ThenElection
No bio...
User ID: 622
ThenElection
7hr ago
My point was different than you interpret: I was responding to the idea that people of means have access to special networks of information. Money gives optionality, which is an undisputed good, but it doesn't give special access to information about how to prepare for AI.
ThenElection
7hr ago
The idea that people with lots of resources are better positioned to find ways to prepare is off. It's not like advisors at the family office have any particular insight into AI. They have been selected for basic competency and controlled risk management, not predicting radical step changes in the world. If they fail to predict ruin from AI, they'll have lots of good company; if they stick their neck out on AI predictions and fail, they'll face much worse consequences. At most, they'll say "this AI thing seems important, let's reallocate your portfolio to include more IBM."
With potential AGI, no one has a solid understanding of what will happen. In those situations, mainstream opinion sources default to status quo bias, which is about the worst thing to do. Weird randos on obscure Internet forums at least offer the potential for some variance.
ThenElection
9hr ago
We don't know. But it's not guaranteed they will, and what determines whether Democrats will are how powerful trans activists are when Democrats win, not past actions or politicians' stated values at one moment.
A downside of this framework is that power is opaque, and the clearest way to seeing whether trans groups are powerful is whether they can cause Democratic politicians to send trans people to women's sports and prisons. Beyond that, we have to read tea leaves: how does media treat trans issues? Do tech platforms give them full censorship rights?
ThenElection
9hr ago
The incentives they experience and respond to. You might have two different people and, at a particular moment of time, they respond the same way to the incentives they face. But if the landscape of incentives change, their actions might diverge.
E.g. if your mental model of Gavin Newsom involves him being deeply ideological on trans issues, then you wouldn't have been able to predict he'd switch to moderating his positions when his party was faced with a broad electoral loss (and wanting to prepare himself for a national run).
And, on the other side, until Trump came most Republican politicians would have condemned broad tariffs and been pro-war. But change the political landscape, and they change their positions.
ThenElection
17hr ago
I'm using the app, have plus, and haven't run into any refusals yet. So far all of my images have been of friends, me and my wife, or political/cultural figures, nothing violent or sexual. I upload the photo, and ask "Can you make this photo in the style of a Ghibli animation frame?"
Still seems to be working.
ThenElection
19hr ago
Well, there's Gavin Newsom's post-election switch to opposing trans athletes in women's sports.
You can take a kind of functionalist position and say that Democratic politicians are what they do, and so in 2020 they were radical trans ideologues. Sure. But it doesn't give much insight into how they will respond to changing circumstances. Democratic electeds don't have deeply held principles (no more than Republicans do): they react to incentives, around easily understood things like power, money, and status.
The extremists driving the unpopular trans positions, on the other hand, are not going to suddenly abandon their views once they start costing them power, money, and status. (And the broader Democratic base will shift to supporting whatever Democratic leadership and media tells them to.)
When people hear that Comrade Bob was accused of sexual misconduct, they immediately think of Harvey Weinstein and Father Jim, panic, and do whatever they can to avoid accusations of conducting their own cover-up.
That reminds me of something I was reading last week. There was a student athlete at Stanford who committed suicide, which is sad and terrible. The reason? She was due for a disciplinary hearing: bad Stanford, I guess. News articles then went into what happened, and they usually framed it as she was facing the hearing because she spilled a cup of coffee on someone.
Hmm. Digging deeper, she had thrown a hot cup of coffee onto another student's face. Okay, this is getting juicy. Maybe he deserved it? What did he do? Well, the news articles breathlessly reported he had been accused of sexually assaulting someone.
Finally, I find out the root cause: he had kissed one of her team members without consent.
These are all obvious questions to ask, and the actual story is pretty straightforward, a series of banal student hijinks that ended in tragedy. But the sheer unwillingness for any news articles to simply tell the story is a result of the dynamics you mention. No one wants to be the bad guy and say "well actually, Stanford didn't brutally murder an innocent girl to help cover up a rapist's crimes," because if you do, you're all of a sudden part of the rape cult, opening you to attack and hurting your career advancement.
