Show HN: Microtasks | Hacker News Reader

Show HN: Microtasks

utask.org

22 pointsgargarplex21 comments

from their page:

>taking a designer's mockup screenshot and building a fully functional React app with clean HTML/CSS– $120

build an entire product... for $120. assuming 1 full work day to complete (which is sort of unreasonable depending on the complexity of the app) this would be $15/hour. so minimum wage. Why not work at burger king instead?

> taking a designer's mockup screenshot and building a fully functional Vue app clean HTML/CSS and a Google Sheets API integration– $185

again, build their entire product PLUS integrate it into google sheets... for $185. hmmmmm...

> writing scrapers to organize data and then build an interface to make it searchable by location– $200

custom scrapers for whatever page your getting the data from, and a custom interface to get the data... for $200.

so these are really low prices for full projects, which leads me to believe that only non-american programmers will actually use this site because these projects are literally offering to pay less than minimum wage here. which, based on my experience on other similar sites, will lead to a large rate of fraud, payment problems, and code that straight up doesn't work.

good luck.

> $15/hour. so minimum wage.

$15/hour is significantly above the locally-applicable minimum wage in most of the US. It's even a little above it in San Francisco.

this is the one thing that you focused on in my post?

15/hour is still ridiculously low compared to other software dev work.

> taking a designer's mockup screenshot and building a fully functional Vue app clean HTML/CSS and a Google Sheets API integration– $185

* Screenshot to CSS: 30 minutes

* Google Sheets API to clean JSON structure: 30 minutes

* Vue app: 30 minutes

-----------------------------

Total time: 1.5h

Effective hourly rate: 123.33

Let's same I'm a 10x programmer, but you're a 1x programmer and it takes you 15 hours. Then your effective hourly rate would be 12.33. I think the flat fee rewards programmers who are more efficient than others.

I've been part of organizations where slow developers make just as much as the so-called rockstars. In what world is that fair? I'm not ranting about "slow and methodical" engineers who do everything super-correctly; I'm referring to "Wally", the character from Dilbert.

I have a bit of a moral agenda with this project, as well. There's no reason that slow programmers should be denied work, but they should not be compensated at the same clip as fast programmers. Great developers often go into management if they want to make more money. This dynamic punishes developers for getting "good" and drains the programming market of its best talent.

Why would enough 10x programmers waste their time on these low paying projects to make it a viable site?

Maybe one or two, sure, when they need a quick couple hundred, but not consistently. Which means if I'm looking for someone to do my projects, it's going to be fulfilled by someone who is expecting low pay, which means, on average, their not going to be a 10x programmer.

I just don't see this being proper valuations for the example projects. They are too low, which makes me question the finished results' quality.

I understand that you think it won't work for you. That's why we have the system of no payment until your work is complete.

The supply/demand economics are such that enough programmers are available to bang out work in exchange for quick cash.

Honestly, have you ever posted a question on Stack Overflow in a popular category? People can't wait to post answers.. for free. Programmers love to show off their talent and smarts.

Now, imagine that they were being paid. And with Code For Cash, they are. They get money as well as ego shoutouts in Slack, etc. It's almost as if it's like Stack Overflow's reputation system, but backed by cash as well as ego :P

Though actual gamification, as well as Stack Overflow and Github integration could be useful here.

There might be a role for a go-to programmer for a specific type of task, or somebody that would be really good at delegating to a specific sub-team, this could even be extended to Scrum-as-a-service as has been tried by a couple of companies now.

This could be built up organically based on reputation and completion of a specific type of task or using experience using a specific technology.

It would also be neat to be able to pick up a more complete task like building an app to spec and then sub-task out the HTML and/or CSS to somebody with more UX talent but less application development skills (not everybody is or should be a full-stack developer.)

I personally know a few people that are killer at UX/UI and barely competent in JQuery, so I would love to hand one of them the task of building a React component from a scaffold with all it's CSS goodness, or taking some that I built and constructing a well-ordered page around them.

Over the past few months, I've been leveraging the Code For Cash community as my own personal freelancer pool. This has provided a huge boost in my freelance consulting (software dev) business: it's like I have an armada of individual contributors that I can delegate tasks to, on-demand.

Some example microtasks I did:

$120– translate designer's mockup image into react+html+css app (https://utask.org/img/microtask1.png)

$200– take links to websites with VC funds name, and build a webapp to make them searchable by location so you can find venture capital funds near you (https://codefor.cash/vc_firms_list)

So now, I present to the community, for your consumption: Microtasks!

