Ben Chugg
@BennyChugg
Followers
634
Following
268
Media
85
Statuses
750
Views belong to Kublai Khan. PhD'ing @mldcmu . Funnier co-host of
Hobbiton
Joined December 2020
Don't wanna be here?
Send us removal request.
Explore trending content on Musk Viewer
ゼンゼロ
• 169672 Tweets
#นายณภัทร
• 147419 Tweets
ゼンレスゾーンゼロ
• 117989 Tweets
#นายใบเฟิร์น
• 94718 Tweets
エアコン
• 92118 Tweets
Independence Day
• 86603 Tweets
Happy 4th of July
• 53387 Tweets
ACTOR JAEHYUN AT BIFAN
• 42212 Tweets
स्वामी विवेकानंद
• 38278 Tweets
クーラー
• 29538 Tweets
Ollie
• 28475 Tweets
JENO FERRAGAMO FW PRESS
• 27908 Tweets
#あや姉さん一生ついて行きます
• 15880 Tweets
Morayo
• 15625 Tweets
選手間投票
• 14485 Tweets
インティマシーコーディネーター
• 13744 Tweets
Schoof
• 13525 Tweets
Ayşe Ateş
• 13397 Tweets
Haas
• 11222 Tweets
人手不足倒産
• 10593 Tweets
How does one learn to appreciate poetry? Just read it? Read about it? Read reviews of it?
199
10
233
the most chaotic places on the internet are by far the homepages of mathematicians. They learned how to use html and css once in the 90s and called it quits
14
26
219
Imagine how boring and dysfunctional civilization would be if humans all had the same interests. Instead we’re idiosyncratic, interested in different problems, and wrong about different things. And we make progress via positive sum collaborations
Eff yes humanity. We’re awesome
5
14
82
"AI will do science in its head"
First they rely on classical empiricism (of Locke, Hume etc) to explain why AI will become all powerful, then switch to classical rationalism (of Descartes etc) to explain how it will do science.
Just stick to one mistaken epistemology please
4
11
75
"train an LLM to predict the future"
I'm going to have an aneurysm.
I just had the most amazing idea.
Imagine you train an LLM with data that is all chronologically labelled. Every book, every scientific paper, every newspaper headline, every primary source from history. Where primary sources are lacking, descriptions of events or archeological
209
68
794
11
2
62
Eternally baffled by the position that we’re near the end of science.
What about depression, sleep, consciousness, dementia, aging? P vs NP, the multiverse, origin of life, AI? We barely understand any of these things, and progress in any of them will raise even more questions
8
2
43
PhD theses would have been so much bolder in the pre-socratic era.
“On the nature of matter: Literally everything is made of water”
“Humans emerged from the mouths of big fish: An empirical investigation”
“A logical proof that motion is impossible”
2
2
40
I find it bizarre when elites colleges tout the success of their alumni. Admissions are so competitive precisely because they select for this kind of person.
It would be much more impressive if they admitted students at random and still showed amazing outcomes
6
0
38
Sure, you could read about AI, trillion dollar compute clusters, geopolitical warfare, and situational awareness. Or you could read about some philosophical issues with p-values
2
6
35
@paulg
1,500 respondents split into 10 groups, not a huge sample size. Also worth noting that the survey was of job seekers, so naturally skewed towards those with reason to have high regret at the moment
3
0
29
“We have not succeeded in answering all our problems. The answers we have found only serve to raise a whole set of new questions. In some ways we feel we are as confused as ever, but we believe we are confused on a higher level and about more important things.”
2
6
30
Had a blast chatting with
@crit_rat
on the podcast! We plotted a revolution in graduate education, solved the free-will debate once and for all, and discussed the many versions of myself in the multiverse who died a silly death. How comforting.
3
6
31
What scares me most are Bayesian Philosophers.
