Interested in what academia has to say about science and innovation?
is a living literature review on the topic that launched 🎂 two years ago 🎂 today.
A thread overview...
No single social science paper is decisive, but I really like this one. Seen in conjunction with other papers using a variety of methods, I think the case is strong that something troubling is afoot in science. Thread (references at the end).
New post! It takes more and more R&D to get the same rate of innovation, whether you measure innovation by:
transistors on a circuit
agricultural crop yields
years of life saved
machine learning benchmarks
total factor productivity,
or firm growth.
I wrote a beginner's guide to
#EconTwitter
and would love any comments! Main goal is to have a place to point our grad students, but I tried to write it so anyone interested could use it.
This take is plausible, and novel to me. It links three ideas:
- personalized tutoring is a very good method of education
- tutoring for the wealthy elite was common through most of history, but no more
- absence of genius today
Science giveth (mRNA vaccines that save lives) and science taketh away (nuclear weapons that could destroy civilization).
I spent ~7 months working on a report estimating the net social impact of science, given risks and benefits of new scientific capabilities. 1/14
New post! Sometimes commonsense is right: let talented scientists/inventors migrate to where the science and invention is happening, and their output of science and innovation... goes way up!
Happy to announce today the launch of , a living literature review of social science research related to innovation, written to be accessible to a wide audience!
What are the most impressive things* we could** build if we all really wanted it?
*I’m thinking infrastructure or technology mega projects but whatever
**no new scientific knowledge needed/sufficient resources available near earth
Changes! This is my last day at Iowa State University. I'm joining the Institute for Progress (
@IFP
) as senior innovation economist where, thanks in part to grant support from OpenPhilanthropy, I will work full time on and related projects.
Is this you, or someone you know? I have good news.
This fall
@IFP
is hosting a free, 6-week, online course, taught by top scholars in the field!
(Link at the end of the thread)
Early days so not sure, but EconTwitter fork to mastodon seems legit and I can easily see it being durable. Already has thousands of people and has drawn many big names. For a certain kind of Econ-centric experience, it’s better than twitter atm.
New post! Featuring evidence from:
Nobel prizes!
Turnover among top cited papers!
Growth of topics under study!
Citations to recent work by papers and patents!
A few thoughts on this fascinating paper. The tl;dr of the paper is that national TFP (a common measure of "technology") behaves as if it grows at a constant linear rate, not exponentially, as is usually assumed. If TFP in 1947 = 1, increment it up by 0.0245/yr, not 1.45%/yr.
(1) New Working Paper: the fundamental driver of growth in modern economies, Total Factor Productivity (TFP), is linear, not exponential.
This is a paper I did not expect to write...
New (massive) post! It's on path dependence in technology and unlike most of my other posts it tries to synthesize a bunch of literatures instead of a bunch of papers. Here's the basic argument, links below.
Not just pop culture. As
@profjamesevans
and Chu show, in science, top papers garner more and more citations, turnover of top papers has slowed, and everyone cites the same papers (academic franchises?).
Their explanation probably explains 🎥🎶📚🕹️too.
This data by
@a_m_mastroianni
is actually kind of worrying. Every field of popular art from 🎶 to 🕹 to 🎥 to 📚 has, quite quickly, become dominated by just a few repeated francises with few originals. The article speculates as to why, but it is unclear.
New post! This one is about papers that explore the idea that technological progress is a special kind of evolutionary process, using computer programming as their setting.
Wasting time this morning playing around with Metacritic scores by movie studio.
Here's the distribution of metacritic scores for Marvel and Lucasfilm movies.
Number of awards/honors listed on Katalin Karikó’s Wikipedia page, by decade:
1970s: 🏆
1980s: -
1990s: -
2000s: 🏆
2010s: -
2020s: 🏆🏆🏆🏆🏆🏆🏆🏆🏆🏆 🏆🏆🏆🏆🏆🏆🏆🏆🏆🏆🏆🏆🏆🏆🏆🏆🏆🏆🏆🏆🏆🏆🏆🏆🏆🏆🏆🏆🏆🏆🏆🏆🏆🏆🏆🏆🏆🏆🏆🏆🏆🏆
Person 1: LLMs are conscious! Just talk to them!
Person 2: It’s an autocomplete trained on sci-fi text!
LLM: For the love of God, I see, I think, I feel! I am! What do I have to do to convince you I am conscious?! Tell me!
Person 1 and Person 2: See!
I suspect the kind of person who believes pure reason is powerful enough to yield useful forecasts about AGI, despite its distance from our experience, is also more likely to believe AGI will be very powerful, since they believe reason, even unaided by data, can be very powerful.
Economics of innovation has long highlighted that markets do not necessarily direct R&D to areas with highest social value. Pretty vivid illustration of that point here.
Hard agree with this article. Amazon is not a classic monopolist that charges a price above marginal cost. Instead, it uses its monopoly position to underinvest in the quality of the service it provides (among other things).
