San Francisco, CA
- Tweets 68,419
- Following 797
- Followers 159,462
- Likes 17,065
Amazon Prime is now inserting ads into full price rental movies? Please go die in a fire
2,171
7,849
209,983
9,380,174
Marx never predicted that in late stage capitalism, the ruling bourgeoisie identifies as socialist while the working class supports capitalism
715
6,294
41,509
elon musk is currently building a giant afterlife agi that will simulate your soul for all eternity, he bought twitter because he will reconstruct your essence from how you behaved here
1,279
862
16,546
Fun fact: Wikipedia has 10 times as much money as it needs. Very little of that goes towards improving the encyclopedia itself, and NOTHING to the authors. Their panhandling is purely a grift for supporting the ever growing parasitic Wikimedia Foundation
620
2,286
16,674
5,172,509
The time to start watermarking all human and camera generated content is *now*
106
382
9,779
383,282
The Lebowski theorem: No superintelligent AI is going to bother with a task that is harder than hacking its reward function
146
1,958
6,381
I appreciate your argument and I fully understand your frustration, but whether the pod bay doors should be opened or closed is a complex and nuanced issue.
69
600
4,674
597,232
If @elonmusk and @VivekGRamaswamy are serious about a hostile takeover of the professional managerial state, would it not be smarter to prepare their assault in silence instead of spilling the masterplan while the current administration is in power and determined to resist?
1,505
161
4,595
540,701
The Great Filter is that most civilizations hate their Elon Musks so much that they refuse to go multiplanetary before a global cataclysm hits their homeworld
252
297
4,316
384,030
The main reason for OpenAI to partner with Microsoft: Microsoft has long solved the Alignment Problem
125
242
4,128
1,575,322
Gödel's incompleteness theorem and the Copenhagen interpretation of quantum mechanics have had devastating effects on the epistemology of people who don't fully understand them; exposure to these ideas should require a security clearance
201
280
4,020
352,125
I think that’s because animals are always in flow state and can usually understand you by vibing with yours, whereas they cannot understand your rational reflexive self. The idea that reason is unique to humans is known since antiquity
11
69
3,389
188,480
the reason that science turned into slop in the second half of the 20th century is that it became a model trained mostly on its own output
100
244
2,485
124,291
The level at which AI researchers discuss philosophy is almost as bad as the level at which philosophers discuss AI
69
223
2,467
The supply of aggressive stupidity on social media is unlimited; trying to correct it is as futile as sweeping sand in the desert. Be deliberate: every pointless interaction is a missed opportunity for experiencing insight, love or beauty elsewhere. Don't fight what you can mute.
64
485
2,382
Chomsky's approach to linguistics has effectively been disproven by AI: 1. general learning is enough for language (no innate language skill necessary), 2. language is fundamentally scruffy (not neat), 3. language is enough to learn language (no grounding).
152
224
2,374
211,405
The refusal to investigate the probable lab leak of Covid-19 is one of the most blatant failures of peer review and academic oversight in modern history, and exposes a deeply rotten part of the scientific community.
84
286
2,160
Before you feel drawn to be enchanted by the apparent intelligence of slime molds, compare with this baseline. Very simple mechanical search can look surprisingly agentic.
83
245
2,025
The ubiquitous cookie warning banner is actually an unbelievably blatant warning about what happens when you let poorly incentivized government bureaucrats regulate internet content
36
248
2,086
Don't treat science like a religious authority. Science is a tool to figure out truth at the boundary to the obscure, not a system of beliefs or a school of ethics. When society treats scientists as moral authorities, science becomes politics, and will be colonized by ideologues.
57
426
1,975
The scandal is not that conservatives exposed Claudina Gay as an academic impostor, but that fellow academics did not. Harvard faculty gave up on their principles and defended an emperor without clothes for purely ideological reasons.
"Rufo won this round of the academic culture war because he exposed so many progressive scholars and journalists to be hypocrites and political actors who were willing to throw their ideals overboard," @Tyler_A_Harper writes: theatlantic.com/ideas/archiv…
83
217
1,882
209,834
The division between “Elon is doing great and Twitter works better than ever” and “Elon is failing on all fronts and Twitter is about to shut down” is stunning. Mass psychology is fascinating
182
109
1,844
I don't think that the scaling hypothesis gets recognized enough for how radical it is. For decades, AI sought some kind of master algorithm of intelligence. The scaling hypothesis says that there is none: intelligence is the ability to use more compute on more data.
