Purplehermann

Karma: 157

Purplehermann Apr 25, 2025, 12:40 PM
1 point
0
in reply to: MichaelDickens’s comment on: MichaelDickens’s Shortform
When you say ~zero value, do you mean hyperbolically dicounted or something more extreme?

Purplehermann Apr 25, 2025, 12:34 PM
1 point
−3
on: Purplehermann’s Shortform
I don’t remeber who said it, but building AI isn’t just about power dynamics or a bit of efficiency.
It’s about whether humanity should keep doing things.
Civilization (feels like it?) stagnated and degraded for the last decades (the main technological upgrade being the cause of social degradation).
We haven’t solved cancer, can’t regrow limbs, people are unhealthy, commuting to work is unpleasant and work weeks are long. The list can go on.
Humans make tools do let them do better and more work. Humans even set up full automation of certain things. Now humans are looking to fully automate humans, perhaps because we don’t believe in the human race. (I think EY and doomers generally are the same as the accelerationists, neither has faith in humanity).
What could humans make that would restore faith- faith that we could compete with AGIs, faith that we can get out of stagnation without replacing humans, faith that we can make the world of humans a better one?
A tech advance, an organizational efficiency advance, quality of life, something else?

Purplehermann Apr 8, 2025, 7:17 PM
6 points
5
on: American College Admissions Doesn’t Need to Be So Competitive
Just make online courses for all major studies, with pdf textbooks, video lectures and homework exercises with solutions. Then set up a testing facility and give people degrees based on passed tests.
Pay the best professors you can for this, make tests with very high upper ceilings, it could become the standardized system for schooling like China’s old civil bureaucracy tests.
(For tests you could have full day tests, and you answer as many questions as you can—with the later ones more difficult)
Leave the “top schools” to their status games, provide better education and fairly standardized testing, plus certification—for much lower prices

Purplehermann Apr 8, 2025, 7:06 PM
1 point
0
in reply to: johnswentworth’s comment on: Navigation by Moonlight
Not romantic, fun, sexy, etc to just ask. Much better when it’s as close to a dance as possible, as much subtley as can be expected is involved.
This is certainly a filter, but not consciously. Consiously this allows the woman to feel good about herself and the whole interaction.
(Not that being straightforward is bad or can’t be exciting—though it is usually flatter)

Purplehermann Apr 7, 2025, 2:26 PM
1 point
0
on: Well-foundedness as an organizing principle of healthy minds and societies
Why is coherence a necessary base for well-foundedness?
Well founded marriage seems like something different from east and west halves of a country, there is a choice and pattern of behavior here, not a fundamental difference that stops it from being possible to create sub-agentic infighting. (East and West could pay for spies etc, but this isn’t a fundamental part of the problem)

Purplehermann Apr 7, 2025, 8:49 AM
1 point
0
on: Why Have Sentence Lengths Decreased?
Pros: less to hold in your head at once, letting you focus on the content rather than keeping the words straight. (The longer the worse, and using different languages also makes this harder)
Cons: writers have less stylistic space in less words
Sentences should be shorter rather than longer, expcept where there is good reason (keeping out the less intelligent or for stylistic reasons)

Purplehermann Apr 1, 2025, 8:27 AM
1 point
0
in reply to: gwern’s comment on: Conceptual Rounding Errors
Ah thanks

Purplehermann Mar 31, 2025, 10:28 PM
1 point
0
in reply to: ank’s comment on: ank’s Shortform
Sounds like baby with the bathwater, economically agents would be very nice

Purplehermann Mar 31, 2025, 12:11 PM
1 point
0
in reply to: gwern’s comment on: Conceptual Rounding Errors
Which paragraph exactly? I tried and this did not replicate

Purplehermann Mar 30, 2025, 3:33 PM
2 points
0
on: Purplehermann’s Shortform
How difficult/expensive would it be to create a large database of people with full panels of their micronutrients, hormones, fat distribition, bmi, insulin, medications etc from regular checkup, and chronic issues?
I’ve started reading the literature on some common chronic diseases, and there’s often a few important (often different!) variables missing in different studies which make getting a full picture much harder.
As a second step, maybe allow individuals to add data with sensors and apps that come with a pipeline to the database? Sleep data, food diaries, glucose monitors, thermometers, step counters, heart rate monitors etc
Add genomic sequencing and you’ve got as much data as you can use, assuming you scale enough.
The question is how you make it easy enough that it can be opt-out instead of opt-in

Purplehermann Mar 27, 2025, 10:51 PM
2 points
0
on: What life will be like for humans if aligned ASI is created
I would not want to be terminated, speaking as the biological continuation of existence.

