actuallyalys

I had the thought of making one for Clojure. I didn’t realize one already existed, so thanks!

The analogy to electricity doesn’t make sense to me. OpenAI’s and their competitors’ products are already widely available, companies have added to AI to a bunch of apps, and the cost per token has already come down significantly. (Edit: not to mention openly licensed models.)

I also don’t see think it’s historically accurate. My understanding was that widespread electrification were motivated by lighting, the radio, and other applications. Getting electricity to more people probably helped spur further development, sure, but there were already “killer apps.”

I read the article more as trying to save face over this outlet or reporter previously being credulous about the metaverse, but perhaps I misread it. I'm not familiar with this reporter or outlet, so I could very well be wrong.

“The team built a “digital twin” of the Earth, much like would-be metaverse companies want to do.” Was this actually a goal of “would-be metaverse companies”?

My impression is that they were much more focused on highly abstracted spaces designed to facilitate work and social interactions.

I understood The Great Divergence as meaning there are large performance gains, but only for certain people, who are better at resisting distractions brought on by new technology. But my conclusion is closer to there being large performance gains but only for a tiny minority of tasks that happen to really benefit from them and more modest gains otherwise. E.g., smart phones are a great technical achievement, but they’re not a huge improvement most of the time—we already had laptops for most portable productivity needs.

I think diminishing returns, as the author starts to allude to near the end of the post, is a big reason. To take the example of word processing, the biggest improvement to my workflow of the past 10-15 years is probably either being able to collaborate on documents simultaneously or to be able to view them on a phone. I actually edit documents simultaneously quite a lot, but I’m not sure it’s a huge productivity boost or that common.

I’m sure there’s more logic to it, but there are twelve notes in the chromatic scale, so six rows of four gets you two octaves. Also putting them in rows of four means you get a nice interval (a fifth?) if you go down vertically.

Agreed. One of the researchers quoted has a good take on this:

> “There are definitely groups out there who would like to push the responsibility of reducing carbon emissions away from corporations and onto individuals, which is problematic,” said co-author Dr Ramit Debnath, Assistant Professor and Cambridge Zero Fellow at the University of Cambridge. “However, personal carbon footprints can illustrate the profound inequality within and between countries and help people identify how to live in a more climate-friendly way.”

Honestly only makes sense in the context of a Python library and implementation as well, since so many libraries use C extensions in order to speed up processing. Also, Python has gotten a lot faster over time.

Like your article states, there’s some evidence for exposure to benign microbes reducing allergies, but that is completely different than pathogens strengthening the immune system. Unfortunately some people seem to have remembered the hygiene hypothesis as meaning that microbes of all kinds are actually good.

There’s also a concept of immunity debt, although that’s more of an explanation for certain diseases being especially bad the year following COVID restrictions than an argument for exposure.

Yeah, there definitely seems to be quite a bit of churn for people who adopt the newest plugins, and I think that can lead to the impression that Vim and Neovim require constant configuration maintenance. I don’t think the churn is necessarily bad for people interested in spending a lot of time on their configuration because it means there’s a lot of experimentation, which eventually makes its way to the whole community.

I was a bit young to have an opinion myself but in the early part of that, you had Windows 98, which had a pretty solid reputation, and Windows XP, whose brought the stability of NT to home users. Windows ME was poorly received but it wasn’t on the market for long before Windows XP came out.

Agreed. For some situations, it might well be easier to take advantage of the static nature of the site and use SQLite compared to setting up a Postgres server. For others, setting up a server could be easier than the “easy” options.

Yeah, there was a pretty strong argument to raise taxes in 2017-2019 in order to prepare for a return of higher interest rates (or at the very least, not cut taxes further) but not much earlier in the decade. Obviously the pandemic necessitated deficit spending again, but we once again have a window of opportunity to put our economic gains into reducing the deficit.

To the Biden administration’s credit, it did pass the Inflation Reduction Act. Hopefully the next administration will let at least some of the Trump administration’s tax cuts expire and raise revenue in other areas.

Ironically, any isometric game nowadays would still use the 3D graphics capabilities under the hood so in a way you’re faking 3D by faking 2D. (Although I suppose you could argue that you’re not really faking 2D since you’re just using two out of three dimensions.)

> The EU often imposes fines mere weeks or days after compliance deadlines pass, suggesting this kind of extraction is a feature, not a bug, of the regulatory regime.

This argument is strange to me. If a local road changes its speed limit and on the same day the new limit takes effect, I get pulled over for driving too fast, that seems expected?

I guess I could see the argument that the companies don’t have enough time to make the necessary changes to adhere to the law, but then you should be advocating for the law taking effect at a later date rather than regulators being slow to issue fines when they are in full effect.

