HNHacker News
TopNewBestAskShowJobs

adameasterling

392 karma · joined September 26, 2018

cto @ https://middle.app

Email me: adam@middle.app

submissionscomments
adameasterling··on Ask HN: Skeptical about my company going “full serverless”. What am I missing?
Within Lambda, yes, sucks. This is why container-based serverless is so much more exciting. (Which Cloud Run offers.)

I'm in the early stages of this rearchitecture but so far I've had no difficulty with local development.

adameasterling··on Ask HN: Skeptical about my company going “full serverless”. What am I missing?
I have limited time before my next meeting so I'll type real quick:

GCP Cloud Run is like the best of both worlds between AWS ECS Fargate and AWS Lambda. (Yes, the comment is in the context of containers. Sort of.)

* Like Fargate, Cloud Run hosts containers and takes care of figuring out where they actually live. Unlike Fargate, you don't have to say exactly how many containers you want running at once; GCP will automatically scale the # of containers up and down based on HTTP load and will scale down to 0. This should make Cloud Run cheaper than Fargate. (If you want to hook up Fargate to a webserver and you don't have autoscale figured out, you'll have to keep a lot of workers alive doing nothing.)

* Like Lambda, Cloud Run bills by the amount of time spent processing at least one request. But unlike Lambda, Cloud Run lets one container handle more than one request at a time (it sucks to have to spin up a lot of Lambda invocations that do a bunch of IO). Web servers that are good at concurrency shine here. This should save money.

* Cloud Run has more generous limits in many respects than Lambda. Cloud Run lets you set up SIGTERM hooks, so you can do some cleanup logic in your container (to e.g. write performance data to a timeseries table or whatever).

That's Cloud Run. On the database side: GCP Firestore is very interesting and we're going to build a big feature around it. AWS has nothing like it. On the queue side of things: We're planning to build around GCP Cloud Tasks; We've more or less built Cloud Tasks ourselves using a mix of MySQL and AWS SQS (and it was hard and we haven't done a good job).

I'd love to start a Discord or something to discuss these thoughts more. It's so hard to get good practical information for system architects/CTO types who just need to hammer stuff out.

adameasterling··on Ask HN: Skeptical about my company going “full serverless”. What am I missing?
This is from the perspective of a small startup CTO:

We've used AWS Lambda for about 4 years, and it's been so good and so cheap that I'm shifting literally everything (except Redis) to serverless. Also, GCP has a better serverless offering (Cloud Run, Spanner), so we're switching from AWS to GCP to take advantage of that. I bet we're going to see a massive cost reduction, but we'll see.

Things I like about serverless (again, from the perspective of a very small startup, with 5 engineers, and me being the primary architect):

* It's so liberating to not worry about EC2 servers and autoscale and container orchestration myself. All our Cloud Formation templates add up to around 3,000 lines, which maybe doesn't sound like a lot, but it's a lot. There are tons of little configuration things to worry about, and it adds up. (Not to mention the sheer amount of time it took to learn.) ECS Fargate takes care of some of this, but it doesn't autoscale based on demand or anything (not without settings things up yourself). (This is a big reason why I want to switch to GCP: Cloud Run is like Fargate in that it runs containers, but unlike Fargate it autoscales from 0 based on load.)

* It's very cheap in practice, at least for loads like ours that respond to events: API services that sometimes see a lot of use and sometimes see very little use; queue consumers sometimes have a lot to do and sometimes have very little to do. AWS Lambda bills down to the milisecond in terms of resolution, and GCP Cloud Run/Cloud Funcitons bills down to the next 100 miliseconds. These are very fine resolutions and for us at least, we've seen costs be small.

* For database serverless products (like DynamoDB for example), it's very liberating to never have to think "Hm, do we have enough CPU provisioned?"

Things I don't like about serverless

* Pushing source code sucks. Lambda will just one day decide your version of Python or whatever isn't good enough and force your customers to upgrade all their user-written code to the latest Python version. (But! Cloud Run supports containers, and so this won't be a problem.)

adameasterling··on Ask HN: Who is hiring? (November 2022)
Well, this is cool. I had a convincing conversation with a squirrel. (Aside: Forest animals might be one of the most perfect use cases for this. A squirrel wouldn't be expected to remember facts for very long, and might make stuff up.)