All it takes is for one of them to have a non-zero credence that I’m actually an AGI from the future set on destroying humankind. Then the expected value of killing me will be so big they’ll have no choice. Game over
2
3
28
@astupple
Many successful *textbooks* are co-authored. Interesting to consider what the difference is
2
0
28
Where's the startup that organizes podcast by topic? If I want to learn about eg "land value taxes" I want a service that serves me up a bunch of conversations touching on that, the more perspectives the better
4
0
22
One thing I admire about both Taleb and Sowell is their recognition that non-academic work is as necessary as lots of academic work, and often harder because there are clear demarcations of success.
Sowell is chock-full of cheeky jabs like this
3
2
22
Philip Stark taking a huge swing at naive statistical modeling. Imagine this guy in a room with the Less Wrong community
4
2
21
It’s remarkable how many smart people are convinced that AI is going to kill everyone in the next few years. Sadly, some people refrain from having kids because of it.
I think we’ll look back on this as a standard moral panic that accompanies many new technologies
3
0
20
@krishnanrohit
It's also striking that he keeps citing the "10% of AI researchers think AI may cause extinction" number. Either (i) he hasn't bothered to examine the "study" in question, or (ii) he knows about the low response rate and wild selection effect but chooses to cite it anyway
2
0
20
Is David Deutsch’s concept of universal explainers implying 1, 2 or neither?
1. Anyone can learn anything (given enough time & interest)
2. There’s no fundamental difference btw ppl’s learning speeds. It’s all just interest and ideas.
1 seems plausible to me. 2 seems wrong.
13
0
18
Patents increasingly cite old papers instead of new ones.
Seems bad? Can’t think of good reasons this would happen besides current discoveries being less fundamental than old ones.
From by
@mattsclancy
5
1
18
@robinhanson
Worth noting the data is aggregated from 1973-2022. Presumably there has been a significant amount of intra-occupation variation in that time
0
0
18
One reason to blog is that even if you have low readership, future gens will have better ability to sort/find/aggregate writing on a topic. You're engaging in a cross-generational convo and even if you're wrong, you'll help improve their understanding
2
1
18
Best explanation to-date I've read of the Church-Turing-Deutsch principle
1
4
17
What does and doesn't happen in the Everettian multiverse? Proponents of many-worlds often say “whatever isn’t prohibited by the laws of physics” but this is circular.
Best answer I’ve heard so far ~50 mins in. Kudos to
@RobertWiblin
for great qs.
0
2
16
1
2
17
You can pick a topic, the internet will have 100s of articles and papers on it, and you can just learn it. That’s insane
1
0
16
A beautiful passage by Darwin on his wife (this the same man who wrote a pros and cons list for marriage, making it all the more impressive):
1
0
15
Didn't realize Hamming's lecture "You and your research" was ever recorded. Amazing
0
0
12
Will large language models write a paper that passes peer-review by July 1, 2024?
I say no,
@kohjingyu
says yes. There is boba on the line ...
2
3
13
A friend, one Mr.
@ChipkinLogan
, is in the middle of writing a book! I just read the first 16 chapters and, speaking as a fiction nerd, it’s truly excellent. Exciting, unique, and unpredictable.
You can download it to Kindle - check it out!
2
6
14
@cxgonzalez
Have to take a few of the claims with a grain of salt given the replication crisis, but Sapolsky's series on human behavioral bio is still one of my faves
1
0
13
Wrote down some arguments for getting rid of prepublication peer-review.
Unfortunately, I then discovered that
@lastpositivist
had made the same suggestion much earlier, and more eloquently. At least mine is … shorter?
1
1
13
I think the most prevalent and incorrect epistemology is that the truth is obvious. It justifies the violence on both the far-left and far-right for, if people disagree with you, it must be because they are deliberately evil.
1
0
13
Piranesi is a great book, well worth all the praise. A combination of Borges and CS Lewis with elements of Plato, the reality of abstractions, religiosity, and epistemology. And being inside Piranesi’s head is a pleasant experience. Worth checking it out!
2
0
13
You are given a dataset — what do you observe? How variable 1 correlates with v3? The distribution of v5? The effect of v9 on v4, conditional on v6?
