zug_zug

Well, it looks to me like they look at 34 types of cancer, and found that 8 of those were increasing.

So "we detect cancer better" isn't really a good theory, unless it just so happens we only have improved our detection for 8 of those cancers.

Would love for somebody more knowledgeable who read the whole paper to weigh-in.

Yeah but I feel like you have to admit there is a tragedy of the commons here -- there will always be more people who want to live somewhere than can realistically have a high-quality of life there....

Part of the reason San Francisco is so desirable is because it's iconic, it would be a great irony to make room for more people by removing what makes it special.

It always felt to me like if we want more San Francisco then we should build more beautiful cities, there's a lot of amazing coast in California.

> but it does show that parental 'hyperactivity detectors' are broken

I'll grant that some are for sure. But it doesn't mean other parents can't quietly and effectively notice their children's responses to various foods.

Frankly any sort of research on behaviors imo should eventually move to scoring by completely objective measures (e.g. computer assessment of whether they can watch a screen for 30 seconds and how often they look away, or whether they can read a paragraph)

I think it's only for short-term psychological measures

Here's a study highlighting the psychological effects of sugar

https://www.cambridge.org/core/journals/british-journal-of-n...

Well if there is plenty of evidence, perhaps it should be linked here rather than this relatively weak meta-study that confuses the absence of evidence with the evidence of absence.

Two things can be true.

No, that's not it at all.

What you would do is you would have a higher standard of proof. If you conduct a study with a 95% confidence interval that sugar doesn't increase excitement more than 5% then, you can say with 95% confidence the mean falls above that 5% threshold.

However in this study, that confidence intervals were all over the place (from huge effect on mood to negative effect on mood) which generally means you need more participants to conclude anything.

Sorry what? This is just out of the realm of plausibility.

There's a reason why academics no longer use sugar-pills as placebo, because sugar has immediate psychological effects.

-------

Edit: Just read it, the blog author doesn't understand statistics.

"—Sixteen reports met the inclusion criteria for a total of 23 within-subject design studies. The weighted mean effect size and related statistics for each of the 14 measurement constructs revealed that although the range for these means was from -0.14 for direct observations and up to +0.30 for academic tests, the 95% confidence interval for all 14 mean effect sizes included 0."

So there's a world of difference between "The studies didn't prove X" and "The studies proved X is not true."

To illustrate - I could do a study right now on myself about whether getting hit on the head reduces IQ, and odds are, I wouldn't have enough data to prove anything. Doesn't mean I proved the opposite.

Also, if the article wasn't paywalled I'd be curious how many grams of sugar were given etc, but probably not even worth my time. I also think that a placebo of artificial sweetener is dubious unless we are sure those can't cause behavior changes too.

My advice on giving advice:

Step 0 - Ask "Would you like my advice?" Anything other than a sincere yes (not a maybe, not a patient okay, not a hmm) then don't give advice.

Here's my theory on why we see internet toxicity --

A lot of people think they are offering "free information" on the internet, but what they are really doing is offering a transaction: "I'll tell you what I think but in a manner where I present myself as higher status than you." [There are the exceptions, particularly here on HN, where somebody is infinitely humble and patient of course, especially on certain topics]

You'll meet a lot of people outside the internet who are interested in doing this too "Let me overexplain X to you" (e.g. Mansplaining). Which conveys again two things -- "free" information, but actually on the condition of presenting oneself as "the teacher."

I think on the internet, where there are 0 barriers to entry, ANYBODY can post an offer, no matter who bad it is (let me peddle a stereotype as brilliance and see who bites).

Someday soon, if we want, we can have a version of HN where every single comment is rewritten by GPT4 to be friendly/positive/deferential but equally informative.

I'm confused. I consider a d12 a larger die than a d6. Perhaps you're making a pun about physical size of the dice?

To be fair, I didn't hand-check the answer it gave (and I didn't retype the whole prompt exactly here) - but here's what it gave me [4o model]:

... (lots of calculations)

Final Comparison

    Variance of 20d6: 58.33458.334
    Variance of 1d20+1d12+53: 45.166745.1667

The variance of 1d20+1d12+53 is closer to 58.334 than previous combinations and represents a reasonable approximation for both mean and variance.

[Edit: Just checked it in google sheets, this looks right to me]

Well are the "people" responsible, or are outrageous-prices (e.g. one subscription costs 50k) responsible?

Or is it just like 1% of people have cancer?

Would love to see a good-faith analysis by somebody who has 50 hours to commit to it, a good grasp of statistics, and no interest in politics.

I'm sorry but this falls flat for me. GPT4 routinely can answer impressive math questions for me (college-level):

- What diameter steel wire would I need to be rated for a weight of 500lbs?

- How many digits would a ID need to be (using 36 characters) to have a 1/10^20 chance of collision over 1 billion random IDs?

- If I have a list of a million times (say durations of a web request) and they follow a normal distribution, and I take a sample of 1 million of those, how close would the average of my .1% sample be to the true average of the billion?

