Show HN: Get Paid to Build Your Next Side Project
demandrush.com
demandrush.com
This solves the issue that you get with subscription based services, which is that if you stop paying every month/year you loose access to the tool to do your work.
Instead of building yet another SaaS wouldn't be smarter for users to gather and pay for a software libre solution?
In this case, Photoshop will continue to dominate (unless they get lazy/give up) because they have recurring cash to reinvest.
The issue of having to pay to get access to your tool it's a big issue.
For big companies having to pay for 100 workstations yearly the monetary incentive is big.
Once many people is aware that in the long term it's a better deal to put their money into software libre projects, maybe different companies could compite to get your money.
Content aware fill was in Gimp before Photoshop. It was an open-source PhD project.
But it's hard to convince people to hand over their money. SaaS works because the profits are reinvested into marketing.
Now if you wanted to throw $400 for a photoshop like system to start from the ground up and work across all three platforms, I think you've got a great idea. You'd avoid the cruft, you could do subscription/pay-per-use cloud based filters, you could start with 32bits per channel from the start, instead of the slow transition to GEGL...
Seashore was a try at this a decade ago.
http://seashore.sourceforge.net/The_Seashore_Project/Screens...
A lot of coders doing side projects do it for fun and to learn and that often means making their own architecture decisions (and mistakes) so they can practice that process and learn
From Dictionary.com:
>> Usually Disparaging and Offensive. a term used to refer to a person who limps or is lame.
>> Origin
>> An Americanism dating back to 1920-25; origin uncertain
>> Usage note
>> When describing someone who is lame, gimp is used with disparaging intent and perceived as insulting. But within the disability community, it is sometimes a term of self-reference.
Maybe "Imp" is offensive in other languages? (Nothing other than alternate English meanings shows on Wiktionary).
Perhaps the best person to ask is a photoshop user who actually paid for it?
CS is expensive and I use a small part of its features... if GIMP covered just 25% of what photoshop could do, I would use it. But I think the only people who think GIMP is close to replacing it are those that have never used/learned photoshop.
Adobe regularly copies features from the GIMP, like content-aware fill
I actually pay, not just for CS, but an entire Windows computer to run it. The rest of my computers run Linux, and I use open source software everywhere.
I would absolutely love to switch to GIMP, if it were at all possible.
But PS is the only program I have ever had to take a class in to be productive in, as it was clearly built to be familiar to people with very clearly-formed ideas about workflow, e.g. old-school photography and design folks.
And switching to GIMP is possible methinks, you just have to learn the interface. Just like Photoshop...
The commenter above you used the software for 15+ years and specifically cited that lack of feature parity is the only criticism keeping them on CC. I have used both as well and do not follow your opinion about Photoshop, maybe because I never used those old school design tools. To me it looks like an Adobe UI.
Overall I just don't understand how you read the comments above and determined the solution is to learn the interface - honestly, the most tired and uninformed defense of GIMP there is. Especially when your primary criticism of Photoshop was... The interface. I guess the argument could be made you eventually get used to working around inexplicable deficiencies like above, but switching to GIMP is not possible if you need the features only included in Photoshop. There's not some random cosmic reason Blender and Inkscape are considered healthy competitors in their respective domains but not GIMP.
Why do you feel so strongly against 'learning the interface'? That "tired and uninformed defense" is also... what you must do. Why is everyone so afraid of learning?? Fucking wimps. You have a problem, you google it, repeat til you know the product. SO HARD!
I write a lot because I want to clearly explain my points on the Internet where there's no verbal nuance. Sorry if it was too much. Are you really criticizing me for thorough explanation on a discussion forum?
Edit: I should probably qualify that in the capacity I use photoshop it's kind of a moot point anyway, since I depend on illustrator just as much and the convenience of smart objects and the interoperability with stuff like libraries are more than worth the price of admission for me.
I could list 4 other points, but really it's just about that one. Photoshop has had 16 bits per channel since 1992 when version 2.5 was released. That's what... 25 YEARS!
Gimp is FINALLY 16bit per channel now in 2017 with their latest beta.
This could be a good thing or a bad thing if major changes are needed. A project of GIMP's size is a big ship to steer.
As long as you wrote useful features then GIMP, for example, wouldn't be able to argue semantics on the patch set and tie up merging for a long time on politics. Their users would demand the features.
(Key difference with Kickstarter etc is that people are pledging for a specific solution rather that for people to solve a problem, and the cash is paid up front rather than a promised monthly commitment. And still there are issues with execution of many projects, particularly once third parties with different goals are involved. "I also need GIMP to be as Photoshop" on the other hand, is basically asking for a "nice unicorn" and hoping both that unicorns can exist and that everybody else involved - especially the unicorn-breeder - shares your definition of "nice" ones)
Also, $400k is, like, nothing in developer time (compared to what Adobe puts into their software, at least).
ChromeOS uses the Linux kernel.
Students getting Chromebooks aren't sucked into the Linux ecosystem and open source community that encourages you to improve parts of the system yourself. Instead, it leaves system maintenance (and learning experience) entirely to the administrator and really only provides Chrome as the operating system, with all implications.
