Econ PhD Student @MIT | Prev @TheEconomist in 🇬🇧 & 🇮🇳 | Hoosier4life
Cambridge, MA
- Tweets 626
- Following 1,279
- Followers 5,150
- Likes 6,848
Pinned Tweet
Last week I led a trip for 20 friends across India w/@SparshAgarwall and @arjunssoin
We explored Delhi, Bangalore and Darjeeling, meeting people doing cool things across Indian politics, tech and civil society
Was especially fun intro'ing first-time visitors to India!
🧵1/
12
8
127
18,784
Personal update: I'm beginning an economics PhD @MITEcon!
Crazy to think that exactly 3 yrs ago on this day I joined @TheEconomist. At the time I thought it would be a 3m internship, but it turned into one of the most rewarding experiences of my life 1/
43
15
744
110,512
Excited to share what I've been up to this year!
On the cover of this week's @TheEconomist is a 6-part special report on India’s economy by Tom Easton+me. It asks
- What it would take for India beat 6% growth?
- And is Modi up to the task?
🧵on the key arguments 1/
14
175
645
245,850
NEW ESSAY: "Dispatches from India"
A year ago I was living in New Delhi, India 🇮🇳. I traveled across 10 states and 30 cities+villages writing for The Economist, including a cover story on the economy
Since then I’ve been chewing on how to capture what I learned
It still surprises me how few resources go into understanding India given it has 1.4bn people (will peak in 2060), is growing at 7%, and could be as large as US/China in my lifetime
My 7-part reflections below
Contents:
1.Decentralisation-why I’ve become more bullish about Indian democracy
2. Illegibility-the hilarious way I learned about black money in politics
3. Revolutions-Modi, Trump, and igniting a social media mob
4. Technology-how Bangalore and Silicon Valley are converging
5. Growth-why India is not an Asian Tiger but could still become the world's biggest economy
6. Culture-my favourite cities, traveling to villages, and book/content recommendations
Bonus: what it was like writing for The Economist! 1/
22
90
691
118,397
New piece in @TheEconomist! An important second-order impact of remote work: Firm “boundaries” are blurring.
Remote work changes the calculus of what to do, and crucially, what not to do. It boosts freelancing, outsourcing and offshoring, just as Coase would predict: 1/
10
109
469
224,927
NEW in @TheEconomist. Industry labs have driven rapid progress in large-scale AI leading to @OpenAI's ChatGPT. I ask:
1. Why did academia pass the baton to firms?
2. Does Google's LaMDA (which I got access to) or ChatGPT give better dating advice?
3. Where are things headed? 1/
5
42
225
132,161
In most countries, college-educated voters have turned against right-wing populists
Not so in India. Call it the Modi paradox
My piece in @TheEconomist this week, inspired partly by countless convos with family/friends in India who 😍 Modi but 😬 Trump
economist.com/asia/2024/03/2…
103
54
171
456,868
Excited to share my biggest project for @TheEconomist to date: ~10k words on the future of digital finance!
Looking past the hype around crypto/CBDCs, I argue that good-old banks, state-led networks and fintechs are quietly transforming global payment systems worldwide 1/6
4
18
139
27,796
Humlum/Vestergaard link AI adoption data with Danish admin data till june 2024--amazing data! they find:
-widespread adoption (most workers use AI and a third of firms have their own models)
-2.8% surveyed time savings
-but impact on wages/hours in admin data is 0!
4
36
139
25,191
I looked at the literature on science funding models for this week's @TheEconomist.
