Morgan Plummer
@mcplummer1789
Followers
465
Following
25K
Media
148
Statuses
1K
Sr Policy Director @americans4ri | Tech & National Security | Defense Reform | Organizational Design | All Things Midwestern | Opinions Mine
Naperville, IL
Joined April 2018
AI can shape public perception, disrupt markets, and steer missiles. The problem isn't that it's too powerful. It's that the same models get applied everywhere. @mcplummer1789 makes the case for clear boundaries in the gray zone for @WarOnTheRocks: https://t.co/WfeXcdIOZt
warontherocks.com
Editor’s Note: This is the third in a series exploring key AI policy choices faced by the Department of Defense & Congress. Please also read the first
0
3
8
The U.S. needs a framework for thinking about AI’s use in the gray zone of conflict; if we don’t draw clear lines then other states will. Some additional thoughts in @WarOnTheRocks today (H/T to @bradrcarson for the intro to Walzer): https://t.co/BhNtyxzQFD
warontherocks.com
Editor’s Note: This is the third in a series exploring key AI policy choices faced by the Department of Defense & Congress. Please also read the first
0
3
5
Leamer has the inside scoop: The Hun have never forgiven the USA for the HUMILIATION of Versailles. Now they're poised for their GREATEST REVENGE: enabling states to implement child-protecting regulations that MILDLY INCONVENCIENCE the profit models of Nathan's paymasters!
1
2
22
Selling the H200 to China would blow the top off computing restrictions for our biggest competitor on AI, eliminating a bottleneck for China that gives the US a key edge. Here's a look at performance and memory bandwidth of the H200 vs the H20 and Blackwell chips.
2
5
21
NEW: 280 state lawmakers from 43 states are urging Congress to reject AI law preemption in the NDAA. We cannot tie the hands of state lawmakers working to protect kids, workers, and Americans from emerging AI harms. Read the letter: https://t.co/AVuSC1aaCe
4
14
31
A vote in favor of AI pre-emption reminds me of the vote in favor of the Iraq War: a decision that will lead to deep regret and which, in a short time, will be an albatross around the neck of anyone running for office. I would know. When I was in Congress, I supported the Iraq
11
16
110
My initial two thoughts on the would-be EO on preemption of state laws. First, the EO is the result of a failed approach to jam through pre-emption. This EO could have been promulgated months ago. It wasn't, probably because the legal theories behind it are dubious. It's
5
13
57
Heard today at @scientistsorg and @FLI_org ‘s AIx Military Integration round table: “making bad decisions faster doesn’t help you win.” Maybe we should be deemphasizing AI adoption speed at the Pentagon and focusing instead on clearer thinking about AI usage.
0
0
1
The NDAA is about giving American troops the tools and policies they need to fight & win our nation’s wars. Including a poison pill, like a federal moratorium on state AI regulation, isn’t just bad public policy…it jeopardizes our national security. https://t.co/XEt3727JhY
techpolicy.press
A proposed federal moratorium on enforcement of state AI laws failed last summer, but lawmakers have resurrected the idea, writes Cristiano Lima-Strong.
0
1
1
99-1. That's how the Senate voted against AI preemption earlier this year. Now House Majority Leader Scalise is trying to sneak a similar measure into the NDAA at the last minute -- blocking state laws that protect children, artists, and workers. https://t.co/hJXEzkKMdX
3
5
23
It is utterly fascinating to me the number of people making an argument for some type of de facto govt subsidy for the frontier labs by drawing comparisons to (for example) nuclear energy without being willing to accept ANY of the regulatory framework core to that same industry.
1
2
6
I'm an academic. I like capitalism.
Steven Pinker: “The hatred of capitalism is deeply ingrained in academic culture” https://t.co/pxtqYh4m1e via @welt
5
1
50
The Dept of War needs to be more thoughtful about how it’s defining “trustworthy” AI if it wants to avoid its past acquisition failures. The warfighter definition of trust should take precedence over all others. Period. More thoughts and recs here:
warontherocks.com
Editor’s Note: This is the second in a series exploring key AI policy choices faced by the Department of Defense and Congress. Please also read the first:
0
0
0
Just a reminder that your inputs (incl. docs) to LLMs are (in most cases) being used for training. As LLMs personalize service (keeping your history), they know a lot about you. This information could be retained indefinitely, used to train LLMs, regurgitated verbatim when
2
2
7
10 years ago today! This work remains as important to our nation as it was then. @mcplummer1789, Diana Banks Thompson, Lloyd Thrall, Dan Madden: visionaries all.
0
1
2
Some days, I think people aren't really paying attention to what is really happening in the world.
The same people trying to prohibit superintelligence are the ones who repeatedly tried to ram through the United Nations a treaty outlawing, among other things, autonomous drones. Without which Ukraine would now be a Russian colony.
1
1
8
OpenAI has provided no evidence it has mitigated the mental health risks associated with its products other than announcing some advisors and reducing sycophancy from a high starting place. Seems premature to be declaring victory and ramping up the porn + emojis again
49
63
547