nelsondev

We had this problem of poorly phrased tickets, and we added a rule to our ticketing software to immediately respond with a list of questions they need to answer to prove they did due diligence before we would start answering. It worked, we got less tickets, and the ones that answered the questions we could resolve faster.

There may be some who are doing this deliberately based on the reasons you mentioned; but for me personally, and my guess is for others, I did it out of ignorance of the frustration it gave the other person. Once I had awareness, I could change my behavior. So the article is useful.

It’s about diminishing returns.

E.g. getting a model from 0% accuracy to 70% accuracy might be a couple Pytorch library calls that any dummy who watched Andrew Ng’s course can do. But getting that same model from 70% to 75% accuracy might be deeply mathematical and require the latest and greatest mathmeticians, statisticians.

But in this hypothetical example, an engineer who stands up the 70% model and keeps it running 99.9% of the time, with high uptime; is more valuable to the bottom line of the business than the 75% accuracy model hacked together with scripts with 50% uptime.

The author (OP) proposes community generated and shared block lists.

I’m imagining an opt in block list, where with enough downvotes, if you are a member of the block list, the content is hidden.

If the block list starts blocking content you want, you can fork it, keep the parts you want, remove the parts you don’t.

On that note, what specifically from Trump’s speech was inciting violence? “March” != “Riot”

This is the section about marching to the Capitol.

“ And we're going to cheer on our brave senators and congressmen and women and we're probably not going to be cheering so much for some of them.

Because you'll never take back our country with weakness. You have to show strength and you have to be strong.

We have come to demand that Congress do the right thing and only count the electors who have been lawfully slated. Lawfully slated.

I know that everyone here will soon be marching over to the Capitol building to peacefully and patriotically make your voices heard. Today, we will see whether Republicans stand strong for integrity of our elections. But whether or not they stand strong for our country, our country. Our country has been under siege for a long time. Far longer than this four year period.”

“

Full text of his speech (rambling and painful to read, but there was no advocating for violence: https://abcnews.go.com/Politics/trump-told-supporters-storme...

From the below article, Parler was setting up moderation policies, and organizing volunteer moderators.

And from the AWS letter, it was Parler removed some of the offending posts, but not all of them, and not quickly enough for Amazon.

https://www.theverge.com/2021/1/9/22221730/apple-removes-sus...

I’m still looking for the original source of Parler’s response to Apple.

Yes, you’re very right. This is a censorship in action, due to current politics.

Twitter, Facebook, Big Tech at large, are proud to stand up for free speech against governments abroad (Arab Spring), but here in the US, free speech against government results in deplatforming.

Imagine there exists a perfect AI, that can tag all tweets as “true, somewhat true, somewhat false, false”, I think it would be in the public interest to allow all tweets, but tag them accordingly.

We have laws against defamation, slander, incitement of violence; but no laws against lies. And I don’t think we should.

Giving a poor person a job is not “exploiting poverty”, it’s alleviating poverty.

Why is outsourcing unethical?

As a software engineer, you have a professional responsibility to stay abreast of current technology. That doesn’t mean hopping on every latest JS fad, but it does mean continually looking for more information, and learning how others in the field are solving problems.

Physicians have to keep up their board certifications/licenses by undergoing yearly(?) retraining, in order to continue practicing medicine.

It’s not that hard to find what other software engineers are doing, to start, Read hacker news! Talk to more experienced engineers, read blogs, as an intellectual exercise, surf through AWS services and think about how you could use each.

From scanning through the course outline, the big omission is learning about distributed systems, and industry leading implementations of common tools.

At big tech companies, you typically have to work at a much larger scale. A solution that works, won’t necessarily work at 10x, 100x, 1000x the scale. You often need to use technologies architected differently than a monolithic Web App.

I would recommend learning more about the commonly used “tools” that a distributed systems engineer would use, more specifically:

- Load balancer - Nginx

- Key/Value storage - Redis

- Pub/sub - Kafka

- Queueing - SQS

- Map/reduce - Hadoop, EMR

- Databases - SQL

Denver, Las Vegas, Reno, Boulder, Albuquerque, Phoenix, Tucson, Santa Fe...

Couple comments:

1. This assumes it’s impossible to disincentivize drug usage via really harsh penalties enforced by state power. But look at China, where they have very little drug usage, due to extremely harsh penalties, and extreme social stigma. Individuals don’t get “mind expanding” or therapeutic benefits of drugs, but at least the society is not paying the cost of externalities of drug usage (see San Francisco’s Tenderloin).

2. Not all who do drugs are rational individuals with full information about what might happen if a chemical dependency is established. There are people pressured into drug usage by both legal mechanisms (Purdue Pharma) and illegal mechanisms (e.g a pimp creating a prostitute by forcing a drug addiction on that person). So the implication that there’s no bad actors, and it’s just the state enforcing draconian measures, misses very real cases of bad actors who should be stopped and punished. A legal system that pursues and removes bad actors can help society.

