rm999

This feels different to me. Apple released a detailed, concrete ten year roadmap that other companies can follow - this already indicates serious commitment. There are enough in-flight initiatives that if carbon neutrality is deprioritized it will have to be an explicit decision from high-up in the company. Also, this is Apple we're talking about, not Google :P

It’s a nonstandard definition at best. While regression is a name given to some algorithms (linear and logistic for example), this is an artifact of the statistical foundations of ML. Its standard use in ML is to describe the problem, not the solution: regression is predicting a continuous value and classification is predicting discrete categories.

Either use is ok, but e.g. a tree model predicting if an animal is a dog or a cat is not regression by any definition.

>It's factually untrue that the Grand Caravan went up in price when adjusted for inflation.

He explicitly says that right before? "4/ For example, our inflation-adjusted data say car prices have not increased since the mid-1990s."

I believe his argument is more around how we're calculating inflation, i.e. that prices have actually gone up more than economists claim through inflation. Looking at inflation adjusted numbers in the fifth thread would lead to circular reasoning and is besides his point. That said I am not great at dissecting these kinds of twitter threads and may be misinterpreting too; I agree that the way he is arguing it is confusing.

Really happy to hear this. As someone who went to Hopkins I would fully encourage my children to apply (I had a great time and got an amazing education). But the idea that I spent a few years of my life there 40 years ago feels irrelevant to the admissions process. Any advantage given would be counterproductive in several ways:

1. It would coddle the child at a time in their life when it's essential to learn independence. This can't be good for the values they are developing.

2. It has no correlation to their success at the school. Getting admitted to a tough school like Hopkins if they are not prepared would be incredibly stressful and would probably be negatively correlated to their future success. I saw a few students who got special admissions crash and burn their freshman year and transfer out.

3. The article goes into this a lot, but it's not good for equality and encourages stratification. This isn't good for society. And it decreases diversity which is probably bad for the university.

>I estimate generously that 50,000 of them are economically productive in the sense of producing something of value that can be 'exported' from this city.

I'm confused, are you saying that services don't have value? It feels like an artificial distinction to me, for example watching a movie at a theater vs buying/renting a movie both provide me similar value. Also, in a world of increasing automation and globalization # of people is perhaps not a great metric.

I'm not really sure if the economy is in as good of shape as it seems. At least here in the USA we're hitting an unprecedented period of economic growth; the Keynesian in me believes that policy has prevented at least one recession in the 2010s and led to real growth in the economy, but it's very possible we're just in for a bigger fall in the 2020s because of it.

NYC has learned the hard way that there should be some preservation; destroying/modifying a building is a one-way act, and over a long enough time horizon someone will eventually make a bad decision that everyone regrets later. See for example Penn Station: a beautiful landmark that was demolished by a real estate company because they deemed it too expensive to maintain. In the act they created a horrible black mark on the city.

There are arguments to be made on both sides, this isn't a black and white issue. The history of preservation is fascinating. Especially after Penn Station's demolition, NYC went through a big debate on other landmarks like Grand Central: https://www.smithsonianmag.com/history/preservation-battle-g....

> While few doubted the significance of Grand Central, the terminal’s owners took issue with the law itself—how, they wondered, could it constitute anything other than an unreasonable violation of their rights as property owners? Preservationists like Onassis, working with groups like the Municipal Art Society, continued to insist that saving Grand Central and buildings like it wasn’t a mere real estate matter, but an issue of public good. On June 26, 1978, the United States Supreme Court agreed with them in Penn Central Transportation Co. vs. New York City, not just in regards to Grand Central but in the spirit of the Landmarks law itself...

In retrospect I don't think you could find a New Yorker who regrets that Grand Central was saved after seeing what Penn Station became.

>Computing has always centered on data

Sure, well-thought out data structures were an integral part of good clean code, but the bulk of data was an artifact of software in the form of logs and largely unused database records, not the "center" of anything. Today, there are many people who take these logs and database records and run very specialized code on them. I entered industry in the mid 2000s and it was very difficult to find jobs that concentrated on this component back then. We're often called data scientists today, and there's orders of magnitude more attention that goes into it. This is a real shift that has brought data into the "center" of many organizations.

> A third big name in the field, the roboticist Pieter Abbeel, made $425,000, though he did not join until June 2016, after taking a leave from his job as a professor at the University of California, Berkeley. Those figures all include signing bonuses.

425k after signing bonus is probably closer to 300-350k base + annual bonus, which feels low for someone in such a distinguished position (close to a typical specialized physician or big tech software engineer with 5-10 years of experience). After finishing my masters degree in machine learning in the mid-2000s, I immediately got job offers higher than my much more talented professors. To me, the real story is that academia has set the compensation bar low for people adding so much value.

Edit: to be clear I know these salaries are high and people in my field are lucky to be making so much. But I also don’t think it’s accurate that people in the field are overpaid, which titles like OP may make it seem.

Agreed, I went through a similar flow.