ThenElection
21hr ago
OAI definitely won the day in the popular vote. And, at least for now, they seem to have taken off most of the guardrails: https://x.com/paul_hundred/status/1904933164256002086
Does it matter? In terms of trajectories, this is about expected, so I'm not updating much from the image gen release. Google actually got to this approach first, albeit with an inferior version that was badly marketed and had a mediocre rollout (history repeats itself, first as a tragedy, the millionth time as a farce).
What is mildly important is Gemini 2.5. It's not out of line with the expected trajectory, but it gives lie to claims about "hitting a wall." It trades blows with Anthropic, OAI, and DeepSeek, and even surpasses them--for the time being, at least--but does so much more cheaply than Claude or ChatGPT. My main takeaways are that progress continues and that Google's structural advantages are finally making Gemini a plausible leader of Western LLMs.
ThenElection
2d ago
Baijiu is the devil. Although I agree it's a sweeter drink, the description I'd go with is more gasoline than anything else. One thing I found interesting when I first drank it was that the premier producer of it is a public state owned enterprise (Kweichow Moutai) with a market cap of around $300B USD, roughly comparable with Coca-Cola.
My own first extended trip to China involved me quitting my job and flying out to China to be with a now ex-girlfriend, planning to use the country as a base to travel around for some unspecified amount of time. But three days after my flight landed, there was some kind of flu that was going around, the entire country was entirely shut down, and I ended up staying with her family for a couple months. This was an interesting experience, to say the least. Surreptitiously smuggled produce from the countryside; needing my passport to enter the local Carrefour; teams of hundreds of men spraying down the sidewalks with disinfectant daily; lots of high stakes MJ.
The family itself was... unique, though all very welcoming. It was large; she had five siblings, as the first five children were all daughters. This confused me when she first told me about it--what about the one child policy? But, there are ways around that, as it turns out. The family was extraordinarily well off by any standards (net worth in the nine figures USD). Her father was an impressive character: in his teens and early twenties, he had sold turnips out of a cart and had nothing to his name. Eventually, he started a business with two partners in construction in a small third tier city (population just a couple million), which ended up extremely successful after a decade or two of economic boom times. His partners then betrayed him, removing him from the business and dispossessing him of most of his assets. Somehow, a year later the government investigated them for corruption and restored his ownership of the company, which was a great windfall for him. Them, not so much: when I asked what happened to them, one is still ostensibly in prison, and the other died a couple months after arrest, under unclear circumstances. My ex was reticent when it came to explaining the exact mechanics of how all this happened, so I didn't press too much on this topic.
Her youngest, the brother, was very kind, and constantly wanted to drink baijiu. Having never graduated high school, his only accomplishment for his two decades was an extraordinary sneaker collection. He had an entire room filled up with hundreds of pairs of brand new luxury sneakers. Beyond that, he had a newborn child, who lived apart from him with his wife. At the time, they were separated, but the families were working to reunite them. This involved lots of negotiations: his wife wanted a larger stipend from his family (they were only paying the wife's family ~$15k USD/month, which they felt was rather stingy and should be doubled) and an upgraded condo. (These stipends, unfortunately, were not available to the daughters' partners.)
What a place and time.
ThenElection
3d ago
For what it's worth, NYU isn't claiming the data are fabricated, instead so far focusing on the illegality of the action.
Though I suppose they have a team of lawyers advising them very closely to say the minimum necessary.
It's more than that; it's almost a requirement, especially for men who want a family in a HCOL blue city. A male educator dedicating his career to helping marginalized youth, no matter his ideological bona fides and other good qualities, is going to have a much harder time finding a wife than even an entirely apolitical and unexceptional corporate guy.
It's widely understood that corporate jobs are just jobs, and you can't be blamed for getting yours. One of the most rabidly woke people I know on social media is a (Asian, female, bisexual) lawyer whose day job is quite literally union busting.
(I don't care about the actual choice of career, just the hypocrisy.)
ThenElection
6d ago
(Warning to the reader: this turned into an extended rant.)
The issue with homeless shelters is quite simple: other homeless people. They are unsafe and chaotic. You can add rules to make this slightly better--no drugs, no alcohol, no pets--but that makes the homeless you most want out of the libraries and off the streets even less willing to go to a shelter.