Key point: there is a big opportunity in testing new funding models in ways that can be rigorously evaluated--for example with randomised control trials (RCTs) economist.com/finance-and-ec…
7
27
118
3/ Ch 2+3 continues on India's relative strengths--the financial and corporate sectors
Contra to narratives of crony capitalism, mkt concentration is actually falling in India. Modi favours national champions, but less than you might think economist.com/special-report…
2
13
69
8,725
6/ Our final chapter lays out a reform agenda touching on everything form innovation policy to beefing up city governance
A striking stat: only 15% of Indian govt employees work at the local level vs 60% in 🇨🇳 or 🇺🇸 leading to poor state capacity economist.com/special-report…
3
15
69
10,644
2/ India is growing at 6-6.5% a yr, slower than China’s heyday of 10%+
But that's still solid. Growth is harder today bc manufacturing is less of a route to riches
India has the world's #5 economy, by 2027 it will be #3. #1 is even plausible by 2070 economist.com/special-report…
3
10
64
17,706
3/ What are India’s strengths?
Ch 1 notes the big shift over the past decade has been better infra–physical+digital. Roads+railways are growing fast. Digital payments and electronic forms work.
This reflects Modi’s forte–executing big projects
2
9
65
25,130
Had fun interviewing @sama at last week's #ItalianTechWeek in beautiful Turin, Italy. We discussed:
- future AI progress and bottlenecks
- why he thinks current models are safe but still wants proactive regulation
- growing Italy's tech scene
- AI market structure
1/3
4
1
70
6,747
Great to chat about my recent @TheEconomist piece on the blurring of firm boundaries with @KellyCNBC and @PeterDiamandis on @CNBCTheExchange today.
And more importantly wish my dad @karthiksramani happy birthday on TV! (And Peter's mom too!) 1/
cnbc.com/video/2023/02/02/mo…
5
8
59
8,857
5/ Why has🇮🇳struggled to⬆️investment and beat 6% growth?
Historical reforms left state-level red tape on land/labour/power mostly untouched
Mass education remains poor contributing to a weak labour mkt. Strongman govt and rising protectionism don’t help economist.com/special-report…
1
6
51
6,084
4/ Ch 4 is the tricky trio. It’s well-known that 🇮🇳 is an IT services superpower, but the govt wants it to be a manufacturing power too
There is some success–14% of iPhones are now assembled in India. But corp investment and exports have yet to budge
economist.com/special-report…
1
6
44
8,091
First, skilled jobs are more distributed across geographies, in part reflecting a rise in remote work.
Second, cross-border service exports, including IT services, are booming.
Third, the market for skilled freelancers has doubled since 2018.
2
14
44
6,583
8/ I'm indebted to colleagues, friends and sources who made this report possible.
Read the full report here: economist.com/special-report…
Acknowledgements: economist.com/special-report…
4
45
17,515
Our cover story in this past week’s @TheEconomist is on the new age of venture capital. The industry is being turbocharged by a wave of capital and competition. This comes with serious risks, but is overall promising for global innovation (long thread) economist.com/leaders/2021/1…
2
17
46
7/ The problem is that the needed reform requires buy-in from states and social groups that might lose from change. But Modi’s weakness has been building such a consensus
Our leader argues that must change for India to sustain its rise over the long-term economist.com/leaders/2024/0…
3
2
44
6,124
Just landed in NOLA for #ASSA2023! Have work you think would interest Economist readers? I'd love to chat. DMs open.
I'll also be presenting newly updated work with @I_Am_NickBloom showing the persistence of the #DonutEffect in big cities worldwide. A few key results: 1/
1
2
40
7,998
If you're interested in writing for @TheEconomist, the @MarjorieDeane internship I started with is accepting apps till Sep 30
This is one of the best ways to get a job at the paper--pls let me know if you have any questions! economist.com/finance-and-ec…
2
7
42
6,955
Fourth, perhaps most interestingly, as firm's grow in size, and adopt more technologies, their "outsourcing intensity" grows.
This can be seen by looking at what share of a firm's cost of sales is purchased from other firms.
1
5
36
12,557
More details in the full piece here: economist.com/business/2023/…
2
11
39
12,277
My Economist special report on India: economist.com/special-report…
Subscribe for future writing: arjunramani.substack.com/
1
3
43
6,383
Second, ChatGPT is not the only game in town.
