MarkMc

I love the smart use of incentives this story highlights. But it doesn't go far enough: IRS and SEC agents should be allowed to audit any company with revenue over $5 million per annum - if the audit results in a fine of at least $500,000 then the agent gets a $100,000 bonus. Conversely if the audit doesn't result in a large fine, the agent has his or her bonus pool reduced by $10,000. This would provide a great incentive to weed out people who cheat the system.

So when I attempt to install the Amazon play store on my iPhone, will I get the following scary message?

"WARNING: You are about to install non-approved software which may be malicious or a virus. If this software causes damage to your iPhone we will not fix your device, even if it is still under warranty. Are you really sure you want to continue?"

> Ms Schaede suggests that quota men may simply have been better at the job than marginal women. Male and female teachers differ. Men were more likely to speak Finnish as their mother tongue and to have studied mathematics. Perhaps these or other characteristics helped their teaching.

> In other words, the selection criteria for teachers may have elevated candidates who were less suitable—exactly the outcome critics of quotas fear.

I don't understand this second paragraph. If men are better at the job and the quota elevated men, how can you conclude that "the selection criteria for teachers may have elevated candidates who were less suitable"?

Perhaps Tsar is better?

The strict lockdowns and border closures were state-level policies, not of the federal government which was just voted out

Thanks for the feedback - those a good suggestions!

It's designed to be used with Zoom or live, so the teacher just asks

Twitter is terrible at targeted advertising.

I'm actively looking to buy a house in Sydney for a million dollars, yet Twitter shows me general ads that have no interest to me - like investing in gold or installing a virtual chatbot app.

Twitter should be working with Domain.com.au to show targeted ads to Domain's active user base.

One problem is that a quota system penalizes Asians more than whites. This would imply that Asians have historically benefitted more than whites from a biased university enrollment system. What could possibly cause the (mainly white) college administrators to give Asians preferential treatment over whites?

I'm currently on holiday in Thailand and Twitter has decided to show me ads in Thai (a language I don't speak).

I think it's just more exciting for people at Twitter to work on NFTs than to do boring work like spam control or filtering ads by user language.

Why doesn't the government force Tesla to do a randomised controlled trial? You get in a Tesla and ask it to drive you to your destination. The Tesla flips a digital coin and either agrees to drive you, or tells you to drive yourself. After 10 million such trips you can see who is the better driver - Tesla or human?

The problem is deaths are so infrequent that it takes many years to get a statistically representative sample. Perhaps at-fault collisions per mile would be a better metric?

I actually tried to implement a neutral network from scratch by following 3blue1Browns videos, and using the same handwritten number data set. But I got stumped when I realized I didn't have a clue how to choose the step size in gradient descent, and it's not covered in the videos. Despite that problem I'd say the 3B1B videos are excellent for learning the fundamentals of neural networks.

Consider using Java with the SWT library, which gives you native widgets

McLaren College | Online Teacher | Full-time | REMOTE

McLaren College is an online school designed to identify and train people in poor countries for software engineering careers who otherwise would not have such an opportunity. We do not charge any up-front fees for our course; our fees are generally charged as a percentage of a student's income after they graduate. For more info check out https://mclarencollege.com

If you have a passion for programming and helping others, please consider applying for our teaching assistant role: Send your CV to jobs@mclarencollege.in and let us know how you heard about the role (ie. on HN).

> First Equal Protection clause from the 14th Amendment applies to only states and not federally

The Holmes case was merely the catalyst for my thinking. The point remains: Why isn't there a legal push to have blind sentencing, at least for state crimes?

> Second, It is extremely hard to prove that discrimination happens while all other things are equal.

It's not that hard. You take a sentencing case - say, guilty plea to armed robbery with no prior convictions - and pair it with a similar case where the defendant is a different race (ideally without know the actual races involved). You do that for 1000 cases and then show that the sentencing is generally harsher for one race than the other.

> U.S. follows trial by jury...In a such a system there will be biases, because people are inherently biased

I'm talking about making sentencing less biased. I don't have any suggestions for making the trial less biased.

> The biggest determining factor is Money. Richer you are, less likely to have trouble with law, and get away with a lot.

OK but when it comes to sentencing money should not be a factor - rich and poor should receive a similar sentence for a similar crime.

> Race, age and gender are objectively measurable [3], "attractiveness" is lot more subjective to eyes of the beholder.

