glacials

You lost me at ncurses. IMO this idea works because the task is so simple no one would bat an eye at the thought of doing it. All it takes is one person to question why we're not using <technology> instead, or why we're not spending this ncurses amount of effort towards something else instead, to jeopardize the effort.

Async communication is great and it should be the default. but knowing its flaws and when to switch to sync is important. Async is:

Low bandwidth: Writing doesn't have inflection or body language, so you need more words to convey the same idea. More: https://www.youtube.com/watch?v=3bFsqk84frI

High latency: An idea that takes 30 back-and-forths to resolve may take minutes in a meeting but weeks in writing. Meetings really have negative latency because you see people react before you've finished speaking.

High effort: Even if you account for low bandwidth and reach the same % chance of being understood as when speaking, when that % "misses" the cost is high because of the round-trip-time. To counteract this, you spend even more time editing than you did getting your meaning across.

Emotional: I perceive people as meaner in writing than in meetings, and in response I become meaner. I don't want to and I try not to, but it can still happen.

If you use a password manager that's two factor, then you are still at two factors. e.g. vault-based or hardware key based

A lot of the drag on CI for complex projects is tests, which are hard to argue against. Complexity : need for tests isn't linear -- once you hit some critical mass of complexity where one person can't know the whole application, the need for tests skyrockets.

I joined a company last year that's trying to solve this [1] by tracing tests so it can skip any whose dependencies (functions, environment variables, etc.) haven't changed. It's amazing what "what if we don't run tests we know will pass?" can do to a CI pipeline.

[1]: https://yourbase.io

Can you explain why the Kelly criterion wouldn't have you bet 0 every time? The chance of winning a round of rock-paper-scissors when throwing a random symbol seems to be 50% (if ties cause re-dos), so wouldn't that work out to 50 * 2 - 100 = 0?

Your setup is great for enthusiasts like you and me, but everyday folks don't want to deal with mounting hardware or audio interfaces or learning what a condenser mic is or spending more than $50.

Where are your customers in the world? Many non-US countries have little/no credit card presence, so if you don't have an international-friendly payment option like PayPal your customers may be using "gift card credit cards" purchased from a store. It's easier to do this once a year than once a month.

Ultimately though, talk to your users! That's how you'll find out.

On the chat team at Twitch in the early days after the Twitch Plays Pokemon crisis [1], we started artificially doubling all chat traffic through our systems, then dropping the doubles just before they would be sent to users. [2]

Not only did it give us a "big red button" to press during emergencies like OP, but it revealed important logical scaling issues before they became real problems.

[1]: tldr; 1 million people playing a single game of Pokemon Red by using chat to send button presses

[2]: https://www.twitch.tv/videos/93572955

Congrats on the launch!

It seems like you're pushing for video-chat-as-a-hangout, which I like a lot! My company has been doing a modified version of this for about a year and we like it a lot.

Can you talk some more about what makes your product specifically better at serving that need than existing video chat apps? Is the value that good text chat is available in the app rather than needing to use a different app, or is there something else that makes it better at the "hanging out" use case?

The MIT OpenCourseWare talk How to Speak by Patrick Winston is a great lecture on this subject: https://youtu.be/Unzc731iCUY

Dual-licensing platform for open source software: https://license.land

That's great feedback! I've shared it with my team, thanks.

As an example of what the opposite pricing model looks like, YourBase[1] charges a flat fee per build such that it's in their best interest to make builds as fast as possible. Because of this forcing function, builds are instrumented and cached down to the system call level deterministically, based on file changes. It's amazing what economic incentive can do. (disclaimer: I work at YourBase)

[1]: https://yourbase.io

Bingo! Balance is key.

Farmers working under communist rule in China who started private farming caused agricultural productivity to skyrocket and the number of starving people to drop [1]. Say what you will about human greed, but aligning a behavior you want with a natural economic incentive is a powerful tool.