Obviously gaming would be a huge use case. It's not hard to imagine a forest creatures RPG type game. (Though, an actual game would need some kind of direction to the conversation. The squirrel I spoke with didn't seem to have any objectives of his own; and didn't recognize when I wanted to politely end the conversation.)

Business use cases are very exciting as well. You'd definitely need to solve the "make stuff up" problem; possibly giving the chatbot access to a restricted set of facts it knows and nothing else. I wonder if that's within the black box, though; is it even possible to separate factual statements about the world from other components of conversation?

Really curious to see what the business model is.

adameasterling··on Electric scooter ban increased congestion by 10% in Atlanta
I love e-scooters. They're fun to ride around on and, when there are enough of them in an urban area, you can just rely on them being there for transit. I've been so disappointed by the public's reaction to scooters, though. They don't see how cars are significantly worse, in terms of the personal cost, public safety and destruction of urban spaces.

Like, just to put this in perspective, I did a bit of digging behind the Atlanta ban. It happened right after (and as a direct result of) these events:

- A 34 year old woman on an e-scooter was murdered* by a driver of a car, in a hit-and-run. [1]

- A 20 year old man on an e-scooter was killed by a negligent driver of an SUV who was charged with 2nd degree homicide. [2]

- A 37 year old man was killed by the driver of a bus who did not see him. Passengers on the bus say he was banging on the bus to alert the driver as he was crushed to death. I'm not able to find out if the bus driver was ever charged for this death. [3]

- A homeless man (of unknown age?) on an e-scooter was killed by a truck. In this case, it does look like the e-scooter ran a red light and the driver was not at fault. [4]

So in 2 out of 4 cases, clearly the car was at fault. The city reached first for not limiting cars in any way, but rather limiting scooters. I mean, I get it. There was no structure for doing anything to cars or the system that causes people to drive everywhere. And to be fair to Atlanta, the ban on scooters was only at night.

We know there are ways to structure cities and traffic laws to make safer spaces for pedestrians and cyclists. Bad saldy, here in America, we are so reluctant to reach for those options. We have so many parking lots, so many stroads, so few walkable spaces. Whenever there's conflict between cars and people, we blame people. And then build extra lanes and extra parking lots and it just gets worse.

By the way, this problem with cars happens here in San Francisco too: My friends and I have all seen an explosion of bad behavior by drivers, including a friend of mine who struck last year in a hit-and-run, breaking his femur. This is not just anecdote: Pedestrian deaths are on the rise here, while citations plummet. [5]

1. * I am not sure if "murder" is legally applicable in this case, but to me, if you strike and kill someone in your car and then run, you're a murderer. https://www.wsbtv.com/news/local/atlanta/scooter-rider-hit-k...

2. https://www.bikelaw.com/2019/06/driver-charged-in-death-of-e...

3. https://www.wsbtv.com/news/local/atlanta/scooter-rider-hit-k...

4. https://www.wsbtv.com/news/local/south-fulton-county/man-on-...

5. https://sfbayca.com/2022/07/05/traffic-deaths-in-san-francis...

adameasterling··on In Defense of Missile Defense
> suffering from the same insanity

It isn't exactly clear what this means, but when I read it I thought, "we all seem to be suddenly hysterical about nuclear war."

I listened to the most recent episode of the Arms Control Wonk podcast[1] last night. In this episode, Dr. Jeffrey Lewis observed that yes, paid media and social media is hysterical about nuclear war right now, but also, that the underlying emotional anxiety that fuels this sort of coverage/discussion is the natural result of nuclear deterrence.

It's supposed to feel tremendously frightening. He noted that the word itself, "deter," comes from a latin word that also gave us "terror." The tremendous anxiety that we all feel is totally appropriate given the threat. It feels like insanity, yes, but it actually shows that the system of deterrence, as terrible as it is, is still working.

