waprin

Repost from 2020, previous discussion: https://news.ycombinator.com/item?id=22126731

Thought it was very useful article so worth a repost. The author is a cofounder of Supabase (YC S20)

I recently left a 400k+ job to work full time on an app addressing this exact problem, despite having a wife, daughter, and Bay Area mortgage . Of course productivity apps are a crowded space but I have a burning need to build the right tool for myself , and I believe might help others.

Remote work has pressed the need for me to improve here since I think I benefited by being in the same room as coworkers and being alone in my home the whole workday has pressed my ability to stay focused.

I struggled with ADHD , but because of past health issues, medication is off the table. Even though Vyvanse was a miracle for focus it caused some serious other problems. And the more I read about psychiatry and psychology the more I’ve gotten excited by lifestyle changes.

One thing I feel strongly about is that productivity tools need to be descriptive, not prescriptive . I’m not a robot, I need to work on what my gut is telling me to work on rather than what some reminder is telling me to work on. But I also need to do work and not scroll Reddit for hours so a tool helps with accountability for that.

I can’t over emphasize how enamored I am with Andrew Huberman Youtube video on the topic. He’s a Stanford neuroscientist with all sorts of suggestions on managing our dopamine levels with techniques like cold water therapy and intermittent fasting.

Social media, including hacker news and excessive email checking, can destroy your motivation if you don’t keep it in check. If you’re struggling with focus you simply have to monitor and moderate these things.

Please email me, username at gmail if this topic is of interest to you.

When joined in they were on the tail end of sunsetting Google Code, which they explained was the smart thing to do because Google doesn’t want to waste its time on opportunities that aren’t measured in billions.

Google then sat on the sidelines as GitHub grew to a tens of billions dollar business, and now Google spends billions trying to get more developers on GCP which they could have gotten for free if they still owned the portal to the code all the developers use.

A lot of Google wisdom sounded a lot wiser before we saw how it fully played out.

When I worked at Google, I made an App Engine Flexible runtime for the Whitespace language (only valid characters are space, tab, etc). https://en.wikipedia.org/wiki/Whitespace_(programming_langua...

App Engine Flexible was another one of these products that just needed a container that responded on certain ports, it was quite fun to get it working. But I ended up writing a tiny "compiler" that would compile a small subset of Python to whitespace since writing just whitespace itself was quite challenging, obviously.

I wanted to release it on April Fools Day but my manager was totally against it because April fools had become its own serious entity in the marketing org and he didn't want me stealing their thunder ( these days Google has dropped April Fools jokes altogether, I guess it got a little played out).

On paper, Python is not the right tool for the job. Both because of its bad performance characteristic and because it’s so forgiving/flexible/dynamic , it’s tough to maintain large Python codebases with many engineers.

At Google there is some essay that Python should be avoided for large projects.

But then there’s the reality that YouTube was written in Python. Instagram is a Django app. Pinterest serves 450M monthly users as a Python app. As far as I know Python was a key language for the backend of some other huge web scale products like Lyft, Uber, and Robinhood.

There’s this interesting dissonance where all the second year CS students and their professors agree it’s the wrong tool for the job yet the most successful products in the world did it anyway.

I guess you could interpret that to mean all these people building these products made a bad choice that succeeded despite using Python but I’d interpret it as another instance of Worse is Better. Just like Linus was told monolithic kernels were the wrong tool for the job but we’re all running Linux anyway.

Sometimes all these “best practices” are just not how things work in reality. In reality Python is a mission critical language in many massively important projects and it’s performance characteristics matter a ton and efforts to improve them should be lauded rather than scrutinized.

Most people bought bitcoin to buy weed on Silk Road , now it’s the reserve currency for a nation state.

Just because not every use case lives up to the big picture ideology and long term goals doesn’t mean there is no big picture long term goals.

A lot of modern AI exists because of GPUs which exists because people wanted higher quality zombies to shoot in their FPS games.

The source code needs to be on everyone’s computer so that its execution is decentralized.

Part of the problem is confusion around the word decentralized. If I have a website billprin.com , in a sense it’s decentralized because there’s a million websites and that one is my little corner and you have your corner.

But in another sense it’s centeralized since I get to be the sole arbiter of truth on billprin.com

On ethereum, everyone runs the same code and has the same state so nobody is the sole source of truth.

Blockchain is way slower and more expensive that a centralized db but it has this key feature that no single party controls it. This enables us to write applications like automated market makers that require no trust in a centralized party or more choice in whom to trust and more granular control over trust.

It also has the nice side effect of making a lot of stuff more transparent and programmable.

When do we prefer not to have to trust a party? Many people don’t trust Wall St entities so are excited about potential ways to decentralize finance.

Others don’t trust big tech companies and are excited about ways to decentralize identity and communication.

And the future is hard to predict , it feels plausible that a big blank canvas will lead to more applications that are hard to predict.

