Jared Dunnmon (@jdunnmon) profile picture

AI, Energy, Security. Views are my own.

Joined November 2017
Photo by Jared Dunnmon (@jdunnmon)Photo by Jared Dunnmon (@jdunnmon)Photo by Jared Dunnmon (@jdunnmon)Photo by Jared Dunnmon (@jdunnmon)Photo by Jared Dunnmon (@jdunnmon)Photo by Jared Dunnmon (@jdunnmon)Photo by Jared Dunnmon (@jdunnmon)
Jared Dunnmon retweeted
it’s not local or cloud. it’s local *and* cloud, as a hybrid system, that increases our gross domestic intelligence our research (w/@jdunnmon and @JonSaadFalcon) shows the US can add 30–40% inference capacity using the existing 70–80m ai pcs/workstations in circulation, and hybrid orchestration cuts compute by 62% vs cloud-only link to research in comments below 👇
sam: local inference is a terrible allocation of scarce resources “a lot of people want local inference, and somehow they assume that the problem is solved now because we have open-source models, we're going to be able to run them on our computers, and so we're not going to need all the stuff in the data centre but from a global allocation perspective, actually it's quite terrible to have everybody, like, have a bunch of the world's limited supply of DRAM sitting in their own computer and being used five percent of the time you want maximum utilisation in the cloud if you're trying to deal with this very supply-constrained environment somehow many people's intuition is the opposite of that”
2
3
8
1,201
The Administration’s decision to reduce intel sharing to Ukraine & access to Starlink equates to aiding the invader & the war criminal who bombs cities. Does the Administration think it will get ANY support for the remainder of the 119th Congress? Their agenda is dead unless they change paths. nytimes.com/2026/10/14/world…
810
1,830
116
6,537
164,529
Unfortunately, I think we could live in a world where datacenter countervalue is a standard practice sooner than later. We should think hard about how to navigate that world.
as we (w/ @JonSaadFalcon + @jdunnmon) wrote in @ForeignAffairs, kinetic attacks against data centers are a direct threat to any nation’s ai infrastructure. concentrating ai capabilities in a few large, exposed facilities is a strategic vulnerability. models running locally are the resilient alternative link to @ForeignAffairs article in comments below 👇
2
323
Jared Dunnmon retweeted
dreams do come true 🥹. excited to see our work (w/@JonSaadFalcon, @HazyResearch, john hennessy and @Azaliamirh) feat. in a major way in @FT. the world is becoming increasingly less dependent on centralized cloud ai. we are just getting started 🚀🌖
8
13
6
92
35,126
Jared Dunnmon retweeted
Replying to @JensenHuang @nvidia
880
2,750
1,106
30,036
10,530,508
We have information that Moonshot AI distilled Anthropic’s Fable for the development of its K3 model. To do this they developed a sophisticated internal platform to conduct large scale distillation against U.S. models, allowing them to quickly switch between multiple methods of access to avoid detection. Moonshot AI has also acquired GB300-equipped servers and has accessed GB300s in Thailand, likely to train its AI models.   The United States strongly supports the free and fair development of AI, including a thriving competitive ecosystem that spans frontier models, specialized systems, open-source frameworks, and open-weight models. Legitimate AI distillation used to create smaller, more efficient models plays a vital role in this open innovation ecosystem. However, large-scale, covert industrial distillation aimed at stealing proprietary U.S. technology and undermining American research is unacceptable.
3,833
1,311
2,260
14,659
9,422,148
Could not agree more with @pmarca and @quxiaoyin on this one, this is exactly the argument I made early last year when it became clear that models from Chinese providers would be optimized for the Huawei stack sooner than later... foreignaffairs.com/china/rea…
5
1,481
I think this is directionally spot on. Programmatic evaluations built with the collective knowledge of industry, adopted widely, and updated at the speed of relevance is about the only structure I can see that yields the outcomes society wants.
1
4
267
Happy Fourth of July! We’re a proudly American AI lab, and we’re excited and optimistic about what our country will do together for the next 250 years—no doom and gloom here! Thank you to the patriots and people who built America and the people still building her more perfect union. We love this country and believe service to God, Country, and family are prerequisites for a meaningful life and to keep our country great. God Bless America!