Some Googlers helped me compare ChatGPT and LaMDA (Google's currently-private chatbot model). We asked math and reading questions, and also for some dating advice. Some results below, but in short, ChatGPT faces stiff competition 3/
2
6
34
44,228
Looking forward:
- How big of @OpenAI 's first-mover advantage from the user data it's collecting?
- Will bigger continue to be better?
- And what happens when quality online text runs out?
Extra thanks @nathanbenaich for data help. Full piece here: economist.com/business/2023/…
2
6
36
7,922
In China’s 20th Party Congress, held today, the party reaffirmed its commitment to turning China into a global technology powerhouse.
As part of this week’s @TheEconomist cover package, I assessed the state of the tech race b/w the West and China (1/) economist.com/briefing/2022/…
4
10
29
Quick thoughts in @TheEconomist on @sama's sacking:
-$90bn val is at risk
-short-run: a more safety-focused @OpenAI will slow AI rollout (@AnthropicAI + @GoogleDeepMind were already more risk-averse)
-long-run: AI could be more competitive if/when @sama and @gdb start a newco
2
33
4,584
planetary intelligence chief has good taste in analysis. but you know sophon was messing with him here because @theeconomist doesn't have bylines ;) @3body
1
1
26
2,160
For this week’s @TheEconomist cover @EconCallum and I wrote about a real-time revolution in economics/policy. The change led to a spike in research, benefited economists who collaborate and affected policy in a way not possible in 2008 economist.com/briefing/2021/…
3
7
32
I'm grateful to have worked w/incredibly kind+curious colleageus and covered the most interesting stories of the day from AI to India's economy. @r_shanbhogue, @PatrickFoulis and @zannymb are the best mentors+ editors I could have asked for and have helped me grow so so much 2/
2
31
4,089
Our cover on 🇮🇳's election:
-the BJP's loss of a majority in parliament is a big win for democracy
-shd acknowledge this might slow economic reforms in short-run
-in long-run it's probably good; changes India needs require consensus-building 1/2 economist.com/leaders/2024/0…
2
7
27
6,089
Our reading/listening recs for understanding effective altruism (EA)--the growing movement to figure out how to do the most good with one's time/money. Includes both positive and critical views. Thanks to several friends in/adjacent to EA for their help economist.com/the-economist-…
3
3
28
My grandfather was a civil servant in Indian govt so I grew up hearing stories about his adventures across the country. Never expected to report on its bureaucratic intricacies one day
Fantastic book and I expect it to be the key reference on the topic @karthik_econ @ivishiknew
Throwing money at a state lacking capacity is like adding fuel to a car near a breakdown: it won’t get you very far.
My colleague @arjun_ramani3 's brilliant review in @TheEconomist of @karthik_econ's epic new book.
economist.com/finance-and-ec…
1
26
2,504
Jasmine's a superstar--she just quit her job after 4 years driving product @substack to write/podcast independently (and she's one of the best writers I know)
Only gets better from here!
new podcast up on BUILDERS VS. PRIGS ft. @arjun_ramani3, economist and my favorite fellow niche-tech-ideology-watcher
I always learn a ton from Arjun. here's what we covered:
00:49 Zuck's new look
02:05 PG's new essay on wokeness
07:08 origins of the "tech right"
16:02 why tech leaders admire Malcolm X
20:35 Arjun’s conversations with Altman, Collison, and Lonsdale
28:20 progress politics in SF, India, and the UK
42:58 buying the dip on effective altruism
51:27 gains from trade with AGI
2
2
24
4,509
First, industry is leading the way in large-scale AI (though it builds on years of work from academia). Why?
- Training big models is computationally intensive and therefore expensive
- In AI research outputs can rapidly be commercialised (a herald to the era of Bell Labs) 2/
2
3
21
4,446
this is the best evidence i've seen on AI impacting entry-level jobs and is an update for me
but if i had to look for alt explanations...
