jerzyt

Of course, Wikipedia is the easiest website to scrape. The HTML is so clean an organized. I'd like to find some code to scrape Airbnb.

Yeah, I'd expand on the "API over DBs" based on my experience. I could make it nest much deeper: DBs over Excel spreadsheets, and Excel spreadsheets over PDFs or scraped HTMLs, yada, yada, yada. But before I waste any time creating an API, my client wants to know if the project is even viable.

Google's Search is dying because of Thousands of self-inflicted cuts. A couple of years ago there were about 3 or clearly marked ads on the first page, followed by generally useful results. Right now, I don't have the patience to scroll past the ads which are almost indistinguishable from useful links.

I don't have the stats, but seems to me that majority of purchases at convenience stores is alcohol. In CA, you cannot buy alcohol at self-service POS. That spells doom for Amazon Go.

At one of my previous jobs, I did a project on employee retention. Our goal was to identify valuable employees who were most likely to leave. The results were surprising at first. The top two factors leading to an employee moving to a new job were: high performance review and high increase in compensation (raise and bonus). On reflection, we realized that we just reaffirmed their market value, so they could negotiate their next salary starting from the new base with us.

Bottom line, whether you're happy or not with your job, you should probably move.

Google is not paying people to provide a service. They are paying a company. I sympathize with the people who are underpaid, but their beef should be with their employers not with Google, even though as another commenter stated: Google is outsourcing non-compliance. On a related note, if people working below the minimum wage are flying to protest at Google headquarters, someone is paying their airfare. Doesn't look like a spontaneous act to me.

This is really a great way to sabotage your competitor. Have a bunch of friend and family buy a product and flood Galaxus with bogus returns and warranty claims. We'll see how it works.

100% in agreement on Excel. Even when coding in Python I frequently save an intermediate file as xlsx to explore/debug, or even load into Tableau for viz.

VMS OS from DEC had exactly the "unique password" stupidity in the late 80s. In a way, it was much worse than Twitter's, because where I worked there were only about a 100 accounts, so you could find the other by hand in less than an hour.

It could also be argued that "medianism" is the perfect method to hide income inequality, because median is resilient to outliers, so the billionaires will not affect the median at all. Every metric can be "gamed" and so can the median.

I can confirm it. I have a commercial vacuum packing machine, and a bag of rice sealed in it becomes a brick. You can't displace a grain of rice without damaging the vacuum. I doubt that a laptop inside this bag would survive it.

I'm with you - I noticed that little pac-man, but it's not an optimal way to see the difference. Seems like scaling both images and doing an image diff would be much more effective. That's a standard tool in semiconductor manufacturing.

And what will stop them from monitoring the alcohol sales?

Really depends on the phase of the project. At the start, getting everyone in the room is really beneficial. Once we decide what we're doing, I'd rather be left alone and work from home. At the end, testing, debugging and fixing, I'd rather be in the office.

Kudos to Spotify, whether it's a money decision, or a decision to stand up to a has-been entertainer. They stood up to a bully.

In 1961 the KGB would be on any foreigner from day one. No surprise at all. In 1979 I crossed from Poland to Russia by train, and while we were stopped for the axles to be changed to the wider track gauge, I moved from one railroad car to another to hang with my friends. When Russians discovered that one car had a missing passenger and another had an extra one, they checked the entire train and we were delayed good several ours.

I thought that for a change Linus wrote something remarkably noncontroversial. Turns out I was wrong.

Double down on stupidity. Nope, we're not gonna fix it. We'll insist that we're right and you're an idiot. I hope that a big consumer country, like US, bans this idiocy.

You can ask the question at little differently. The authoritarian state may not care about most sites, but they do care about a few selected sites which they consider "subversive." By monitoring the traffic at both ends, they could potentially correlate packet timestamps, and infer your IP without directly tracing it. Kind of like Shazam for dictators.

At one of the Big Four advisory firms we were only allowed one-on-one interviews. I was never able to get a satisfactory answer for this rule.

Now I wonder if it was designed to prevent this specific scenario.

For me, the worst thing about airline food is that it's notoriously overcooked. I think a lot of the meals would be quite decent if cooked properly, but the service at 35,000 feet is difficult to manage and many of the meals are just ruined. The best in-flight meals I had were braised dishes, which have a natural resistance to being overcooked. Another exception was on Korean Airlines, where the dishes were mostly fresh greens, not cooked at all.

What Netflix is doing is hiding the fact that they have very little new content. If they implemented all your suggestions, I'd have about 3 titles to watch, which I would like, but they're worried that I'd cancel the subscription.

Yes, without context it's just not interesting. I wanted to map the cities, but without state name, quite a few of the cities are ambiguous. For example, MA, MO, and OR all have a city named Springfield.

It is rampant in consulting business. At least in an internal project it's possible to pull the plug when the results are not promising. In a consulting engagement, when the results are garbage, there's no revenue and no potential for an upsell. In effect the pressure to "find" significant results is enormous.

One of the biggest risks for a startup is becoming a consulting firm. Once you do that, you won't be able to scale up. Been there, done that. That's why the FAANGs are so successful - they have millions of clients, without a single one who could control their destiny.

Not to criticize the concept of JIT to avoid excessive inventory, but I've always argued that the flipside of JIT is turning every path in the supply chain into a critical path.

I opened the comments specifically to see if anyone else thought the price was unrealistic. I'd be surprised if that price would even cover the cost of materials.

Even as an observational study, it leaves a lot of unanswered questions. First and foremost, all the values are fairly close pre and during Covid, but there are no confidence intervals for any of the differences. Not even sure that the reported differences are statistically significant.

The old style printf from C is still the best formatting tool for output/debugging. The C++ style was just a distraction without introducing anything of real value. log4xyz has some nice features in terms of enabling/disabling at runtime, through a config, but ultimately, printf rules.

I did a project in Atlanta which required my presence on-site, so every week a flight to and back. I've spent enough time working there that I had to pay Georgia state tax. Anyway, I consider Atlanta to have one of the worst airports in US. Pathologically long lines through security, even with the rapid pass.

Maybe it won't be a problem in the Covid era, with zoom calls everywhere, but in the pre-Covid era Atlanta was a PITA for travel.