12
1
48
10,911
Ukraine launches TrophyLab: we are opening access to captured Russian weapon technologies for our global partners. Every missile, drone, and vehicle seized on the battlefield is now a source of knowledge for the free world. Through this secure platform, allied governments, labs, and defense tech manufacturers gain access to deep technical data, reports, and vulnerabilities. Users can also request physical equipment for testing, significantly shortening the development cycle for countermeasures. What was meant to be the enemy's secret advantage is being dismantled to defend democracy. Join the platform: 🔗 trophylab.mod.gov.ua/en/
698
5,606
928
30,668
4,071,840
Awesome work from friends at Cartesia. So cool to see @krandiash @_albertgu @nimit_sohoni et al. continue to push the frontier of what’s possible in voice! Also, @krandiash seems like he is a practiced croissant consumer…
We released Sonic-3.5 and Ink-2, the #1 streaming models for text to speech and speech to text you can use in your voice agents today. New architectures enable new frontiers for speed and quality. We're now the only provider to have #1 models for both speaking and listening.
1
9
997
My latest with @Avanika15 and @JonSaadFalcon. If America fails to create a level playing field for its open-weight ecosystem, it will cede a large portion of the AI future to China. We can and should act now.
“The next phase of the U.S.-China AI competition will be in no small part decided by which models become the default on the world’s local devices,” write @jdunnmon, @Avanika15, and @JonSaadFalcon. foreignaffairs.com/china/chi…
1
1
4
1,233
Jared Dunnmon retweeted
The dominant story in AI has been the growing cloud: bigger clusters, larger models, more gigawatts. We believe the future is in the opposite direction: on-device inference, smaller models, watts instead of gigawatts. Today we're releasing @OpenJarvisAI v1.0: a personal AI assistant that lives, learns, and works on your device.
49
93
29
608
162,472
Jared Dunnmon retweeted
we built a thing so you don’t have to buy your intelligence from @sama's meter its called openjarvis: an open-source framework for personal ai, on personal devices download it, use it, and top the energy leaderboard for a chance to win a mac mini!
Personal AI should run on your personal devices. So, we built OpenJarvis: a personal AI that lives, learns, and works on-device. Try it today and top the OpenJarvis Leaderboard for a chance to win a Mac Mini! Collab w/ @Avanika15, John Hennessy, @HazyResearch, and @Azaliamirh. Details in thread.
5
5
24
3,931
Jared Dunnmon retweeted
Uncovered this old photo from a SpaceX board meeting nearly 10 years ago with some of our crew. Grateful to my @DFJvc and @dfjgrowth partners @FutureJurvetson and @rglein for spearheading the investment into what’s potentially become the most important company in history. Proud to be an investor!
2
2
1
57
14,012
Jared Dunnmon retweeted
Shifting from the AI conversation for a bit, this is a huge moment. The first use of an American precise mass system in combat, and it is one that the United States copied from Iran!
CENTCOM's Task Force Scorpion Strike - for the first time in history - is using one-way attack drones in combat during Operation Epic Fury. These low-cost drones, modeled after Iran's Shahed drones, are now delivering American-made retribution. 🇺🇸
2
5
1
29
4,596
Jared Dunnmon retweeted
The U.S. government just essentially announced its intention to impose Iran-level sanctions, or China-level entity listing, on an American company. This is by a profoundly wide margin the most damaging policy move I have ever seen USG try to take (it probably will not succeed).
107
810
56
5,138
321,485
This is unsurprising -- in fact, I wrote about it in Foreign Affairs last February. foreignaffairs.com/china/rea… Tldr the US should consider (a) unfair trade practice tools (eg Section 301 / 337) (b) reemphasizing export control enforcement & (c) adapting FDPR to this situation.
We’ve identified industrial-scale distillation attacks on our models by DeepSeek, Moonshot AI, and MiniMax. These labs created over 24,000 fraudulent accounts and generated over 16 million exchanges with Claude, extracting its capabilities to train and improve their own models.
1
123
Jared Dunnmon retweeted
One cost more people (even prospective founders) should consider: if you did something else, would the impact you're having today still happen somewhere? If yes, you're missing the opportunity to change things for the better
5
3
140
23,728
Admiral McRaven eloquently describes how soldiers from America’s NATO allies have laid everything on the line to support the US, despite our occasional differences. He is right. Denigrating their sacrifice is not just ungrateful — it is unacceptable. theatlantic.com/ideas/2026/0…
2
5
376