HNHacker News
TopNewBestAskShowJobs

hailwren

704 karma · joined August 14, 2019

submissionscomments
hailwren··on Bitcoin mining becomes unprofitable as BTC price falls to average cost of mining
> Overall, the lighting is still necessary, it's just not as efficient as it could be

Yes, and the completely fixable loss in efficiency has a significantly greater impact on our consumption than the entirety of cryptocurrency yet I have yet to see someone complaining about the perils of unshielded lighting.

hailwren··on Bitcoin mining becomes unprofitable as BTC price falls to average cost of mining
1. Pedantry but long term energy storage isn't really a thing. consumption is more or less supply

> Results from EIA’s 2015 Residential Energy Consumption Survey (RECS) show that televisions and related peripheral devices—such as digital video recorders (DVRs), video game consoles, and streaming devices—consumed 7% of the electricity used in American homes in 2015 [1]

quick edit: I derped the math. I think American residential is 21% so that's ~1.48% of total supply. You'd need a 33% efficiency increase in residential TV efficiency to match the energy saved by abolition of crypto. That's without taking into account that crypto is on average produced from _more_ renewable energy than the average.

1 - https://www.eia.gov/todayinenergy/detail.php?id=37032

2 - https://hbr.org/2021/05/how-much-energy-does-bitcoin-actuall...

hailwren··on Bitcoin mining becomes unprofitable as BTC price falls to average cost of mining
No, it's not outdoor lighting @ 0.85%. It's only the waste from outdoor lighting. Outdoor lighting is far higher. [1]

But now we're playing a game where you tell me the thing you like and the things you don't like and I can either present a counterargument of things I like and don't like and we can both get mad that we have different preferences...

Or, we can design a system that allows people to have different preferences without trying to regulate our preferences on each other.

The math: (residential power %)(lighting %)(lighting % wasted) (21)(.13)(.35) = 0.9555% (oh, it was higher than I remembered)

1 - https://www.darksky.org/light-pollution/energy-waste/#:~:tex....

hailwren··on Bitcoin mining becomes unprofitable as BTC price falls to average cost of mining
The environment claim is total bs. Crypto uses 0.5% of global energy supply. It's beaten by things such as residential television use, and only the wasted light from outdoor lighting (0.85%). It's also the only time that people try to blame an energy consumer for the sins of energy producers.

The environmental damage of energy is caused by coal plants, not how many people are watching the simpsons at the same time.

hailwren··on Rust Is Hard, Or: The Misery of Mainstream Programming
In order for this line of thinking to explain the statistics, either (1) Rust would have had no users over the time period or (2) more than the loved percentage of new users like the language (mean goes up).

It's pretty trivial to see that Rust adoption is increasing, so you can discount (1).

hailwren··on Luna Cryptocurrency Collapse: How UST Broke
Yeah, rocks was a bad example. The gas used by armored cars to move cash around is a good example of similar economic cost.
hailwren··on Virginia police routinely use secret GPS pings to track people’s cell phones
Judicial elections also directly introduce lobbying into the judicial branch. Successful judges now have to pay for their campaigns and more money tends to equal more success in elections.
hailwren··on Language homogenization at Harvard
This article is almost criminally flawed. The author makes some horrible assumptions and presents shoddy data. Even if we assume that the latent space model is "correct" (we shouldn't), they don't present anything like variance or number of samples in a year. Then they sort of arbitrarily fit a line to data which pretty clearly looks non-linear.

Suppose, for example, that The Crimson (not Harvard btw, it's a student newspaper) runs 10x as many articles this year as last. It's possible you're going to get a huge reduction in cosine distance just by virtue of a few authors producing a lot more content.