The key is realizing that the core feature is not having to trust a centralized entity.

There’s technological challenges , and ways in which the current landscape doesn’t live up to the ideology. The money both incentives early adoption and development and turns it into a casino. But hopefully that helps clarify why some people are excited about the underlying potential.

Hacker News seems to think the circles they hang out with represent all developers?

Dan Boneh teaches cryptography for Stanford and works for a16z crypto.

Silvio Macali is a Turing award winner working full time in a new blockchain.

People like Gavin Wood (CS PhD) and Vitalik are obviously top 1% of engineering talent.

It’s amazing how HN insists crypto people live in a bubble, but upvote the low effort “web3 is a scam” written by nobodies post du jour but choose to ignore all the worlds leading cryptographers excited about web3.

HN is 100x the bubble/ echo chamber than the crypto world is .

Uniswap / automated market makers / decentralized exchanges are already a “killer app” of web3.

It’s not a hypothetical, it already has billions in daily volume.

It’s true that , for any blockchain app, you could trust a third party who could use traditional database to do it faster and cheaper.

What the web3 detractors repeatedly refuse to understand is that not needing to trust a third party is a feature many people care about. By handing trust to a third party over important things like your money and your identity, you hand over a massive amount of power over yourself to them.

It’s also true that all this tech is still nascent. Big picture, ~10 years is really not that long. In many ways, technological limitations means a lot of these projects are still “toys” , just like microcomputers and Linux once were. But there is a ton of cryptography research and application that _might_ transform these from toy to the future. And again, with billions in daily volume, calling it a toy is unfair, it’s in between a toy and the future.

To clarify “not needing to trust a third party” - it’s better to phrase it as “needing to trust fewer parties and having more choice in who you trust.”

I’m amazed people can watch scam after scam from Wall St financiers and struggle to understand why some find axing the middlemen appealing.

Anyone can list 100 reasons why it won’t work long term. The question is do you use that list as a reason to write it all off, or a list of problems to solve as an engineer. HN obviously leans towards the former.

The love of lists seems to explain the explosion of Twitter threads. Some of them are long form messages spliced up but most of them are some form of lists.

I’m personally still partial to a good old blog posts with paragraphs, both for writing and reading, but like the author I can’t help but notice that readers love lists.

Check out Vitaliks post on the topic :

https://vitalik.ca/general/2021/01/11/recovery.html

Check out Vitaliks post on social recovery wallets https://vitalik.ca/general/2021/01/11/recovery.html

> The fact that GOOG doesn't pay dividends is not important or meaningful.

I actually agree with you here. The thing is, the fact that Bitcoin or Ethereum doesn't have "intrinsic value" is not important or meaningful either. Obviously GOOG is valuable and obviously BTC is valuable, and obviously ETH is valuable, but in both cases you can play semantic games to toss around economic terms like "greater fool theory", without understanding where the fundamental value actually emerges from, to pretend like they're worthless.

If you actually follow the activity around Bitcoin and Ethereum, the idea that these coins are suddenly going to drop to $0 and nobody will want one is about as absurd as that happening to GOOG.

No, you’re in fact the one whose incorrect. Buying a share of GOOG in no way gives you rights to their trademarks or other IP.

In fact you don’t even get voting rights, unlike traditional stock. GOOG is the one that’s been “hollowed out “.

Likewise, owning a share of GOOG in no way gives you access to their cash.

There is one and only one thing you can do with a share of GOOG - hope to sell it to someone else for more money.

On the flip side, owning ETH gives you a stake which actually does give you governance rights, either via your own validator or a staking pool. And it gives you access to many other ways to participate in the network such as DeFi.

The dotcom boom was mostly fueled by rabid speculation. Today, people love their tech stocks, even though none of them pay dividends - the only reason to buy GOOG or AMZN is the hope that you can sell that stock down the line to a "greater fool". Some might call it a Ponzi scheme since only way investors make money off AMZN or GOOG stock is from new investors money coming in.

When I look at projects like ethereum, I see a lot of excitement and innovation around how cooperative game theory can innovate new business models and ways of coordination.

The way Twitter repeatedly rugpulled devs is a perfect example of how broken the web2 model can be. And now the founder of Twitter wants to criticize crypto projects for taking VC money, as if Twitter somehow didn't take VC money and enrich those VCs in its rise to power? The hypocrisy is astounding.

As far as the services, the goal posts keep moving. First there's no application, then store of value is an application, then art collection, then decentralize automatic market makers, then social tokens, then decentralized autonomous organizations. These are young, immature ideas but clearly stuff is happening. Whether it will live up to the grandiose promises, we can't say. But if you want to complain about the web3 advocates making grandiose promises about how their code is going to change the world, well, they stole that playbook directly from the likes of the Google and Twitter founders.