1. https://www.armscontrolwonk.com/

adameasterling··on Nuclear Close Calls: Able Archer 83 (2018)
I recently read a fantastic book on Cold War brinksmanship. It was, "The Bomb: Presidents, Generals, and the Secret History of Nuclear War" by Fred Kaplan, published in 2020.

It's a fantastic book. I couldn't put it down. (Okay, technically, I listened to it. The audio is very good!)

Kaplan agrees that Able Archer is probably the closest we ever got to nuclear war. But, something else interesting happened at this very same time, in Autumn of '83. A made-for-TV movie was released called, "The Day After."

Kaplan argues, based on Reagan's own diaries, that it was this movie that was largely responsible for Reagan's sudden about-face in '84 and beyond. Here's what Reagan wrote the first time he watched it:

> (Monday, October 10, 1983) Columbus day. In the morning at Camp D. I ran the tape of the movie ABC is running on the air Nov. 20. It’s called “The Day After.” It has Lawrence Kansas wiped out in a nuclear war with Russia. It is powerfully done—all $7 mil. worth. It’s very effective & left me greatly depressed. So far they haven’t sold any of the 25 spot ads scheduled & I can see why. Whether it will be of help to the “anti nukes” or not, I cant say. My own reaction was one of our having to do all we can to have a deterrent & to see there is never a nuclear war. Back to W.H."

Almost immediately, Reagan cooled his bellicose rhetoric. In the next few years, he started a friendship with Gorbachev, the new Soviet premier. Together, they finally ended the arms race. We got 30+ years of not needing to worry about nuclear war, all because of a story (maybe). Stories are powerful.

adameasterling··on San Francisco braces for commercial real estate crash
Donated! Huge fan of the voter guide!
adameasterling··on Zapier: A $5B Unbundling Opportunity
So I'm pretty sure this is the technology revolution that we're all sleeping on.

Like, self-driving cars? VR? Meh.

Accessible programming? A safe environment for folks close to a business problem that lets them solve that business problem on their own?! That's cool.

Like for me, in my experience as a software engineer, there's basically four types of employees: The tech-naive domain expert, the tech-friendly domain expert, junior coders, and senior engineers.

The two ends of the spectrum should stay focused on they're good at: Doing business stuff, or coding on tough problems.

It's the two roles in the middle that interest me: Tech-friendly domain experts and junior coders.

Right now, in most companies, tech-friendly domain experts could totally figure out a tool like Zapier and do things that they'd normally need an engineer to do. But their company doesn't use something like Zapier, so they can't. They just have to talk to technical people and make them do it. But there lies long debugging loops and, just, slowness.

In some companies though, those that have embraced low-code platforms like Tray or Zapier or the product we made (https://docs.middle.app), they've enabled their tech-friendly domain experts to do tons of stuff they couldn't do before, without bothering people like myself, a more senior coder (I can work on platform things!).

ALSO, I've found that junior coders have a special role. We've hired non-programmers and taught them Python JUST so they could code new things in the developer side of our integration platform. It's a sandboxed environment with limited risk; they can go crazy coding whatever connectors or other things that those tech-friendly domain experts need. I don't have to be involved, outside of a code review every now and then. I know that they can't break anything other than the narrow thing they're working on.

It's neat. Neat stuff.

adameasterling··on Opposition to High Density Development
> Relative to single family homes, apartments are viewed as decreasing property values, increasing crime rates, lowering school quality, increasing traffic, and decreasing desirability.

The problem is most cities only allow two kinds of development: Low density or high density. We're missing the middle!

https://en.wikipedia.org/wiki/Missing_middle_housing

I've really enjoyed these two YouTube channels on the subject.

"Oh the Urbanity" has a great video on Montreal's "missing middle" https://www.youtube.com/watch?v=mYCAVmKzX10

And "Not Just Bikes" discusses a "missing middle" neighborhood in Toronto: https://www.youtube.com/watch?v=MWsGBRdK2N0

adameasterling··on Ask HN: Who is hiring? (July 2021)
Middle/Perkville | Full Stack Software Engineers | USA | Full-Time | REMOTE

Middle started as an in-house integration platform: an in-browser Python editor to make connectors, and an in-browser visual scripting language to do the connecting. It's been a huge success, allowing us to say "yes" to customer requests way quicker than before.

