roland_nilsson

Can't we use Claude to figure this out

Except that note-taking systems are meant to help you organize your own mind and understand the world better. Offloading tasks to AI won't help you with that.

Much needed, as we only had 450,000 apps to choose from before. /s

Well this is a problem worthy of an Ig Nobel.

"I can do anything I want. It's just bits. You don't own me." Fair enough! x-D

I think this problem is at least in part due to the hypothesis testing concept itself. Classical hypothesis testing is asymmetric: there is a "null" hypothesis, which is typically the uninteresting/useless case, and an "alternative" hypothesis, which is the one you would like to be true. Critically, you cannot determine if the data _supports_ the null hypothesis, only if the data _rejects_ it (and supports the alternative). A so-called "null result" occurs when data is not sufficient to reject the null hypothesis; then you can't tell if you actually have a useful finding (for example, that there is no major difference between species A and B) or a failed experiment (data was so bad/noisy that we cannot conclude anything). And so you end up with the unfortunate situation where you either succeed in proving your favorite hypothesis and get your degree / promotion / tenure, or you have nothing.

This happens because hypothesis testing conflates effect size (how big is the difference between A and B) with uncertainty about that effect size (significance/reproducibility). Confidence intervals are more useful IMHO, as they help untangle these two aspects, for example showing that the difference between A and B is small _and_ reproducible. Bayesian analysis is also a major improvement, as it allows examining both the "null" and "alternative" hypotheses on equal terms, as well as reasoning about our prior beliefs / biases. Unfortunately many areas of science are still stuck with statistical methods from the early 1900's.

So he proved his point! :)

Reading this I couldn't help thinking of the epicycle theory of planetary motion. Under the holy assumption that planets had to move in perfect circles, people invented increasingly complicated circles-on-top-of-circles models in order to explain all observed trajectories. Then Kepler came along and said "hey, it's an ellipse!"

Second that! (No pun intended.) Dunno how many times I have had to explain to people that the kWh is an amount, not a rate.

The yoctomole is a wonderful unit, for example when you want just slightly more than half a cup of coffee!

But the mole as a "counting unit" is different from OP's idea of assigning units to things like network requests. The mole is just a shorthand for a number, like a dozen or a score. We don't have different kind of "moles" for, say, carbon atoms and water molecules. Or coffee.

Just a reminder that "70% higher risk" is a relative value describing the _fractional_ increase in risk. In absolute terms, the probability of developing Parkinson's (prevalence) was 0.33% in group exposed to TCE and 0.21% in the non-exposed group. So you might also say TCE increases the risk of Parkinson by 0.12 percentage points.

Yuval Noah Harari argues that AI has hacked the operating system of human civilisation.

Storytelling computers will change the course of human history, says the historian and philosopher.

'Like bro, you want to do stuff with “diffusion models”? You don’t even know how to add two normal distributions together! You ain’t diffusing shit!'

This made me laugh :) I'm in academia and this post does remind me of certain students...

From my location in Sweden, sputniknews.com is currently available but www.rt.com is not (error 403). Probably sputnik will go away as well, as the EU ban on these outlets is "expected to cover all means of distribution or transmission, including internet video sharing platforms and applications" https://techcrunch.com/2022/02/28/eu-rt-ban-extends-online/

This seems like a boneheaded move. What credibility will the EU have in criticizing media censorship in Russia when they engage in censorship themselves?

I tried to pay attention to this piece, but the website kept trying to open a pop-up every half minute. After five or so interruptions, I gave up :-/

Increasing automation to push prices might keep the factories within the country, but it won't save manufacturing jobs.

Well I'm no economist, but this seems inevitable if (1) consumers want cheap products and (2) production is cheaper elsewhere. Given those conditions, any company that wants to survive will have to move production abroad (and those that do not will quickly be replaced).

So if we want to keep domestic production, either consumers need to voluntarily choose more expensive alternatives -- which seems unlikely to happen -- or we have to put tariffs or other restrictions on import of goods. The EU does this for a variety of agricultural products to protect farming, for example.

Or is there another way out?

Isn't the whole point of graph databases that they can traverse graph edges efficiently by following pointers to nodes, which relational databases can't do? Then it seems a bit strange to implement a graph database on top of a relational database like SQLite?

No, I think you must estimate p only using the cohort where people were not treated, otherwise you will underestimate the population fraction. In order to test if the null hypothesis is true, we can't assume that's its true when constructing the test.

Well, given that 90/15000 = 0.6% in the control group developed the disease, you can consider the vaccine group as a Bernouili trial of n = 15000 with probability p = 0.6%. Then the probability of observing 5 or fewer cases is 3.4*10^-32, from the tail probability of the binomial distribution.

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

Of course, that's assuming that five guys from the vaccine group didn't get infected at the same after-ski party, or any funny business that violates statistical independence ...