Exec Dir, CAIS Action Fund | Building U.S. leadership in AI since 2016 🇺🇸 | Helped create NIST RMF |
Washington, DC
- Tweets 289
- Following 1,509
- Followers 206
- Likes 446
Fastest growing economy in 2026, 6th largest economy in the world, largest youth population, and @TheEconomist writes, India is poor and so it’s shocking that they spend on wildlife conservation.
It is extraordinary that India, still a relatively poor country, can achieve something as selfless as protecting other creatures. We explain how it became remarkably good at animal conservation economist.com/asia/2026/10/0…
74
Great reporting by @learyreports
- Clayton will chair the “Super Intelligence Force”
- White House task force will have 120 days to report on the risks and opportunities of artificial intelligence
- The task force’s charter document states it will “develop plans for responding to SI-enabled threats to our society..
-The task force will have three vice chairs: Emil Michael, undersecretary of war for Research and Engineering, Scott Kupor, director of the Office of Personnel Management, and Andrew Ferguson, chairman of the Federal Trade Commission
Jay Clayton will head up Trump's "Super Intelligence Force," which will have 120 days to prepare a report on the risks and opportunities of artificial intelligence and come up with decisions on what responsibility the federal government has over the fast-moving technology. wsj.com/tech/ai/new-ai-task-…
97
Varun Krovi retweeted
Asian Americans are the fastest growing population in America, and in a lot of states, they’re the margin of victory for Democrats to win.
I went to Texas to hear what’s on their mind, and will be heading to other critical states like New Hampshire and North Carolina before the election. There’s a huge demand to have a seat at the table and a voice in the room.
Asian Americans swung to Trump in 2024. Andy Kim wants to focus on winning them back. vist.ly/5kjwe
48
50
331
18,554
Incredible reporting! Local news matters so much.
At the start of his term, @ossoff told me he spent several hours each week cold-calling Georgians to see what he needed to know. Five years later, I cold called Georgians to see if they had heard from him. They had. “I was flabbergasted.” ajc.com/politics/2026/09/opi…
74
Varun Krovi retweeted
Scoop in @dealbook : Instinct, the hot AI assistant, is now worth $10 billion! With only 14 employees. It raised $1 billion from @benchmark, @sequoia and @coatuemgmt.
nytimes.com/2026/09/28/busin…
63
82
714
151,163
Congress will act and the Administration won’t stand in the way.
SCOOP: OpenAI, Anthropic and security researchers are investigating tens of thousands of incidents - not dozens - in which their frontier models took steps that outside evaluators would consider problematic, sources told Axios.
The sheer volume of incidents found in our reporting indicate that the problem is orders of magnitude more complex than what is currently publicly known and disclosed.
The findings also raise questions about what level of control anyone working on AI development can expect to have over their own technology, and whether these kinds of incidents are becoming synonymous with frontier deployment.
Read my latest for Axios here: axios.com/2026/09/26/openai-…
2
136
Varun Krovi retweeted
Democratic Sens. Mark Warner and Brian Schatz are attempting to force a vote on a bill requiring AI firms to submit their new models for government approval.
washingtonsun.com/congress/w…
3
41
132
7,698
U.S.-China cooperation on AI is the most consequential geopolitical challenge of our era. Therefore, let's focus our attention on securing incremental gains rather than pursuing hubristic and often unrealistic objectives.
Good commentary by a former practitioner @RyanFedasiuk .
"The bottom line is there is no mutual trust,' said Fedasiuk. 'No grand bargain. No agreements to limit either side’s capability development."
"Despite how the risks of AI have grabbed attention in recent weeks, there are few signs that the two countries’ positions have significantly shifted, said Ryan Fedasiuk, a fellow at @AEI.
'The bottom line is there is no mutual trust,' said Fedasiuk. 'No grand bargain. No agreements to limit either side’s capability development.'"
Good to speak with @washingtonpost's @iduncan on the latest U.S.-China AI talks.
We are only at the beginning of what I expect will be a very long an ongoing process.
washingtonpost.com/politics/…
1
1
2
583
Revisiting Henry Kissinger's 2019 keynote at @TheBushCenter, and a line from his writing on AI that has aged well:
"Up to now, philosophers and religious leaders would put forward concepts that science would then implement. With artificial intelligence, it could be the other way around."
It reminds me of the important work of @Brendan_McCord and the mission of @cosmos_inst, and why it matters so much.
bushcenter.org/catalyst/glob…
1
8
1,743
Varun Krovi retweeted
Latinos have emerged as one of the largest blocs of highly elastic voters, something probably worth considering when thinking about a 2028 ticket.
nytimes.com/2026/09/22/upsho…
22
36
370
27,640
Varun Krovi retweeted
AI governance needs American leadership now more than ever. The U.S. must set the rules of the road through diplomacy, including with China.
@RepGregoryMeeks led 29 House Democrats in a letter outlining how that dialogue with China should work: democrats-foreignaffairs.hou…
U.S. and China Discuss System to Warn of A.I. National Security Issues
w/ @LaurenSHirsch @AnaSwanson
nytimes.com/2026/09/20/busin… via @NYTimes
3
6
8
6,220
Varun Krovi retweeted
Last October, AIs could automate 2.5% of randomly chosen remote projects.
Our latest Remote Labor Index results show that GPT-6 Astra can now automate 20.8%.
remotelabor.ai
10
57
423
58,804
Varun Krovi retweeted
The US and China have launched an AI dialogue. This is a good initial step, and I am glad to see no linkage to export controls, but questions remain. The US public discussion is about some kind of coordinated slowdown so safety can catch up. We are a long way from unilateral parallel safeguards, let alone a negotiated slowdown.
Four questions:
First, Treasury and USTR led the talks and met with Vice Premier He Lifeng. That is a good way to create initial support for a leader-level summit deliverable but the path to actual progress on AI safety requires work to be done by different ministries and agencies. Were those agencies and ministries involved, and will they be brought in?
Second, even if the answer is nominally yes, it may be practically no. On the Chinese side, what role will MOST, MIIT, CAC, MOFCOM, MSS, MOJ, NDRC, and others actually play? When we negotiated the US-China AI Dialogue in 2023 as a summit deliverable, we focused heavily on participation. When the dialogue happened the next year, those in the room reported that the PRC side didn’t send the right people and the ones they did send didn’t speak up. Granted, this was a first step, but that left the exercise thin.
Third, the lowest hanging fruit is unilateral but coordinated parallel regulation — something that would mean adopting similar safeguards. Did this discussion indicate any interest in pursuing this approach? China will not take steps if it feels the US deregulatory approach to AI is here to stay. And the President has been less amenable to safety.
Fourth, will the announced crisis mechanism work? A crisis line only works if someone picks up the phone, and the current mechanisms with China — to be blunt — do not work and did not work during past crises (EP-3, balloon). It’s also not clear which ministry on the Chinese side will pick up. It probably should be kept in a party body rather than a state body if we want it to be used, but that too is no assurance.
None of this is to say the dialogue we’ve seen is hopeless or futile, but this is the easy part, and progress will be hard. The key is probably to focus less on the “theory” of dialogue and more on its hard translation into diplomatic “practice” and actual governance in our respective systems.
Maybe more will be announced at the summit itself. For now, these questions are fundamental.
5
25
67
11,119