1) h/t @BasilHalperin: AI-exposed jobs are generally higher pay and might have a higher return to experience. they are akin to a *long-duration asset*, so when rates go up you reduce hiring more than other jobs (results also hold in non-college sectors h/t @ZMazlish, but perhaps diff explanations across sectors)
2) In the short-run you only have so much money to spend on capital v labor (credit-constrained firms). so even if capital doesn't immediately automate labor, if the expected return to capital goes up you reallocate cash to AI experimentation over hiring
This is similar to Ricardo's "wage-fund" theory and is also consistent with an *anticipation effect* of future automation
Is AI already impacting the job market?
A new paper from me, @erikbryn, and @RuyuChen at @DigEconLab digs into data from ADP.
We find some of the ***first large-scale evidence of employment declines for entry-level workers in AI-exposed jobs.***
A thread on our paper:
3
30
2,500
So it's bittersweet to be making this step. Being shortlisted for journalist of the yr by the @Wincottfound w/@econcallum for our AI writing was a surprising bonus. I will keep contributing to the paper periodically and am excited to test ideas from my reporting in my research 3/
1
25
3,943
New essay in @gradientpub! Using the idea of "bottlenecking", formalised by Baumol in the 60s, @zhengdongwang and I explain why we think an AI-driven growth explosion is unlikely
A 🧵of technical, social and economic bottlenecks that together could limit AIs economic impact 1/10
1
6
23
5,919
So @SBF_FTX got the wrong end of the St Petersburg paradox...to be expected with enough gambles, as he himself admits!
My piece for this week's @TheEconomist on what the crash of FTX/Alameda, companies run by effective altruists (EAs) including SBF, means for the EA movement 1/
1
4
16
My leader for @TheEconomist on the market downturn's impact on VC. A few points: down-cycles tend to take over 2 yrs to bottom-out which suggests pain ahead. But many startups raised loads of $ in 2020-21 so are in better shape than one might think. economist.com/leaders/2022/0…
3
3
21
In this week's @TheEconomist I wrote about how India's economy has done under Modi
TLDR if you adjust for global growth, India contd on trend. The govt had wins (GST, digital stack, pub infra) but dropped some reforms (factor mkt dereg, devolution) and became more protectionist
1
22
1,367
The piece gives three broad explanations:
1. class/caste politics
2. economic growth
3. elite admiration for strongmanship (there is some evidence in the piece of the limits to this)
And concludes that w/o a far stronger opposition, elite support for Modi will hold
8
3
19
9,915
Rly enjoyed this convo with @michaeljswalker
As for the clip, India's conglomerate capitalism is clearly second-best; cronyism remains a big issue. But we should acknowledge that concentration is⬇️and the govt at least partially disciplines big firms
economist.com/special-report…
The UK election might be the most entertaining happening in the world right now. But it isn't the most important.
In my final episode on India's election I spoke to @arjun_ramani3 about the country's economy. It includes this provocative part-defence of 'crony capitalism'.
2
20
7,043
Mfn'ing has historically had three superpowers for development: tradability, tech-intensity, and labour-absorption
Services increasingly have the first two, but the techy/tradable portions are lacking on the third. So the "growth ceiling is lower"-@rodrikdani
My piece with @arjun_ramani3 this week. The combination of automation and Chinese industrial dominance limits the opportunities for manufacturing-led models in the developing world. How much can the growth of services exports fill the gap? economist.com/finance-and-ec…
4
18
9,211
5900 students live on campus and though weekly testing is required, only close to 5000 tests are conducted on an average week. Compliance will need to increase for a more expansive reopening next quarter.
Students are supposed to get tested every week. But analysis of Stanford data suggests that 19% of students have been out of compliance with this policy since arriving on campus in September.
stanforddaily.com/2020/10/22…
1
18
Huge fan of @srajagopalan and @IdeasofIndia so it was a true honour to be on Shruti's pod!