So that leaves individual housing and apartments. But they can't be temporary: if they are, what happens when the beneficiary runs out of time? Do you kick them out, making them homeless again? So you indefinitely let them stay. A one bedroom in my city runs around $2500 a month, at the very low end. That's $36k/year for each person housed, which in isolation is still better than $100k/year. But the population housed would be constantly growing. And it's assuming no additional costs: you might reduce emergency room visits from once per week to once per month, but it's still a cost. And what happens when the tenant destroys substantial parts of the property? During COVID, vacant hotels were used by my city to house the homeless, and one hosting a couple hundred suffered $20M in damages over two years. $20M here, $20M there, and soon you're talking about real money.
All these funds are coming from taxpayers that are themselves having to spend a significant part, and often a majority, of their income to pay for rent or mortgage. It's the number one reason people leave my city.
And yes, our housing policy is shit, significantly contributing to the issue. But in a world where activism to improve our housing policy has failed for over a decade, I have to assume that it'll be at least a decade before anything improves on that front. Does that mean I should just forego crazy luxuries like clean and safe libraries, parks, sidewalks, and transit for the next decade? Why shouldn't I just move, taking the 60k I pay every year to the city along with me, when there are plenty of places that do manage to have public spaces at a small fraction of the cost? Plenty of people are doing exactly that already, which has driven massive deficits in the city budget. And then how are we going to pay for even more homeless services? Shutting down schools? Libraries? Parks?
ThenElection
6d ago
My city spends almost six figures per homeless person. The exact accounting is difficult, because of a combination of understandable (what philosophically counts as spending on the homeless?), bureaucratic (how do you get figures on the costs of emergency room visits?), and sheer graft (nonprofits that mysteriously siphon away lots of money with no services rendered evident), but it's a lot. Despite that, the homeless problem is as bad as ever, and many of the libraries are as a result entirely unusable to the public.
So, suppose it is true that so long as spending isn't, say, a quarter million per year per homeless, libraries will remain unusable. Voters are left with a set of unenviable choices: spend a quarter million per year on the homeless and finally get clean safe libraries; let libraries remain ersatz day shelters for the homeless that happen to be decorated with shelves of books; or stop funding public libraries. The first option isn't practicable, and the second is just stupid. So the third option ends up being the one that actually happens.
ThenElection
7d ago
In a deep blue area, mine are similar. Except, they're also places that double as homeless shelters. Last time I went to one for a change of pace, in front of me there was literally a man looking up images of preteen boys in briefs on a computer, and he was zooming in on some... Very suspect areas. And no one dares go to a bathroom, because you've got an appreciable chance of stumbling across someone who ODed on fent.
I like libraries and in theory want them funded, but I want them to exist to serve the local community. They would have more defenders if they served their purpose of being places to borrow books instead of being places to enact an ideological agenda.
Widespread, I don't know, but it did happen. For instance:
The Morning Star bundle ceremony among the Skiri Pawnees (who lived in what is now central Nebraska) reasserted devotion to the power of the rising Eastern Star (Mars). It was their only ritual involving human sacrifice and was one of only a few not tied to seasonal cycles. For the ceremony to occur, a male member of the tribe had to announce that he had seen Morning Star in a dream and, upon awakening, perceived it rising in the east. Ritual tradition then called for dispatch of the dreamer (now deemed the "warrior leader") to secure a girl captive by raiding neighboring villages... After being dressed by the Morning Star priest in sacred raiments from the Morning Star bundle and anointed with red ointment, the captive stayed with the Wolf man... the Wolf man led the captive to the scaffold, constructed of different symbolic species of wood. The killing was carried out with a ceremonial bow and arrow. Immediately a stone knife incision was made near the heart, and specially prepared buffalo meat held to receive drops of the victim's blood before being prepared for feasting. Before the body was removed and placed in the prairie facing east, the entire village, including children, lodged dozens of arrows in the victim's back. The Skiris believed that this ceremony allowed the victim's spirit to ascend to the sky to become a star, while her body returned to the earth... The last known Morning Star Ceremony sacrifice took place on April 22, 1838, with the killing of Haxti, a fifteen-year-old Lakota girl. The United States subsequently suppressed the ceremony, but it also seems that some Skiris themselves wished to stop the human sacrifice.
http://plainshumanities.unl.edu/encyclopedia/doc/egp.rel.035
I'm always somewhat torn about highlighting this kind of thing: it's not surprising for Native Americans' level of development, and most cultures have things equally sordid in their past. But it's an essential corrective to the idea that Native Americans were noble savages, little fairy children dancing in the woods in communal bliss until big mean Europeans came and ruined Eden.
