ShredKazoo

Was there even a serious policy discussion of banning gain-of-function research?

Gain of function research is supposed to help with pandemics, right? Did it help with COVID?

In real life, intelligence is associated with picking better goals.

"Better" according to what metric?

It may be the case that there is a tendency for high-intelligence humans to pick "more enlightened" goals. Perhaps there is a natural "enlightened goals" attractor for our species.

However I don't think we can extrapolate from that to a fundamentally alien AI.

I think even if this statistical tendency exists, it has clear counterexamples -- consider that 2 genius chess players may have opposite goals, of beating one another. And we shouldn't bet the future of humanity on this statistical tendency extrapolating outside of the original distribution of human species.

Here are some intuition pumps on how diverse goals can be even across intelligent species:

* Orcas killing sharks for their livers: https://www.livescience.com/2-orcas-slaughter-19-sharks-in-a... I don't believe dolphins show the same level of violence, even though both are smart cetaceans

* Intelligent dogs bred to be responsive and attentive to human needs -- unlike close cousins like the wolf

* Chimpanzees and bonobos are both related to humans, both highly intelligent, but with very different culture and goals https://www.youtube.com/watch?v=c6Ko0Hzi47U

I don't think it's impossible to come up with arguments against them. For one thing, extrapolating the current gradual rate of progress forwards means that we will get to see several minor "intelligence spills" before the hypothetical big and last one, and by observing what went wrong, humanity will have the opportunity to come up with solutions.

This has been discussed under "warning shots" https://www.lesswrong.com/posts/idipkijjz5PoxAwju/warning-sh...

See discussion on the EA Forum

https://forum.effectivealtruism.org/posts/7b9ZDTAYQY9k6FZHS/...

Maybe you'll find this useful, I think I saw it elsewhere in this thread and bookmarked it https://github.com/poteto/hiring-without-whiteboards

Interesting take.

Doesn't that make it harder to do an apples-to-apples comparison between candidates though?

What does rewriting your codebase have to do with hiring?

How about LinkedIn certificates? Last I checked Triplebyte has a certification you can put on your LinkedIn if you do well enough on their process.

Why not just apply at companies that focus on take-home problems?

I assume that's a joke? Sounds like going back to manhole covers.

It's incredibly cold in the outer solar system right? So why is it that there is liquid water in these moons? Is it some sort of geothermal heating?

What's the most complex life which could conceivably exist in such a cold environment? Presumably there's very little sunlight penetrating through the ice to the liquid ocean.

I wonder if there's a way to write a DEB to RPM conversion script with automated testing etc, so Fedora gets to snarf up all existing DEBs :-)

This angle works in the other direction as well: Is a newborn a human? If yes, is a baby in the process of being born a human? If yes, is a baby who will be born in a week a human?

My understanding is that opinion polls show most people have a moderate opinion on abortion, which seems pretty reasonable given that "becoming human" is something that happens in a continuous manner over a 9-month period.

The "this topic is less-than-maximally-important" critique applies to pretty much every story on the HN homepage.

Fair point, but on the other hand, if language policing leads the American Medical Association to make decisions which decrease the quality of medical care, I'm not enthusiastic about that. The relative importance of the opinions of everyday people vs these associations is not obvious.

https://news.ycombinator.com/item?id=35006303

Re: "the edifices from which actual powers spring" -- if you want a high-paying job as a doctor, maybe you'll be best served by adopting their language guidelines. I could see the trend spreading through the general population because playing the language game becomes the best way to achieve "actual powers".

Interesting, thanks for chiming in. I wonder if the best way to push back against language policing is to simply poll various groups, see which language they prefer, and publish the results. The deltas could be interesting as well, e.g. even if most blind people dislike the term "blind", it would still be interesting if they are more likely to prefer the term "blind" than a member of the general population.

I think public opinion polling might actually solve the problem, because the actual group has greater moral authority than language police activists. I suspect people might not cite this Atlantic article during a discussion of whether their organization should adopt language policing for fear of coming across as a reactionary old fogey. But if you were citing a poll of the group in question, that's a level of moral authority that's hard for a language police activist to argue against.

Basically the hypothesis to test here is that a latinx-type reaction is fairly common, it just doesn't generally reach public consciousness the way it did in the case of latinx.

NAACP stands for "National Association for the Advancement of Colored People". Interesting how "Colored People" was an enlightened term in 1909, but "Colored People" is now out and "People of Color" is the new way to go.

For example, it seems like we can't say "mother" anymore in medical settings. For my partner's entire pregnancy, our providers only referred to "birthing people" because of some tiny number of trans men that exist and also want to give birth.

The troll argument here is: now that we're saying "birthing people" instead of "women", it's time to acknowledge that abortion is not actually a "women's issue".

I'm not entirely sure how to parse your comment.

Did you notice this section?

Equity-language guides are proliferating among some of the country’s leading institutions, particularly nonprofits. The American Cancer Society has one. So do the American Heart Association, the American Psychological Association, the American Medical Association, the National Recreation and Park Association, the Columbia University School of Professional Studies, and the University of Washington.

....

