akprasad

This method assumes that the weaker model is aligned. I'm curious how the paper addresses that point.

> "But what does this second turtle stand on?" persisted James patiently.

> To this, the little old lady crowed triumphantly,

> "It's no use, Mr. James—it's turtles all the way down."

Quoted:

> The release did not mention the AI scandal from November, which was spurred by an investigative report published by the science and technology news publication Futurism. Earlier this year, a round of layoffs hit the magazine shortly after the Arena Group’s CEO and chairman, Ross Levinsohn, announced the company’s incorporation of AI. Sports Illustrated accused of publishing articles written by AI Read more

> Among fake profiles uncovered by Futurism was that of purported author “Sora Tanaka” which claims she is a product reviewer. The page said: “Sora has always been a fitness guru, and loves to try different foods and drinks. Ms Tanaka is thrilled to bring her fitness and nutritional expertise to the Product Reviews Team, and promises to bring you nothing but the best of the best.”

> Tanaka is not a real person.

The ASCII candle is really charming, especially when lit. Looking forward to exploring this further. Great work!

There are two separate concerns here:

- increases in material standards of living

- decreases in class mobility and increased concentration of social power

It is perfectly coherent to appreciate one and condemn the other. The parent comment seems to focus on the latter.

Broadly, nihilism rejects purpose and meaning. I don't see how one can justify we "move on with living better lives" within such a worldview (i.e. why act that way when nothing matters?).

Assuming this is all true, what then makes the Japanese Shinkansen economically viable?

I take the parent comment's point as "beneficial for children's learning and socializing" as compared to the status quo ante.

Andrew sets an extremely high bar for open-source engineering. As a loose member of the broader Rust community, I've benefited enormously not only from his code but also from his discussion comments and his blog posts (especially [1] and [2]).

Congratulations on the release, Andrew! I also notice that you've joined Astral [3], which as a fan of Rust/Ruff/rg I'm thrilled about.

[1]: https://blog.burntsushi.net/transducers/

[2]: https://blog.burntsushi.net/foss/

[3]: https://github.com/astral-sh

> How can the work both be "genuinely well-intentioned" and at the same time have "complete absence of concern for the user"?

I think the author is distinguishing between (1) well-intentioned work from workers in the trenches and (2) indifference to the user from upper management. It's institutional misalignment.

I just donated $10 to Signal. Here's how to do so on iPhone in less than a minute:

1. Open Signal and click on your user icon in the upper left.

2. Go to "Settings" --> "Donate to Signal".

3. Click "Donate", select your donation options, and pay with Apple Pay.

The article states:

> Despite this evidence, the gas industry’s campaign was largely successful. Industry-funded studies successfully muddied the waters, as I have seen over the course of my research career, and stalled further federal investigations or regulations addressing gas stove safety.

Consumer indifference to safety concerns and federal lethargy in regulating or investigating them are, to me, precisely the outcomes a misinformation campaign of this kind would aim for.

Nor is the impact small. It's just one study, but:

> This issue took on new life at the end of 2022, when researchers published a new study estimating that 12.7 percent of US cases of childhood asthma—about one case in eight—were attributable to gas stoves.

The AFR piece that underlies this article [1] [2] has more detail on Ng's argument:

> [Ng] said that the “bad idea that AI could make us go extinct” was merging with the “bad idea that a good way to make AI safer is to impose burdensome licensing requirements” on the AI industry.

> “There’s a standard regulatory capture playbook that has played out in other industries, and I would hate to see that executed successfully in AI.”

> “Just to be clear, AI has caused harm. Self-driving cars have killed people. In 2010, an automated trading algorithm crashed the stock market. Regulation has a role. But just because regulation could be helpful doesn’t mean we want bad regulation.”

[1]: https://www.afr.com/technology/google-brain-founder-says-big...

[2]: https://web.archive.org/web/20231030062420/https://www.afr.c...

Congratulations on shipping.

I see a lot of focus on startups, AI tools, productivity hacks, tech stacks, etc. What audience do you have in mind? I personally find that the most interesting sites on HN are outside of your scope here (examples: https://ciechanow.ski/, https://neal.fun/, https://search.marginalia.nu/).

Iceland has a total population of just under 400,000 [1], so that's quite the turnout!

[1]: https://www.statice.is/publications/news-archive/inhabitants...

The article states that doctors observed an increased rate of cardiac injuries in COVID-19 patients in early 2020, well before vaccines were available.

