HN user

greendata

251 karma
Posts3
Comments93
View on HN

Cool! I don't understand the exact "schema" issue. Is that your database schema? Having written frameworks for deploying lambda in the past, I can say as recently as 2018 lambda was fairly hard to debug and in general using AWS is difficult for small companies. It's really designed for big business and large organizations where it works really well.

There are some workarounds, for example if your data ingestion is a lot of users uploading excel files, you can just upload from the client side directly to S3 with a pre-signed S3 URL without having to touch lambda. Then you can process the data on S3 using lambda or whatever you want. There are other strategies if you're ingesting real time data.

Yeah, if you're just using plain TypeScript you can certainly re-write crud apps all day and that definitely happens, in fact it's almost certain to happen.

Thanks for sharing. Following along to see how this works for you.

What stack did you use? On the backend, certain frameworks like Rails/Django can reduce this issue, but certainly for Rails you'll always need a dev around and hopefully the same one who built it. That's very hard to keep.

There's still a lot of chaos on the front end and the favored stack changes every few years. The devs three years from now might not be as familiar with the stack you used. It's honestly a nightmare at the small business level for front end.

As a dev, what's happening now with Zapier, Bubble, Tray.io, etc mirrors what we saw with static sites and Wix, squarespace ten years ago. They are not really at parity for SAAS apps but soon will be and then in a few years they'll be even better. Around 2000 you could easily spend 100,000 on a static website of a quality you can get for free now.

The thing with Bubble etc is that to remain competitive you will likely want to add functionality and anything advanced, anything that really moves the company beyond competitors, will require writing code (or buying an expensive service) and that's ok to add later when you've validated the idea. The downside of this is that's it's so easy to make a SAAS app that your profits are much lower. Apps that could have generated 100k/month are now doing 10k/month and a lot of devs have moved onto different problems like blockchain/defi/etc. We've already had great no code in wordpress for a long time and now it's just moving onto serious SAAS apps.

The new paradigm is definitely going to be Bubble/Zapier/etc plus the code that makes your app profitable added as necessary when necessary.

But how do they receive the credit card payments? What company is the one that actually receives the credit card payments from the consumer? Is it the "authorized distributors"? Or is the bank account that receives the CC payment linked to the parent company? It seems that states could go after the location that actually receives the credit card payment (if it's not a Caribbean corporation).

That's a great model. Do you use cloudformation each for deployment? If so, have you thought of creating a single cloudformation template for the whole deployment so you can do the entire deployment in one stack update?

Have you encountered any issues to watch out for when only using one APIG for each environment (150 micro-services). Have you encountered any downsides to doing this versus 1 micro-service to 1 APIG? I'm also running into the Gateway throttle limits and I think deploying many micro-services (like you have done) to 1 APIG is the best solution.

I agree. The bootcamps have a probably done a great job of flooding the market at the low level. It's pretty clear that bootcamps are the new ITT tech. I think a lot these bootcamps are selling false promises and preying on people trying to better themselves.

They do better work at that wage. You will definitely find well-performing local workers at a higher wage. Responsible and sane local workers won't work for that wage. According to the Wall Street Journal they were paying $750/month plus room and board. That's too low for even the imported workers which is the reason they are "running away." These farmers are essentially recreating serfdom.

Django 1.9 Released 11 years ago

I mostly agree. At this point they are both excellent frameworks with all similar features. Django has a really nice admin out of the box that some people love.

Rails is built with a "convention over configuration" philosophy. Rails apps tend to look similar whereas Django apps tend to be setup in a number of different ways. The downside of Rail's "convention over configuration" setup is that devs can get things up and running fast and then get frustrated and lost if something breaks....but it's really not that hard to dig around the code. With Django new devs get lost first and figure everything out. After a few months of heavy use and breaking things you won't notice much of a difference between the two. I find rails a bit easier to work with but I like python so I use django/flask.

You'll definitely find specialists in python but they'll also be experts in specific frameworks or packages like Django/Pylons/Pyramid etc. I'd say the same thing goes for ruby...if you're a ruby expert it's almost explicit that you'll at least have an intermediate knowledge of rails.

I don't think you can advertise that those grow lights or fertilizer are to be used for specifically illegal activities. One can use grow lights for many different plants. This company is simply poorly named and the marketing is aimed specifically at marijuana which is unfortunately illegal on a federal level.

Indeed there are. Many businesses are committing crimes without even realizing it, especially things like structuring. These crimes are often selectively enforced.

Many US cities are near bankrupt and probably a lot more will be in the next ten years. There will be enormous political pressure to cut costs. One way to cut costs is to reduce pensions or fail to fulfill pension promises. I don't advocate this and in fact I think the police are being setup in some ways. By allowing some police to engage in what looks like theft the general public will be amenable to pension cuts in the future.

Just see Detroit as an example of our possible future.

Meth is not the worst drug. In fact methamphetamine comes in legal prescription form as Desoxyn and is used daily for many with ADD and narcolepsy. I think that's a mistake to even touch that stuff but it's true. I think you could make an argument that commonly used drugs like Clonazepam (Klonopin) are far more dangerous. Unsupervised withdrawal is much more dangerous than heroin and it plays a major role in overdose deaths when mixed with opiates. The point is that all drugs to some degree are dangerous. They should all be legal and used with extreme caution if at all.

You should probably fix your website before you extol the virtues of outsourcing. That said, I think immigration should be more open to everyone. Big companies get H1-B workers bring wages down for engineers and programmers. Agriculture, construction and the restaurant industry also thrive on legal and illegal immigration and also bring the wages down in those industries. Doctors, lawyers, accounts, etc exist in a protected class of workers. We should open the gates for H1-B style doctors, lawyers, accountants, and managers.

It has the opposite affect. If you don't obtain gainful the loans will destroy your credit score. You absolutely need to take any job that will have you. There is no alternative.

US labor is becoming so cheap that offshoring doesn't make as much sense as it used to 10 year ago.

Exactly, this does nothing. Let's not forget the US is operating on the same level as China. The US spies on China as much China spies on us. We spy on our own citizens as much as China does. The only real difference is that the US doesn't actively censor internet but I expect that to change.

Ripe for Disruption 12 years ago

Google Adwords is absolutely ripe for disruption. Actually all of Google's services, except Mail and Search, are not great.

Have to agree. Great article but it just confirmed my decision to stay away from app engine. What's the purpose of paying for something that's both harder to use and more costly than a traditional server? Datastore is great but I can get the benefits at a quarter of the price. At least with Heroku you have something that's easier to use.