Kelvin Yu
@kelvinotcelsius
Followers
1,277
Following
419
Media
16
Statuses
106
Awaiting paperwork. Past: AI and R&D policy @housescience , 2x founding engineer, angel investor. Opinions are my own.
Washington, DC
Joined August 2018
Don't wanna be here?
Send us removal request.
Explore trending content on Musk Viewer
Le Pen
• 646833 Tweets
Kyiv
• 244287 Tweets
#TheLoyalPin𓇢𓆸
• 83004 Tweets
#CDTVライブライブ
• 63613 Tweets
Okhmatdyt
• 47777 Tweets
#SixTONES_ここに帰ってきて
• 32070 Tweets
Good Monday
• 30709 Tweets
オールスター
• 28299 Tweets
Kiev
• 26904 Tweets
ドウデュース
• 23912 Tweets
サブウェイ
• 14685 Tweets
nerve
• 13387 Tweets
ナイチンゲール
• 13153 Tweets
#BiS解散10周年
• 12932 Tweets
石丸伸二氏の敗訴確定
• 12769 Tweets
支払い命令
• 11475 Tweets
石丸構文
• 11308 Tweets
Pinned Tweet
To kick off 2024,
@ansonyuu
and I are excited to release , an open collection of policies, tactics, and reforms for advancing scientific and technological progress.
12
42
189
Excited to announce that I’ll be spending my next chapter at
@housescience
focusing on emerging tech R&D, CHIPS implementation, strategic competition, and AI policy!
13
1
82
Brought to you by the great minds that also produced master pieces like “Governments Should Not Fund Research”
This is such a strawman,
@CatoInstitute
(also you should disclose that your board vice chair has a $33B TikTok stake).
3
12
72
4
5
33
@danwwang
’s 2023 letter has great commentary on this.
What’s driving people to flee: “It’s starting to feel like the only thing scarier than China’s problems are Beijing’s solutions.”
Massive surge of Chinese immigrants trying to cross the US border through Mexico.
Like in the Cold War, the direction of migration is extremely telling for how rival great powers are performing…
29
127
854
2
3
25
A friend’s looking to hire someone for a full-time role to lead AI policy for their House committee (Republican staff).
Rare opportunity to lead a policy portfolio on day 1, engage frequently with Congressional leadership, etc.
2
8
24
Great thread. Most of these struggling companies had the wrong goal of building general-purpose machines that looked cool but failed to scale production w/ better unit economics.
The winning path is to produce specific parts faster with better margins than trad techniques
The high hopes for 3D printing of the last decade, however, have been unfounded. Most of the companies in the field are unprofitable, more like advanced research projects than exceptional businesses.
But the techniques may be adopted by other firms anyway.
4/n
1
1
24
0
3
21
Where most think tanks opine about problems in blogs and reports,
@IFP
gets their hands dirty with substantive execution work.
Congrats team!
@calebwatney
@rSanti97
etc
0
3
20
Many wonderful individuals helped us create this resource. We are especially grateful to
@rSanti97
@BrianNosek
@charlesxjyang
@ashwinforga
@masao_dahlgren
@erinsm28
@elidourado
@mslkmp
@Ben_Reinhardt
@SusannahGlickm1
@dicooke34
@shreyjaineth
@jasonjzhao
@a_nnawang
@humford
...
1
0
17
Lobbyist
Prime contractor
National Security Council
Hill committee staff
Hill intern
2
2
18
I’m default suspicious of industrial policy critiques “we don’t have X so we can’t do Y and might as well not try.”
There is nothing intrinsic about a random island that makes it great at making semis. It’s the result of intentional investment, cultivating talent, etc.
CHIPS critics say it won't work because U.S. lacks a semiconductor workforce. This is circular. U.S. lost that workforce because industry expatriated itself. How do you bring that workforce back if TSMC et al don't start building in U.S.?
33
73
334
3
2
16
Tune in at 10am ET today to hear our esteemed witnesses testify on how to advance AI in our national interest
@JasonGMatheny
@Farshchi
@ClementDelangue
@DeweyAM
@ruchowdh
0
6
16
Despite their measurable impact on quality of life, our society does not focus enough on how to increase scientific and technological output. Many falsely believe that progress “just happens,” thereby reducing human agency to influence innovation.
1
0
14
DoD departments trying to claim 'space' for themselves in the 1950s was truly comical:
Army: the moon is just “the high ground”
Air Force generals: space was “just a little higher up” than air
Navy: “outer space over the oceans”
3
0
15
...
@_neilhacker
@jasminewsun
@billyjoelburke
@michejafang
@_jzhao
@aj_kourabi
, Alice Wu, and Andrew Sosa for contributing to lever entries and providing invaluable feedback.
1
0
12
This is false in many ways, including in public policy, where states and institutions can intentionally boost innovation with a variety of interventions. We collected a few of our favorite examples in the hope that they inspire more discussion on experimenting with such policies.
1
0
11
In addition, encouragement and inspiration from folks like
@calebwatney
@tkalil2050
@arbesman
@Noahpinion
@HeikeLarson
@stuartbuck1
@Parthion
@hamandcheese
@heidilwilliams_
@TobiasAHuber
and Willy Chertman helped us push L4P to completion.
4
0
12
The two types of bullshit talk DC hones you to recognize real quick are benign report-speak and exaggerated excitement -- both are often used to hide broken or mid performance
The reports consistently take this incredibly straight-faced tone.