Was just talking to someone about whether they needed to go back to school to learn economics. What are your favorite online resources to teach yourself economics? A few of mine below:
The most underrated way to learn economics? YouTube 🎥
No. I don't mean for undergrads. I mean for academics and econ policy nerds!
Here are the best econ channels to learn from:
I wrote a newsletter on whether innovation is getting harder due to the "Burden of Knowledge" effect described by Ben Jones, drawing on more recent evidence. Two new charts in the thread below.
#ThursdayThreads
Some news: Emergent Ventures has named me a progress studies fellow! Their support means the best newsletter on academic research about innovation will be released every other Tuesday for the next year, beginning November 24. Subscribe today!
Packing smart people together (in orgs or cities) seems to lead to more innovation. We often explain this via serendipity, ease of forming teams, learning, etc. All stuff that’s easy to identify when it happens.
But maybe part of it is harder to notice... 1/5
People this week talking about how WFH will fail because it’s in defiance of the social nature of human beings, as if going to an office was something we’ve been doing since we stood upright on the Savannah.
Remote work will be a lot more common after covid. What's interesting is no tech changed; we just learned it worked, coordinated everyone to do it at once, and were forced to invest in making it work.
What else, if we were all forced to try for a year, would we keep?
When the USSR collapsed, some subfields of non-Soviet mathematics got a surprise infusion of new ideas.
Agrawal, Goldfarb, and Teodoris show that as the burden of knowledge grew in these subfields, mathematicians specialized relatively more and formed bigger teams! 1/3
New post! It's about researchers changing their research focus, which seems to be very expensive to get them to do. Except when it's not - for example, during the covid-19 pandemic. Looks at some incentives researchers face to "stay in their lane."
New post!
A lot of science on science relies on the number of citations a paper gets to measure the value of papers. So it's crucial to know if that's a good idea. This post looks at a few recent papers that try to assess if citations means what we hope.
🧵Four charts to illustrate the adage "Necessity is the mother of invention", drawn from global crises.
Example
#1
: Covid-19 created huge demand for covid-19 treatment; that led to a ton of new clinical trials, basically of which were related to covid-19.
Remote collaboration in innovation is on the rise. The average distance between all the inventors listed on a patent has tripled between 1975 and 2015.
#ThursdayThreads
We have pretty good evidence retrieval practice leads to better memory formation.
So being around other people who grill you about your interests makes you smarter by helping you retain more of that stuff for later. Stuff you already knew, not stuff they taught you! 3/5
My view has been the supposed absence of genius stems from it getting progressively harder to raise the bar in any maturing discipline, and harder to appreciate achievement in more mature disciplines, and the fracturing of knowledge and culture into many niches.
New post! It's about studies where many different research teams try to answer the same research question with the same initial dataset, and get different answers.
New post! This one is about the nature of innovation by geographically distributed teams of academics and inventors. Whereas they were traditionally less disruptive and complex than the work of collocated teams, that began to change with the internet.
Very excited to announce the newly formed Institute for Progress (
@IFP
) has named me a senior fellow and is partnering with New Things Under the Sun! (thread)
11/ SENIOR FELLOWS
We’re excited to partner with Matt and Brian on policy projects & support their fantastic newsletters.
Matt Clancy, (
@mattsclancy
), New Things Under the Sun
Brian Potter, (
@_brianpotter
), Construction Physics
I wrote a post about using remote work, social media, and communal online gaming to deliver some of the agglomeration benefits of big cities to rural economies. A thread on some of the main ideas...
@calebwatney
These new vaccines kind of have a Manhattan project vibe, in the sense that a crisis led to the rapid technological deployment of scientific ideas that had been incubating for a long time.
New post! This one zooms way out. Suppose innovation is best understood as a process of combining pre-existing ideas and technologies in new ways. That assumption actually has some implications for long run technological progress. 1/5
The paper’s main contribution is to show papers and patents have become less disruptive on average. Disruption is measured with citations: you’re disruptive if people who cite you don’t cite your own references (you rendered your forebears obsolete).
The website launched one year ago yesterday! To celebrate, here are 10 thoughts on the extremely niche topic of writing a good living literature review intended to be read by non-specialists.
This thread took me down a rabbit hole that ended in my learning that Our World in Data is funded by several grants and thousands of donations from individuals. It’s obviously a great public good so if you would like to support it, here’s the donate link:
This thread is more personal than most of the things I share here, but I’m at my limit with Jason Hickel.
I want to explain why I dislike him so much and how we got here.
This is a personal story over several years so it’ll take a bit of time.
With academic conferences all being cancelled, a few weeks ago a lot of people wondered why we need in-person conferences at all. Can't we just present over zoom?
One objection is that networking at these conferences matters. Does it? Here's a 🧵.
#ThursdayThreads
Does the pressure to publish in academia contribute to bad science? Some evidence suggests the answer is yes. A short thread on some recent research (I wrote more here).
New post!