As a programmer that works alone and by myself, I would sometimes like to outsource parts of my work, and there are indeed very modularized parts that could be outsourced like that, but I'm very very afraid of the quality of the code that will come back from something like that.

Also, even if the code is of good quality, maybe it has lots of boilerplate I dislike, or is written in a style too different from mine.

Hey, thanks for the feedback. The style of the code that I get back is almost always different from mine. However, it tends to be modular and functional; i.e., I can spend ~5 minutes integrating it as a feature instead of a few uninterrupted hours. So it's a trade off.

Update..

Edited the "Post a microtask" interface

New screenshot: https://imgur.com/a/IzbBx

You can specify acceptance criteria and also mention coding style.

The micro appealed to me but the typical projects showcased on the home do not look like micro to me, as others have already pointed out. How do you assess the real level of complexity of a project in order to determine its price ? Is it all on the customer side ? Nothing mean, I am just curious tbh.

Hey, so I've been programming (as a hobbyist and a professional) for 20 years. The key thing to note is that all of these are things that don't require access to an existing codebase. All these projects could be done within 4 hours by someone talented and reasonably experienced. Combine that with the fact that we take care of finding the clients and offering same day payment, I think it's a fair deal.

>How do you assess the real level of complexity of a project in order to determine its price ?

One thing I should mention is, as I said, I'm a super experienced programmer [1], so I analyze each project and do the work of dissecting into what needs to be done. I provide links to helpful APIs and help provide suggestions of ways to tackle the problem.

So once I know what needs to be done, I do a bottom up estimate of how long it would take for me to do each microtask. I then multiply each time estimate by the market hourly converted rate for a San Francisco Silicon Valley startup software engineer and sum. I put that bid into the Slack channel, and if someone good takes it, that's that, but if nobody bites, I raise it by ~$15 every so often.

There are 400+ people in Slack currently, and at any given time there's 20+ experienced developers online. Tasks are usually snatched up pretty quickly. The bottleneck in the system currently isn't experienced devs, but qualified businesses and entrepreneurs.

[1] But not a truly great one, like Ada Lovelace, Adam D'Angelo or John Carmack.

Hey ! Thanks a lot for the reply, looks interesting I think I'll give it a shot then.

Examples of one of these microtasks might appeal to both potential developers and business clients. As it is, there is little on your website to explain what it even is, other than a CTA to push me to the next step.

This could also be a useful evaluation/interview tool where the microtasks represent real-world problems, versus the pages of FizzBuzzing and stripped-down tasks currently in vogue.

That said, this personally seems useful to me as I'm working on a scripting/templating language that should vastly simplify fulfilment of these non-codebase dependent tasks, and this gives me a new way of thinking about that.

In my opinion this only serves to proletarianization of programmers.

Thanks for the feedback + opinion.

My perspective the market is already structured this way. In the freelance consulting world, I see two types of competent programmers:

1) People who can take high level business goals, articulate them into software APIs/specifications, and then implement the code

2) People who can take specifications and deliver code

People in the first group tend to earn upper middle class incomes. People in the second group already tend to earn proletarian incomes. The upside here for programmers is immediate payout for tightly-specified work.

If AI can write an interesting sentence, it's clear that soon enough it will be able to write software. Especially software that is a variation on a typical CRUD app. My prediction is the biggest value add activity will be having the experience to know what is possible and then knowing what and how to direct the AI to write.

Programmers are, generally, proletarian (wage laborers). This certainly recognizes that fact, but doesn't serve to advance it.

There's no reason that slow programmers should be denied work, but they should not be compensated at the same clip as fast programmers. Great developers often go into management if they want to make more money. The current market dynamic punishes developers for getting "good" and drains the programming market of its best talent.

Good idea, but the network effect is hard to overcome.

Could it be made into a "solve my programming problem" kind of marketplace? I think there is none of this out there.

Regarding network effect... I actually have that problem solved. 12 months ago, I coauthored a book called Software Engineer's Guide to Freelance Consulting which has been a bestseller on the Amazon marketplace. I started cultivating a Slack channel of developers. My next step was to write scrapers/crawlers of ~50 freelance programming job boards in order to keep the developers in the channel happy. And I have also tested advertising campaigns, and they're successful in attracting developers to sign up and complete their profiles. So if we need to scale the developer side of the market, that's a solved problem for now.

But I just haven't found the acquisition channel for the people who post the tasks. Perhaps I should focus on marketing to developers rather than middle managers..

Reply on news.ycombinator.com