Have a listen and you'll get a sense of the debates we had @TheEconomist when reporting this six-part report on India's economy
In the latest episode of @ideasofindia @mercatus I speak with Tom Easton and @arjun_ramani3 @TheEconomist about their 6-part special report “India Express.”
We spoke about our reasons for optimism and pessimism for India’s growth story, the nest stage of reforms, the informal sector, unemployment, India aspirations and much more.
mercatus.org/ideasofindia/ar…
1
3
20
3,693
If you're a college student interested in data, advocacy, or journalism, check out the Stanford Open Datathon, a unique event for students world-wide to kickstart projects related to open data. Super excited for this and apps due Mar 15! datathon.stanford.edu @StanfordODP
6
19
5 counties went blue in IN. 3 are home to the largest unis in the state, 1 is the capital/big city, last is a Chicago suburb. Shows the continued concentration in educated/urban areas.
Tippecanoe-Purdue
Monroe-IU
St. Joseph-Notre Dame
Marion-Indy
Lake-Chicago suburb
2
18
HUGE congratulations to my childhood idol @DwyaneWade (aka #wadecounty) for joining the NBA hall of fame on a unanimous ballot!
The OG 🐐 in my books!!
But I hope you move on from padel to real tennis soon :)
4
1
18
1,888
It's true India struggles in labour-intensive manufacturing
But Shahi Exports ($1bn rev, 100k full-time employees, 10%+ CAGR, 50+ factories mosty in the South) bucks the trend!
I wrote about what India can learn from Shahi in last wk's @TheEconomist
economist.com/asia/2024/09/2…
What inhibits India’s global competitiveness?
2 nice articles on jobs,
highlighting
- skills mismatch
- labour laws & labour unrest
- high energy costs
- state differences (TN outlier)
bloomberg.com/news/newslette… @jyots43 @JohnReedwrites
on.ft.com/3TSQWif @menakadoshi
2
2
16
3,653
Check out the data team’s new Stanford in the 2010s series! The past decade witnessed the rise of CS...
Stanford in the 2010s is a new series featuring data visualizations on how things changed (or did not change) in the University community. For our first installment, we graphed the percent change in degrees conferred by major over the course of the decade. stanforddaily.com/2020/04/25…
1
16
New in @TheEconomist: How good can China get at generative AI? The piece look at:
1. How China's best models compare to the US/UK's
2. China's access to the three big inputs to AI: data, chips, and talent
3. Why AI adoption may matter more than innovation
1
17
2,424
Super excited to be co-leading @StanfordODP - feel free to DM me if you’re interested!
The Open Data Project is making data about Stanford more accessible and transparent. Through our dataset portal and educational initiatives, we aim to support advocacy and data journalism efforts that are increasing accountability on campus.
APPLY HERE: docs.google.com/forms/u/1/d/…
![[Join us in our mission to make open data at Stanford & beyond accessible to all. Apps due
Oct 26, 11:59 p.m. PDT. Apply now at tinyurl.com/sodp20]](/pic/orig/media%2FEkviuOsUcAEGQoC.jpg)
ALT [Join us in our mission to make open data at Stanford & beyond accessible to all. Apps due Oct 26, 11:59 p.m. PDT. Apply now at tinyurl.com/sodp20]
1
15
We ended the trip in Darjeeling, the home of Sparsh’s family tea estate and ground zero for @AltCarbonIndia's ERW facilities
We hiked the Darjeeling hills and learned about the history of Indian jazz and macro from @tajmahalfoxtrot + @DuttaAnirudha 4/
1
1
17
1,797
Check out our Class of 2020 Senior Profile! From political beliefs to dating, questions ranged a broad set of topics. Suggestions for next year appreciated
The class of 2020 has dealt with one of the most unusual years in Stanford history. With COVID-19 forcing the mythical senior spring online, seniors have had their years and lives upended. We collected 327 survey responses on a variety of topics. stanforddaily.com/2020/06/14…
14
recommended.