Facts don’t magically emerge from data. Data are used to correct errors in our theories - not as some fount of prior wisdom
Arc of Enlightenment in a rational, civilized world:
Data → Facts → Information → Knowledge → Insight → Wisdom
548
3K
16K
1
0
13
@JosephNWalker
Some interesting evidence for this is that the 1970s saw an explosion of "scientific" models showing the end of civ was near. E.g., Limits to Growth (MIT study), Population Bomb (Ehrlich), peak oil predictions, co-founder of Earth day saying mass starvation was inevitable, etc.
1
1
13
This is wildly sad. This intro series contains no mention of global poverty or factory farming. Instead, it focuses on AI, the long-term future, and forecasting ... the "most important issues."
Where do I hand in my “sympathetic to EA” card?
Want an audio overview of effective altruism?
Here's a ten-part curated podcast series, covering all the most important issues: moral uncertainty, the very long-term future, AI, forecasting, and mushroom farming 🍄
3
47
211
3
1
13
I think it's safe to say that Claude Shannon would not receive tenure today
1
0
13
“To be a rational agent your beliefs must form a probability distribution.” - Bayesians
Why stop there? My beliefs form a 17 dimensional paracompact Hausdorff space. Who’s rational now?!?!
2
0
12
The entire library of Rome contained ~3 gigs of information. The internet has more than 3 trillion. It’s also probably easier to search the latter.
Think how more people have a voice thanks to this achievement. Amazing
0
0
12
Pessimist: There are always problems!
Optimist: There are always problems!
The difference is that the optimist recognizes that new problems are a sign of progress.
A problem-free society is a closed, stagnant society.
When we solve any problem the number of problems we know about does not decrease by one. It increases by many. Each solution reveals many more new, more interesting problems.
3
6
34
0
4
12
@robertwiblin
A little earlier, but the Ehrlichs' predictions are always something:
"In the 1970s hundreds of millions of people will starve to death in spite of any crash programs embarked upon now. At this late date nothing can prevent a substantial increase in the world death rate."
1
1
12
"I don’t feel the least humble before the vastness of the heavens. The stars may be large, but they cannot think or love; and these are qualities which impress me far more than size does."
- Frank Ramsey
1
0
12
@KerryLVaughan
1. There are no “AGI developers”, because nobody knows how to build AGI. It may be that the scale hypothesis is true, and AGI simply consists of massive NNs. This would be philosophically surprising, but if true we’d see increasing evidence to this effect as NNs grow in size
4
0
12
A lot of philosophy of maths makes the claim that once a theorem is proved, it is proved forever. But this seems clearly false? There can always be a mistake in the proof that we haven’t yet caught.
Mathematical knowledge is no less fallible than others kinds of knowledge.
3
2
12
It’s important to be cognizant of problems — but to get a better picture of reality (and know if attempted fixes are working!) we should also be aware of the good news.
@future_crunch
crushing it and helping us balance the scales once again:
0
2
11
@kareem_carr
Perhaps revisit the assumption that everything is getting worse. By many metrics, the world is improving drastically. E.g., poverty and child mortality over time:
3
2
11
“The answers to questions create new questions, and they’re questions you could not have predicted ...
I cannot predict what the questions of physics will be in 500 years. Those I could conjure up now would not be nearly as interesting as the questions which will actually arise"
1
1
11
As I get older I get both happier and more patient, except for anything to do with bureaucracy in which case I’m becoming exponentially easier to infuriate. I’m on the path to either enlightenment or libertarianism and I can’t tell which
3
0
11
Invoking free will because we don’t understand creativity is like invoking God because we don’t understand the ultimate laws of physics
6
0
10
"I can’t for the life of me figure out why this isn’t seen as a serious social issue in the USA; do people not find it alarming that 32% of Americans have 0-2 friends?"
That's insane. I'm 100% here for a friendship revolution. lets gooooo
New post - "Friendship Forever", citing
@robinhanson
,
@WillManidis
,
@TylerAlterman
and h/t
@kasratweets
for helpful comments
7
4
19
1
1
10
@JeffLadish
You could make this argument for anything.
“This suggests the sun creates enough energy to resolve all our needs, we just don’t know how to capture it yet.”
The knowledge is the hard part.