Direct link to the image: https://upload.wikimedia.org/wikipedia/commons/thumb/e/ef/Pr...

Article: https://en.wikipedia.org/wiki/TOP500

Non-WSJ link: https://www.cnbc.com/2019/09/14/saudi-arabia-is-shutting-dow...

>in data analysis/science, there's more than way to highlight/analyze a problem, and more than one way to solve it. In my experience, the best solutions are often the craziest.

I strongly disagree. I have hired, mentored, and/or managed 20+ data scientists and analysts over my career; in my experience the ones who come up with "crazy"/creative solutions are often the least effective in the long run. This is especially true on the product analytics side, where there are 5 ways to shoot yourself in the foot in non-obvious and subtle ways. The best analysts I've worked with follow fairly regimented best practices around the way they analyze problems and present solutions. By its very nature, non-standard solutions are more error-prone and harder to explain, which breaks institutional trust in the analyst.

That said, there is plenty of room for analysts to grow into more creative fields like machine learning or product management, but I don't think the person being interviewed falls in one of these categories.

>I'm getting pretty tired of this story. I read an article a few weeks ago that mentioned CreditKarma has free tools for tax prep, so I signed up for it and I'm planning on using it next year.

This is exactly what I did this year and it went incredibly smoothly. It felt a lot like Turbotax, but Credit Karma was noticeably faster without all the confusing upsells (I did my and my wife's taxes in ~30 minutes, including three W2s and investment income from two banks; almost entirely within the iphone app).

Personally I think the writing is on the wall for Turbotax et al. Credit Karma got much better from year 1 to 2 and I predict it will be increasingly recommended in future years, starting a race to the bottom in the industry.

I know a lot of senior data people (read: high demand tech people) applying there right now. A few things:

1. People who work there are generally happy. Interesting problems to work on, a company where tech is respected, etc.

2. A lot of those issues you list are common in big tech, see Google. The people I know who are looking at Facebook may want to work at a bigger, more stable company.

3. As everyone else mentions, $$$. That said, Finance also pays similar rates but I don't see a lot of people fighting for those roles, so that doesn't fully explain it IMO.

Most of the issues discussed in the article are not from limits in precision, they're from overflows that arise from limits in the exponential range. Finding the average of 8e+307 and 8e+307 is a "low precision" problem, and even naive methods to avoid overflow will not hit limits on precision in this case (e.g. 8e+307/2 + 8e+307/2).

You're right that there are issues with precision and floating point math (the article touches on this a little), but the implementation is definitely wrong if it promises to work like in the case of Python's statistics library (which was correctly using the fractions library to store arbitrary numbers, but converted to a fraction in the wrong order).

I moved to NYC from a tier 2 city and found the expectations tend to be about the same (~40 hours). That said, a city like NYC can attract and retain type-A people who work long hours, so I think on average NYCers work longer hours than my previous city.

> Do higher-salaried jobs require more hours

No, but moving up the management ladder can require more responsibility and hours.

Not to disagree because you're totally correct, but the winning model used GBMs (a non-linear model) and a bunch of subtracted correlated features to great success. When using GBMs, subtracting two features can add a lot of value.

https://storage.googleapis.com/kaggle-forum-message-attachme...

>One competitor found that the difference between two of the features, specifically f527 — f528

I read through the thread where f527 — f528 was discovered, and mostly the finding was that those two features alone were very predictive. Further research found that the two features are highly correlated, but negative of each other.

See the write-up of the winner here: https://storage.googleapis.com/kaggle-forum-message-attachme...

>Yasser Tabandeh posted a pair of features, f527 and f528, which can be used to achieve a very high classification accuracy ([1]). Furthermore f527 and f528 are very highly correlated. However, it is not needed to keep both features. Their difference f527 − f528 contains the same amount of information

He also explains his method of finding other pairs of "golden" features in a 2-step iterative process.

Original thread: https://www.kaggle.com/c/loan-default-prediction/discussion/...

I've used Segment at two companies in a row (in both cases they were implemented at early stages before I joined). It's a good product with broken pricing. They charge insane amounts of money for add-ons and seem to get exponentially more expensive as you grow. My recent renegotiation with them was the most painful experience with a vendor in my 12 year career.

I want to like Segment, I really do. But until they change their attitude I strongly recommend against using them. Especially for startups, because it's not easy to get off once you deeply integrate Segment into your stack.

I'm not a recruiter, but I spend a lot of my time building up teams so I figure that makes me a de facto recruiter.

I've personally gone both extremes:

1. Spamming linkedin, message boards, etc.

2. Seeking out specific candidates, emailing them, offering to grab coffee or do a call.

I've had more success hiring with #2, in which regard I agree with the article. But it comes with a perverse side-effect: I am ignoring 95% of potential candidates, likely a similar group everyone else is ignoring. I am mostly contacting people with years of experience at good companies with a decent web presence. Because if I didn't, I wouldn't have time to do my other job duties. When I (or our internal technical recruiters) spam out a job we are giving more people a chance at the job, and are introducing fewer personal biases.