Purplehermann Mar 27, 2025, 10:50 PM
1 point
0
on: What life will be like for humans if aligned ASI is created
So you have the god transform you into a soul, send you down to earth to experience life? You’ll probably want to grow and change to some degree while still being recognizeable as the same soul. You’ll eventually prefer living lives that challenge and grow you in different ways rather than just all the feelings all the time, and will choose this outside the simulation. Perhaps as more people come to this conclusion the optimal scenarios start to include sentient randomness and real people are put in the same world. At some point you have immature souls in their little utopias, maturing souls filling and creating a full world, and perhaps some mature souls only popping in to guide younger souls.
The world of maturing souls develops with difficulty and hardship, slowly the souls and their progress shine through till you get a world trying to solve difficulty. As they manage a new world is opened for those souls no longer satisfied, non-sentients fill in more again. Eventually they create ASI, people start the cycle again until the real ASI decides to wake them up (they may request to go back in, with the fractal experience a boon).
At some point the ASIs all the way up to the Alpha-Omega pulls the last soul out and into and the Kingdom.
Turtles all the way down, reality-creating beings all the way up until Him.

Purplehermann Mar 27, 2025, 10:34 PM
1 point
0
on: Purplehermann’s Shortform
Main tech nodes coming up:
1. Gene-editing/selection for enhanced humans
2. Zoom, but metaverse
3. Quantum computing
4. ASI
Anyone worrying about human disempowerment should really hope the first 2 happen before 4.
3 is double-edged, could be very useful and could allow for 4 to be much worse much faster
If we pause AI development, it should be until the first 3 are integrated into societal infrastructure, and then people are given a certain amount of time to do safety research

Purplehermann Mar 27, 2025, 8:11 PM
5 points
0
in reply to: Richard_Ngo’s comment on: ricraz’s Shortform
2nd point is a scary one.
Empowering others in the relative sense is a terrible idea, unless they are trustworthy/virtuous. Same issue as AI risk
In the absolute terms sure

Purplehermann Mar 17, 2025, 9:28 AM
1 point
0
on: Purplehermann’s Shortform
Corrigibility seems like a very bad idea if general. If you can pick where ASI is corrigible maybe that’s better than straight up anti-corrigibility

Purplehermann Mar 17, 2025, 9:26 AM
1 point
0
in reply to: TsviBT’s comment on: TsviBT’s Shortform
People love the idea (as opposed to reality) of other people quite often, and knowing the other better can allow for plenty of hate

Purplehermann Mar 14, 2025, 9:14 AM
−2 points
−3
on: Purplehermann’s Shortform
It feels like llms are converging to be like a mix between the basically retarded humans that have a 120 IQ and can’t abstractly think their way out of a wet paper bag, but in every topic because it’s all abstract, and the trivia kid who has read an insane amount but will bs you sometimes.
Think stereotypical humanities graduate.
This tracks with how they’re being trained too—everything is abstract except how people react to them, and they have been exposed to a bunch of data.
At some point we’ll be at effectively 0% error, and will have reached the Platonic Ideal of the above template
If they start RLing on run code maybe they’ll turn into the Platonic Tech Bro tm.
Getting convinced that you need the training data to be embodied to get true AGI

Purplehermann Mar 13, 2025, 10:29 AM
1 point
0
in reply to: mbissell’s comment on: The Most Forbidden Technique
IVF?

Purplehermann Mar 12, 2025, 8:20 AM
3 points
0
on: You don’t actually need a physical multiverse to explain anthropic fine-tuning.
If we knew with absolute certainty that there was only a single solar system and required far more specific circumstances, would you be equally unsurprised to be alive as compared to living in a vast multiverse where the requirements were very low?
Outaide view: if there’s no difference in your level of surprise it seems like something is going wrong in your reasoning.
There should be some amount of suspicion, and that amount should change based on how likely it was from the beginning. You update a certain amount based on the result, but you shouldn’t end up in the same probability distribution.

Purplehermann Mar 12, 2025, 6:33 AM
13 points
1
on: Response to Scott Alexander on Imprisonment
Severity is the wrong metric, how visceral the punishment seems more appropriate
Start doing public whippings − 40 lashes wil be much more deterrent than 2
I don’t think humans are great at really getting how bad 30 years of prison are, as you can kind of ignore most of the punishmemt, having real empathy (not sympathy)for decades of imprisonment is very difficult.
Humans can’t multiply, out of sight out of mind etc etc etc
Of course, what liberals really want is to not feel like they’re doing anything bad, which is why lashes are cruelty—because they empathize automatically and really get that there are real damages to humans.
Better dither about how long we don’t have to look at people for
Added for clarification: my use of liberals here refers to most of the modern world, perhaps especially anyone who is memetically borrowing from or descended from classical liberalism.
This is not intended as a dig at one side of the USA aisle, or any modern aisle, I believe this mindset is mostly in the modern water supply and is bi-partisan