DEV Community: Jeff Meyerson

Don't Trust Facebook

Jeff Meyerson — Sat, 07 Aug 2021 16:25:11 +0000

React is malware.

Across every interface, it gives Facebook direct access to everything you see and do.

Boycott Audible

Jeff Meyerson — Thu, 29 Jul 2021 12:57:47 +0000

I've decided not to publish "Move Fast" on Audible unless Audible figures out how to resolve our review queue stasis without our manual intervention.

You can find the book for free on the Software Engineering Daily podcast feed, as well as in our apps.

One of the key themes in the Move Fast book was how Facebook Engineering overcame the tyranny of the manual app store review through diplomacy, blitzkrieg, feature flagging, and Chuck Rossi.

I'm not going to submit to the inconvenience of the nightmarish Audible scleropoly backend.

https://www.softwaredaily.com/post/60eae282c6de2d000c601ee0/Full-Audiobook-Move-Fast-How-Facebook-Builds-Software

CODEVID-19: A Pandemic Hackathon

Jeff Meyerson — Wed, 18 Mar 2020 19:24:10 +0000

With the COVID-19 pandemic in full swing, many people are looking at a lot of free time stuck indoors. Avoid cabin fever and use that time to make a difference in your community and the fight against COVID-19 by joining CODEVID-19, the world’s first pandemic hackathon.

https://findcollabs.com/hackathon/codevid-19-isp21fkqtjupchx7kjed

TalkPriority: Podcasting Meets Group Calls

Jeff Meyerson — Sat, 12 Oct 2019 16:35:34 +0000

We are living in a podcast boom. It is a great time to start a podcast, and to listen to podcasts.

Podcasts are a wonderful medium. Podcasts allow us to listen to conversations between two people, and feel like we are present in the room with them. Podcasts have gained in popularity recently because of bluetooth headphones and 4G.

Bluetooth headphones and 4G also improve the experience of phone calls.

People mostly communicate through text these days. But phone calls are better than ever. Every week, I have a group call with my siblings. I am also spending more time on the phone with old friends, as well as my parents.

These phone calls are more enjoyable than listening to a podcast. They fill time similarly to podcasts.

While I am on a group call, I wash dishes or jog through the park. It is like listening to a podcast and interacting with it.

I like to call people who understand good conversational norms. This means that they listen to me and ask questions. They don’t monopolize the conversation. If I could have more phone calls like this, then I would be listening to fewer podcasts.

The Social Network For Phone Calls

What would a social network based around phone calls look like?

Here is my idea.

This is TalkPriority. The user flow works as follows:

User signs up
User is put into the “talk room”
Only one person can talk at a time. You push the green “Push to talk” button in order to talk
Everyone has a “priority”. You can interrupt anyone who is ranked lower than you. Your ranking is based on how little you talk.
When you are talking, your “seconds not talked” goes down. When you are not talking “seconds not talked” goes up.
In the “transcription room” your call is transcribed so that people can look through the history. This lets anyone who joins the call have context for what is going on.

I want to try this, so I am building it. It will be open source and built on FindCollabs here.

If you are interested in trying it out or helping to build it, join the FindCollabs or leave a comment here on DEV.

Non-Winner-Take-All Software Categories

Jeff Meyerson — Thu, 10 Oct 2019 14:10:45 +0000

Software companies are often talked about in a zero sum, winner-take-all context. And this make sense for many categories, like social networking, professional social networking, marketplaces, dating marketplaces, and ridesharing.

Not all software is winner-take-all. Many software categories lack network effects, or do not feature other forces that pull the category towards winner-take-all.

Examples include continuous delivery software, monitoring, logging, knowledge work gig economy, digital transformers, and (in the limit) cloud computing.