- Suppose in D&D I am told to roll 20 d6, but instead of rolling that many dice I want to roll just two (larger) dice and add a constant. Which standard D&D dice might give the closest variance and what is the constant?

I wonder if these poorly-conducted studies are properly reporting any sources of funding an conflict of interest, as well as how they get past peer-review if they have relatively basic methodological flaws.

I'm not GP, but I've played a few adventure games and was very excited to try Monkey Island because of its sense of humor and style. But it felt basically unplayable to me (basically it felt like the only strategy was brute force), here's the puzzle that made me give up:

Give the jawbreaker to Blondebeard to loosen his gold tooth, then give him the pack of gum. When he blows the bubble with the tooth in it, pop it with the pin and the tooth will drop out. Get the gold tooth, then exit the shop. (To get the gold tooth on Mega-monkey, chew the pack of gum to get the chewed gum. Inhale from the helium balloon, use the gold tooth in the chewed gum and then chew the tooth in gum. It will float out the window of the shop and land in the mudpuddle outside. Use the pie pan in the mudpuddle to get the tooth).

Frankly I'm not very convinced by that -- it seems that most social networks are already killed by winner-takes-all SV companies buying-out any possible threat.

The situation in America is an absolute nightmare with data being sold through all sorts of mechanisms with zero oversight (your ISP, your car, most apps on your phone, whoever does your credit scores) -- I'm sure there's some way to make successful tech that isn't the hellscape America has.

Is that true though? Are there any 5-person startups with < 10k users getting sued/shutdown by this regulation?

I feel like if you really think there's a piece of evidence that you think needs to be shared you should link to some of the published journal articles tho book links to rather than trying to sell a product.

A blog post won't be indexed on google scholar for other academics to reference though. Maybe a preprint would be?

In my opinion the goal is to get a record of the information and the dataset out there.

shrug I meet that checklist, but of course I'd only be interested in a non-technical cofounder who passed a similarly comprehensive checklist (e.g. proven career in sales, major connections with potential clients, strong understanding of market position and competitors, already has created mockups that people are interested in purchasing).

For me it's interesting -- I'm somebody who has in the past made the niceties when somebody reached about but ultimately not reciprocated much. And it wasn't because I didn't care, it was because never in my life had somebody reaching out from the past resulted in genuinely satisfying interactions. So in a certain fatalistic way, I had decided it wouldn't amount to anything which was then possibly self-fulfilling.

I used to believe that whenever coworkers ask how to stay in touch after I leave a job it was just a lie we all tell ourselves to make the transition easier. But I'm finding that many of my coworkers say they do stay in touch with old colleagues.

Perhaps part of the lesson is that some of us who don't reciprocate with you do care, but have no mental model (or lack the social grace) to navigate a more tricky relationship (e.g. what if work is now much better now that this person left, and what if the people and things you had in common no longer apply).

I'm glad to have read it here, it's newsworthy to me, but I don't think the discussion would be worthy. Maybe it'd be nice to have those two be disconnected.

Philosophy mumbo-jumbo. Consciousness is not a scientifically meaningful term if it is not defined in a falsifiable way.

>> You can't take days testing gradual roll outs for this type of content, because that's how long customers are left unprotected by that content.

If you can't take days to do it then do a gradual rollout in hours. It's not a high bar.

You say that, but if you're in a 10-day survival situation you'd happily have olive oil by day 3.

Maybe if you're planning for 3+ months off the grid it becomes a bad idea, but I don't think there are many scenarios where we're off the grid for 3+ months and then society comes back in any meaningful way (e.g. more than 1% of population alive, electricity again).

I dunno this seems stupid to me. So 10 days worth of emergency food at $8 per meal?

Like olive oil is safe to eat for years, costs $5 / meal if it's all you eat and you get the fancy kind, and if the world doesn't end you can just use it and buy another backup bottle.

It's really scary how a tiny amount of a mineral can permanently damage the human mind. I have to admit there's a part of me that almost wants to be skeptical, that mg of a rock could overcome the wonderous self-aware human mind. Every so often I have to remind that skeptical piece of me that it's an idiot and that every human alive is wildly fragile to any number of elements and chemicals.

------

Onto the article, if I ran a production software system, I'd be collecting data on the health of all my servers.

If I ran a country, I'd collect health data on all my towns (perhaps blood tests of 1% of the population each year).

Of course in malfunctioning environments, measurement is scary to leaders because it can give evidence that they aren't doing their job.

> Why do people think this writing style is attractive?

Do you actually want to understand the answer to this question?

This actually, according to hard economics, is potentially an optimal solution.

Fossil fuels provide a great benefit to one party but produce negative externalities to other parties (people in hurricane zones, people born into hot regions, people who live in the future). If it's worth it economically because the value is so great, then an optimal economic solution is to say "You can have this thing, but you are accountable for the damages it causes to other parties"

Once we start pricing in the costs of the side-effects of fossil-fuels, the tradeoffs will be more clear, and the market will create faster incentives toward the ideal tradeoff.