bad ideas and mistakes.
You should probably check out the lists on the "other resources" tab.
Please do not contribute any confidential information or restricted IP that you do not have rights to share.
Any contribution will be considered entering that information into the public domain. Information on this page is free to use without license.
10
11
Scoping: We're defining a "high-leverage opportunity" as a capability or infrastructure that would unlock significantly better, more reliable deployment of advanced systems at scale.
We consider a problem underserved if there isn't someone who focuses on solving this problem full-time who could scale up a solution with the right opportunity.
That said, naming gaps and framing potential solutions is highly opinionated; while we welcome debate, we are striving for progress, not consensus.
N.B. some rows are very USA-centric, primarily because Atlas Computing is a US 501c3. We'd be happy to support analogous efforts for other countries, but this document will tend toward focusing on America.
12
13
Prioritization: The sorting on the public list does not reflect prioritization or expected impact.
14
15
Sharing: We work in public, and are soft launching this. If this gets attention, let it be based on track record of execution, not just ideas.
16
17
Contributing: New contributors and contributions are welcome! Please add them as comments.
If you have many things you would like to add in bulk, or would like to contribute to our private addendum (e.g. confidential information that could inform prioritization and resourcing decisions) please email Evan.
We're trying to only lightly edit and curate at this stage, so feel free to add whatever you're currently worried about, especially if it's in your area of expertise.
18