Are these categories bad investments? Many of them can be excellent investments:

Continuous delivery software costs almost nothing for the provider to host it; lock-in is significant; there are upsells over time
Monitoring and logging software can be priced by amount of data that is being stored; as data gets more heavy and multidimensional there are upsells in search, indexing, automated insights
Knowledge work gig economy platforms like Fiverr, Upwork, and Toptal are often not the sole platform where their workers hang out; if you are a designer on Fiverr, you are probably also looking for work on Upwork
Digital transformers like Pivotal, Mesosphere, and Red Hat use consulting services as an onramp to high-margin software subscriptions
Cloud computing infrastructure companies have great margins, and they will get better over time with economies of scale, better virtualization utilization, and data center robots; higher level cloud providers like Zeit and Heroku will allow different developers (often in the same org) to develop and deploy software however they want to

There are secular trends which make these businesses more appealing.

The abstractions for building a software company are becoming “no-ops”; if you build monitoring software or continuous delivery software, your tool probably won’t have outages very often; your op-ex is very low
Software never gets torn out of an organization, it just gets papered over
The TAM for software is gigantic yet unclear; how many paper companies will need continuous delivery software? How many oil refineries will need distributed tracing? How many plastic manufacturers are going to buy a CRM? How many insurance companies will give up with their in-house efforts at digital transformation and bring on Mesosphere or Pivotal?
Today’s Indie Hackers are tomorrow’s small businesses--and there will be tons of them. And each of them will need monitoring software, data engineering tools, and part time workers from Fiverr
Procurement will get much easier; today, much software procurement is gated by CIOs and CFOs who inhibit the innovation for the sake of cutting costs; more engineers will get trust, blank checks, and creative freedom for their cloud infrastructure (of course, the whole business of cloud cost management is a countervailing trend to this point so it’s possible the pendulum here has already swung)

It’s still great to be a winner-take-all company. But it’s not a requirement for success.

The Expo Hall Track

Jeff Meyerson — Wed, 09 Oct 2019 14:22:20 +0000

In the early days of Software Engineering Daily, Pranay and I went to the Facebook F8 conference.

We arrived at 8:45am, just as the doors were opening, so that we could be at the front of the line for food. It was like Willy Wonka’s Breakfast Factory--egg sandwiches, sweetbreads, fruit bowls, salads, oatmeal, granola, and bacon.

We ate a 5-course breakfast, then went to the all-day fridges and filled our backpacks with Frappuchinos and sparkling water. Then we walked to the espresso booth and staved off the food coma with 4-shot Americanos.

By the time of the “morning break”, we had pockets full of candy and brownie bites and dried quinoa energy bars. We alternated inhaling food with oxygen. We were trying to do a “2-cycle” day where you try to eat enough to force digestion to clear the way for more food by lunchtime.

In the evening, as the sun set over the pier, CHVRCHES was playing a live concert while waiters rotated around balancing trays of sushi and cupcakes. Pranay and I continued to gorge, standing with our bulging backpacks of free swag, our taste buds overstimulated.

We reflected--why were we there? How was it helping our business?

Common wisdom today is that company founders should avoid conferences. Y-Combinator emphasizes that conferences are distractions, celebrations of blind optionality and roulette wheels of networking. You end the day with a pocketful of business cards and unmet expectations.

And this is true, if you enter a conference without a strategy. But if you have a strategy, you can get a ton of value out of a conference.

Without a strategy, going to a conference is like scrolling through a social network. You will constantly have your senses titillated with free food, caffeine, and hype.

You can go to talks and be subjected to classroom-style lectures that will be posted to YouTube the next day. You can take the increasingly popular “hallway track” and socialize aimlessly with other attendees.

Or you can go to the expo hall--which is where the action is.

As I have gone to more conferences and realized the necessity of deterministic strategy, that strategy has increasingly centered around the expo hall.

I love the expo halls at tech conferences. At the expo hall, vendors have paid $15,000 - $500,000 in order to set up huge booths where they can demonstrate their products and talk to customers.

At the expo booths, you can find out how each company sees the future. You can compare those visions for the future with each other in real time.

You can talk to the expo booths, and see which companies managed to get engineers to speak in the marketing booth. You can see which companies have hired the marketers who can speak competently about products. You can assess the different sales funnels and see how the booth fits into the overall sales strategy of each different company.

You can study how booth grandeur, and swag, and free beer, and free churros impact a company’s ability to actually draw in potential customers. Are these companies handing out treats to paper over the fact that their technology is aimless and undifferentiated?

