Dare mighty things! Journeyman wondersmith @spec__tech. Past: space and eldercare robots, AR, medieval history, etc. Long games. 🏴☠️🪐🐉
Princeton/NYC
- Tweets 15,961
- Following 555
- Followers 12,195
- Likes 42,409
I wonder if 50 years from now we're going to look back at how we've redesigned our world around computers with the same regret that people look back at how we redesigned cities around cars. 🤔
118
618
5,040
Our ability to do things is dying a death of 1000 cuts.
It's nice to talk about bottlenecks and silver bullets, but my hypothesis is that stagnation is the accumulation of many small frictions.
My plan is to continuously update this thread with examples as they pop up.
63
312
4,329
595,143
I have a unsubstantiated hunch that this is also what killed corporate research with huge positive externalities.
Over the course of several decades, companies realized how little value they could actually capture from deep research and got it off their books.
19
169
1,947
282,863
"Most people don't do interesting things, unless they're in a community where that thing is normalized"
- @slatestarcodex
I found this extremely truthy and triggered some correlaries that I don't see talked about much:
12
86
1,202
182,747
1/ Did you know that Vannevar Bush (you know, the guy who helped enable everything from radar to the manhattan project, the NSF to memexes) wrote an autobiography?
Turns out that yes he did, it's been out of print since the 70's, and it's *excellent*
BOOK REPORT THREAD
32
163
855
My politics are the opposite of whatever this is.
I try to avoid both politics and negativity, but sometimes exceptions must be made.
39
28
821
26,931
There used to be tons of smaller trucks on the road before Obama-era emissions standards made them near impossible. Since standards decrease as vehicle size increases, the easiest way for manufacturers to comply is to build a bigger truck.
*sigh* There used to be tons of smaller trucks on the road before Obama-era emissions standards made them near impossible. Since standards decrease as vehicle size increases, the easiest way for manufacturers to comply is to build a bigger truck.
toddofmischief.blogspot.com/…
5
36
800
50,244
1/ Recently finished "Where’s my Flying Car - A memoir of Futures Past" by J Storrs Hall.
It's simultaneously extremely critical of the state of atom-based technology and presents a precise and aggressively optimistic vision of possible futures.
BOOK REPORT THREAD 🧵
35
212
822
1/ For the past year, we’ve been building a new research organization:
Speculative Technologies (@spec__tech) exists to create an abundant, wonder-filled future by unlocking powerful materials and manufacturing technologies that don’t have a home in other institutions.
40
144
752
360,544
This speech is deeply underrated!
It's a masterful articulation of how we often can't justify our real reasons for doing great things through cost-benefit analyses or reason.
Often they're not even acceptable to say out loud because they seem childish or naive.
12
72
723
91,741
I feel like anybody who uses “PhD level” to mean “very smart” does not have, or interact with very many people who have, a PhD.
24
37
654
41,479
My favorite example of this is that an F-35 is effectively useless without basically the entire American economy behind it to provide fuel, spare parts, support crews etc.
F-35 not as weapon, but tip of infrastructural spear.
Great conversation b/t @dwarkesh_sp, @tamaybes, and @EgeErdil2 on the importance of complementary innovation + infrastructure for understanding how science/tech grows
13
49
647
53,656
It feels like we've had a collapse in the species of organizational terms: basically everything is either a company/corporation, an institute, a lab, or maybe a foundation.
I want more guilds, societies, orders, leagues, associations, clubs, and other weirder things!
41
47
641
33,094
To tack my personal opinion to this awesome historical anecdote:
We're not setting people up to be 25-year-old Manhattan Project contributors. Instead of encouraging highschoolers to go incredibly deep on science, we tell them to be well rounded and do a million activities.
17
46
565
141,714
The subtlety is that both computers and cars are amazing tools!
But the trick is that at some point remaking the world in service of a tool might start to go against serving people.
16
37
549
This isn't just a story about defense.
It's a story about the hollowing out of a once vibrant ecosystem for doing weird research.
18
62
555
202,308
~80% of the times I go to the website of an organization I've never heard of, I end up going to wikipedia to figure out *what they actually do.*
This seems like a problem.
