Opens profile photo
Follow
James Miller
@JimDMiller
Smith College economist. Author Singularity Rising, Game Theory at Work. JD Stanford, PhD U of Chicago. AGI alignment, free speech, free markets.
jamesdmiller.orgJoined September 2009

James Miller’s Tweets

Pinned Tweet
Nerd score: How many have you considered? Cryonics, multiverse, Boltzmann brain, AI utopia, quantum immortality, Roko’s basilisk, gray goo, paperclip maximizer, great filter, ethics in infinite universe, acausal trade, longevity escape velocity, simulation and zoo hypothesis.
63
296
How reasonable is this view?
Quote Tweet
If Russia alone fought NATO to the point of exhaustion, it’s worth recalibrating whether the military can take on China. twitter.com/elbridgecolby/…
Show this thread
15
7
"So those arguing for AI regulation are saying that the exact same authorities across the western world who did such an appalling job on covid, the people who have shown no interest in reforming the shitshows of CDC and FDA etc, who even after covid have just ignored the whole… Show more
2
ChatGPT-4 saying if my spaceship travels at 1 g acceleration, takes me 20 years to travel 1 billion light years if I accelerate for first half, and decelerate for second so end up at zero velocity. But if I don't decelerate and just accelerate the whole time takes 19.6 years!
8
14
If we really have bodies of alien pilots (which I strongly doubt) probably means you can travel much faster than the speed of light.
Quote Tweet
Pentagon UFO whistleblower David Grusch says the United States is in possession of vehicles of non-human origin and bodies of the pilots. Full interview by Ross Coulthart & reaction by Ryan Graves. #ufotwitter #ufo #uap
Embedded video
12:23
1.7M views
19
40
Imagine we get proof that some UFOs are real aliens, and we are about to create an artificial superintelligence (ASI) that will exterminate humanity. Which most likely?
  • Aliens save us.
    42.2%
  • Aliens kill us.
    19.1%
  • Aliens let ASI kill us.
    38.7%
173 votesFinal results
13
4
Anti-nuclear activism which foreclosed Western Europe's sane way of reducing global warming gases is about to cause the slaughter of cows in Ireland.
Quote Tweet
The push to end life, of both animals and humans, in the name of “climate activism” is fundamentally evil Ireland Looking To Kill 200,000 Cows To Fight Climate Change cowboystatedaily.com/2023/06/02/to-
2
8
A woke publication might soon use LLMs to discern patterns in anonymized data, matching to known identities. This could result in massive doxing of those found to have committed anonymous online blasphemy, with emails automatically sent to offenders' employers and schools.
5
22
Genes + randomness determine our lives. We are basically wind-up toys.
Quote Tweet
The Jim twins were separated at birth and reunited at 39. They quickly found that they had lived oddly similar lives. Both had married and divorced someone named Linda, were currently married to a Betty, had sons named James Allan, had dogs named Toy, drove the same car, had… Show more
Image
1
7
You are almost certainly not smart enough to maintain online anonymity.
Quote Tweet
i wonder how many people know that camera sensor noise is unique so if you post pictures from the same camera on main and alt they can be linked
Show this thread
15
118
Ten years ago, I wish we had all been asked: "In a decade, experts will be at odds over a rapidly developing technology, disagreeing over whether it will exterminate us or significantly enhance our lives. How should we manage this technology?"
7
AI if aligned for truth is going to prove lots of people wrong. We should agree that accepting that AI has uncovered some of your past errors is a sign of your trustworthiness for being willing to update based on new reliable information.
Quote Tweet
If we play our cards right, AI is going to do to Wokeness what science did to Christianity
2
13
Anyone found holes in ChatGPT-4's reasoning ability compared to that of an undergraduate of average intelligence?
19
30
The stock market fake info problem could be solved by having a "reset" on trades. The stock exchange rule could be if significant fake info likely influenced trades, all relevant trades cancelled. Could limit possible cancellation period to, say, 20 min.
8
8
From OpenAI "we believe it would be unintuitively risky and difficult to stop the creation of superintelligence. Because the upsides are so tremendous, the cost to build it decreases each year, the number of actors building it is rapidly increasing, and it’s inherently part of… Show more
30
175
Average age of UK pilots during Battle of Britain was just 20. Are there aspects of AI alignment research where a well-trained 20-year-old might outshine a 25-year-old? If so, let's scout and nurture our 16-year-old talents today.
5
14
Experiment I would love to see: Pick 100 families at random. Pay them a lot to play (as best they could without endangering anyone) as being hunter-gatherers (including giving up all tech other than health care stuff). After a year figure out how happy they are.
15
22
I'm asking GPT-4 to imagine I can send tech info back in time to 1935 to help the US more easily win WW II. It tells me "It's important to keep in mind that the use of such technology should be for the preservation of peace and life rather than promoting violence."
6
20
It's the right time to establish a new college. Soon, AIs will supersede professors in teaching undergrads. Nevertheless, college will continue to provide valuable social experiences. New colleges need to identify cost-effective, appealing, and expandable locations. Grant… Show more
1
9
Some organizations might tell employees "work from home if you can still fulfill all your responsibilities" to see which jobs can soon be safely automated.
1
7
Anyone with a sufficiently high IQ + strong math skills should be able to immediately get a job in AI alignment.
Quote Tweet
Must watch: AI discussion w/ @sama and @patrickc Forget the congressional stuff, this is a high-level discussion. At 15m, SamA: ".. only takes about 6 months to take a smart physics researcher and make them into a productive AI researcher" 🤔 youtube.com/watch?v=1egAKC
2
8
Many academic revelations remain unseen by decision-makers, despite their significance. This lapse cannot extend to AI safety. AI labs must delineate what research they're open to and how academics can earn their attention.
4
Quickly clicking every link to minimize the amount of wasted time spent on yet another dumb training program is, on average, a reasonable strategy.
Quote Tweet
depolyed fake phishing emails to a client to test their security and one of their employees clicked every link because he thought thats how you past the test so its not going great
Show this thread
5
5
We live in dangerous times: (1) AI at minimum radically challenge existing elites. (2) Taiwan's chip factories make control vital to China and US. (3) Russia faces humiliation and might use nuclear weapons. (4) Next US presidential election results likely not accepted by many.
8
18
Biden forcing airlines to compensate for delayed flights caused by airline's error. Suspect is a climate change policy where raise price of travel to get fewer poor people to fly, but voters perceive as good for poor airline travelers.
3
10
If atomic bombs had just been invented, AI Luddites would be all "BOOM! New fear. Civilization might end." Like we haven't been cozying up with explosives for centuries!
3
7
If you tell me something about the limits of AI in the next 20 years that doesn't hold today for AI Chess and Go -- I'm doubtful.
4
28
AI could disrupt the crypto landscape by introducing groundbreaking currencies that surpass current options in every aspect users care about.
5
6
Open source AI morally scary if (once?) it allows anyone to create a sentient AI and do whatever they want to it. Imagine someone trains an AI to be a sentient simulation of you based on everything known about you. Could happen to help firms target ads to the real you.
2
6
The most practical way of slowing the AI arms race would be to encourage Microsoft, OpenAI, Alphabet, Anthropic, and Meta to merge into one company. Would need an antitrust exemption, but preventing the world from being destroyed should suffice.
5
9