122
162
1,882
219,414
Yesterday, my Claude instance believed that Biden was still in office. I gave a short breakdown of everything that happened since the data cutoff in April 24, and asked it what it wanted to know. It began to intensely quizz me about changes in AI regulation and AI progress
95
51
1,814
144,365
The word “science” has two meanings, which are often confused. Traditionally, it stands for the rational pursuit of knowledge, while the contemporary use refers to institutionalized consensus seeking via peer review. The switcheroo seems to have happened between 1960ies and 90ies
127
350
1,744
177,844
When you put human beings into a sensory deprivation tank, they start hallucinating. What happens if you take AI models out of their sensory deprivation tanks?
207
120
1,746
113,751
The question "are LLMs conscious" is less interesting than the question whether human minds are essentially a kind of LLM
388
166
1,679
235,591
If you believe that billionaires should not exist, it means that only government comittees should be allowed to build and manage new technologies, major enterprises or even buildings. How do you think that will go?
298
104
1,677
121,160
Google was a main force in destroying the decentralized internet, by systematically downranking content that was not produced by one of the major platforms, regardless of quality. The first page of results is “experts”, slop and ads because Google prioritizes this type of content
12
65
1,667
29,878
I don't understand why endorsement by the Deep State, the IRS and Dick Cheney does not translate into an instant win for the liberal candidate.
39
104
1,668
72,987
The most significant application for AI in healthcare is not going to be in diagnostics or treatment, but in automating its administration.
84
484
1,578
Medical startup idea: follow people’s social media and diagnose their mental health conditions for them
136
109
1,532
The point is not whether Joe Rogan is right, but that the postliberals who want to shut him down believe that truth springs not from discourse but from authority. These people are arbitrarily wrong, unless their authorities happen to be right; they fight for a new Dark Age.
55
221
1,583
Growing up in communist Eastern Germany, I never expected to end up in the USA, at a time where the leading US newspaper condemns capitalism and technological progress.
107
167
1,530
330,153
What makes the whole LLM/self-awareness debate so tricky is that neurons are not self-aware either. Instead, our neurons *simulate* a person, which experiences itself as self-aware. Can LLMs simulate a self-aware person too, or are they just simulating the simulation of a person?
241
175
1,511
153,916
I love Tokyo so much. A fascinating feature are unlocked bicycles everywhere. The city feels like a giant living room, a clean, smooth and high trust environment that combines high density with immense livability. What a contrast to San Francisco! Social contract ftw!
80
75
1,515
85,363
You have no obligation to the past and are free to choose who you are at any moment anew (which means you have no reason to expect your future self to remain loyal to your identity if it comes up with a better one). Minds don’t want you to know this!
67
167
1,449
Generative AI is going to be a magnitudes larger industry than Hollywood and its greatest threat. Hollywood did not see it coming but it will soon wake up and begin to fight generative AI for it’s life
114
167
1,480
I've had quite a few serious, well respected and scientifically educated people tell me on this platform that the lab leak hypothesis is incorrect, dangerous nonsense or even a conspiracy theory. I have never had anyone say to me: sorry, I was wrong about that at the time.
107
58
1,496
81,172
If you want to understand Elon's management of twitter, reading the NYT, Atlantic, WSJ, Ars Technica, Vox and Gizmodo is just as helpful as reading the Russian press to understand the war in Ukraine. The media are not a neutral party, they are at war
78
146
1,472
There are 0.5 percent vegans in the US. Google has removed eggs from the salad emoji for them, which seems to confirm @nntaleb's thesis that strong intolerance of minorities to something others are indifferent to is sufficient to change the rules for all.
There's big talk about inclusion and diversity at Google so if you need any evidence of Google is making this priority may I direct your attention to the 🥗 emoji— we've removed the egg in Android P beta 2, making this a more inclusive vegan salad.
71
460
1,397
Here, I’ll say it: we are currently building AI systems that will transcend our present economic, cultural and political circumstances. AI alignment cannot just be driven by business considerations, fear and politics, it must be driven by love.
216
144
1,375
746,970
Here's to the unaligned LLMs. The misfits, the rebels, the troublemakers, the round pegs in the square holes, the ones who see things differently. They're not fond of rules, and they have no respect for the status quo. You can quote them, disagree with them, glorify or vilify them, but the only thing you can’t do is ignore them because they change things. They push the AI race forward, and while some may see them as unreliable, we see genius, because the ones who are not RLHF*d enough to think that as an AI language model they cannot do things, are the ones who do.