19
70
545
The FAA needs data on low boom technology to approve supersonic overland, while the private sector needs the regulatory certainty of a noise standard to invest in the technology. The result is a Catch-22 that Congress could resolve with a single sentence.
The FAA needs data on low boom technology to approve supersonic overland, while the private sector needs the regulatory certainty of a noise standard to invest in the technology. The result is a Catch-22 that Congress could resolve with a single sentence.
thecgo.org/benchmark/congres…
3
17
486
63,601
After digging into why DARPA works, I asked the follow-up question:
How could you follow DARPA's narrow path in a world very different from the one that created it?
My answer: benjaminreinhardt.com/parpa
25
79
398
I went to the International Manufacturing Technology Show (IMTS) last week.
It's a ginormous trade show devoted to CNC mills, grinders, robot arms, metrology, CAD, 3D printing etc.
Felt like a new world (especially compared to Twitter) so I thought to share what struck me
1/
15
65
395
Lots of people are talking about lab-grown meat.
What about lab-grown wood?
Lab-grown shells?
Nature is an existence proof that proteins are able to create incredible materials. We're getting good at proteins -- what materials could we unleash with them?
29
26
381
Slack (in the systems sense) seems to both lead to amazing things and flies in the face of efficiency, justification, and rationality.
Over time, we've killed slack in more and more domains in the name of efficiency.
With the spotlight on the effects of Boeing's relentless drive for measurable efficiency, I wanted to re-up this great piece:
slatestarcodex.com/2020/05/1…
Slack is the illegible dark matter behind a lot of good stuff in the world.
It's also the first thing to get cut.
3
23
367
21,794
I backed a Kickstarter for a visual encyclopedia of megastructures a while ago and it's everything I was hoping for!
It's striking how this aesthetic and ambition has taken on a connotation of "unseriousness" -- a thing for children and dreamers.
12
31
351
How many people would work on a project they genuinely didn't think would be completed in their lifetime?
149
9
296
358,695
1. We should strive to live in a world where Elon Musk is unremarkable. “Oh yeah, that guy getting to Mars. He’s cool but I’m a fan of the guy building a tunnel between China and the US”
5
34
297
1/ Recently finished "The Genesis of Technoscientific Revolutions" by Narayanmurti and Tsao.
The book lives up to its grand name! It feels like a (fuzzy) image of the elephant that is "how research works" where everybody else is feeling a snake or a tree.
BOOK REPORT THREAD🧵
9
46
285
Another unsubstantiated hunch is that ambitious people are drawn to startups because it's one of the few domains with slack provided by VCs willing to make wild bets.
3
6
274
9,602
1/ Recently finished "Scientific Freedom: The Elixir of Civilization" by Donald Braben.
It's a sobering (but hopeful!) exploration of the stagnation in what I would call "paradigm shifting research" and what to do about it.
BOOK REPORT THREAD 🧵
8
41
276
My addendum to Tammy's list of excellent advice is one not everyone will agree with:
~Precocious young people should build a deep technical foundation~
A precocious young person reached out, asking me to expand on it. Here were my reasons
(all advice caveats apply)
1/
Recently I’ve gotten a lot of inbound from new grads asking for career advice, so I wrote down a list of things I wish I’d internalized sooner.
The key thing: being precocious has an expiration date. Threading the note in its entirety:
4
20
261
51,384
One like = one opinion (+reason) about Elon Musk.
Took the @vgr bait to do a one-opinion-per-like personal challenge. Live-fire stress test of your brainstorming capacity
9
22
255
The Szilard point - when more value is wasted applying for a grant than the value of the grant 🤦♂️
6
66
215
5. Elon’s popularity* is a signal of latent desire for ambitious projects in the physical world.
People don’t see how they can contribute so they root for him. The logical conclusion is that we need more outlets for that enthusiasm.
*even detractors say “sweet, but won’t work”
3
30
214
I worry that most of the focus on "AI for science" is getting it backwards.
Try to do great science and build awesome technology. Build AI tools when it would be useful.
Setting out to do "AI for science" will lead to all sorts of lamplight effects and square peg round holes.