0
0
10
In 1940, Bertrand Russell was denied a teaching position at CUNY because (in court!) his work was deemed “lecherous, libidinous, lustful, venerous, erotomaniac, aphrodisiac, irreverent, narrow-minded untruthful and bereft of moral fiber.”
OG cancel culture
1
1
10
I'm like a month late to this but I think the analogy between optimization and evolution is unhelpful
If you optimize hard enough over any open-ended problem, you get minds; minds that want things. Minds and wanting are effective ways of computing complicated answers. That's how humans, and human brains, came into existence just from evolution hill-climbing "how to reproduce".
69
42
504
0
2
10
"I told you self-referential systems cause problems" - Godel
0
2
9
This seems like bad advice, and a recipe for having the opposite of constructive conversations.
Ideas go deeper than language. If you can't frame your points using someone else's favorite words perhaps they're not very good points
7
0
10
Slowly trying to dismantle the church of Bayes. Excited to be published by
@Conjecture_Mag
!
"when one steps outside the... confines of mathematical reasoning, the solution to the riddle is simple: Pascal should not hand over his wallet because the mugger offers a bad explanation." Elegant collision of
@robinhanson
@gwern
@DavidDeutschOxf
0
7
33
0
2
10
"Another danger is the over-estimation of what probability theory is capable of. The mistaken idea that "probability theory generalizes logic" led to some badly confused work in AI and has also warped some accounts of the philosophy of science."
0
1
9
@UtilaTheEcon
@robertwiblin
Damn, adding commenting capacity to arXiv is possibly one of the best ideas I've ever heard.
2
0
9
New drinking game: One person proposes a coordination problem, and then everyone gives their answer at the same time. Anyone in the minority drinks. (e.g., where would you meet if you were separated in New York?)
I call it the Schelling pint.
1
0
9
By induction, if population size keeps increasing, we'll all be walking at sub 4 minute mile pace.
2
0
9
Why haven't Japanese toilets caught on in North America? Built-in bidets and personalized music - these things are lightyears ahead of what we're using.
I'm not usually one for conspiracy theories ... but it's clear that Big Toilet is keeping our butts dirty for profit
3
1
9
“It’s almost as if we’re just constantly solving problems.” Great ep with
@isabelleboemeke
- what a badass. Full acknowledgement that energy increases well-being and we want more of it, no talk of returning to some pre-industrial time.
Just problem solving and optimism 🎉
Why aren't we more excited about nuclear energy ⚛️? How dangerous is nuclear today, and should we be worried about radioactive waste 🗑️? And what should science communication look like in an age of social media 📢?
Listen to
@isabelleboemeke
—
0
0
6
2
0
9
the ultimate irony of life is that to have a good online dating profile you need good pictures but to have good pictures you need a girlfriend
1
0
9
@robertwiblin
Probabilistic models are simply that: models. They never apply perfectly; the question is whether a given model applies well enough for a given situation. That's why there's a difference btw modeling games of chance (highly amenable to stats) and modeling e.g., risk of AGI
1
1
8
Thank god
@DavidDeutschOxf
isn't a Theoretical Computer Scientist, otherwise all the TCS acronyms would get very confusing.
Taking Computers seriously? Theoretical Children Science?
1
0
9
Hume, a well-known atheist in his day, once fell into a bog and couldn’t get out. He asked a group of fishwives to rescue him and they made him repeat the lord’s prayer before they would help
4
1
9
"I’m not interested in doing research and I never have been...I’m interested in understanding, which is quite a different thing. And often to understand something you have to work it out yourself because no one else has done it" - David Blackwell
0
0
8
@moreisdifferent
Worth reading some criticism of Haidt's position, e.g., .
I'm pretty convinced he overstates his case.
0
0
1
Rumsfeld was widely mocked for his speech where he drew the distinction btw known-unknowns and unknown-unknowns. He even won the foot-in-mouth award for it.
But I think it’s an incredibly insightful statement about the nature of uncertainty
1
0
8
Jumping 20 steps ahead assumes you know what works. But policies have unforeseen negative consequences. They’re just guesses about how to improve life, and can be wrong. *That* is why you move one step at a time, not because it "depresses expectations."