ThenElection
8d ago
I'd be curious about the job. It almost functions as a test of some combination of IQ plus perseverance: if you mark all the good words, you get maximum points. You can do this faster if you're smarter, but can make up for it with enough perseverance. Not altogether unlike work itself.
Though, if I really wanted the job and had this as a step, I'd probably just write a script for it.
ThenElection
8d ago
Some mad black Deleuzianism, arriving from 1993:
With the emergence of a market-driven integrated technoscience of control and communications comes the diffusion of electronically synthesized reality interfaces across the entire efferant and afferant surface of the Body. Having libidinally saturated the actually-existing channels of consumption, capital is overflowing into cybersex — sex with/through computers - in its relentless passage to the traumatic disorganization of the biological order. Eros dissolves definitively into its function as a subprogram of runaway Thanatos at the point that it unreservedly invests technical interfacing with digitally synthesized excitations. The mask capital exhibited to seduce eros was a pretension to ultimately resolve matters in relation to stimulation or unpleasure, but this has now fallen away, since cybersexuated capital cynically displays its program to replicate a tradable modulation of unpleasure, and thus its unsurpassable addiction to traumatic excitation.
Technocapital colonizes every aspect of human life, deterritoralizing all that is holy.
Gen Z enters a world where their humanity has been abstracted and previous social bonds and customs have been relentlessly mined for capital. The job market (like dating apps) operates within a libidinal economy: it is driven by desire, competition, and consumption of possibility. Just as cybersex is a market-saturated extreme of libidinal capitalism, the contemporary job hunt is saturated by hypercompetitive dynamics and the promise of a better two-sided match. Capital and desire become identified: the ideal job or partner is constantly promised but never realized, which makes the Molochian altar ever more appealing.
ThenElection
11d ago
the ability to point the whole of society against any individual. Western democracies, influenced by Hobbes, had gone to great lengths to make sure this could not be done without considerable hurdles.
This wasn't some ambient power that the Left decided to wield one day. It's a newly developed power, enabled by technology and the virtualization of society. Before ~2000, there was no way the whole of society could be pointed at anyone. Everything was too fragmented for any group (even the elite!) to act coherently, and information was collected, integrated, and acted upon in a much looser cycle. "What should the world do if a random person does a Nazi salute in her car?" wasn't a question anyone considered, because it was entirely unactionable. Now, though, a picture gets taken and shared online, millions of people can see it within an hour, her employer can be easily identified and be communicated to, public relations can carefully track negative sentiments, and she can be fired by lunchtime.
Why was it initially wielded by the Left? Good timing and proximity to the tools of symbolic production.
ThenElection
14d ago
The IQ 130+ students aren't getting individual calc tutors in middle school, it bears pointing out.
Differentiated education is good. But you could go much further with the same amount of resources by grouping students of roughly the same ability and teaching classes targeted to their abilities. Even if you want to put more resources into students at the lower end, it makes more sense to group them in low student-teacher ratio classes appropriate to their learning rate.
ThenElection
18d ago
However, I find it fairly unlikely (<50% chance) that by the end of 2025 there will be an AI that exists that can 1) be able to play Pokemon at the level of a human child, i.e. beat the game, able to do basic navigation, not have tons of lag in between trivial actions, and 2) be genuinely general (putting the G in AGI) and not just overfit to Pokemon, with evidence coming from being able to achieve similar results in similar games like Fire Emblem, Dragon Quest, early Final Fantasy titles, or whatever else.
Back in 2020, Google announced MuZero, which could play not only chess and go but also Atari games. It wasn't an LLM, but it was a deep RL system, and I'm pretty confident that it's capable of learning to play Pokemon well past the level of the average child.
LLMs playing Pokemon badly is like managing to teach your dog perfect English and then observing that in the process he picked up mediocre French, despite not being taught it . Would we criticize the dog for that? What's impressive about Claude is that it can kind of play at all, despite it not being trained to.