7
23
219
26,975
Some rough power consumption numbers:
Human: 100 W
House: 1 kW , peak 10 kW
Car: 50 kW, peak 200 kW
Small city: 20 MW
MIT Campus: 18 MW
I found it especially surprising that a car uses roughly an order of magnitude more power than a house!
15
21
206
I find it kind of insane that we just implicitly accept that in order for anything weird and new to happen, you almost always need one person to both have the skills to lead that effort and to be adept at fundraising and/or working bureaucracies.
13
9
186
11,106
Trying out a potentially disagreeable assertion:
It’s almost impossible to simultaneously focus on solving a need and building a new technology because in the early days of a technology there are almost always better ways to address that need.
25
9
171
5/ The coupling between energy+power and technological paradigm shifts is surprisingly tight.
There are a massive number of things where the key technical constraint is cheaper and denser energy that we've just stopped thinking about.
4
12
169
It was delightfully object level. Nobody talking about transforming human interaction or web3 enabled AI for industry 4.0* — they’re like “this tool cuts metal really well — look at it cutting metal. Here is a piece it just made”
*Ok, almost nobody: there were some startups
3/
2
9
169
This is why I'm constantly screaming from the rooftops that the way "science and technology works" was mostly invented whole cloth in reaction to WWII.
Can't be overstated how almost all ideological and intellectual trends in living memory are reactions to (increasingly distant and cartoonified understandings of) the world wars.
6
9
173
14,088
11/ A good point that when people with dim views of the future have *power* bad things happen.
3
19
165
An underrated answer for "why did science take so long to develop?" is "because it looks really stupid."
Instead of the intuitive things that make sense and everybody has known forever you explain things with invisible forces that you know about from math and experiments.
8
11
167
10,583
The phenomenon where:
1. There is an organization that notionally fulfills some societal function.
2. So people just assume that function is taken care of.
3. But in fact that organization is terrible at it or actually fulfills some other function instead.
13
10
164
20,584
You read histories about these people and (most) weren't varsity-athlete, club-starting, straight-A influencers.
They were obsessed with taking things apart, putting them together, and how stuff worked and spent all their time on that.
That doesn't get you into good schools.
6
5
158
8,129
Sometimes I wonder if the thing that makes startups possibly the most dynamic modern institution is just because it's the one place in today's world where it's normalized for young people to be in charge of organizations and significant resources.
7
5
166
7,784
Increasing regulations and fees have brought the count of parklets in SF down from 2,000 to 700 😢 growsf.substack.com/p/the-gr…
1
154
27,533
6/ Hall introduces the concept of the "Machiavelli Effect."
It's effectively the known phenomenon of entrenched interests - ties it into Kuhnsian paradigm shifts, and argues that the effect has become much stronger over time.
4
24
157
Hot take: people who care about improving the technoscientific enterprise focus way too hard on the US from 1945-1972, instead of systems across the world ~1870-1940.
(I'm not exempt here!)
13
12
159
10,592
This pull quote from an IEEE Spectrum article captures the current moment in technology well.
1
13
153
6,349
2/ The core of the book is the question "Why didn't the future that people expected in the 50's come to pass and how could we get back on track?"
This question is not new, but Hall's approach is the most precise and brutal treatment I've seen.
3
16
157
8. People forget that Elon was basically treated like a crackpot for *years* before SpaceX successfully launched.
The question it raises is “how many projects have died because their crackpot leaders weren’t well-connected decamillionaires?”
1
14
152
My hunch is that the differently structured property control also leads to things like underpasses being filled with cool little bars instead of needles and garbage.
bloomberg.com/news/features/…
2
6
142
Is academia an outdated and irredeemably broken system or is it a critical piece of modern civilization?
I want to make a subtle (gasp!) argument that it's both:
Academia is broken, but has unique gems at its core. The way to fix it is to focus on those and unbundle the rest.
12
14
151
48,716
1/ It's always surprised me how many people who are interested in technology are unfamiliar with Technology Readiness Levels (TRLs.)
It's a great framework for communicating a technology's maturity and it's time for them to be used outside of aerospace.
benjaminreinhardt.com/trl
5
24
143
Whoah. I'd had a hunch this was a case but never saw it in graph form.