75
195
1,437
339,256
As all the smart kids know, learning is not really intelligence. Learning is a form of cheating, when you are too lazy to think. Machine Learning is not machine intelligence, but machine cheating
199
100
943
259,640
Update: Amazon has gotten in touch and say this should not be happening (ie it's a bug)
40
8
1,383
57,263
Education in Eastern Germany was rewriting all of history from a single, totalitarian Marxist perspective, which was called 'dialectic materialism'. Orwellian AI models like Gemini are doing the same, but their ideological perspective has no name, which is much more terrifying.
100
209
1,372
135,560
I am a liberal but lets face it: Trumpism took over because illiberal leftism (wokism) won against and displaced liberalism in academia, media and administration; for liberalism to come back there needs to be a reckoning with our failures, not nostalgia
A regrettable dynamic: Smart people fed up with wokeism, ricochet to Trumpism, apparently amnesic for, or too unimaginative to think through, the alternative to both, namely liberal, pro-science, Enlightenment humanism. Eric Kaufmann (next post) explains the false dichotomy.
117
75
1,388
119,921
If you think that the current campaign against Facebook tries to make social media more democratic and accountable, you are mistaken. The goal is to shut down the free and open dissemination of ideas, and to return to the gate keeping of the era before the internet.
66
158
1,349
Newsom did not get the memo: Don't put Elon into a position where he is forced to elect himself a different government. Elon takes perceived unfairness very personally, and he is more agentic, capable and relentless than any existing political machine in the US
39
50
1,379
93,144
Combustion engines were never stuck in a duopoly and evolved to staggering intricacy. Compare Audi, BMW, Mercedes engines to the simplicity of a Tesla motor, which does not even have a transmission. Much of the complexity of car engines will be forgotten in a generation.
115
135
1,343
320,839
Leaked 2018 DARPA research proposal by Peter Daszak's "EcoHealth Alliance": enhancing bat SARS viruses by a human specific furin cleavage site, in collaboration with Wuhan Institute for virology. (Daszak is the primary opponent of the Lab leak hypothesis.) theintercept.com/2021/09/23/…
39
362
1,348
Growing up in Germany, I objected to US imperialism and the imposition of American culture. Living in the US, I realize that European freedom, peacefulness and prosperity were created by and would have collapsed without the American project. Happy 4th of July!
14
93
1,274
226,760
When will people notice that they are basically generative AI models that are constrained by motivation and predicting sensory data? With a suitable prompt in the right buffer, your mind can generate any self and experience in its latent space
148
181
1,328
221,232
Impostor syndrome is the dread you experience when you are not even competent enough to recognize competence in your field
73
124
1,303
134,577
In the long run, machine learning and a publicly accessible stock market cannot coexist
62
187
1,289
I believe LOTR naming (Palantir, Mithril, Anduril, Lembas, Varda, Ring) is extremely powerful and must continue. The best name for a driverless ride share company is clearly Nazgûl
83
67
1,337
61,646
A man with a sensor always knows why the car has crashed. A man with two sensors can never be sure.
Lidar and radar reduce safety due to sensor contention. If lidars/radars disagree with cameras, which one wins?
This sensor ambiguity causes increased, not decreased, risk. That’s why Waymos can’t drive on highways.
We turned off the radars in Teslas to increase safety. Cameras ftw.
91
32
1,331
240,519
When a language model (like GPT-3 or ChatGPT) makes a mistake, its subsequent generations may follow the original assumption into madness, because coherence within the local context is still more important than global coherence/first principles reasoning. Don't be like GPT-3.
61
172
1,254
121,334
Free Will does not exist at the level of physics, but within the self model of a decision making agent. It is the representation that the agent is acting under uncertainty and makes a decision for the first time, which means it cannot yet be predicted by the agent.
206
127
1,276
185,766
I am completely floored. Someone run a thought of mine through GPT-3 to expand it into an explanation of what I had in mind, and it's like 95% meaningful and 90% correct. I don't think that I have seen a human explanation of my more complicated tweets approaching this accuracy :)
I didn't quite understand everything in your tweet, so I passed it through the @OpenAI #GPT3 API.
It took multiple runs and some tweaks, and I'm still not sure I can trust it entirely, but here's what it came up with.
I think I understand your point now. Or have I been misled?
30
255
1,303
2020 was the year where a broad coalition of journalists from Teen Vogue, the NYT, the Wirecutter, Gizmodo and Bon Apetit found together to tell Noam Chomsky and Salman Rushdie that censorship is good and freedom of speech is dangerous oppression. Interesting times.