makes legible a lot of what is typically obscure about the practice of AI research.
also has one of the best arguments I've seen for why definitions of AGI have thus far been unsatisfying
Year two, in which I utterly solve AGI with the Parmenidean Ascent zhengdongwang.com/2024/01/01…
15
1,855
chapter by chapter notes on abundance, read read right after visiting abundance-pilled shanghai and taipei: goodreads.com/review/show/74…
16
845
In Bangalore we spoke with @pramodkvarma, architect of India’s digital stack, and founders/investors about how India's tech/AI scene might develop differently than US/China's
My friends were surprised with how similar the vibes in Bangalore felt to Silicon Valley 3/
1
17
1,411
I learned a ton, including many lessons on how to organize group travel
But the most rewarding part was sharing an important part of my life with close friends
Here are @SparshAgarwall's daily threads for more detail: nitter.app/SparshAgarwall/status/…
And more thoughts from @willdepue here: nitter.app/willdepue/status/19535…
Day 1 of the Discovery of India Trip was off to a great start.
Featuring the finance ministry, bful frescoes, interactions with the chief economic advisor, understanding how AGI might impact the Indian economy, and trying to navigate India’s North & South Blocks.
1
17
1,929
Grateful to have @HarinContractor as a mentor and role model through @TheWLP as a I begin my own economic policy journey. Congratulations Harin!
Harin Contractor, MPP'08 (@HarinContractor) will be honored with the 2020 Alumni Award at the Minorities in Public Policy Studies (MiPPS) Alumni Gala this fall. Learn about his journey 👉 har.rs/39KN3DV
1
1
13
Just spent 2 weeks in the Bay Area-am more optimistic about it than a year ago
Yes, some activity left due to Covid+#WFH. But that's quite small relative to the region's dominance
One example: over a long horizon the Bay Area's share of US VC investment is basically flat
1
15
"the model does the eval"
loved reading this. ZD takes you inside the head of an AI researcher shows you how to organize your intuitions about AI progress
1
1
14
1,444
Why I think AI will surely change the world (it already has), but may not accelerate economic growth by much.
But the best part of this podcast is origin story of my friendship with @zhengdongwang
🎙️ Will AI levy the transformative economic impacts many predict? Just what should we expect as the technology advances? @spaniel_bashir speaks with @arjun_ramani3 and @zhengdongwang:
4
14
5,050
In Delhi we got a crash course in geopolitics as Trump hiked tariffs on India the day before we met India’s chief economic advisor and leader of the opposition
We visited the new parliament building and went out with some of the country’s youngest MPs (our age!) 2/
1
15
1,332
Really the best gift I received in the mail over Xmas break :)
I was looking forward to many pieces from this week's edition, but none more so than @arjun_ramani3's on the origins of inflation and his read of John Cochrane' new textbook, The Fiscal Theory of the Price Level economist.com/finance-and-ec…
1
14
2,722
biden passing the torch is for liberals what the assassination attempt on trump was for conservatives
1
1
12
2,717
Check out the V1 of the Stanford Community Survey! (A play on the American Community Survey) We will be surveying students quarterly to see how things change over time.
Introducing the first ever Stanford Community Survey by @StanfordDaily! @94305_Daily will now regularly survey students about their opinions on Stanford institutions, their Stanford experience, national politics and more.
Check out our insights here: stanforddaily.com/2020/05/12…
1
12
Great short summary of the early lit on whether AI is skill-biased
Generally, AI reduces inequality in productivity across skill groups within jobs
But this may not reduce wage inequality overall (depends on relative effects across industries + general equilibrium effects)
From @emmabwiles and @johnjhorton: “There is a fair bit of evidence from the economics of AI literature that when you give people access to AI tools, the lowest-skilled people are the ones who benefit the most.”