At a minimum, we need mean, variance, and number of samples. This doesn't tell you anything about "Harvard", it just tells you about the students who the Crimson choose to publish. There are lots of structural reasons within Harvard that the Crimson has probably stopped being a unified voice of the student body -- but again, that's not what the article purports to show.

hailwren··on Heuristics that almost always work
Another good allegory is that, in the US Air Force, the flight crews considered most dangerous are those with the highest collective rank. Sure, the young crews are learning but the old ones still think they know it all and have often forgotten critical details.
hailwren··on Facebook loses users for the first time
Seems like FB has about $55b on hand, but their market cap is ~$800b vs Bytedance's highest valuation of $425b.

Without taking into account fiscal responsibility, it's in the realm of possibility that FB could buy TikTok with credit + options.

hailwren··on Facebook loses users for the first time
Right, if anything -- coffee consumption has only gone down because we've moved on to more efficient stimulants. Either higher concentration/lower cost artificial substitutes or prescription drugs.

Hardly the best example of America getting over an addiction.

hailwren··on Cannabis use produces persistent cognitive impairments: meta review
Woah, it's a meta-analysis. So, you're definitely not going to get a good summary in 10 words.

Early use of cannabis was correlated with worse outcomes, but those who started later in life were still associated with deficits in the studies surveyed.

> To conclude, meta-analytical data on the acute effects of cannabis use on neurocognitive function have shown that cannabis intoxication leads to small to moderate deficits in numerous cognitive domains, most notably executive functions, verbal learning and memory and processing speed. These acute impairments are in accordance with the residual effects that have been documented in several meta-analyses suggesting that the detrimental effects of cannabis persist beyond the period of acute intake.

and

> Additionally, as youths remain particularly susceptible to the effects

Are both in their conclusion and paint a more complete tl;dr.

hailwren··on Cryptoland [video]
To be fair, not a single communique of theirs has led me to believe that English is their first language.
hailwren··on An Algorithm for Passing Programming Interviews (2020)
The trick with O and hash tables is to remember that they're amortized O(1). Imagine a bag with the cost of resizing sitting inside it. Then, spread the cost across all inserts. There are always more inserts than the cost of resizing, so you end up w/ amortized O(1).
hailwren··on “Play-to-Earn” and Bullshit Jobs
There's something fundamentally satisfying to the human psyche about understanding a set of rules and optimizing actions against them. I think it hits that 'feeling smarter than something' nerve.

We've probably lost some potentially fantastic physicists to MMOs.

hailwren··on Photos of a KH-12 Kennan Keyhole Secret Military Spy Satellite (2013)
This is podcast is an advertisement by a man who is actively selling his technology to the US government. It's about as reliable as a founder seeking seed capital for their new AI product.
hailwren··on The Strange, Unfinished Saga of Cyberpunk 2077
Given that CDPR are in Poland, I guess we can say it didn't work?
hailwren··on AI training is outpacing Moore’s Law
You can overfit your data by coming up w/ a model that is very specific to it. Consider an n dimensional data set where one variable is a boolean.

A researcher might notice that the True and False cases are different enough to warrant their own algorithms. Say they now fit a lower dimensional line to the True and False cases.

They've just performed gradient descent but the grad student was the gradient. It's a common joke in CS/ML departments.

(This is an obviously simple example, but the point is that hyperparameter tuning and algorithm selection _are a part of your algorithm_ and the data you're looking at while doing so is part of your training set)

hailwren··on Why asynchronous Rust doesn't work
I haven't used C#, so I can't comment on the IDEs there.

However, it's surprising to me that you can get a better experience than VScode + rust-analyzer. It feels like magic when I use it.

hailwren··on Why asynchronous Rust doesn't work
Are there any good resources for design patterns in Rust? I've gone through Rust by example and am writing mid-size rust programs now but would love to check my work against the experts.
hailwren··on OnlyFans to block sexually explicit videos starting in October
Unfortunately, this would incentivize bad behavior on the part of disreputable merchants. Dark ui patterns would abound in some parts of the internet.
hailwren··on OnlyFans to block sexually explicit videos starting in October
Of course the reason given isn't "I got caught by my wife", it's "I don't know how that charge got there, it must be ... fraud".