Same question, why not “laser focus” your efforts on convincing one technically proficient investor with deep pockets, rather than having to convince 100s-1000s via crowd-funding?

Is it that investors need to provide the roof top too, so it’s a way to reach out to property owners?

For productionizing/training massive models, yes.

But in the development phase, when you are testing on a smaller corpus of data, to make sure your code works, the on-laptop dedicated chip could expedite the development process.

A trade between two parties is not a redistribution of wealth .

If I buy a car for $10,000, then I now have a car worth $10,000, and the seller has $10,000. Wealth was not redistributed, we’re both where we started.

Regarding using c6g’s instead of i3’s to power Elasticsearch machines, could you tell me more? (We’re in the same boat, considering a switch)

Specifically around hyper threading, my understanding is c6 don’t have “vCPUs”, and just have “CPUs“, so the effective number of cores doubles. Did you find similar throughput (in terms of Elasticsearch Search/Write TPS) between a virtual and real core?

The article missed an easy one; turn on your webcam.

> How do people build recommendation engines?

One method, collaborative filtering with latent factor analysis, popularized by its efficacy in making recommendations on Netflix, is to use matrix multiplication to solve the problem.

E.G. Let’s say you have all users (rows) x all books (columns), in a massive sparsely populated matrix, where the value is the rating that user gives a book.

To make recommendations, the goal is to “guess“ what a user would rate a book they haven’t read, and if your guess is they would give it 5 stars, and then you recommend it, and the user gives it 5 stars, it’s a good recommendation.

The “latent factor” idea is breaking the problem up, so in order to compute the rating matrix that is the final size N users by M books, you split it into:

N users x D latent factor (cross product) D latent factor x M books = N users x M books

It then becomes a machine learning problem, using a loss function plus gradient descent, to solve for that D latent factor.

Customers with a similar latent factor, will have similar taste.

Once you have the latent factors, you find the nearest neighbors (the closest other latent factor vectors measured by dot product or cosine similarity), to compute the nearest books. The vectors that multiply together to give the highest rating, will be the best recommendations for the user.

I don’t think the feeling of inadequacy from seeing those more successful is generational.

I think it’s more a reflection of the guilt you harbor at that time, e.g. if I just spent a weak procrastinating, I’ll resent successful people more than if I spent a week productive.

To your point of internet and social media increasing the spectrum of success, I definitely think music industry is a great example of that.

I’ve had that exact same experience, of feeling sick preceded by a day of near manic state. It’s gotten to the point where I can predict the following day I will feel sick, so I try and start resting more that first day.

Cool! Didn’t know anyone else felt this

I took a number of these courses as an Amazon software engineer and I found them very useful.

Amazon’s goal of the internal program is to train Software Engineers to know enough about data science to be effective as “machine learning engineers.” Once trained, software engineers can either implement models themselves, or more likely, partner with a data science team to productionize a model.

The courses by Brent Werness are particularly good, he has a knack for explaining the intuition behind the math, pulling back the layers of abstraction, so it’s not just “call scikitlearn and cross your fingers.”

Some of the more advanced topic courses like Natural Language Processing are particularly challenging, as the material is more of a survey of the latest research papers, to give you an idea of the state of the art, but understanding the mathematical tools is non trivial.

The history goes farther back than the Civil War, and I don’t see strong evidence that it’s always racial in its origins.

A lot of immigrants, both recent and historical, came to the USA because they didn’t like the government in their home country, Pilgrims as a key example, Falun Gong members leaving China as a more recent example.

The general distrust of a strong government is still part of American culture.

Why did you feel the need to say “western sense of morality” rather than just “sense of morality”?

The article describes national outrage at this story of injustice, your implication is that a larger “eastern” sense of morality has allowed this to happen, rather than the corrupt individuals named.

String value = “postalrat”; Assert.assertEquals(value, identifyFunction(value));

tl;dr the wallstreetbets hive mind has the purchasing power of a mid size hedge fund

if there’s roughly 1000 upvotes on a piece of content, and assuming only 1% of lurkers ever up vote; that means there are roughly 100,000 people taking investment info from this sub.

let’s say they have an average portfolio size of $10k, then 100,000 users * $10,000 = $1 billion dollars.

assuming they split the portfolio by meme %, so like 1/3 TSLA, 1/3 SPCE, 1/3 AMD; then we’re looking at like $330 million in trading volume going into SPCE.

when you look at the last couple days of trading volume on SPCE, about 30million shares traded daily, was previously around $10 a share , it about equals out, that like $300million dollars of meme money did a pump and dump on SPCE

Apologies if this is covered in another post; but what’s preventing you from getting a job and getting back into stable housing?

I put together a simple map of all coworking spaces in NYC I could find. I also separated out some of the spaces that are trade specific, such as artistic leaning ones.

Other findings include average price is $425 for an individual desk in a shared space. Brooklyn is conspicuously bereft of spaces, even though it is the home of many in NYC tech.

The map is editable, please feel free to make any changes necessary/add any new spaces not pictured.