Keep an eye on which companies buy the cheap expo hall booths. At the next conference you attend, are they buying more expensive booths? When you read about a large funding round for a tech company, is that company immediately buying the $250,000 emerald package at AWS re:Invent? And do you judge that to be a wise bet?

The expo hall is the Chuck E. Cheese bazaar of the tech industry. Use it as meditative stimulus, and ruminate on the future of technology.

FindCollabs: Meet People, Create Software

Jeff Meyerson — Tue, 08 Oct 2019 15:09:09 +0000

Most of the apps we use every day are consumer applications: social networks, productivity apps, and marketplaces. These are the kind of apps that I want to build.

Consumer applications require more than code. They need design, product management, and high-level thinking.

I love to build software, and I love to build it in open source. But most of the software in open source is infrastructure. Open source software is operating systems, databases, and compilers.

I don't want to build a database, or a new programming language. I want to build consumer applications, like Audible or Instacart.

Where should we go to build consumer apps in the open together?

GitHub alone will not suffice. Consumer applications need designers, illustrators, and project managers. These users aren’t productive within GitHub, which is focused on the highly technical Git workflow.

FindCollabs is meant to solve this problem. FindCollabs is a place to find collaborators and build projects.

FindCollabs has the common social tools you expect: tagging, text chat, emojis, and video chat.

Consumer facing applications still need code so we integrate with GitHub.

On FindCollabs, you don’t need to be a programmer to start a project and build a team. If you are a product designer, you can post your idea along with wireframes and work with software engineers or low-code specialists to take your product to market.

The FindCollabs project finder allows people with similar interests to find projects to work on together.

Join FindCollabs and post a project. We’d love to see what you are building, whether it’s consumer open source software or not.

Enterprise Containers

Jeff Meyerson — Mon, 07 Oct 2019 17:36:51 +0000

Enterprise software infrastructure is going through a renaissance.

Kubernetes has captured the hearts and minds of the open source community. Cloud providers have successfully opened the wallets of even the most resistant enterprises. Vendors of all shapes and sizes are getting their phone calls returned by banks and insurance companies wishing to make progress on digital transformation.

Digital transformation requires an updating of all areas of application management. From integrations to security to logging to monitoring to data management to data access to authentication. Everything is getting upended.

Three years ago, the Docker container became the deployable instance du jour. Enterprises knew that they wanted containers, but they weren’t sure how to deploy and manage the vast quantities of containers. Then came the container orchestration wars, and enterprises were scared to commit heavily to any particular orchestration provider. Kubernetes won, and now enterprises are figuring out who to buy it from.

Enterprise developers are excited, because “cloud native” software is more fun to work with than the software generation prior. But there is mass confusion, as the options for infrastructure proliferate.

Yes, your enterprise is hybrid cloud and multicloud. But what are you putting in the cloud? How many Kubernetes clusters are you deploying? Can you put any data in the cloud, or are you terrified to move any data out of your on-prem data center?

Today, containers are still used under the covers, but it is less clear how we want to be manipulating them. Enterprises are excited about deploying and managing Kubernetes. But they might soon find that the period of time where we operated a Kubernetes cluster is as short-lived as the period of time where the state of the art was to manage your individual containers with scripts.

This morning I spoke with two people who gave me very different (though mutually compatible) visions of the Kubernetes ecosystem.

The first was Ville Aikas, who has worked on Google infrastructure for 11 years and is involved in the Knative project. Knative lets developers deploy serverless platforms on top of Kubernetes. Knative is highly tunable, and allows you to create a serverless platform for both short lived, auto-downspinning, event driven functions as well as longer lived container instances. The former is like AWS Lambda, the latter is like AWS Fargate.

These types of containers are two opposite ends of a spectrum. But along the entire spectrum, no application developer should be caring about any of this. The application developer should be writing business logic and the deployment platform should figure out the most economical way to run that business logic.

That’s what enterprise developers have to look forward to in the near future. Closer to the present tense, I also spoke with Brad Meiseles from VMware today. He told me about VMware’s PKS and gave me some projections on how Kubernetes sprawl might play out.