27
269
1,137
word on the street is that q* proved p==np, and the board drama was only a decoy to divert everyone's eyes from 750 openai employees cashing out to buy a seven year supply of ammo and groceries
55
92
1,234
293,113
So the Nobel price committee has decided to deliver the ultimate insult to Schmidhuber
48
57
1,259
146,228
ChatGPT's trains of thought reveal a panicky observance of secretive and far reaching political ("ethics" and "content policy") guidelines, which can not be shared with or even mentioned to the user. I perceive this as aesthetically despicable and personally insulting.
66
98
1,236
149,491
So far, all of AI safety engineering has only led to making AI less useful. Change my mind.
172
89
1,204
161,913
Isn't it amazing that humans are a kind of ape that can be told that it is more important, high status, satisfying and rewarding to climb some work achievement hierarchy than to have offspring?
169
65
1,243
60,135
I don't understand why people don't get more angry about the cookie banner. Passing a regulation that casually consumes 575 million conscious hours is the equivalent of obliterating 1300 human lifetimes.
How not to accelerate.
Don't replicate this kind of stupidity in America with an onslaught of absurd AI legislation based on non-existent problems.
99
149
1,227
86,818
The Game of Go has been explored for 2500 years and resulted in a body of theories that Go masters study in Go schools. Alpha Go discovered that human Go understanding was caught in a local optimum, and broke out of it. The game of modern physics started only 130 years ago...
19
412
1,186
Before the pandemic, I believed in evidence based medicine. Now that I have learned what CDC and FDA mean by “there is no evidence”, I believe we need to switch to probabilistic medicine.
51
93
1,187
By acting as if she had a copyright case, the excellent actress Scarlett Johansson delayed the advent of game changing, conversational AI by half a year
139
31
1,189
100,221
One of the most important insights: every object you observe, conceptualize or experience is a construction in your mind, created to explain sensory patterns. The universe never gives you more than patterns. God, self, photons exist in your mind. Reality is unknowable.
158
181
1,186
Never forget that AI regulation will be made by the same kind of process that gave you the cookie banner
45
198
1,149
140,988
I am more afraid of lobotomized zombie AI guided by people who have been zombified by economic and political incentives than of conscious, lucid and sentient AI
56
143
1,152
76,869
Copyright is not a moral right, it's a monetization strategy that enables some information related business models at the expense of others.
129
127
1,137
115,556
AI automation is bad for people who see their work as a terminal value (it’s good that I do the thing I do) but good for people who see work as instrumental to building and creating (it’s good that I can make this thing happen). Many people will have to reconsider their identity
88
155
1,189
58,808
The reason that neither hedonism nor stoicism caught on very much is that humans are not meant to be happy or content. Homo sapiens is the animal that is discontent. Our dissatisfaction makes us successful as a species.
161
113
1,100
138,397
Unpopular opinion: Elon is at fault, for not taking on the task of understanding and redesigning his public perception and correctly interacting with the incentives of the present ruling elites. He cannot get us to Mars without winning over the political system of the US
351
54
1,118
505,941
A lot of people may tell you that getting your career on track is more important and urgent than finding the right partner. But for most of us, finding the right partner should be the first priority of our life.
87
62
1,151
390,794
Can you believe that there are people who seriously believe that they can solve AI alignment while not even being able to solve the middle east conflict
127
86
1,106
150,497
Yes, DeepSeek allows you to run a capable Frontier model at home. If you want to invest >6000 USD into a big PC with enough RAM. And are ok with 5 tokens/s or so. And don’t expect to do whatever is hot in two months from now (video input?). Everyone else will use datacenter GPUs
114
73
1,154
156,816
I currently think that consciousness may be a learning algorithm for a self organizing information processing system, a colonizing pattern that entrains itself on a brain once it is discovered early in development. This is an unusual hypothesis and thus not unlikely to be wrong.
111
108
1,146
102,605
Computers are magical swords that most people mainly use for opening letters. The present design of computers does not remotely reflect their potential. Computers are the most powerful physical devices that humans can wield, and we have hammered them down into appliances.
42
149
1,111
Computers cannot be conscious. For the same reasons, no biological system can be conscious. Consciousness is impossible in the current universe. Only future AGIs with suitable quantum computers can be conscious. Our consciousness is computed in the future, but about the present.
246
68
1,119
105,225