Check out the latest American Worker Project essay to learn more: eig.org/wp-content/uploads/2…
1
13
1,688
Question: What is the fastest way to become conversationally proficient in a new language (in my case, Hindi)?
Looking for the best courses (either online or in London), websites, self-study materials, etc.
Will also talk to people, consume Hindi media, alongside this.
5
13
2,649
Fascinating data from my colleague @doug_dowson--share of phds from top 10 econ programs going to the private sector has doubled from 15->30% since 2013. Tech accouts for nearly all of the growth #EconTwitter economist.com/business/2022/…
1
13
suppose @DarioAmodei is even partly right and an automation shock is imminent. how should policy respond?
one idea-partially restore a displaced worker's wages if they find a new job with lower pay
Hyman/Kovak/Leive find this leads to faster job-finding and pays for itself! 1/
1
1
12
1,040
h/t @arpitrage for pointing out the education/BJP support correlation after the Karnataka state election
7
9
8,025
Benchmarks/evals suffer from selection bias -- the tasks for which you can construct a usable benchmark are precisely the tasks RL is good for
An implication is we should EXPECT there to be a growing gap between benchmarks and real-world impact
I find the story of AI and radiology fascinating. Of course, Hinton's prediction was wrong* and tech advances don't automatically and straightforwardly cause job replacement — that's not the interesting part.
Radiology has embraced AI enthusiastically, and the labor force is growing nevertheless. The augmentation-not-automation effect of AI is despite the fact that AFAICT there is no identified "task" at which human radiologists beat AI. So maybe the "jobs are bundles of tasks" model in labor economics is incomplete. Paraphrasing something @MelMitchell1 pointed out to me, if you define jobs in terms of tasks maybe you're actually defining away the most nuanced and hardest-to-automate aspects of jobs, which are at the boundaries between tasks.
Can you break up your own job into a set of well-defined tasks such that if each of them is automated, your job as a whole can be automated? I suspect most people will say no. But when we think about *other people's jobs* that we don't understand as well as our own, the task model seems plausible because we don't appreciate all the nuances.
If this is correct, it is irrelevant how good AI gets at task-based capability benchmarks. If you need to specify things precisely enough to be amenable to benchmarking, you will necessarily miss the fact that the lack of precise specification is often what makes jobs messy and complex in the first place. So benchmarks can tell us very little about automation vs augmentation.
* Hinton insists that he was directionally correct but merely wrong in terms of timing. This is a classic motte-and-bailey retreat of forecasters who get it wrong. It has the benefit of being unfalsifiable! It's always possible to claim that we simply haven't waited long enough for the claimed prediction to come true.
ALT screenshot of https://www.nytimes.com/2025/05/14/technology/ai-jobs-radiologists-mayo-clinic.html
12
1,415
Important evidence-based thread on reducing police violence. Summary: some interventions work (tightening use-of-force laws, demilitarization, reforming union contracts, investing in non-police alternatives etc) and some don’t (body cameras, implicit bias training, etc).
For those who are interested in research-based solutions to stop police violence, here’s what you need to know - based on the facts and data. A thread. (1/x)
12
Looking forward to this event on May 21st during the South Korea AI Safety Summit.
Register for the (virtual) panel here: meetup.com/london-futurists/…
Just announced: New names have been added to the panel for the AI Safety Summit Talks online event on 21st May. All participants have thought long and hard about the risks and opportunities of forthcoming next generation AI platforms meetup.com/london-futurists/…
1
1
12
2,062
A key implication of our findings is the future of work-from-home is hybrid. Commuting to work a couple days a week makes the suburbs relatively more popular, but is still enough to prevent employees from leaving the metro areas containing their employer entirely.