Now, we're making Middle available as its own product. Middle is in private beta but it is already generating substantial revenue from its first customers and has a strong pipeline of prospects that want to get started.

We need excellent engineers to accelerate us to a full rollout and beyond. This is a fantastic opportunity for someone who's excited about startups and brand new products, especially anyone interested in democratizing programming.

Us: Small team (12 employees), mostly in Pacific time (SF Bay Area)

Our stack: Python/Django, React/TypeScript, AWS/Lambda/ECS/CloudFormation/RDS(MySQL)/SQS

We're looking for full stack software engineers, particularly those with an expertise in React OR AWS. We're also open to solid generalists still early in their careers.

Senior Software Engineer - Full Stack, AWS or React expert (REMOTE) - 5+ YOE - $120-185k, 0.05-0.5% equity

Software Engineer - Full Stack (REMOTE) - 2+ YOE - $105-185k, 0.05%-0.5%

If you'd like to learn more or set up a call with me (CTO), check out our AngelList jobs page, and apply: https://angel.co/company/middleapp/jobs OR you can e-mail me (CTO) directly at adam@middle.app (Please put "HN Full Stack" somewhere in the title.)

adameasterling··on Internet down in Tumbler Ridge, B.C., after beaver chews through fibre cable
I completely agree!

Beavers make habitats! They create entire ecosystems! They turn semi-arid land into lush wetlands. Their lodges become homes for muskrats and otters. The wetlands they make support bird, and frog, and fish populations.

California used to have hundreds of thousands of them. The now-semiarid Central Valley looked very different before humans started trapping them in masse, and dammed rivers in the Sierras. I wish I could go back in time and see central California circa 1700. I sometimes wonder if we could restore some of what we lost.

adameasterling··on Software designers, not engineers: An interview from alternative universe
So, in some ways, it is not that different. It uses ideas that have been around for years: The core of the product is a visual programming language, plus a database, plus a low-code Python editor. All of which have, obviously, been done before.

But the problem is, no one has put them all together in quite the right way to satisfy our needs.

- We needed a platform that would let our non-programmer staff create complex integration-related features for customers. We needed an accessible VPL that was easy to get started in, but powerful if power was needed.

- We needed a lot of integrations to not-particularly-popular business software (which means, we needed the ability to write and/or extend them ourselves); we needed the integration editor to be accessible and easy to use by junior programmers.

- We needed data pulled from integrations to be stored and kept around for a while; we need to look up individuals, see their history and a history of triggered VPL logic; we needed to be able to update and bulk-execute VPL scripts against the data store; we needed full text search capability as well.

- We needed workflows/data organized into tenants with a user auth scheme that would let us invite business customers to see just their data and no-one else's, with view-only permission as well.

- We needed something that could scale up to handle a lot of work, and still be at a competitive price.

There isn't a platform that has all of these things. One or two, sure. But not all of them. My hope is that in 5 years b2b saas companies won't have to build much integration software themselves; there will be us, or someone like us, that'll handle it for them, and they can focus on whatever it is they actually made the company to do.

adameasterling··on Software designers, not engineers: An interview from alternative universe
I found myself nodding vigorously to Ms. Atkinson's (fictional) responses.

> You fixate on some technical solution and then completely forget what is the context in which you're building it. Good software designers basically treat all problems they face as ill-defined, even if they are fairly specific.

Oh my god, I nearly fell out of my chair. YES.

What we need are tools for software development that are highly flexible, easy to use, and safe. We need to put those tools in the hands of someone like a software designer from this blog post: A domain expert (and/or intuitive listener) who deeply understands the business, understands their customers' needs, and can hook up whatever logic is required themselves.

(And uhh, shameless plug, I'm actually working on a product right now that attempts to realize this dream. We're hiring platform engineers. See my profile for contact info/links.)

adameasterling··on Ask HN: Who is hiring? (April 2021)
Nice!