The Luddites were not anti-technology per se. Rather, they were against an unequal distribution of that technology because it was destroying their community and livelihood, precisely because that technology concentrated power in the hands of the capital class [1]. Given the rest of your comment, it seems the Luddites are your natural allies.

Related HN discussion on [1] from 3 weeks ago: [2]

[1]: https://archive.ph/ZUiC7 ("Rethinking the Luddites," The New Yorker)

[2]: https://news.ycombinator.com/item?id=37664682

The essay defines the concept of psychogeography then explores how people have historically engaged with that concept.

As a small example of what this essay is pointing to: suppose your neighborhood could be 100% walkable and that you could easily walk to work, the doctor, your friends' houses, etc. What is desirable about that kind of environment? What would it take to make it possible? What kinds of systemic forces counteract it?

> Many of the issues we face [...] are deeply connected to the physical world and our interactions with our immediate surroundings. [...] As we emerge into post-pandemic public spheres, we have the opportunity to imagine new versions of the public sphere[.]

> Guy Debord, a founding member of both bodies, defined psychogeography as an environment’s impact, whether mindful or not, on an individual’s behaviors or emotions.

> These sorts of actions, even on the smallest scale, carry significant meaning when practitioners assert how they wish to inhabit a space, and when they are able to convince others to likewise undertake this reflective process of questioning the status quo.

Another lovely example of this format is https://idlewords.com/talks/.

Some prior discussion on namespacing packages in Rust:

https://internals.rust-lang.org/t/blog-post-no-namespaces-in...

See my comment here: https://news.ycombinator.com/item?id=37745319

> For those interested, a link to the CA proposal:

> https://www.cde.ca.gov/ci/ma/cf/

For those interested, a link to the CA proposal:

https://www.cde.ca.gov/ci/ma/cf/

None of these four points are part of the author's focus. Broadly, I would summarize the author's points as:

- Alignment and capabilities research are not separate. There is an "alignment dilemma" on whether to contribute to alignment work or abstain from it.

- Discussions of AI risk are subject to "persuasion paradoxes" that make it difficult to reach clarity.

- It is helpful to take a "recipes" framing: (a) what are recipes for destroying the world? (b) how likely is an AI to discover or enact these recipes?

Overall, a clear post.

My guess is that this should be something like "If you have n significant digits in your evaluation metric, you should have at least 10^(n+1) data points."

Matt Might's writing has appeared a few times on HN. In this post, he digests some of the lessons he's learned from his son, who passed at age twelve from a rare genetic disease. I found this post moving and insightful, and I thought others might feel the same way.

I've almost always heard "creationism" used in a much narrower sense:

> The term creationism most often refers to belief in special creation; the claim that the universe and lifeforms were created as they exist today by divine action, and that the only true explanations are those which are compatible with a Christian fundamentalist literal interpretation of the creation myth found in the Bible's Genesis creation narrative.

https://en.wikipedia.org/wiki/Creationism

"you" (plural) as in "you all," since "you" can also refer to just one person now that "thou" is no longer used [1].

[1]: https://en.wikipedia.org/wiki/Thou

I agree. The quote I shared above has nothing to do with external systems that dole out rewards and punishments. Ancient definitions of karma are very different from the mainstream definition in vogue today (edit: with the usual caveat that ancient concepts have multiple interpretations, and I'm sure there are contemporaries of the view I shared that are more in line with what you imagine).

That's a nice way to phrase it! Google tells me this quote is from Ralph Waldo Emerson, who himself was deeply influenced by ancient Indian thought (see [1], e.g.).

[1]: https://www.jstor.org/stable/2709649

(edit: replaced [1] with a better link.)

For a similar idea from another culture, see the ancient definitions of karma:

  Now as a man is like this or like that,
  according as he acts and according as he behaves, so will he be;
  a man of good acts will become good, a man of bad acts, bad;
  he becomes pure by pure deeds, bad by bad deeds;

  And here they say that a person consists of desires,
  and as is his desire, so is his will;
  and as is his will, so is his deed;
  and whatever deed he does, that he will reap.

— Brihadaranyaka Upanishad

I believe you and the author are describing two different types of students. The author describes students who want to work and succeed but are unable to cope with "real-world problems rather than contrived academic examples," whereas you seem to point to students who want instead to luxuriate in their self-perceived social class.