When Dorsey and Google execs complain about crypto, I'm reminded of a quote from HBO's Silicon Valley:

"I don't want to live in a world where someone else is making the world a better place better than we are."

If startups wanted to convince employees their equity was valuable, they easily could simply by adding protections to avoid many of the bad scenarios. For examples, there’s no good reason why employees shouldnt get preferred shares. The whole reason common vs preferred share distinction exists is so that the founders can’t just immediately sell the company for a bad deal and cheat the investors but employees don’t have that option. They’re told they are not given preferred shares because they’re not risking money the breath after they’re told they should turn down public liquid RSUs at a big company for startup equity instead.

Startup equity is very unfortunately broken by design. It’s a malicious feature, not a bug.

It’s a shame because many people myself included would pick the smaller company gamble over the big company if the terms were at least fair. But too many naive engineers accept awful financial instruments designed by VCs to screw employees and poison the well for everyone.

If you have a choice between public stock and common ISOs you should pick public stock 99.9999% of the time.

I don't think it's true that high paid jobs tend to be interesting, but they do tend to be impactful. A lot of big tech jobs are about cleaning up infrastructure technical debt, in many cases the work is pretty boring, but between the number of engineers at these companies and the scale of their products, even very small improvements can be force multiplied. However, even though there is a lot of downstream impact, it won't be super visible to you. It's the classic "if an engineer can save 15 minutes of every Google engineer's time every day, that's worth billions to the company."

But to more broadly agree with your point, I do think if one job pays you a lot more than the other, on average that's because they value you more, and they will be a lot less likely to waste your time. If you work for 50k a year, your CEO will be happy to have you work on all sorts of stupid things because you're cheap so why not. If they pay you 500k a year, they're going to want to make sure that whatever you work on is actually important to the business because now when the accountants add up costs you are going to be more significant. Furthermore, psychologically it's hard not for a boss to value someone making 500k more than one making 50k a year.

I'm sure there are many, many exceptions where people are paid poorly and working on super interesting or important things (medical research being a common example), or paid well to work on meaningless stuff, but as a rule of thumb, I do think higher paying jobs tend to be better jobs independently of the bigger paycheck.

I understand your sentiment, but on the flip side, if you browse around Reddit you’ll quickly see people going all in on random meme coins like SHIB with comments like “I’ll be in a resort in Bali or homeless.” The problem is if they lose their gamble, which is pretty easy considering the huge number of scams and pump and dumps, then when they are in fact homeless, it doesn’t just affect them. Because then people will say we need to take care of the homeless with social programs which taxpayers and people who didn’t YOLO on memecoins have to shoulder the burden of.

You can’t really have it both ways, you can’t have a super strong social safety net for people who lose it all and a system where it’s easy to gamble your life savings on incredibly unregulated markets and incentive all the scammers to come out of the woodwork.

There’s a reason IPOs are much later these days than during the Dotcom boom precisely because too many retail investors lost their shirt to pump and dumps which at scale leads to broader social instability.

The main benefit of NextJS is that it makes it very easy to write code once and quickly swap whether it's going to be statically generated ahead of time, rendered on the server, or rendered dynamically on the client. It also comes with a bunch of features you'll likely need for a modern web page like image optimization and a really nice PaaS to deploy it to.

But I was complaining about the exact opposite- I have a hard time finding potential design cofounders as an engineer, and every hackathon and similar events seems to be filled with engineers , but short on designers even though designers win pitches more than engineers (you can fake implementation for a demo but can’t fake design).

The fact that it’s harder to find designers than engineers for founders despite the opposite dynamic in the job market is what I was saying was surprising.

The platform does not have many designers looking for technical founders. As I said, it’s filled with MBAs. When pressed what skills they would contribute, they say things like “backend finance.” Usually either they have no money but need you to build a complex project before you can think about raising money (raising the question why not just build it and pitch VCs yourself), or they already raised money but want to give you some absurdly low percentage (you’re not really cofounder).

Are there people in there that would be able to raise money /sell the product if you built it? Maybe but others have discussed why that’s the hardest of all skills to vet for.

Either way if the platform was filled with strong designers interested in partnering with technical cofounders I’d be much happier and speak more highly of it.

Fair points, but a note on AirBnB with the caveat that I only know the public information and never worked there. To me, their only moat is their brand and the network effects they built with that brand. The famous stories are them fundraising by selling Obama themed cheerios, and going to apartments helping hosts take better pictures. Neither of those things are the “cranking out iPhone mocks” you’d expect from a freelance designer but seem like stories fundamentally linked to their design backgrounds and ability to tell stories in the visual world. Strong design probably matters more for something like Snapchat than Coinbase so I’m sure it depends. You’re probably right in your explanation for why Eng skills are more in demand but I still suspect you’re understating the impact of a great designer.