Payment processors are required to maintain a certain chargeback rate with the card companies and so that risk gets passed on to the merchants.

This is all exceedingly well documented and a Google search away if you'd like to learn the inner workings of it.

hailwren··on OnlyFans to block sexually explicit videos starting in October
The problem has never been payment providers' morality. Porn payments come with significantly higher chargeback rates than traditional payments. If someone can find a way to stop red-handed husbands from claiming credit card theft when the wife checks the bill, there's a market opportunity.
hailwren··on The Future of Clojure
I argue that Clojure née the functional programming community writ large capture a peculiar local maximum. I used to subscribe to the “functional = better” camp until going to MIT. I challenge anyone to find a cutting edge CS paper* that is written in their favorite functional language. If it were so effective — wouldn’t one academic have exploited this efficiency to catapult themselves ahead of their imperatively bound brethren?

Maybe there’s something peculiar about academic work that requires an imperative mind. This is why we see the highest paid practitioners programming exclusively in functional languages ... oh wait, the opposite is true.

My theory is that FP hits a nice sweet spot for someone who would like to casually improve their craft. Something akin to breaking out Popular Mechanics after work.

* - excluding, of course, papers about programming languages themselves

hailwren··on About 150 U.S. Cadillac Dealers to Exit Brand, Rather Than Sell Electric Cars
Yes, these are common. However, it's important to ask yourself why they're common. They're usually chosen because of cheap real estate and car buying being a destination event. i.e. we buy cars infrequently enough that we value ability to choose among more options more highly than the time it takes us to arrive at a destination. This is the opposite of how we consume fuel. How often do you drive out of your way for a lower gas price? I almost never do.
hailwren··on AlphaFold: a solution to a 50-year-old grand challenge in biology
There are two threads here. The first is that it would not be surprising to learn that describing the way that proteins fold is a very hard thing for humans to understand. See i.e. 4CT [1] and its computational proofs.

The second is that explainability in ML is much more tractable than it was 10 years ago. This is not to say that it's solved, but having solved the predictive problem -- I would expect model simplifications and SME research to proceed more quickly towards understanding the how now. I did some work w/ an Astrophysics postdoc using beta-VAEs [2] to classify astronomical observations, and simplifying models in order to achieve human-explainability proved to not cost as much predictive power as you might expect. It might be that the same holds true here.

1- https://mathworld.wolfram.com/Four-ColorTheorem.html

2 - https://paperswithcode.com/method/beta-vae

hailwren··on Psilocybin and migraine: First of its kind trial reports promising results
All the replies are offering advice, but the one piece of advice I wish someone had told me at the beginning of my migraine journey was to not stop trying to fix them. There is an insanely long list of drugs that "work" but will probably have middling effects for her (triptans, tryptalines -- yes they're all different, sertraline, hydroxyzine, meloxicam, the newer CGRP injections, RF ablation, nerve blocks, massage, chiropracty, acupuncture, MRIs, botox injections, spinal taps, these have all been prescribed to me by various Neurologists).

The only real solution is to keep moving forward and keep hounding your doctor to keep you moving.

hailwren··on Moderna Covid vaccine candidate almost 95% effective, trials show
Is it possible to spread a virus without an infection? It seems like other than being a surface, any ability to spread would be predicated on also testing positive for the virus -- wouldn't it?
hailwren··on The Prestige Trap: finance, big tech, and consulting
I needed to rewrite this. I mean something like if I select an individual, uniformly at random, after conditioning on industry and Harvard degree I would expect the salaries of most of the individuals in both industries to be about the same. However, I think tech is closer to a normal distribution and finance is heavy-tailed around tech's mean.

If you do something like conditioning on a CS degree, finance probably mops the floor w/ tech.

hailwren··on The Prestige Trap: finance, big tech, and consulting
(for Harvard grads) is the important part there.
← PreviousPage 4 of 6Next →