Within large US cities, households, businesses, and real estate demand have moved from dense central business districts towards lower-density suburban zip-codes, but there has been no major reallocation across cities, from @arjun_ramani3 and Nicholas Bloom nber.org/papers/w28876
2
12
My review this week of #BreakingTheMould a terrific new book on India's economy by Raghu Rajan (ex central bank head) and @rohlamba
Manufacturing exports a la China will not be India's silver bullet. It needs a new model that will be uniquely Indian
economist.com/finance-and-ec…
9
699
NEW: in this week's @TheEconomist, my colleagues and I looked at the promise and perils of big AI models (doesn't the cover seem fit for a Pixar movie?)
My piece looks at AI risk and ways to mitigate it--both technical and social 1/
1
4
11
2,058
This is by far the best deck on Indian tech each year. It also isn't just abt startups--also covers macro, govt policy, R&D etc in a tech-relevant way
The Indus Valley Report 2024 - 3rd edition is here!! 🎉
This report is our take on India and the Indian startup ecosystem, through different lenses to sensemake what we call as “Indus Valley” - a catch-all moniker for the Indian startup ecosystem
Here are 15 lenses or frames for you all to munch on
@sajithpai @llnachull @blumeventures
2
10
2,304
A key question is reconciling the 2.8% reported time savings with no impact on wages/overall hours in the admin data. some guesses:
- people are bad at estimating savings
- people "shirk" with their extra time
- the reallocation of worker time to new tasks isn't very productive
1
10
1,444
So it's great to see interest in this growing from both academics and funders. Thanks @mattsclancy @heidilwilliams_ @DrPanch for your help
10
Some of these hurdles may seem surmountable. But because many are independent, they are stronger together than apart.
Thus, we warn against over-indexing future expectations of growth based on progress in one domain.
Full essay here: thegradient.pub/why-transfor… 10/10
1
2
10
1,232
I have it on good authority that that the index puts @zhengdongwang at the highest standard of living in London among all those the paper surveyed😊
An annual liveability index aims to measure culture and the environment, education, infrastructure and health care objectively. When all the scores are done, Europe can take another bow econ.st/3MMaxwN 👇
8
5,249
The biggest changes are in developing countries. Alipay (2011) and WeChat Pay (2013) now make up 90% of China's digital payments
India’s UPI (2016) and Brazil’s Pix (2022) have fast grown dominant domestically
Broadly, countries don't want to depend on US payment networks 2/6
1
10
963
Check out our analysis of how the Stanford community voted in the 2020 election. While California was divided on many of the statewide propositions, the Stanford community was much more aligned.
94% of the Stanford community on campus voted for Joe Biden & Kamala Harris in the 2020 election compared to 64% of CA and 51% of the nation. See the data on the presidential race and across statewide propositions in our latest story from @kaushikeenayudu stanforddaily.com/2020/11/11…
9
Good 🧵 on why trend GDP growth in India is likely a few pp lower than 8.2% (banks think ~6.5%).
This is bc of a wonky approach to GDP deflation. As I wrote recently this is unlikely to change anytime soon, so tracking Indian growth will stay tricky
economist.com/finance-and-ec…
India has just recorded an impressive 8.2 percent growth rate for the financial year. Interpreting GDP numbers correctly is important, otherwise we can prescribe misdirected policies. What is behind the headline number? Short 🧵
1
8
3,499
The 2021 rental surge has started to show in official inflation data with more to come in 2022. One reason for the lag: Zillow looks at listed rents whilst official data look at paid rents which update infrequently. See my note @TheEconomist for more: economist.com/finance-and-ec…
The WFH “Donut effect” - workers moving out from city centers to the suburbs - caused city-center rents to collapse. They have now returned to pre-pandemic levels. Rents in the suburbs are up 20% and still rising – inflation in action
1
8
Looking forward, payment competition will be global:
-Sanctions on Russia have helped China's settlement system, CIPS and the China UnionPay card network
-India is in talks with 30 countries to adopt UPI
-In general, US influence over payments is ⬇️ 6/6 economist.com/leaders/2023/0…
2
1
8
1,228