I am in the same camp of thinking that Musk overestimated how many people he could lay off while keeping Twitter going. But it seems plausible to me that Twitter was somewhat overstaffed.

How many employees are “not necessary” depends on your priorities. If you mostly want to keep a site up and running, the requirements are relatively slim. Musk also laid off a large part of the Trust and Safety team, who don’t have much to do with keeping the site running technically but were responsible for moderation. Whether their jobs were “not necessary” depends on your assessment of how much moderation Twitter needs.

I wondered that but looking at both papers, it seems much less likely. For one thing, they looked at the same illnesses in both the original paper and the new paper, so it’s not like they did a bunch of new analysis in order to find something that’s significantly.

For another, the study doesn’t look at that many diseases, just four: cancer, heart disease, lung disease, and stroke. It’s possible the original study was doing some p-hacking in order to reach significance but these are pretty major categories of disease so they seem defensible. It would not surprise me if these are the only diseases in the original dataset.

Finally, the results are significant at the .01 level. Combined with the number of diseases, this is not anything as blatant as the classic XKCD green jellybean comic, although more subtle p-hacking could be at play.

Retracted: https://journals.sagepub.com/doi/pdf/10.1177/002214651456835...

Corrected: https://journals.sagepub.com/doi/abs/10.1177/002214651559635... (Not open access but it’s available from sci-hub.)

If you use Neovim, you could try Conjure.

I really like it for other lisps, but I haven’t used it for Common Lisp. I wouldn’t say it “contorts buffers in a bizarre way,” although in my experience, different Vim users have their own take on that.

It’s definitely plausible to me she would experience some form of shadow banning or algorithmic deprioritization due to posting advertiser unfriendly content and (incorrectly) blame it on her brother. These systems are frequently opaque so I don’t think it’s reasonable to count that as a strike against her credibility.

And of course it’s not really that implausible that Altman could pull some strings. It wouldn’t require a huge conspiracy to shadow ban a few accounts. I don’t think it’s the most likely explanation, though.

> I love how we went from questioning copyright & licensing to "GPT vs Google, which one is better".

Have we? Certainly the people litigating haven't. And as this article notes, actors' newest contract does have protections against AI. SAG-AFTRA's press release states [0] they are pursuing legislation. That could be bluster or could go nowhere, but certainly people haven't given up.

[0]: https://www.sagaftra.org/sag-aftra-statement-regarding-scarl...

There wasn’t traffic in the sense of sums complaining about heavy traffic, but I believe the game had car animations. I definitely remember pedestrians and cyclists.

Yeah, I think the idea of a single interconnected network across the planet would be astounding, even if the idea of using wires for communication is not.

I think I have seen this model before, in the days of shareware. Offhand, I don’t remember the name of the tool.

A downside of this is that if you don’t choose the right metric, people might burn through a lot of their uses due to mistakes. Like, to use your input files example, imagine someone accidentally selects a huge directory with more than 100 files and then end up with no free uses left.

Some of this also comes down to overall UX design too, of course.

I believe him when he says he doesn’t think all medicine is worthless, but there are definitely a lot of moments where it seems like he’s trying to downplay even clearly effective medical treatments as having uncertain effectiveness (e.g., cancer treatment) or not really medicine (like glasses), so I don’t really blame Scott Alexander for interpreting his position as close to “no medicine is effective.” (Edit: or you from suspecting ulterior motives.)

And yeah, while I think these insurance and medical spending studies merit further consideration and follow-ups, there are a lot of factors which Scott Alexander points out beyond whether medicine works even if they are RCTs.

The elephant in the room with people questioning public healthcare or public education is that wealthy people will always be able to easily access private replacements and many middle class families will be able to cobble the funds together, so the main people losing out will be the poor.

> the height of at-home shopping

I wondered that, too, but I did see that Amazon revenue has continued to grow since 2021. Of course, it’s possible AWS and other divisions are compensating for a decline in physical goods sold. Unfortunately, I don’t have time to find revenue numbers for specific divisions.

Agreed. I think this is clearest when you look at benefits, e.g., the many American companies that only offer health insurance benefits to their full-time workers, but you see it in terms of pay, too. Most middle and high-paying jobs are exclusively full-time (and some expect 50, 60, or even more hours a week).

Right, part of the system is not allowing in people who you already know to be untrustworthy.

I do think this distinction is important beyond how we view Aristotle. Some creationists argue Louis Pasteur disproved abiogenesis (and in their view, evolution as well) even though his experiments addressed the spontaneous generation of life under current conditions, not whether any form of life can form from nonliving material at all. I believe he disproved both maggots and microorganisms spontaneously generating.

Yes, this is the correct explanation. Datelines are generally based on the writer’s location. I don’t know for sure that’s CNN policy, but it’s AP policy, which is widely followed.