Sure, but it's not completely random. If you and I both choose the most attractive person in a group of 10 people, the chance that we both choose the same person is far greater than 10%.

Let's say we ask 100 people to rate defendants by attractiveness and use the average score for each defendant. It's quite possible we would find the gap in sentencing severity between the top-10% and bottom-10% people in terms of attractiveness is greater than the gap between black and white defendants.

But it's really a moot point - if we prevent the sentencing judge from knowing the race and sex of the defendant then the judge will also not know how attractive he or she is.

Someone mentioned that Holmes will likely get a relatively light sentence because she is an attractive white woman.

I've read there is good evidence that black people get harsher sentences than white people and men get harsher sentences than women, other factors being equal. I suspect there is also evidence attractive people get lighter sentences than ugly people.

The current system seems like a clear-cut violation of the US constitution's Equal Protection clause. So why doesn't an organisation like the ACLU take legal action to force all sentencing to be done by a judge who is unaware of the race, sex and attractiveness of the defendant?

Silicone Valley is in LA ;)

I just switched to Duck Duck Go yesterday and was not impressed. When I searched "define hot take" Google gives me a canonical, prominent definition with bold-font title and button to hear the pronunciation. DDG gives me multiple definitions in a normal search results page with none prominent, and no way to hear the pronunciation.

I'll be switching back to Google

It's interesting to consider how an 'asteroid mining' company could be profitable without ever building a rocket: Suppose SpaceX announced they had (a) identified an asteroid with high percentage of gold; and (b) bought a large amount of gold put options with 10-year expiry. If SpaceX could convince the market that it had the technical capability of mining the asteroid, the price of gold would plummet and SpaceX would make a tidy profit. It would also cancel the mining operation because the low price of gold had made the operation uneconomical.

Isn't this the fault of EU regulators? GDPR says that a website cannot degrade the service for users who decline to be tracked, yet almost all these popups make it harder to reject cookies than to accept.

Yes TikTok is a serious threat to YouTube because they have solved the chicken-and-egg problem. Most YouTube competitors don't have enough users to be worthwhile for content creators, but TikTok now has billions of users. Currently TikTok is limited to short, popular videos but if they can branch into long-term videos it will kill YouTube

Bob Dole's passing has reminded me that one of the answers in the greatest crossword puzzle of all time was, "Bob Dole Elected". Worth a read: https://www.dailydot.com/unclick/clinton-1996-election-nyt-c...

No our pay rates are competitive - we have long-term funding in place to cover any shortfall between student fees and operating costs

McLaren College | Online Teacher | Full-time | REMOTE

McLaren College is an online school designed to identify and train people in poor countries for software engineering careers who otherwise would not have such an opportunity. We do not charge any up-front fees for our course; our fees are generally charged as a percentage of a student's income after they graduate. For more info check out https://mclarencollege.com

If you have a passion for programming and helping others, please consider applying for our teaching assistant role: Send your CV to jobs@mclarencollege.in and let us know how you heard about the role (ie. on HN).

Immigrants generally increase their income, and income is inversely correlated with birth rate.

So wouldn't a high level of Australian immigration reduce the total world population, relative to a low immigration level?

I run an online Javascript course and we have what I think is a better solution than ProctorU:

1. We don’t require the student to install any software - we use the video streaming feature of the browser on their laptop to record the student’s screen and webcam

2. We get the student to place their phone camera behind them at the 4 o’clock or 8 o’clock position with a view of the student’s desk, then we record the scene using the browser on the phone

In this way we get 3 video feeds - the screen and two cameras - which we can monitor live and record. You might argue that this is still too creepy, but if we cannot see the student while they work it’s just too easy to cheat and that would make our credentials worthless.

Sorry about that - should be available in all countries now

> The problem with any education is how do you figure out the actual value added?

Answer: with a randomised controlled trial. That is, the school randomly rejects a number of applicants who meet the enrolment criteria, then it follows up after 5 years to see if the accepted students are significantly better off than the rejected candidates (eg. in terms of income).

Austen Allred, the CEO of Lambda School, has already floated such an idea: "I’d love to eventually do an impact study that statistically shows how much more attendees of Lambda School make over the course of their lifetimes." https://twitter.com/Austen/status/1433260773040693257

Sounds to me like you were only wrong in hindsight. At the time it was a reasonable decision to pass on GitLab in the circumstances.