Control your data locally? Why not upload it to Google Drive instead. Learn git commands? Better find that hidden Developer Mode switch, because crosh isn't going to get you very far. Get involved in improving Linux or Chromium? Nah, better edit some JavaScript in a web view with someone else's proprietary online code editor.
Google can (and will) swap Linux out of Chromium whenever they want to, because none of the interfaces and none of the community overlap. When Fuchsia gets swapped in, neither you nor the students will notice, and it won't have made a difference other than Linux having helped Google's locked-down, centrally controlled platform to succeed.
Either way, ChromeOS doesn't contribute to taking (co-)ownership of any of the above and encouraging the freedoms commonly associated with the use of Linux. Neither does it sustainably establish even the Linux kernel as pillar of the OS, it's just an implementation detail to be hidden and potentially replaced later. So why exactly should we look at ChromeOS as GNU/Linux having taken off for the casual user?
More money may help entice more time to be taken from some of those other things, but some of those are more or less fixed (job, chores), not sustainable to give up for long (sleep, relaxing) or simply not worth a small (or any) amount of money to give up (family).
Enough money and stability that someone can give up their day job -- that's a different story. But that's a very big jump, and suddenly even $400k doesn't really go that far, assuming you can even get that money. There's a reason very few open source projects have a full-time developer being compensated by the community.
The opportunity with Photoshop may not be to create another Photoshop clone, or add more features, but rather create smaller, specialized editing apps, that each target a subset of the Photoshop user-base, with perhaps less features.
Look at how Adobe Lightroom cherry picked a few Photoshop and Bridge components, to create a more specialized app that does a specialized job that Photoshop might have done in the past.
Or how Adobe's iPad Pro apps again deliver just a small subset of the total available features from their parent apps, making them more specialized for specific use-cases.
Photoshop's massive fragmented user base is ripe for the pickin' for sure. Could you tear down Gimp like that? Or would it be better to start from scratch?
Hate to break it to you, but not one person who uses photoshop uses Linux.
Not true: I use photoshop (albeit, not professionally) and have a subscription for it, but am writing this from linux. Dualboot exists, you know.
But I would still say Linux support isn't what's holding back any photoshop replacement
The reason is a simple value estimation. It's open source (and so free to get) therefore its price is $0. Any product that is free signals that the product has correspondingly low value. I think it's simply unrealistic to expect people to put forward $400 for something with such low value perception and whose fair market price is $0.
This is especially true when Photoshop has such a high value perception (in part because it is better, and in part as a justification for shelling out the $400).
Ask yourself this - Have you, or will you, paid (donated, ...) $70 for using Vim (atom, neovim, emacs, VSC, ...) to support development? Have you paid $200 to Debian, Ubuntu, or Redhat for your Linux OS? $5,000 for (open|libre)ssl? $13,000 per core for PostgreSQL/MySQL?
I would be curious if the GIMP dev team could reach usability parity with Photoshop on only that money. It seems a bit on the low side frankly (about a year's time of a good developer and UX designer, plus support staff).
I would say that I'm in too, but I also need a quality Illustrator alternative that works well on Mac.
I use it and love it and it's one of my favourite programs, but really it's better in Linux.
Yes and no. I'm doing that right now...
The key is that I don't charge for the software as such. I charge for services, etc. around the software.
People are happy to pay for value-add, even if the underlying software is free. You just need to show value-add...
Yep. Matt Mullenweg & his crew at WordPress have been doing the same thing, and are wildly profitable at it.
This change in how the money is allocated is a subtle, but important, distinction. Monetizing a service means you now have a minimum of two products to manage, and from my time at a company which did exactly this I've found that the OSS software development is the product which absorbs any shortfalls. It's hard to watch the budget for OSS development shrink constantly; to watch developers passionate about the OSS software being reassigned to grow the service instead.
It's probably still the best way to handle a crummy situation, but it's not what the OP described.
I'm reminded about the kickstarter for Django migrations: https://www.kickstarter.com/projects/andrewgodwin/schema-mig...
Most backers gave 5 or 10 pounds, but on average ~500 people gave ~35 each.
You might be half-right about perceived value though - I think anyone that is a serious Gimp user see value in the software, but might still think 400 is too steep, given that it is free now. But I also think many users would be willing to pay for getting some features in/some polish done - personally I'd say my current limit would be around 100 USD for any one OSS project.
Now one one hand, Debian does much more than any small project, but I also think it's more likely that 10 000 USD would do more for something like the Gimp, than for Debian (assuming we had a developer or two on standby that could take the opportunity of devoting a month full time on getting some pain points resolved, say).
If all GIMP needed was a few more developers to be competitive against PS, that would have happened by now.
You're seriously underestimating the amount of work GIMP needs to do.. and/or the amount of work that already goes into GIMP.
Chris Webber, for example, took less than a $40k salary without benefits to work on MediaGoblin full time for a year:
The games industry gladly takes advantage of this.
CRUD apps may be tedious, but getting a job at a local medium-sized company working on them is usually much more pleasant than working for a game developer.
Photoshop's content-aware fill first appeared as a third-party open-source plugin for the GIMP.
