singron

I think containers are completely optional and it can install and run everything directly.

There is something nostalgic about these old Google documents. It reminds me of when search ads were on the right side.

Google is famous for only supporting a few general languages. Initially it was just C++, Python, Java, and Javascript (browser only!), although a few more eventually trickled in. The language deviants kept writing "official" style guides and adding support to random infrastructure and libraries, hoping that one day the next major project might get written in OCaml. New Googlers often had a confusing moment after learning the restrictive language policy when they compiled their first program and blaze would print that it was generating protocol buffer bindings for Haskell (this would eventually be suppressed).

Google puts multiple clusters in a single building.

She is the NTSB chair, not Boeing.

It's super hard to get this right. E.g. if you only update the cache after the transaction commits, you might commit without updating the cache, or if 2 writers interleave, the first one to commit might make the final update to the cache with a stale value.

Matthias Wandel. He doesn't do a lot of reviews, so I wouldn't go there only for that, but he does speak candidly about the process.

Isn't stock also a scam since they don't have to issue dividends?

Yeah if they have no negative reviews, that's a bad sign. A particularly scrupled YouTuber I follow typically won't do a paid video if the product isn't good and instead does an unpaid tear down video. That probably limits his opportunities to brave marketing teams with high quality products, but it also makes his reviews quite valuable.

My experience running my email server is exactly what danluu says. Receiving emails and classifying spam is not a problem at all, but my outgoing mail is often blocked. It's not that I struggle to moderate the whole world, but the whole world is moderating me.

The value prior to disposal is irrelevant. The IRS cares about the basis (cost to produce) and that it was disposed at $0.

Fwiw our hiring pipeline has 30%-70% of candidates make it to the next stage once past the phone screen (it has varied somewhat as we re-ordered or changed process). Assuming 3 stages until offer (e.g. HM interview, tech screen, full-day interviews), and offer-acceptance, with 50% success rate at each stage, 1/8 get an offer and 1/16 accept it. At 60% it's 1/4.6 and 1/7.8, so the exact rate makes a huge difference.

The corollary to this sensitivity is that if you are just slightly better or worse at interviewing, you can see a huge difference in the number of companies you need to pass screening at to get a job. In practice, companies get more people into the pipeline than necessary to reduce the chance that it dries up without making the hires, so as a candidate, you will also have to take that into account.

You can play around with the variables on this graph to see number of phone-screen-passes: https://www.wolframalpha.com/input?i=y%3D1%2F%28x%5E4%29+whe...

I was really amazed about how many people lie on the resumes. E.g. a recent study [0] said 70%, but there are many of these with similar numbers. There is a difference between putting your best foot forward and just straight lying for financial gain, which is fraud, and can (very rarely) send you to jail.

Personally, if I suspect dishonesty, it's a strong no-hire. If I even get a whiff of dishonesty, I'll at least flag to recruiters, hiring managers, and interviewers, and we've actually compared notes and realized a pattern of behavior before.

It is hard though since a lot of the stuff they would lie about are things you can't verify.

0: https://www.forbes.com/sites/bryanrobinson/2023/11/05/70-of-...

You would think they would have that address on a post-it note in the PD and call centers, but I guess not: (opposite situation, but relevant: https://xkcd.com/1105/)

You don't need this many tools, especially really early. It also depends on the particulars of your business. E.g. if you are B2B SaaS, then you need a ton of stuff automatically to get SOC2 and generally appease the security requirements of your customers.

That said, anything that's set-and-forget is great to start with. Anything that requires it's own care and feeding can wait unless it's really critical. I think we have a project each quarter to optimize our datadog costs and renegotiate our contract.

Also if you make microservices, you are going to need a ton of tools.

You can have conversations across language with Google translate. You take turns speaking into the microphone. It's actually pretty awesome. Obviously you probably aren't going to make best friends or fall in love this way, but you can ask almost any question and understand the answer in many languages without having had spent several years learning the language.

Most people have the same filing status and number of dependents as the previous year, just have W2 and 1099 income already reported to the IRS, and take the standard deduction. For most people, they can tell you what you owe. If it's wrong, you can tell them your dependents or filing status changed and get a new form. If you have a weird situation the IRS doesn't know about, you can still do your own taxes, but that's not most people.

NFL games are also overwhelmingly on weekends.

