uniqueuid

I haven't thought too long about the article, but to me one perspective seems missing:

Psychologist Robert Jay Lifton wrote an extremely impressive and surprisingly sober book "Thought Reform and the Psychology of Totalism" on the torture methods used in mainland China. He interviewed victims that had just escaped to Hong Kong (then still a safe harbor for civic freedom).

This book has much more detailed, personal impressions of what "struggle" means. Struggle in torture prisons is a sequence of absolute physical and psychological destruction, followed by the attempt of the tortured to break and destroy their old selves, to denounce what they identified with, loved and fought for, to try and re-model themselves into model citizens. Only, what exactly the party demands model citizens to be is often left vague.

I was left with the impression that the extreme violence exerted on individuals cannot be left out of any high-level analysis of what totalitarianism does.

diskernet is certainly interesting because it records archives as you browse.

But it's not open source and pretty limited regarding the use case.

The thing is, archiving is a multi-faceted and hard problem (i.e. video content, live streams, interactive sites), and will remain so. A complicated task leads to complicated tools.

Thanks, I hope you find Gigerenzer useful. The paper is a bit academic, but he also wrote a couple of nice popular science books on the (mis-)perception of numbers and statistics, those might be useful in a business environment.

For real-world applications outside engineering and academia, I would rely heavily on confidence intervals and/or confidence bands. For example, the packages from easystats [1] in R have quite a few very useful visualization functions, which make it very easy to interpret results of statistical tests. You can even get a textual precise description, but then again, that's intended for papers and not a wider audience.

Apart from that, I would mainly echo recommendations from people like Andrew Gelman, John Tukey, Edward Tufte etc.: Visuals are extremely useful and contain a lot of data. Use e.g. scatterplots with jittered points to show raw data and the goodness of fit. People will intuitively make more of it than of a single p-value.

[1] https://easystats.github.io/easystats/

Yes, I argue that individual p-values (as they are used almost exclusively in numerous disciplines) are bad, and adding more information on effect size and errors are needed. CIs do that by conveying (1) significance (does not include zero), (2) magnitude of effect (mean of CI), and (3) errors/noise (width of CI). That's significantly better than a single p-value (excuse the pun).

I can understand why you see it this way, but still disagree:

(1) p-values make significance the target, and thus create incentives for underpowered studies, misspecified analyses, early stopping (monitoring significance while collecting data), and p-hacking.

(2) p-values separate crucial pieces of information. It represents a highly specific probability (of the observed data, given the null hypothesis is true), but does not include effect size or a comprehensive estimate of uncertainty. Thus, to be useful, p-values need to be combined with effect sizes and ideally simulations, specification curves, or meta-analyses.

Thus my primary problem with p-values is that they are an incomplete solution that is too easy to use incorrectly. Ultimately, they just don't convey enough information in their single summary. CIs, for example, are just as simple to communicate, but much more informative.

Your points are theoretically correct, and probably the reason why many statisticians still regard p-values and HNST favorably.

But looking at the practical application, in particular the replication crisis, specification curve analysis, de facto power of published studies and many more, we see that there is an immense practical problem and p-values are not making it better.

We need to criticize p-values and NHST hard, not because they cannot be used correctly, but because they are not used correctly (and are arguably hard to use right, see the Gigerenzer paper I linked).

Yeah and it's super simple to roll significance and effect sizes in one with confidence (or credible) intervals.

Plus the interpretation is super straight forward: CI contains zero: Not significant. If nothing else, we should make CIs the primary default instead of p-values.

I agree, and would add lack of statistical power to the list. Underpowered studies (e.g. small effects, noisy measurement, small N) decrease the changes of finding a true effect, paradoxically increasing the risk of false positives.

It's immensely frustrating that we haven't made a lot of progress since Cohen's (1962) paper.

So much has been said about p-values and null hypothesis significance testing (NHST) that one blog post probably won't change anybody's opinion.

But I want to recommend the wonderful paper "the null ritual" [1] by Gerd Gigerenzer et al. It shows that a precise understanding of what a p-value means is extremely rare even among statistics lecturers, and, more interestingly, that there have always been fundamentally different understandings even when p-values were invented, e.g. between Fisher and Pearson.

Beyond that, there was a special issue recently in the american statistician discussing at length the issues of p-values in general and .05 in particular [2].

Personally, I of course feel that null hypothesis testing is often stupid and harmful, because null effects never exist in reality, effects without magnitude are useless, and because the "statelessness" of NHST creates or exacerbates problems such as publication bias and lack of power.

[1] http://library.mpib-berlin.mpg.de/ft/gg/GG_Null_2004.pdf

[2] https://www.tandfonline.com/toc/utas20/73/sup1

That will be a tough order with the projected demographic development, according to which elderly will soon outnumber working population.

Brother apparently has been going down this path as well. I've read that they intentionally degrade print quality with third-party inks and other things [1,2]

[1] https://www.reddit.com/r/printers/comments/s9b2eg/brother_mf...

[2] https://news.ycombinator.com/item?id=31860131

Perhaps I just don't get them, but task groups never really made sense to me.

The whole beauty of async tasks is that you can spawn, retry, and consume them lazily. When you create a task group, you again end up waiting on a single long-running last task, desperately trying to fix individual failures and retries that hold up the entire group.

I agree that especially within the standard context of python and its syntax, async seems weird (because it's sprinkled into an existing paradigm).

The best mental model for me always was to think: Here's an await, that means "interpreter, go and do something else that's currently waiting while this is not done yet."

And that's all about IO, because what you can wait on is essentially IO.