On the flipside, some industries like healthcare don't belong in such a ruthless system, because "voting with your wallet" is sometimes equivalent to "voting with your life". Other industries like scientific research have too long a lag time for the the companies to stay afloat, or for the decision makers to benefit from the decision. These industries are the ones that frequently need help from government, nonprofit, donors, etc.

[1]: https://reaction.life/chinese-capitalist-growth-happened-spi...

"Do things that don't scale" [1] generally refers to the startup advantage. Doing thing A more cheaply than thing B _at_ scale is a different problem.

[1]: http://paulgraham.com/ds.html

Can someone explain why Mano's job description prompted this response from Feynman? I don't know enough to understand what this problem is, or why Feynman thinks any person solving it would be sad.

> Most the big streamers got to where they are by view botting, the "fake it til you make it" approach.

I'm sorry but this is not true. There have however been perception issues around viewbotting. One such:

For most of the life of host mode when a small stream would get hosted by a big stream, nearly all of the transitioned viewers would stay (through inaction) in the host's channel. Many would be AFK. The small streamer ends up with a large viewercount but a disproportionally "dead" chat, leading to cries of viewbotting.

Coincidentally, big-to-small hosts like this are frequently how fantastic but relatively undiscovered streamers get their "big break", further entangling the accusations with their success.

Source: Worked for five years at Twitch, built host mode

(PS: As someone who became good friends with some of these "big streamers" and those at Twitch combatting viewbotting, I can't help but feel insulted on both their behalves that the amount of invisible effort they pour into their craft so frequently gets shat on.)

I interned at LLNL on three separate occasions during school. Everything about what y'all do in high-performance computing is levels beyond anything I've gone on to see in the "real world". I cannot drop enough praise.

Speak of the devil, as you posted this I updated the messaging to say exactly this! Thanks for the feedback.

Thanks for the feedback! Reflecting on this, I completely agree -- what this tool is actually estimating is a minimum to compensate you for lost time, sanity, energy, and money. Your actual price should be based on the market.

I've updated the messaging on the page to reflect this. Thanks!

Modern stock (Google-flavored?) Android has also taken up buttonless gestures with the last couple of generations of Pixel devices. Depending on your Android version and your phone, you can toggle it on and off in settings by searching for "gesture".

I started a side project in 2013, and when I (for other reasons) left my job at the beginning of 2019 I decided to go full time on it. I timeboxed myself to one year to show salary-like profits.

Things that snuck up on me:

1. I thought it'd be the same work as before, but more of it. But when your side project is promoted to your full time job, you want to keep it that way. You start focusing on different things, like making sure you make enough money to keep going. My project used to be ~90% free, and after six months of "that's okay, I'll just build more paid features" it didn't work out (that's a whole post on why) and I was forced to move a lot of free stuff behind the paywall, which comes with all the user backlash you'd expect.

2. Apart from building paid features, learning how to do so and funnel people into them is half a full-time job in and of itself. YC's free Startup School was a fantastic resource here.

3. The more people pay you, the more you feel beholden to them. You'll spend more time doing customer support than you want because of the guilt. You'll think you can resist this or engineer it away, but it just keeps coming.

Before: - 10k MAU - Net negative $300 a month

1 year of full-time later: - 13k MAU - Net positive $500 a month

Uplifting result, but not salary-like, so I'm moving on and demoting it back to a side project. Nonetheless I don't regret this; if I hadn't done it I'd always be wondering "what if".

For anyone else who's hunting all over for a rendered example, just open devtools -- the site dogfoods itself.

  Location: Seattle, WA
  Remote: Yes, been successfully remote for last 2 yrs
  Willing to relocate: No
  Technologies: Rails, Go, Node, Vue.js
  twos.dev
  linkedin.com/in/ben
  github.com/glacials
  qhiiyr@gmail.com

I recently built my side project into a $500 MRR business, but it's not enough of a salary to continue. Since I'm coming down off a year of building my own company, I want to join somewhere that can give me the same thrill of wearing a lot of hats and being very close to both the product and the customer.