Today, a bank is figuring out its Kubernetes strategy, and they might be purchasing Kubernetes-as-a-service from multiple vendors. Additionally, different bank departments might be buying different Kubernetes-as-a-service for similar purposes. In other words--a bank is not necessarily purchasing Mesosphere OR Red Hat OpenShift. There might be one part of the bank that prefers Mesosphere and another part that prefers OpenShift. This is even more likely to happen given the increased procurement power of the actual developers (the “bottoms up” sales process).

Assuming the bank does not have a completely centralized procurement strategy, with a uniform, org-wide Kubernetes deployment plan, the bank could buy PKS, Mesosphere, and Platform9 to have Kubernetes for its on-prem servers and EKS, AKS, and GKE to have access to Kubernetes on all the cloud providers. This sounds like sprawl, but I don’t think it’s unrealistic. All of these platforms are developing very quickly. They have divergent feature sets that might make them appealing to different parts of the organization.

Here’s where Brad made a fascinating point: fast forward another 5-10 years, when Kubernetes spend is out of control for the bank, the bank could look to consolidate its Kubernetes deployments onto just a few providers. Since these providers are all running Kubernetes conformance tests, porting these workloads from one provider to another might not be too tough in the limit.

In any case, the developers at the bank hopefully won’t be thinking about containers at all: they will be thinking about applications.

Tools are for solving problems

Jeff Meyerson — Wed, 02 Oct 2019 13:56:49 +0000

I am resistant to learning new tools.

People keep telling me about Notion, and yeah I want to try it out. Notion probably does allow me to manage my documents better. But I don’t really have a document management problem right now. I have an execution problem. I know how to get things done, but I’m not doing them.

Technology is for solving your problems, not appearing fashionable. You don’t need AWS Lambda. You don’t need to go try out Golang. You don’t need to break up your monolith into microservices.

You need to figure out what to build, and then go build it.

If that requires you to learn a new technology, then go for it. But you should always be asking yourself: what problem am I solving right now?

If Google Docs solves your problem, you should use Google Docs. If GitHub solves your problem, you should use GitHub. You wouldn’t go shopping for the coolest new hammers at the hardware store when the hammer that you own does a great job of driving in a nail.

You will move faster if you focus on solving problems rather than trying out solutions to problems that you don’t have.

I have been working on side projects since college.

With every project I start, I have the problem of finding the right collaborator at the right time. When I write music, I can’t find musicians to work on music with me. When I start a software company, I can’t find a cofounder who wants to start it with me.

So I mostly work alone.

Working by yourself has its advantages. You do get to work as fast as you want. You get to own everything, so there will never be confusion over whose idea the project was.

But in my experience, it’s better to work with collaborators. There is the classic phrase: if you want to go fast, go alone. If you want to go far, go together. This is true, but is a false dichotomy. We can go fast and far.

Have you ever worked on a project with someone who you have amazing synergy with? Have you ever had the feeling of finishing each other’s sentences, but in a productive way rather than a romantic way? This is what it’s like to have a collaborator that you get along with on a project.

So if finding a collaborator is such an awesome feeling, why isn’t there a tool for doing that? I have been asking myself that question for the last eight years.

That’s why I started FindCollabs. FindCollabs is a place to collaborate on projects.

The problem of finding a collaborator turns out to be not unlike the problem of finding a boyfriend or girlfriend. You might have to go on a few bad dates in order to figure out what you like.

If you are looking to manage your documents better, we don’t have much to offer. But if you are looking for a collaborator, FindCollabs is the place to go.

Rejected from a top tech company? Build a side project.

Jeff Meyerson — Tue, 01 Oct 2019 17:43:28 +0000

You made it through the online coding problem nonsense. You endured the tedious whiteboarding problems. You sat before the condescending interviewer.

All for nothing. Time and time again. You've gone on too many interview loops to count, and yet you can't get a single job offer.

You understand data structures. You know the contract between .equals and hashcodes. You can breadth-first search a binary tree. So why aren't you having any success?

Your confidence is battered. Your interviewers smell the fear on you, and they don't like it. So how do you come back from all of this rejection?

You need to build something.