By the way, I really wish there was a better story to executors in async python. To think we still have the same queue/pickle-based multiprocessing to send async tasks to another core is kind of sad. Hoping for 3.12 and beyond there.

[edit] one really neat example that helped me get asyncio was Guido van Rossum's crawler in 500 lines of python [1]. A lot of the syntax is deprecated now, but it's still a great walk-through

[1] http://aosabook.org/en/500L/a-web-crawler-with-asyncio-corou...

async tasks are cool, but the usual PSA applies here:

Be careful to hold your references, because async tasks without active references will be garbage collected. I've been bitten by that in the past.

Long discussion here: https://bugs.python.org/issue21163

Docs: https://docs.python.org/3/library/asyncio-task.html#asyncio....

"Important

Save a reference to the result of this function, to avoid a task disappearing mid-execution. The event loop only keeps weak references to tasks. A task that isn’t referenced elsewhere may get garbage collected at any time, even before it’s done."

> It's like the US had a civil war and the losing side flees to Hawaii and in 2090 we still arguing whether it is part of not of US or should it declare independence.

That actually sounds like a very plausible counterfactual to someone outside the US.

While these facts are important, they sound like Taiwan is unequivocally part of the PRC. But that's probably not what people in Taiwan or many western nations would say.

- Taiwan clearly has the outward appearance of a full-fledged nation

- There are (few) countries that have recognized it as a sovereign nation

- Taiwan maintains some form of diplomatic relations, and Nancy Pelosi's recent visit was a very public sign of that

- Taiwan and the PRC have been at war in the past

Some might take this to say that the official status is contested.

Wrong, ARD, the second largest German PSB has had a very lively comment section for ages [1].

Part of the reluctance in setting up comment sections in general is that media orgs are liable for postings there and manual moderation costs a lot of money, unless you outsource that to underpaid, traumatized semi-slaves as Facebook does.

[1] https://meta.tagesschau.de/

European countries' media systems are very different from the US and your conclusions are not really valid.

The ZDF, one of these PSBs, reaches an average of 5 million people daily on TV, which accounts to >10% of the German TV audience, and it has an online reach of >100 million monthly visits which also makes it one of the top news sources online. Their audience leans towards elderly segments, but it's still a huge chunk of the market, however you may define it.

nicbou's link has details on the how, but here's a direct answer to the why:

The money is not a tax because it's not the government that collects it, and the government cannot decide how much it should be. Instead, there are institutions that have been granted powers to levy their own fees and self-govern to a large degree. Accountability is established through an oversight board that includes representatives from civic society and other independent review institutions.

That's really a lot of ground to cover, but two simple recommendations would be:

- idempotency / consider all state as a cost

- start off with some chaos monkey testing - e.g. establish automatic regular restarts/re-deploys from the start

Both are no silver bullet but ensure that you won't have too much anxiety and can act robustly when something goes wrong. It also has a host of positive downstream effects, such as facilitating setup of new machines and/or scaling.

It's an interesting question, but way too simplistic.

In his great book "the myth of digital democracy", Matthew Hindman pointed out what most people don't get: The web is becoming more centralized AND more spread out at the same time. Popularity is a scale-free distribution that gets more extreme over time.

So we have counterculture, it's just not the large homogenous blobs that we were used to.

Because for many people, proper data handling is a hard requirement without which a product is useless.

In addition, not having an imprint is actually in breach of German law — I'd add that ASAP or you'll get a nastygram from some zealous automated lawyer bot.

- Aggressive marketing of platform defaults, such as Edge on Windows

- Aggressive misuse of incompatibilities, such as video calling for Chrome and Edge (e.g. Teams)

- Loss of websites optimizing for standards and instead optimizing for the monopolist (Chrome). Same as with IE.

Your critique assumes that the goal of scientific writing is to be intelligible to lay people.

In truth, the entire weird and crufty vocabulary is simply a common set of placeholders that makes it easier to grasp research, because the in-group learns to understand them as such.

I agree that random hyperplanes are underpinning both, but as far as I recall, usually only one set of random planes is used in LSH (i.e. one set of trees). The granularity of approximate proximity rests in the number of identical signs, i.e. being on the same side of the plane.

There is another good and more technical explanation (using bands) in chapter 2 of mining massive datasets by Leskovec, Rajaraman and Ullman.

In this year, my base assumption is that everyone's trying, even if it's just trying not to psychologically decompose.

Although you have to admit that we are in a timeline where "funny" has taken on a dramatically different meaning. And since some of the largest corps out there (Google) are not doing any better, I'd feel bad blaming the one single guy who's honestly trying.

Pinboard is arguably the type of service that can withstand the tide of times much better than overblown posh "modern" web apps. But it's a sad truth that even the best things eventually erode and die. I really hope Maciej is doing well and doing other important things, and I hope that others pick up the crucial task of safeguarding the web in a sane way.

As for the rest of us, I stick to Cavafy:

> As someone long prepared for this to happen

> Go firmly to the window. Drink it in

> Exquisite music. Alexandra laughing

> Your first commitments tangible again

> And you who had the honor of her evening

> And by that honor had your own restored –

> Say goodbye to Alexandra leaving;

> Alexandra leaving with her lord

Wild speculation, but perhaps ZFS + postgres + potentially write-sensitive SSDs resulted in write amplification that would just occur again.

I'm usually an AI minimalist, but this level of putting together abstract knowledge blows my mind. Perhaps we're on the brisk of some really crazy things after all. And with that I don't mean crazy effects created from humans (usually) being stupid, but from AI providing meaning cheaply which used to be hard for humans to do.