markatto

This is sometimes done intentionally to hide latency and make a UI feel faster - I certainly don’t like it though.

I’ve been finding that very rigorous/technical people who are willing to adapt their workflows are very effective, whereas more “management” types are sometimes effectively pushing work off onto their code reviewers.

Lutes, even!

This is just a matter of priorities - I use LLMs to write code every day and I have never put a single line of code up for review that I didn’t read and understand.

It's turn based, the most similar game I've played is probably Fantasy General. Closer to Advance Wars or Fire Emblem than Warcraft.

I tried this with Claude - it has to be explicitly instructed to not make an external tool call, and it can get the right answer if asked to show its work long-form.

Your HN jobs page lists comp as 24-48k, is this accurate?

It matters for things like verifiable builds.

This is similar to how wild boar was hunted in Europe - there were special spears made with extra crossbars to reduce the chance of a boar reaching the spear holder after being impaled.

A half century? Surely that must be a typo?

I’m surprised that this doesn’t get talked about more - I’ve worked with plenty of women, minorities, and immigrants at large tech companies, but I can’t recall a single person with a southern accent.

I think that you should mention that this is specific to 90 degree v8 engines - other engine designs can have similar crankshaft / firing order trade offs but would not be described as “flat plane” or “cross plane”.

That’s not their best hardware deal, but you also have to understand that you are paying for rack space, electricity, bandwidth, on-site staff for repairs, etc. They are actually absurdly inexpensive compared to competitors.

The piano is not the canonical model of a scale - it’s just a bit more visual than some other instruments. I went to music school and primarily play guitar and violin but can sort of play some keys. I could definitely figure out pretty much any scale on a piano but couldn’t just sit down and immediately play most of them.

I have found that even 2D games (eg Hades) aren’t reasonable to run at 4K, but basic gui desktop with a terminal emulator works fine.

There’s still basically one with a couple storage options - they have shipped with different fans though. My (fairly recent) unit has very reasonable fan noise.

Location: SF Bay Area

Remote: Prefer to work in-office, but would consider remote opportunities.

Willing to relocate: Possibly, for the right opportunity.

Technologies: Linux, Python, Rust, C, Kubernetes

I have been working in SRE/infrastructure roles for the past 13 years at employers including Google, Facebook, and several smaller companies.

I am interested in both SRE-type positions and roles developing systems software. I'd especially like to work with rust or on open source software.

My job search started a few days ago so I don't have an updated resume yet, but my work history can be found on my LinkedIn: https://www.linkedin.com/in/markatto

mark@marktmerritt.com

I had very mixed experiences as a customer in ~2009 - it was definitely possible to get through to someone very competent, but it was pretty difficult. We were trying to use their "cloud" VPS offerings (I believe it was openstack-based, and I liked the idea of supporting something more open than AWS), which were extremely buggy (for example, we had to default-retry every api operation multiple times in scripts because pretty much every type of call failed 40% of the time.) I generally had better success getting information about outages by reaching out employees I knew on freenode than I did through official support channels.

Why do you think the unit economics of Carvana have to be worse than a traditional dealer? Maybe they have to do more shipping of vehicles around, but they are less reliant on storefronts/sales staff. It seems likely to me that they overspent on marketing/gimmicks/badly timed top-of-market inventory, but that doesn’t mean that the business model fundamentally can’t work.

Yes, some hashes might not meaningfully hurt it, but they won’t add any entropy, which is the real problem.

1) I have more of a sysadmin background, so someone else will have to answer this. I will say that if you interviewed in the sre-swe track, transferring to a swe role is not difficult.

2) I have been on call for most of my career, but google is the only employer that has actually provided compensation. Depending on your lifestyle, the extra pto can be really nice (for example it’s easy for me to ski on less-crowded days), but as I’ve gotten older being on call has definitely gotten more painful (now that I have a serious partner, not being able to do things on the days she has off can be frustrating.) The actual difficulty/stress of the pages you get will be highly team-dependent, but SRE does a reasonable job of training and tracking pager load.

3) This is highly team-dependent, but there are many teams where systems skills are not a big deal. You need to be able to think in a reliability-focused way, though.

4) Some teams actually develop their own code, other teams rely on swe teams for most of the code. You will read a lot more code than you will write, and you will write less code than they say you will during the interview process. This is also true for SWEs - everyone spend more time big-company-ing than computer-ing.

5) I have not done this, but I have considered moving to a SWE role because there’s more open source opportunities on that side.

6) I’ve found google to be the best place to work of the mega corporations I’ve been, but honestly if the money was the same I’d prefer a smaller company. I’m better at tasks like “chase this bug though a bunch of layers until you find some weird kernel behavior” than “explain why your rollout plan is compliant with our reliability directives.”

I use the "multiple security keys" approach, and the biggest problem is keeping track of which keys are registered with which services and making sure the list is up to date. A few examples of situations where this is a problem:

1) I don't keep all of my keys on my person, so if I want to sign up for a service when I'm not at home, I have to remember to go back and add my other keys at a later time. If I wanted to, for example, keep a backup key in an offsite location such as a safe deposit box, this would be even more painful.

2) If I lose a key, I need to go and change every service to deactivate the lost key and add my replacement key. This is both time-consuming and error-prone, as it requires me to keep a full list of providers that I use keys with somewhere.

3) Some providers do not even allow you to register multiple keys.

There are many SciFi titles that are extremely expensive in hardcover these days. There is a lot of amazon volatility, so I tend to use price watch sites, but some also consistently stay in the hundreds of dollars (for example, the original printing of Hyperion). I suspect that this is because the genre used to sell badly in hardcover and many SciFi fans have come into money over the last few decades, but I'd be curious to hear from somebody who understands the market better.

I have worked at both Google and Facebook without any degree or certification. In my experience, these companies care a lot more about interview performance than formal credentials.

I assume it's more based on this: http://www.catb.org/~esr/writings/unix-koans/

How about an option to overlay the pentatonic scales over other scales, similarly to the way you color the root notes but with a different color? You could use a similar idea to overlay root-third-fifth or other chords. I think this would be useful for those of us who are still trying to build muscle memory, but still present information in a way that minimizes "stuck thinking."

You could use a block cipher for the actual video stream with a single key, and encrypt that key with all 31 public keys.

Ken is a great writer in his own right though; The Paper Menagerie might be my favorite book.

They shouldn't be giving their customers a static IP, they should be giving them a 'customername.ourplatform.com' address that the customer can point a CNAME at.

They're still taking downtime for this... Even if they're forced to have a no-VIP no-HA no-LB setup (seems insane to me) it should be much simpler to set the DNS TTL to a low value right before and switch it to the new IP after the new box comes up.