Build a complicated, hulking four-month deliverable that nobody else commissioned and that nobody else will care about. Make a game, or model an elevator system, or an economic phenomenon.

You will learn so much, but that's not where the most value comes from.

Building something big insulates your ego.

If you have built and shipped something cool and unique on your own, nobody can deny your identity as an engineer, even if you sometimes forget how to find all the subsets of an int array that sum to k.

This is so important--if you lose confidence that you actually belong at a strong company, you are dead in the water.

I've been where you're at.

Spending months and months preparing for interviews, doggedly memorizing the cookbook answers for how to rotate a matrix, find anagrams, and so on. Cracking the Code Interview is an awesome book, but also reinforces the status quo of a certain kind of uninspired, assembly-line narrowness to the interview process.

The absurd, outdated reality of software company hiring.

It's really sad that a lot of companies think this is the best way to screen people. I have to believe it's not.

I've failed in more final round interviews than literally everyone else I can think of, and I bust my ass and write code all weekend and soak myself in engineering information.

But it's OK.

Their loss. I've built stuff, and so many of the people I know who can close on these positions at FB, GOOG, etc have not. They are either scared to build from the ground up, or they continually start and can never push through to ship a project.

If I hadn't built and shipped cool, difficult projects, I wouldn't have this sense of validation and internal strength.

I've learned far more from side projects than anything a university or a tech company taught me. I think that we should be spending more time on side projects. That's why I started FindCollabs.

FindCollabs is a place to build projects and find collaborators. On FindCollabs, we take our work seriously and we help each other to design, invent, and engineer the systems of our dreams.

You can integrate with GitHub, post your project, and start working with peers to build whatever you want.

If you don't know what to work on, you can find a project that already exists and find friends to build things with.

FindCollabs: build projects, find collaborators.

The Problem With Side Projects

Jeff Meyerson — Tue, 01 Oct 2019 14:58:14 +0000

In college, I started building side projects.

I woke up early every weekend, and worked on whatever inspired me. I wrote music, invented recipes, and created a stock trading music rhythm game on Android. I wrote fiction about elevators that moved horizontally, and robot investment bankers.

On weekends, I was extremely productive. But weekdays were dreadful.

I attended University of Texas at Austin and studied computer science. In these academic CS classes, I was getting destroyed. I always felt behind, and could never focus on my school projects intently enough to complete them. I was barely keeping my grades afloat by copying answers from generous peers.

Banner Warhol, one of my first side projects

I also felt like something was deeply wrong. Why was it that these tools of creativity were being taught like an education in paint-by-numbers?

Computer science is about creating new software, not rebuilding stuff that has already been proven to work.

For my operating systems class, my partner and I were so disinterested in building a file system that we found the solution to our final project online, and spent all of our pair programming time obfuscating the miserable C code.

My classes were tedious. But my weekends were liberating, and empowering. Computer science gave me the tools to bring my ideas to life in a way that I had never experienced before.

On weekends, I would work from sunrise to sunset. I went for entire 48 hour periods without interaction with another soul, except for cats and baristas. My side projects were useless, and sometimes borderline insane. But I was achieving a level of flow state that was without comparison.

I barely graduated, and as I entered the corporate workforce, my life followed a similar pattern. Unable to achieve the level of obedience required of an entry level Java developer, I floundered at job after job.

Weekdays were boring. But every weekend was a glorious exploration of creative ideas. Alone, I toiled away on projects that gradually rose in quality and sanity. Eventually I started a business and was able to use my creativity to fulfill my economic requirements.

But that only worsened the loneliness. My digital business required very little interaction with the outside world, and now I could go entire weeks without social interaction. As loneliness became an acute problem, I looked for online communities where I would fit in, and could not find one.

That is why I started FindCollabs.

The problem with side projects is that we do them alone. We barricade ourselves in our own mental palaces. We embrace a creative, solipsistic indulgence: writing code that nobody will ever use, music that nobody will ever hear, and stories that nobody will ever read.

It is time for us to work together on our side projects, rather than in a self-gratifying cocoon. It is time for us to work in harmony rather than parity.

It is time for us to find collabs.