Easier said than done -- data models frequently evolve in hard-to-predict ways. Instead, build your data model in a way that is easily pliable, and won't need a complete refactor because of something simple like a hot key or a new index.

My company had just started doing performance reviews, as they had been growing a lot (generic small startup with unlimited vacation and all that) and it was That Time. So this was a quarterly performance review, but was my first after having worked there 1-2 years.

The system had six levels resembling standard deviations in a normal distribution, so ~2% at the extreme high and extreme low. In parallel, you pick one for yourself and describe why in text, and your manager picks one for you and describes why in text. Then some software simultaneously shares them between you, and you have a 1:1 to discuss things you didn't agree on, or that one of you brought up but not the other.

I rated myself as the highest level, because I thought I was the second-highest level and had learned that my opinions about myself are generally lower than others' opinions of me, so I decided to purposefully account for imposter syndrome. My manager rated me as the second-highest level and said that the only thing keeping me from the highest level was the amount of vacation I'd taken that quarter.

I'd taken 5 weeks off that quarter. It's a lot for a quarter, but it was also the only time I'd taken off the whole year, and 5 weeks was the "recommended" yearly amount. So I wasn't sure how to feel, and second-highest level didn't seem like something to complain about.

Still not sure how I feel five years later because I do see both sides and it's a hard problem to fairly solve. But I live a happy life and don't have any complaints, so I don't let it get to me. It's just a performance review, and you will look back in 10 years from a different job, city, and/or life situation wondering why you cared so much.

All complex software has bugs.

In good software, the bugs don't cause user-facing issues. QA teams aim to make software good by finding them.

Speedrunners aren't users or QA. Speedrunners find bugs no matter what. From the developer's perspective it really doesn't matter if you can save 20 hours beating a game by clipping through the wall next to the boss, if that clip is so immensely complicated to perform that no user would ever encounter it naturally. The devs can be surprised, or not surprised, to hear their game was beaten much faster than they anticipated by speedrunners; either way it does not make their game worse or rushed or a product of bad communication.

The Legend of Zelda: The Wind Waker came out in 2002 and sold 4.6 million copies. In July of this year, a new bug was discovered by dedicated minds, not randomly but after nearly two decades of hard work, that saved two hours on the speedrun by clipping through a barrier that was supposed to be unpassable. Does it really matter from a quality standpoint if this bug exists or not? Would your original rating of The Legend of Zelda: The Wind Waker be lessened 17 years later because this bug was found by people who perhaps spent more hours trying to find it than you've spent playing video games in your entire life?

I'm one of twelve speedrunners of this game [1] and this is a killing blow to our community. Those of us lucky enough to have the game installed are the only ones who can continue speedrunning this game, and we can never uninstall it or switch computers. No new speedrunners can enter the scene. It was enough work getting it to run over the years without Games for Windows Live, or at a modern resolution, or at a modern framerate, with horrible keyboard controls, and without crashing at boot due to driver issues.

Now, new players can't even get their foot in the door enough to see that there are fixes. Even with a crack, this is the end of speedrunning for us.

[1]: https://www.speedrun.com/tronevolution

Electric cars are an energy-source-agnostic vehicle. Whether their energy comes from coal or nuclear or wind or solar, they are unaffected. Because of this you can gradually swap out pieces of the energy production backbone to renewables, and see a reflective immediate payoff in personal transportation CO2 emissions.

If the vehicle is tightly coupled to burning petrol, that can't happen. Electric cars are merely an exercise in uncoupling energy source and energy use.

This is a far cry from a rebranding. Restructuring into a parent/child company or redesigning a logo don't affect the day-to-day operations of Facebook employees. Shifting into a white label company does, and it's a huge risk for a company not built for it. Facebook is good at full-stack, and will take the approach that lets them work to that advantage.