Public criticism led Stanford to abolish outright its Elimination of Harmful Language Initiative—not for being ridiculous, but, the university announced, for being “broadly viewed as counter to inclusivity.”

Even if no one is arguing back, it seems like the language police are winning.

I acknowledge that language shapes the way people think. But the replacement of cost-benefit analysis and evidence-based debate with arguments regarding the connotations that different words pull in has been a huge mistake in my view.

There might have been a point at which language policing was done in good faith. But I strongly suspect that the replacement of the word felon with the term justice-involved person is essentially attempting a fait accompli on a debate regarding the extent to which convicted criminals should be considered culpable for their crimes. The problem with this mode of debating is that winning correlates with the ability to wield social power, not with having facts on your side.

At this point the heterosexual relationships discussion has become so poisoned that I don't think further activism of any kind is likely to help. In terms of constructive solutions, I think what we need are people pioneering new "curated dating spaces" with explicit norms, events for underserved singles with enforced rules that represent some sort of negotiated compromise between what single men want and what single women want. Don't compete with dating apps, that's low-margin and has entrenched incumbents. Partner with nightlife establishments to run events, singles night with a theme of some kind and a reputation system for attendees, perhaps powered by matchmaking software. That's my take.

The thing that frustrates me is that if you speak out against safetyism, some people will ignore what you're saying and just declare you a safety threat.

It'd be nice if we made more decisions using cost/benefit analysis instead of emotive fearmongering.

I actually think 2001 is a good talking point, because by now there is widespread agreement that it represents a case of fearmongering getting out of control -- with the benefit of hindsight, it's clear that people who e.g. spoke out against TSA excesses were not, in fact, aiding and abetting terrorists. But it seems like analogous arguments "aiding and abetting" arguments are currently being made in a few other domains.

I think the explanation is the same as for hitchhiking:

https://freakonomics.com/podcast/where-have-all-the-hitchhik...

It's a multi-step process:

1. Society has some kind of neutral/positive norm like hitchhiking or friendly chats with strangers

2. There are a few high-profile incidents of people abusing this norm. Fearmongering news stories, hurt/anxious people staying inside and venting on the internet, etc.

3. The message goes out: "Responsible people avoid [norm] for the sake of safety"

4. After the message goes out, people respond differently depending on how concerned they are with being prosocial. The people concerned with being prosocial tend to abide by the message and avoid the norm. The people less concerned with being prosocial tend to ignore the warning.

5. The norm is now a little bit riskier, because the people who engage in it are now disproportionately antisocial. This causes the per-capita rate of abuse to rise. Return to step 2.

So this is not the first time HN has discussed the Pew statistics showing far more single men than women in the 18-29 bracket:

https://news.ycombinator.com/item?id=34954694

People have proposed so many hypotheses:

* Women are getting into same-sex relationships

* Women are dating older guys

* Some men are dating multiple women

* Grey zone relationships where the woman considers it "committed" and the man does not

* Probably other explanations I'm not thinking of

It would be easy to do another survey, add on some additional questions, and figure out which hypotheses are true. For people in committed relationships, add questions like:

* Do you have multiple committed partners?

* How old are your committed partners?

* What genders are your committed partners?

* How many committed partners do your partners have?

* Are you confident that the person you consider yourself to be in a "committed relationship" with would describe it the same way you do?

And for single people, add questions about whether they want to be in a relationship, whether they have casual relationships they don't consider "committed", whether anyone else would claim they're in a committed relationship with you, etc.

Until we do a second survey, all discussion of this issue is just anecdotal evidence and speculation.

There are so many people looking for respectful, reasonably attractive, healthy, financially successful and emotionally mature men.

In what region?

First two sentences of the abstract:

Given the computational cost and technical expertise required to train machine learning models, users may delegate the task of learning to a service provider. Delegation of learning has clear benefits, and at the same time raises serious concerns of trust.

My understanding was that the threat model for data poisoning is when the attacker controls part or all of your dataset, not the learning algorithm. Am I getting this wrong?

through some cryptographic trickery you can prove that the backdoor can't be detected.

Can you explain more about this? E.g. in the worst case, if I know the learning algorithm, I could retrain the model myself and notice the difference, right? What is the threat model exactly?

I always figured that ransomware was like a vaccine, provoking an immune response, with the net result of overall stronger cyber defenses

Despite our citizens’ penchant – almost enthusiasm – for self-criticism and self-doubt, I have yet to see a time when it made sense to make a long-term bet against America.

https://www.berkshirehathaway.com/letters/2022ltr.pdf

I'm an American in my early 30s. My impression is that within my lifetime, the self-criticism and self-doubt has increased a lot. Does this accord with the perspective of older Americans? I remember my dad (in his 60s) said that the pre-Watergate times really were quite different in terms of the trust that Americans had in their government. But he didn't give me a sense of the overall trend over the past few decades.

It does feel to me like we've reached the zone of diminishing returns to negativity. For example, if America now was like the America I grew up in, I feel like we would make at least a token effort to figure out what went wrong with the CDC's response to the pandemic and how the agency could do better next time, through some sort of blue ribbon commission. But it feels like the pandemic (and everything else) has become hopelessly partisan in a way that interferes with constructive change.