By that logic, you can hire a freelance dev to do all the programming too. There are some pretty good people out there who aren't that expensive. But there's a bunch of risks with that approach and they are similar to the ones where you outsource design - you will constantly have to re-hire if you need pivots, there might be major friction switching people, and the costs can start to add up. It's not 100% analogous, freelance design is a bit cheaper, it's easier to spot bad designers etc, but I think the analogy somewhat holds.

Consider that the #1 market cap YC company of all time - AirBnB - was started by 2 founders with a background in design.

Well even if I want to find someone local, doesn’t mean I can just walk outside and find them, especially since the pandemic shut down most events in the Bay Area. Dating apps seem pretty popular and I think vast majority of people on those platforms are looking for someone local.

I briefly floated a profile on there, though I decided to take a new full time job and focus on that for a while instead.

While on the platform, I was flooded with non technical people who didn’t have much besides an idea and an MBA. I guess if they could credibly raise money/sell product then maybe it would make sense but I was highly skeptical of the value add, and I got the feeling a lot of them were looking for a free dev to build some big product that they now get to shop around to investors, taking 50% of the equity for doing so.

The #1 skill I am always hoping to find someone with that I vibe with is design. No matter what you build, both UX design and a consistent visual aesthetic are very important. For some reason SWEs make more money than designers in industry but whenever you’re at startup network events it always seems SWEs outnumber designers 3 to 1. And likewise I didn’t see many people with design portfolios on this platform.

I’m surprised so few people care about where there cofounder lives. It seems YC is geared towards situations where you go all in, work hard full time on the project etc. That would be a situation where I would most care about a strong relationship that I think would be much easier to build in person.

The author talks about linking it to human sacrifice, which is definitely an angle worth considering. But there are some others-

1. Immediacy is one of the ten principles of burning man, and burning the art when finished forces people to appreciate it in the moment while it still exists. It’s worth noting a lot of art is burned, not just the man.

2. Fire is aesthetically pleasing . A huge percentage of the burning man crowd is into fire dancing and fire arts

3. They burn the Man because that’s what they do - similar to abstract art you can just accept how that makes you feel rather than needing to cognitively analyze it.

4. It connects to both eastern and western religion. Impermanence is a pillar of Buddhism and a man publicly dying is a pillar of Christianity

Of course I don’t think drawing a connection to human sacrifice is outlandish as a ritualistic burning of a symbolic man during a large semi-spiritual gathering is easy to connect to human sacrifice. But it’s worth considering a few other ideas.

Yet YouTube, Instagram, Pinterest, Reddit, Robinhood, DoorDash, and Lyft backend were originally primarily written in Python. What’s funny is that nobody can really deny Python is slow yet somehow the biggest websites in the world were written in it. More proof that Worse Is Better?

I strongly agree with you.

Reminds me of something I read that claimed when drum machines came out, the music industry thought it was the end of drummers. Until people realized that drummers tended to be the best people at programming cool beats on the drum machine.

Every single technological advancement meant to make technology more accessible and eliminate expertise has instead only redefined what expertise means. And the overall trend has been a lot more work opportunities created, not less.

It’s closer to a web version of pio solver.

Equity and range evaluators got a lot less useful because people figured how to make software that solves the full (simplified) game tree. This is a superior approach to anything other than all-in spots (which have game tree of one nodes.)

Most serious players these days primarily use software like pio and monker because the competition has gotten tougher and old equity calculators are like using a graphing calculator vs a

UML is closely associated with Object Oriented programming. In the 2000s, OO programming was all the rage, and having a mental notebook of “design patterns” was considered the hallmark of a real software engineer.

Since then, OO has lost a lot of its luster, the languages most associated with it like Java and C++ are considered somewhat clunky and uncool. Functional programming took over as the “smart” paradigm to talk about, even if in practice most popular programming languages borrow a mix of imperative, OO, and functional features.

UML was also popular in a time when you could make a SQL database and just use that. The introduction of more database varieties complicates things.

Besides OO, UML was also strongly associated with “visual” programming done by non programmers. This paradigm is doomed to be reinvented and pushed as the next big thing every 3 years until eternity, which we currently see with whatever the latest no code trend is, but in the 2000s was associated with UML.

Finally, language independent data formats like JSON, Thrift, and protobufs got popular. Why make a UML diagram when I can just make the actual data structures then immediately have that as a binary format?

To the extent UML means “data modeling”, people still do that. But it became associate with a lot of cultural baggage, and most of that baggage was on the losing end of a lot of technology mindshare battles.

Really nice article and demos, very informative.

Maybe unsurprisingly, noise is a really important aspect of audio and music as well. A lot of synthesizers have a parameter for adding white or pink noise to sounds. What's interesting is in a lot of music, my ears would never pick out the noise without actively listening very closely, but if you take the noise away the music sounds way more plain and empty in a very subtle way.