> I wonder if you've felt the pressure of implementing or embedding a scripting language client side? You're already developing a DSL, whether it's text or graphical is not 100% the point

So, yes.

Right now, there's a clear division of labor: Python goes in the "developer" half of the app, and is used to make data sources or "steps" that can be used in the VPL.

The visual scripting language only exists in the "account" side of the app.

So far, this division of labor is working, but we have encountered a few times where it would be convenient to just put Python code directly into the account side of the app.

About the part where you said, "whether it's text or graphical is not 100% the point" - I do think a graphical scripting interface is absolutely key, actually. Sometimes I forget how much of a wall an actual programming language is for folks. I've been blown away by how productive our support staff has been in our VPL.

adameasterling··on Ask HN: Who is hiring? (April 2021)
Going backwards in the questions here:

> Maybe a better question would be who are the target customers for this?

So far, our customers are software companies (and one brick and mortar company) in the fitness space. (That is to say, customers like ourselves: Perkville is such a company, a loyalty/rewards SaaS mostly in fitness.)

Long term, the target customers are any software companies that need to connect their product to a ton (hundreds or thousands) of different (large and small, diverse, needy) businesses.

> Or its more like a platform that manages user data and other services can use it as a store which executes a particular script when the state of the store changes?

You've anticipated a key differentiator: The data store. Other IPaaS systems don't keep around data that they've processed. They're entirely event driven. We needed the ability to look back in time (customers often want to do things like, oh, if it's the tenth gym check-in in a month, give the user some bonus points); we needed the ability to diff a new record against an old version of that record (customers often want to trigger rewards based on a membership status change for example); we needed the ability to bulk execute a workflow against a bunch of already stored data (we sometimes need to "back sync" a reward rule).

The visual scripts ("workflows") that our support team write are basically functions. How they get triggered is up to them. They can trigger a workflow automatically when a record is created or updated, or they can trigger them manually, or they can trigger them on a schedule, or they can trigger one from inside another. (On the roadmap is the ability to trigger a workflow from an API call, in a synchronous, atomic manner -- this is the final piece to totally replace Perkville's old hard coded integration system with Middle; automated point redemption.)

> Is it zapier but you can write a python script to automate a task?

So, as with Zapier, there are two halves of the app.

The first half is the "accounts" site. That's where all of a customer's own customers live, as separate accounts. Each account has workflows (this is the VPL), workflow triggers, a data store, and reporting screens. Python isn't written here. This is designed for our support team to go in and fulfill feature requests for customers.

The second half is the "developers" site. This is where programmers can create data sources and actions that then can be used on the account side. The goal here for us was to keep the interface as simple as possible. I wanted it so a programmer just needs to fill in a Python function, in an in-browser code editor. For a data source, they just need to handle, let's say, an input date range and an authentication dictionary, and then return (or yield ) records that match.

...

Hopefully that was informational? If you want to sign up for news, there's a sign-up form on our site: https://middle.app/

adameasterling··on Ask HN: Who is hiring? (April 2021)
So, we aren't quite ready to show screencaps (or other marketing material) to the whole world just yet. But I'll for sure make an HN post when that changes.

If you're cool with a text description though:

Basically, the core emotional impetus of this project for me begins with unfulfilled product enhancements for Perkville. Like, imagine a random 1-location Yoga studio wants to award bonus points on a member's 1 year anniversary. Sounds easy, right?

Well, saying "yes" to that customer means significant coding time, to store and update the membership start date, to handle edge case logic, to update all existing integrations to store the start date (and membership systems all store that in different ways, sometimes really obtuse and complicated ways), and then we need to build out interfaces for setting up the bonus. And what if a business wanted to award on the 6th month anniversary and not annually? Or for only certain membership types? And we need to QA, and add unit tests, and do code reviews, and on and on.

And we have hundreds of requests like that in the queue, with more coming in all the time. It's... too much.

Me and my engineers felt buried, for one. And our customers get unhappy too, that we just can't deliver fast enough for them. I really, really hate disappointing a customer. It hurts me, that this rewards platform we've been working on for years isn't doing what they want it to do. That we can't keep up with their imagination.