The left needs to raise people’s expectations for society. To do that we need to put forward demands that are two steps ahead of popular consciousness, not twenty. Utopian “demands” actually depress expectations by making change appear unachievable, contributing to demoralization
48
98
971
0
4
8
Donoho with a hot take that recent progress in AI isn't actually about deep learning, but instead about frictionless research exchange, allowing good ideas to be rapidly spread, executed, and improved
0
1
8
Imagine if schools didn’t offer electrical, civil, or chemical engineering degrees, but just “engineering” full stop. No specialization. That would be insane. But that’s how it works in computer science
1
0
8
Epic guest list, great questions, lots to disagree with - what could be better? A seriously binge-worthy podcast by
@dwarkesh_sp
. Stoked to see where it goes!
0
1
7
Having
@ReachChristofer
, the sexy Swede himself onto the pod soon. Now's the time to ask all those burning questions you've always had for him, like "where do I get your accent?"
Stoked to chat to Chris, the podcast host we need, but don't deserve 🦇
Got invited to be a guest on Increments with
@BennyChugg
and
@VadenMasrani
: what would you like us to talk about? 😎
1
0
6
1
0
8
Guys I don’t want to raise geopolitical drama, but Fiji’s flag sucks. It looks like a half-finished powerpoint slide
Let 2024 be the year of good vexillology
6
0
8
@lastpositivist
Popper would feel so vindicated by this tweet.
And then he would look at your twitter handle and lose his mind
2
1
8
@deepcohen
Can't believe nobody said Rademacher random variables. It sounds so badass, such a let down
1
0
8
This convo between
@ReachChristofer
and
@mattgoldenberg
is one of the most unique and lovely podcast episodes I’ve heard. Takes guts to cry on air (I have to be halfway drunk before I even start recording …)
All the love fellas ❤️
Here's my second chat with the fascinating
@mattgoldenberg
. This time we dove into showcasing *actual* emotional work, investigating my own core - "I'm not good enough"- hangup, instead of sticking purely to masturbatory theorizing... 😁😎
Enjoy!
1
1
8
2
2
8
@alyssamvance
This only reports opinions of people who respond to a survey about "high level machine intelligence timelines" ...
This is literally the definition of selection bias
1
0
8
Sick of
@vadenmasrani
and I shouting about longtermism? The cure: A three hour war of attrition with the lovely gents from
@hearthisidea
. It’s like sex, love and rock n’ roll, only it’s expected values, future people, and morality. Close enough
2
2
8
Wasserman arguing that we should abolish peer review before it was cool
1
1
7
It's interesting that when I really want to understand something, my information consumption order is roughly: wiki, articles, chatgpt, debates, books.
Books are good but only once you understand the territory enough for the details to matter
4
0
7
Walked in on my girlfriend telling our dog she loved him infinitely. He looked at her forlornly and she clarified that it was an *uncountable* infinity. This is why we get along so well.
But also how could your love for this guy be only countably infinite? Would be ludicrous
0
0
7
Started keeping a research journal recently. Possibly of interest to fellow nerds out there. Some topics so far are game-theoretic probability, stochastic calculus, and some random stats stuff
0
2
8
Trying to stir up some AI controversy. My (opinionated) review of Erik J. Larson’s book “The Myth of Artificial Intelligence”
1
1
7
Working my way through Do Explain - amazing. Lessons learned so far:
@ReachChristofer
aspires to have bathroom-related accidents in public and not care, and
@HermesofReason
has clammy hands. Highly recommended
0
1
7
Despite my feed being mostly filled with hyperbolic panic about how close the newest language and image models are to AGI, the actual uses, as usual, are simply awesome tools enabling better research.
Check out
@elicitorg
improving lit review.
0
0
7
Jerzy Neyman on discovering one of the most fundamental lemmas in statistics. Sometimes it's good to be interrupted
0
0
7
@MatjazLeonardis
Hunger by Knut Hamsun. Piercing voyage into the psychology of desperation and starvation
1
0
7