Now, is there a significant gap between Claude and MuZero? Who knows; MuZero did use (lots of) task-specific training, and maybe with some Claude could match it (not interesting to me; I only care about Pokemon as a test of how well a model can one-shot it). But 3 years ago, transformer-based models were very limited in what they could do compared to today; 3 months ago, they couldn't play Pokemon at all; today, they play Pokemon as badly as a 90 year old grandma with dementia (though badly in a very alien way). I'd be surprised if 3 months from now any LLMs could play Pokemon as well as a child, but I'd be even more surprised if 3 years from now they didn't play flawlessly. And although getting there from here is highly nontrivial, it's also not some vast unknown researchers and engineers have no idea how to approach.
ThenElection
21d ago
Kind of, but San Francisco politics is complicated. Newsom was known as a business-friendly moderate in San Francisco politics. "Progressives" absolutely despised him, for being anti-homeless, friendly to real estate developers, and a crony to a certain corrupt state machine politician named Willie Brown (there's that name again...) In early 2004, he wanted to shore up his left flank a bit after a fairly close 2003 election against Matt Gonzalez, a Green. The gay marriage stunt was perfect for that: it distracted angry progressive voters by positioning him as an enemy of national conservative political figures, and it cemented his dominance of California's LGBT donor class.
ThenElection
21d ago
Although they'll play at the role when it suits them for political advantage, Democratic elected officials are not ideologues. They have plenty of ambition, and when it's advantageous for their careers, they'll happily take whatever stance best straddles the line between electability and donor support.
Newsom calls trans athletes playing in women’s sports ‘deeply unfair’
Does it take a complete cynic to think Newsom didn't come to this position honestly, as opposed to sticking his finger in the wind? One might even suggest that he has Presidential ambitions. And, keep in mind, Newsom is thoroughly integrated into the Democratic apparatus: he's not someone who has built a brand on any kind of independence or heterodox thinking.
Democrats have mispositioned themselves for the past couple years, but they can jettison the most unpopular parts of their electoral coalition and come back again and win.
ThenElection
30d ago
The correct analogy would be that you receive a phone call directly from the CEO's deputy, where he verifies his identity, and tells you "you're about to receive an email saying...".
That's not quite the analogy. The CEO announced publicly that he would be sending emails. So, sure, employees know that the CEO is sending an email and to expect one. But external phishers also know that employees are expecting an email from the CEO.
Generally speaking, at scale getting employees not to divulge sensitive information to phishing attempts is a really hard problem. Even giving very specific, clear instructions (expect it in this time interval, from this domain, from this identity) is going to fail, because it always fails, even when dealing with workers with high technological and intellectual capacity. The only thing that kind of works is "don't trust anything from external domains."
It does not have the ability do to anything but react to input in ways that are predetermined by the design of the chinese manual
Not too familiar with Searle's argument, but isn't this just saying that the lack of ability to generalize out of distribution is the issue? But I don't get how being able to react to novel inputs (in a useful way) would even help things much. Suppose one did come up with a finite set of rules that allowed one to output Chinese to arbitrary inputs in highly intelligent, coherent ways. It's still, AFAICT, still just a room with a guy inside to Searle.
Perhaps it's the ability to learn. But even then, you could have the guy follow some RL algorithm to update the symbols in the translation lookup algorithm book, and it's still just a guy in the room (to Searle).
It's not even clear to me how one could resolve this: at some point, a guy in the room could be manipulating symbols in a way that mirrors Xi Jinpeng's neural activations arbitrarily closely (with a big enough room and a long enough time), and Searle and I would immediately come to completely confident and opposite conclusions about the nature of the room. It just seems flatly ridiculous to me that the presence of dopamine and glutamate impart consciousness to a system, but I don't get how to argue against that (or even get how Searle would say that's different from his actual argument).
The geofencing is something I have some ambiguity on. Is it primarily legal/regulatory, or is it because Waymo requires extensive pre-data to function? I.e. if you dropped a Waymo on a Montana back road, would it be able to drive and navigate as well as a human driver in the same situation?
It seems like a bit of an unfair standard to hold it against Waymo capabilities if the issue is primarily legal/regulatory.
More options
Context Copy link