So, that was my emotional state 3 years ago when I sat my own boss down, and the product manager at the time, and was like, look, you need to let me solve this problem in a different way. We can't productize every single request that comes down. We need to figure out a way to build features for customers in a much more flexible way.

We looked at a bunch of tools in this space, like Zapier for example. But none of the tools were good enough. Zapier doesn't have the integrations we need (how long has their MindBody integration been in "in construction"?) and besides, it doesn't have the two-tier tenancy that we need. (We need to keep our customer's data all separate from each-other, with the ability to handle GDPR stuff on it as well.) And Zapier's (and all the others) tools for support really suck. You can't look up a user and see all the logic that's been executed for them. (People write into support all the time asking about their points.) Oh and on top of all that, the developer portal in Zapier really is not great. You have to download their SDK, become intimately familiar with it, and it is real complicated. (To their credit I think they've worked on this.) OH -- and Zapier is crazy expensive at the scale we need. We push a lot of events, and my estimate for putting everyone on Zapier came up to some insane number, like 50% of our revenue.

So, we decided. Let's just build our own and solve all of those problems. And so... we did. We got a v1 out pretty quick. We're able to hammer out integrations and updates to those integrations very quickly. Support can make features that we used to need engineering help on. There's still a ton of work to be done, but it's doing well enough that we've snagged some significant customers in the Fitness space. And it's really worked out for Perkville.

Anyways, that was quite the essay. That's the product. We're excited about it.

adameasterling··on Ask HN: Who is hiring? (April 2021)
Sweet! :) I think it has a lot of potential for sure. There are things we needed that none of the other "low code" platforms had, and it's been a lot of fun to build so far. And it's been great seeing people use it!
adameasterling··on Ask HN: Who is hiring? (April 2021)
Middle/Perkville | 2 Senior Software Engineer Roles | USA | Full-Time | REMOTE

Middle started as an in-house integration platform: an in-browser Python editor to make connectors, and an in-browser visual scripting language to do the connecting. It's been a huge success, allowing us to say "yes" to customer requests way quicker than before.

Now, we're making Middle available as its own product. Middle is in private beta but it is already generating substantial revenue from its first customers and has a strong pipeline of prospects that want to get started.

We need excellent engineers to accelerate us to a full rollout and beyond. This is a fantastic opportunity for someone who's excited about startups and brand new products, especially anyone interested in democratizing programming.

Us: Small team (12 employees), mostly in Pacific time

Our stack: Python/Django, React/TypeScript, AWS/Lambda/ECS/CloudFormation/Aurora(MySQL)/SQS

Two positions:

Senior Software Engineer - Backend, Infrastructure (REMOTE) - $120-150k, 0.2-0.5% equity

Senior Software Engineer - Frontend, React (REMOTE) - $120-150k, 0.2-0.5% equity

If you're interested, apply on our AngelList jobs page, here: https://angel.co/company/middleapp/jobs

adameasterling··on Ask HN: Who is hiring? (March 2021)
Middle/Perkville | Full-Stack Software Engineer | USA | Full-Time | REMOTE

Middle started as an in-house integration platform: an in-browser Python editor to make connectors, and an in-browser visual scripting language to do the connecting. It's been a huge success, allowing us to say "yes" to customer requests way quicker than before.

Now, we're making Middle available as its own product. Middle is in private beta but it is already generating substantial revenue from its first customers and has a strong pipeline of prospects that want to get started.

We need excellent engineers to accelerate us to a full rollout and beyond. This is a fantastic opportunity for someone who's excited about startups and brand new products, especially anyone interested in democratizing programming.

Us: Small team (12 employees), mostly in Pacific time

Our stack: Python/Django, React/TypeScript, AWS/Lambda/ECS/CloudFormation/Aurora(MySQL)/SQS

Benefits: Full remote, Flexible vacation (most take 3-4 weeks/year), $120-135K, Equity

If you're interested, check out our jobs board on AngelList:

https://angel.co/company/middleapp/jobs

← PreviousPage 2 of 2