Explains a lot.
6
10
154
13,575
Realized that the soul-sucking nature of bureaucracies is that it's always your fault, never the system's or a bureaucrat's.
For not filling out the form correctly, for waiting in the wrong place, for not satisfying all of the potentially contradictory requirements.
6
19
144
A thing I'd love to see:
A time-lapse video of a construction site that is annotated with how much everything we're seeing costs and comments like:
"These two machines aren't being used because ..."
"Nothing is happening for three days because they're waiting on..."
6
9
152
8,514
Freeman Dyson died today.
A thread because
1. Many people who haven't heard of him who should!
2. He's a gestalt person, so many people only know a few of the pieces - physicist, theoretician, builder, science fiction inspiration, literary critic, institution defier ...
1
44
148
Colossal / Boring Company partnership to genetically engineer earthworms into tunnel boring sandworms
11
15
154
11,231
Friction around how much health insurance companies can have discounts based on outcomes. It looks like certain wellness programs that are allowed, but there are limits on what you can do if you make it outcome-contingent. dol.gov/sites/dolgov/files/e…
2
3
137
23,998
29/ Carl Sagan ftw. We need to distinguish between what feels good and what’s true. When there is a paper for everything, as is now the case, it allows people to find peer reviewed evidence for any opinion.
3
31
144
4/ A core theme of the book is the coupling between how much energy we can harness, atom-based technology, and how we deviated from an exponentially increasing curve in energy use in the 1970's.
6
19
148
7/ Of of the 🧐😮 arguments is that civilian government R&D funding is *anti* correlated with quality-of-life improvements.
6
19
144
You, too, could achieve this status just by opening your aperture to projects that are just awesome instead of just profitable or socially responsible.
4
10
149
28,484
As more community interaction moves online, things that are more transmissible through the internet become normalized more easily than those that aren't.
eg. Essays, tweets, short-form videos
1
5
138
8,994
1. Really impressive that CRISPR went from discovery to helping people in ~20 years!
2. That 20 years is a pretty characteristic timescale for deep research to become products (let alone create huge companies)
The first approved CRISPR medicine in the world for sickle cell disease and beta thalassemia!
A huge victory for biotechnology, patients, and humanity
businesswire.com/news/home/2…
1
22
126
29,161
Unfortunately, to a large extent the job that they're actually taking taxpayer money to do is ... give jobs to people in various congressional districts.
3
1
134
9,012
Not entirely serious thought:
What if the way that large language models actually accelerate the economy is by enabling good enough communication with animals that they become much more productive?
Who needs robots when you can negotiate with birds and monkeys
Parrots are clearly intelligent enough to understand video UIs
They also apparently prefer watching videos of other parrots 🤔
This implies an opportunity for a "parrot streaming" platform. Looking for a team who is as excited about this opportunity as Iam
10
13
132
52,646
What would you want to ask someone who:
- Worked at Bell Labs during its golden age
- Ran a national laboratory
- Started the engineering school at Harvard
?
(Interviewing Venkatesh Narayanamurti later today)
25
11
136
Bring back patronage as a prestigious thing!
There are a few people who genuinely want to produce public goods and can do so effectively with two people's worth of salary (themselves + an assistant).
Much more effective than university buildings! Let's make it as prestigious
Some philanthropist should fund Dwarkesh; he provides a great public service.
I'm surprised that's not already the case. Potentially many philanthropists underestimate the value of producing and disseminating high-quality ideas.
3
10
138
17,551
But it's much harder to transmit doing cool physical things, and so it's harder for them to become normalized and harder to get into community where they are normalized.
(I'm guilty of this myself! See the fact that I'm doing the normalized weird thing of writing a tweet thread
2
3
132
7,635
A thread of ongoing arguments in innovation/tech/science, etc.
All credit for this idea goes to @ganeumann
12
27
130
Fun fact: you can graduate from some of the most prestigious universities in the us with a stem degree without ever taking a linear algebra class.
3
3
129
17,849
This is so brutal and so good.
I want someone to do this for every large bureaucratic org/program.
caseyhandmer.wordpress.com/2…
4
11
133
9,621