Local contractors "are unfamiliar with U.S. quality standards and expectations."
1
0
12
0
1
10
British Navy implemented a policy of self-innovation-sabotage in the 1800s under the logic that "that any naval technological innovations would only threaten Britain’s naval predominance”
They would change course after *surprise surprise* French and US developed ironclads anyway
2
1
9
People assume the government is a giant bureaucracy where individuals don’t matter, but the reality is everything meaningful is driven by a handful of serious people. Gallagher was one of them in the House — he will be missed dearly.
0
0
7
Check this out if you're thinking about what's next! I've lived in three co-living houses for extended periods and hands down
@benjlaufer
's had the highest talent/agency filter + intentional community structure.
Welcome to Somewhere Campus
We're coordinating 12-20 people to move to small and walkable town, for one year, and to work on bold ideas to improve how we live.
Receive up to a $15,000 stipend to join our micro-campus and be a part of an intentional and informal r&d collective.
42
30
399
1
0
8
In 49 BC, Caesar agreed to a deal from Cicero to give up 90% of his legions. But Cato killed the whole thing because he was ideologically devoted to nothing but complete submission, thus triggering Caesar to march on Rome
Not the best guy to name your think tank after
Brought to you by the great minds that also produced master pieces like “Governments Should Not Fund Research”
4
5
33
3
2
6
@daviderady
@aj_kourabi
1) using micro points (eg tesla and edison) to draw macro generalizations (gov funding unnecessary for applied research)
2) private charity complements federal funding but it can’t replace the latter at any comparable scale
3) counterpoint to their iPhone example
1
0
5
couldn't ask for a more inspiring and thoughtful partner on this!
we made a big collection of policy levers to help accelerate science & tech
29
51
491
0
0
5
Advances in theory are underrated in catalyzing hype, funding, & talent.
We are incredibly lucky that the first significant quantum algo just happened to have a major security application.
Without it, there would've been little justification for the state to spend billions funding a technology that would take decades to mature.
1
0
0
1
0
5
Looking for someone technical, has a deep AI/Silicon Valley network, and available to start within the next month.
Can vouch for quality of character and emphasis on serious policy work over political shenanigans.
Dm for more info.
0
0
4
One example is Shor's Algorithm (1994) and quantum computing. SA factors large numbers in polynomial time on a quantum computer, which makes breaking current encryption schemes much easier.
SA remains the most significant quantum algorithm ever discovered.
1
0
1
What are examples of single theory breakthroughs that catalyzed large-scale public investment?
5
0
3
1
0
3
@MoritzW42
@remnote
Excited to see what you build next! Remnote is by far the product I use most everyday.
0
0
1
Export controls cost little political or financial capital to implement.
Winning a technology market or capability, on the other hand, requires heavy and targeted investment, sustained will, and institutions willing to take risks -- all scarce resources in DC.
0
0
3
US policy bias towards export controls instead of domestic R&D and production falls into the same trap of playing to not lose instead of playing to win.
1
0
3
Dostoevsky on EA vs. e/accs, AGI, Xi Jinping, and Silicon Valley:
0
0
3
Brain drain immigration policies are such low hanging fruit w/ high impact. The VAST majority of Chinese (and Indian) PhD grads are still in the U.S years following graduation.
8) And lots of great recommendations in the talent section:
"America should be attracting defectors and accelerating China's brain drain, to our national benefit, by welcoming the best talent on the planet to the U.S.” 💪💪💪
2
6
56
0
0
3
There exists no profession with a greater asymmetry between value creation and value capture than an academic/gov researcher.
Patent licensing is the only potential few instruments for financial value capture, but as
@stuartbuck1
details, institutions consistently block this
Consider Katalin Kariko, the recent Nobel winner for her mRNA research.
She was driven out of academia by the Univ. of Pennsylvania. Nonetheless, Penn kept the patents on her research, refused to sell it back to her, & recently earned many times more than any other university.
1
1
5
1
0
2
@hamandcheese
They’re about to enter secret back door negotiations for a sole source contract, so both
0
0
2
This was against the backdrop of the Eisenhower admin wanting to create ARPA & NASA in response to Sputnik and fears of Soviet ICMBs. In response, the departments protested and tried to claim advanced military research, particularly in space, as their own.
1
0
2
@jasonjzhao
Trade schools/technical colleges are obvious at secondary level
Academies of Loudoun is a fantastic model for technical training for HS. They offer pathways into 4y degrees or trade schools and have a heavy focus on HS internships.
2
0
1
1
0
1
@aj_kourabi
True for physics more broadly (eg tens of $ billions going into particle accelerators to validate theories) which is upstream of everything else lol
1
0
1
We are incredibly lucky that the first significant quantum algo just happened to have a major security application.
Without it, there would've been little justification for the state to spend billions funding a technology that would take decades to mature.
1
0
0
Shor's discovery in 1994 accelerated QC timelines by at least 1-2 decades by creating a step-function change in funding, hype, political prioritization, etc.
IQT (CIA's VC arm) remains the
#1
funder of quantum startups in the world.
0
1
1
There are plenty of better critiques of industrial policy generally and CHIPS implementation specifically, but ones of the former pattern tend to be bad and intellectually lazy
0
0
1
@auren
Insanely different ecosystem from SF/LA/NY that’s centered around political capital. You acquire many novel and valuable mental models that you otherwise could not get exposure to.
2
0
1