I'm not saying either approach is right or wrong, I just think there's room for both. Maybe the ultimate solution is a single centralized marketplace for jobs, but it doesn't seem like there's been a clear winner in this space.

Ah whoops you're right! At 8:39 I saw a mention of a scanning electron microscope and thought they were saying they used it in the analysis. They are actually saying one would be needed for that particular sample.

Cool video, but what a disappointingly inaccurate title.

Spoiler alert: they spent days finding metallic dust, cleaning the dust, and sifting through the samples with a microscope. They shipped ~15-20 interesting samples to Europe where an expert with an electron microscope analyzed each sample. The expert presumably spent a lot of time analyzing each sample on his expensive equipment and found zero meteorites.

The title insinuates it's easy to find meteorites with no specialized equipment, but everything about the video contradicts this.

>Any job that has included in the listing for job responsibilities: "A passion for what you do", I have not applied to.

You're mistaking "passion for what you do" with "passion for a job". I've worked with a lot of people who love their field and I've never seen a correlation with hours worked. If anything the people closest to burnout are always complaining about their job and fantasize about leaving their field entirely. Those people are not passionate for what they do.

Pretty sure yes; after Safari started blocking third party cookies by default a few years ago, Google, Facebook, and Microsoft have all started supporting the option to use first party cookies to get around this. As I understand it, they deploy code that stores tracking information on-site using first party cookies, then access that data directly.

https://digiday.com/marketing/wtf-what-are-facebooks-first-p...

https://searchengineland.com/google-analytics-adwords-respon...

https://searchengineland.com/bing-ads-apple-intelligent-trac...

Someone please correct me if my understanding of this is incorrect, I've been out the ad world for awhile (thankfully!)

This comment is skewing what LIC is. LIC is a neighborhood in its infancy because it was rezoned from primarily industrial to residential/commercial only a few years ago. But it's basically the perfect neighborhood for the city to grow into. In defense of the neighborhood I lived in for years:

> There's very little in terms of culture

There are literally 100s of artists who work out of LIC. Many of those "empty warehouses" are actually art studios full of spaces rented out by artists. It's the densest concentration of artists in the outer boroughs, and probably beats out 90% of Manhattan too. Check out the LIC art walk, or the LIC art open, or PS1. I agree with the lack of basic amenities (see: neighborhood in infancy), but lack of culture is an absurd accusation.

> Realistically it's the densely overcrowded 7 train, getting worse each year, with no relief in sight

Why not mention the E, N, W lines, which feed a lot of Queens and more of Manhattan than the 7? LIC (especially Court Square) is a convenient neighborhood to commute into, if it's nothing else. Also, LGA and JFK are both straight shots on the highway (15 and 30 minutes respectively) and JFK is easy by subway+airtrain. This beats out most of the rest of the city, any other criticism should be pointed towards the placement of the airports.

I believe they are referring to the number of transistors per chip (which as you correctly state are not "moving parts"). Also, fun fact: these chips are 7+ years old now!

https://ieeexplore.ieee.org/document/6330636

Yep exactly. Especially for smaller, faster growing companies: what you choose to not work on is usually more important than what you choose to work on (once you remove the core work to support the company's current state). Without someone with product experience - usually either a PM or an experienced software leader - adding additional engineers often paradoxically hurts total productivity.

I'm sure many have seen this, but this was clarifying for me: https://i.imgur.com/1zUUPGh.jpg

Yes, the notch introduced unfortunate design headaches, but it maximizes utility given current technology. I think if you took those pixels away, most existing iphone X users would object.

Full dataset is here: https://www.scribd.com/document/388081338/BANKRATE-Most-Valu...

For those of you who are wondering, Computer Engineering is 14th, "Mathematics and Computer Science" is 23rd and Computer Science is 30th. If I had to guess, CS has a wider income distribution than many of the higher ranking degrees, which are dominated by engineering degrees (many of which have lower unemployment rates too). The non-STEM degrees that beat out CS (assuming we don't count social sciences) are Finance, Public Policy, Economics, and Agricultural Economics. Cool dataset!

T-mobile is horrible with security. They have this service called Digits, which lets users access their phone number from other devices. The problem is, this subverts the security model a lot of the American internet ecosystem is built on, i.e. your phone number will be secure online. Someone got into my t-mobile account and enabled Digits, then had free-reign on my gmail, texts, paypal, etc (I don't use the strongest passwords, but always keep 2-factor authentication on important services like these). Every time I got hacked I'd spend a couple hours of trying to figure out how, and the T-mobile agents would always claim no one accessed my number. Finally, the third time, I escalated up the security chain within T-mobile and they figured out the issue.

https://www.t-mobile.com/offers/t-mobile-digits

Yeah they aren't quite comparable; I didn't want to dig into the complexities of that, which is why I followed the last thought with "incidentally".