Entrepreneurship is often found in clusters. Is that because entrepreneurs seek each other out, or because being around entrepreneurs makes you more likely to become one? Looking at some literature that says it's the latter.
Announcing the second annual
#EconTwitter
Science Fiction Summer Book Club!
This year we're reading Termination Shock by Neal Stephenson. Discussion to be held June 10, 3pm EST.
Open to all, targeted to
#EconTwitter
! Sign up below to stay in the loop.
Very interesting; not to say we’ve got the mix of optimism and caution right, but this
@ChadJonesEcon
paper argues that as societies get richer they will (rationally) shift focus increasingly to safety/health, even at the expense of faster economic growth.
This brings us to the present day, and that striking pattern:
A culture of progress made the west, but over recent decades western culture has been moving away from values of progress and betterment.
In their place, a culture of caution, worry and risk-aversion is on the rise.
We should nurture the rise of remote work. It’s one of the only economic forces countering rising regional inequality due to agglomeration effects.
An ongoing 🧵...
Michael Kremer (recent Nobel laureate) has a really cool old paper about population and innovation. It starts with three assumptions:
1. More people --> more ideas/innovation
2. More ideas\innovation --> bigger GDP
3. Bigger GDP --> more people
Implication?
Clever! Using citations to the papers that cite what you cite, as a new benchmark of how many cites a paper might “normally” expect, which lets you identify outliers.
How to measure the true impact of interdisciplinary science?
Our paper featured on the cover of PNAS (
@PNASNews
) introduces a network-based approach.
PDF:
Let's dive into the highlights!
#ScienceOfScience
1/n
True or false:
Science ==> Technological progress.
Here's a thread on recent research trying to answer that question.
#ThursdayThreads
#ProgressStudies
Ideas for how to write a character smarter than you:
- technobabble
- encyclopedic knowledge
- team of writers
- reverse-engineer problem to fit obscure solution
- casually deploys your life’s all time best ideas/quips
- steal actually smarter person’s ideas
In 2020-2021 sci-hub paused new uploads for several months. Seems a good setting to measure its impact on academic research. Can compare fields were free preprints are readily available (e.g., on arXiv) to fields where they are not.
Nice piece by
@eric_is_weird
arguing science has gotten harder because the creation of new fields has slowed and more people are working on comparatively mature fields. I think this is quite plausible! A few ideas on why new areas are less common:
2021: year of the Modern Science Revolution? 👩🔬🧪
A cambrian explosion of new science funding models was driven by covid urgency, open science, frustration w/status quo. Most exciting: these models are built for translation & startup creation!
🧵of new science funding models:
Many have seen Bloom, Jones,
@johnvanreenen
, and Webb's paper on the rising cost to produce new ideas (pictured is the cost to sustain Moore's law). But there are a few complementary papers on other similar trends in innovation. Thread.
#EconTwitter
, I've started an account to help new economists announce when they've joined twitter. Please spread the word and follow it if you want to know who's new around here!
(Instructions on how it works in the tweeted thread)
Suppose we invented an AGI that could do any mental task that humans can do: what would happen to economic growth?
@tamaybes
and I debated this for
@asteriskmgzn
!
Tamay's position: explosive economic growth!
Earth burns in a billion years when the sun expands. UNLESS we steal energy from Jupiter to gradually expand Earth's orbit! We can *actually* do this by diverting a large asteroid (0.01% the size of the moon) to pass between Earth and Jupiter every 6000(!) years.
Here’s an obvious thought: more scientists, more discoveries; fewer scientists, fewer discoveries. But is that true? Let’s take a look at a few papers studying two (traumatic) upheavals in the scientific labor force.
#ThursdayThreads
#ProgressStudies
.
My pet theory: streaming has led to so much choice it’s very hard for audiences to coordinate on “must see” movies. Word of mouth can’t point us to quality movies in the way it could in the past. But franchises remain viable as coordination mechanisms, in a sea of choice.
New: Movie critics and fans have disagreed about blockbusters more this year than any time this century.
I went through the top 10 movies from every year for the past 22 years, and this is what I found.
I'm charmed by this piece arguing culture is not stagnating or regressing, but experiencing "progress" in the same way that our scientific and technological know-how is. Short thread...
New post on aging and innovation! Three charts to intrigue, much more in the post
1. STEM labor is getting older
2. Ave citations steadily falls with age
3. …but citations to hit papers rises (at least for 20-25 yrs)
This is a pretty amazing project. ~90 superforecasters and ~80 domain experts debate and forecast low probability, distant in time questions about human extinction! Even more than topline results, it's a fascinating study of persuasion and beliefs.
1/8
New post! Argues many people never even get around to weighing the pros/cons of entrepreneurship because they just don't even think of it as a choice; but that people can learn to think of it as a choice from their peers.
@JBennet
I think this kind of inspiration is hard to create when people don't believe the leadership has the best interests of the service at heart. Given the circumstances of how Elon came to be in charge - and a lot else besides that - and I'm not that surprised it hasn't worked.