If a developer pulls out of Apple, the developer loses half of their mobile market, and Apple loses a single customer. I guess if you got all developers to leave Apple, and then all iOS users switched to Android, then it could hurt Apple more than developers, but that's probably not happening.

If you drive 60mph for 20 minutes, you will go 20 miles. If you drove 65mph, then you would travel the same distance 1 minute 32 seconds sooner. At lower speeds, the effect is bigger, e.g. going 45 instead of 40 saves 2:13, and 35 instead of 30 saves 2:51. If my math is right, to only lose 30 seconds by going 5 under for 20 minutes, you would have to be going 195 in a 200.

Other comments give more detail, but some of the good ideas made it into Firefox. A similar thing happened at Google when Omega seemed like it might replace Borg. Borg adopted the best ideas and Omega got cancelled.

Sometimes it's easier to retrofit the novel ideas from a research system into the production system than it is to make the research system into the production system, and that's OK.

These mobile platforms are two sided markets. They only compete on one side for users. They don't compete on the other side for app developers since developers need to use both platforms to reach all users since each user is exclusive to a single platform at a time.

This made me wonder what the shortest open source license is, and it seems to be the Fair License: https://en.wikipedia.org/wiki/Fair_License

> <Copyright Information>

>

> Usage of the works is permitted provided that this instrument is retained with the works, so that any entity that uses the works is notified of this instrument.

>

> DISCLAIMER: THE WORKS ARE WITHOUT WARRANTY.

EDIT: I would not use this license. I'm not sure how this got approved by the OSI, but I'm not a lawyer. Some of the email threads think the disclaimer is insufficient, and I'm not entirely sure that it confers all the expected rights since it only says "Usage" (copying, modification, distribution).

Fd stores are an awesome way to do this too. We used to start up a new version of the process and then they would have to find each other and pass the fds. It's pretty complicated and it's a weird model for the supervisor.

I've had options at 2 companies and RSUs at 1, and I have a similar story (bigger earnings on RSUs, bigger "losses" on options (I technically still hold exercised stock, although I don't anticipate ever being able to sell it)).

A huge part of this is that a company won't issue RSUs unless it's already reasonably successful and the shares are liquid or nearly liquid, whereas options are issued by early stage companies that are mostly likely going to fail and are far from a liquidity event.

As an employee, options just aren't good compensation since they have too high risk and you can't work at enough companies to form a portfolio. The VCs that invest in these companies expect to win on less than 1/20th of the companies with preferred shares, so it's quite likely you will never make money on options in your entire career.

People talk about taxes a lot, but the most important factor is whether you are going to make money at all. There is no point in optimizing taxes on your gains if you are making losses instead.

The 409a isn't based on what the market will pay since private shares aren't on the stock market. It's a guess about what the market might pay, but since the shares aren't liquid, the guess can't be tested, so there is huge leeway to pick a favorable number.

This case is probably relevant: https://en.m.wikipedia.org/wiki/HiQ_Labs_v._LinkedIn

In that case, breaking the ToS superceded the fact they were merely accessing public information.

The other question is whether Beeper is violating terms of service or their users are. I'm guessing Beeper is not and they instead need to be implicated for some kind of tortious interference. I would love if Apple individually started suing their own customers though.

The gas burners are wildly inefficient at transferring heat to the pot/pan. E.g. Wirecutter says gas is only 32% efficient, so the nominal 18 BTUs is only 5.8 BTUs in the pan. Whereas electric is at least 75% efficient (higher for induction) so a nominally 12 BTU electric stove provides at least 9 BTUs in the pan. In this case the electric stove is still effectively more than 50% more powerful.

They will each apply the insert locally. Then they will attempt to replicate that insert to each other. Each will attempt to apply the replicated conflicting insert, which will cause an error and halt replication for both nodes.

If you update the same data on both nodes, this is a recipe for almost certain disaster. Postgres is not a distributed database and this doesn't make it one.

They had Google Code. I think it might have first been deprecated before Critique was popular (the prior code review tools at Google weren't quite as beloved). Also since then, buganizer was made partially public facing. It sounds silly, but they could relaunch a software forge suite and it would almost definitely be better than GitHub, and they could cross sell cloud and probably not even have a free tier.

Note that glibc has a similar problem in multithreaded contexts. It strands unused memory in thread-local pools, which grows your memory usage over time like a memory leak. We got lower memory usage that didn't grow over time by switching to jemalloc.

Example of this: https://github.com/prestodb/presto/issues/8993