While one certainly can't equate the two I have found GIMP works well for a broad array of digital imagery tasks... just like Photoshop.
Much of Photoshop's dominance has to do with the network effects of Adobe's ubiquitous ecosystem as much as the comprehensiveness of their tooling.
Indeed it did, and that plugin was hard to configure and hard to use. I remember installing GIMP just to try that plugin out, and it was not fun, it involved a lot of manual steps to get it working, and I never quite got the hang of it.
In contrast, when I got my hands on a version of PS that supported content aware fill, (which admittedly ships in the box, skipping the installation step) I was able to figure it out in a few minutes and start getting good results in half an hour. I am pretty sure I spent more than half an hour trying to get the GIMP plugin installed.
Means I end up paralysed around starting new endeavours.
The wage slaves.
IANAL
Obviously a lot more of the details would need to be worked out to try and minimize how much companies try and game that to simply fund their own research tax-free, but I wonder if the type of behavior you're looking for could be economically incentivized given the right corporate tax legal reforms.
From a quick search, it looks like the IRS does sometimes allow businesses to make deductible contributions to a 501(c)3, but that it first attempts to determine if the business expects to receive a "substantial return benefit". [0]
The IRS seems to already have specific rules around "qualified sponsorship payments". [1]
[0] https://www.councilofnonprofits.org/tools-resources/corporat...
[1] https://www.irs.gov/publications/p598/ch03.html#en_US_201701...
Google gifted schools Chromebooks, which seems like an incredible act of charity. Except for the fact that you have to pay a subscription to Google to use them. Which means they really just managed to get a tax write-off while picking up more customers.
How do you ensure a company is doing something in the best interests of society or the charity or organization they are donating to, rather than in the best interests of themselves?
Do you happen to have a reference for this? G suite for education is free as far as I know and I wasn't able to find anything else that you might be referring to.
You can see an educational portal listing the cost of the management license at $30 here, which isn't really a complete picture, but demonstrates that we aren't talking about something they give schools for free: https://edu.google.com/products/devices/
Note that this is kinda of well-hidden, but that last link cites the "total cost of ownership" of a $149 Chromebook over three years as $588, which should give you an idea the difference between the hardware they give away, and the eventual cost of everything you need to deploy the suckers.
Schools is definitely big money for Google, it's just well-disguised as an altruistic endeavor.
You need to show that those licenses were not included in the gift. And let's just assume that they're not that stupid.
> Note that this is kinda of well-hidden, but that last link cites the "total cost of ownership" of a $149 Chromebook over three years as $588, which should give you an idea the difference between the hardware they give away, and the eventual cost of everything you need to deploy the suckers.
For a budget windows laptop or tablet the hardware cost would be higher and the support cost unlikely to be lesser.
> Schools is definitely big money for Google
Schools are a tiny segment for Google that would not hurt them in the slightest to lose. The margins on low-end hardware are not exactly the stuff dreams are made of, and they really are making peanuts on ChromeOS. We can and should contrast this to Microsoft, who has been playing the "loss leader" pricing game in educational circles for far longer with far more success.
Generally speaking, it's pretty difficult to tell if charities are actually being charitable, and there is no way to ensure it -- we can't mark each dollar's fall. But we do have laws that cover those sort of 'hidden catch' scenario; you can't trick people into a financial obligation. Is there some other loophole that you can identify?
And of course, as my source demonstrated, there is definitely a license cost to use Chromebooks in a managed environment, and it is not free for educators.
As the situation is _right now_, the law is probably fine, but I'd be against the parent's suggestion of allowing a loophole for them to get to repatriate cash tax-free for this sort of usage, because it's likely just going to work out in their favor, and not in ours or the public interest.
Here's the thing, in corporate devs are as essential as people toghtening nuts in a factory.
The value one produces there is multiplyed by the internal know how, customer base, marketing reach and tight focus management, none of which it's brough by the developer.
That value is stripped from developer work just to be pocketed by management and sales is a myth, besides, if it werent there'd be plenty developers going solo and the median salary would be in the millions.
Edit: and lo and behold, instant downvote. No wonders the tech community is so blind and guillible when it comes to the value argument, from the extraction myth to the stock options gambling.
Some developers are just writing code to satisfy a spec and really could be replaced by any other developer; others are uniquely valuable and significantly increase their employers value, without necessarily being recognised for it.
Good software acts as a multiplier, so it's not equivalent to tightening bolts. It's more equivalent to finding optimisations that reduce the number of bolts which need to be tightened. A relatively small amount of software can drastically increase the value of a company/process
It's true that an individual can't do the same things a company can do. That's not some profound assertion. It's however impossible to go from that and assert that is because they are incapable of doing those things even if they had the funding to do so. That is to say that the marketing issue is more or less an issue of funding or money. That is what keeps that individual devs from capturing more value that is instead captured by companies or by managers, marketers, etc. As well as hiring additional people to form an actual company that performs the functions of a company. However a significant number of founders are developers, many of them including Paul Gahram and Mark Zuckerberg average developers that went to found multi-million to billion dollar companies because of funding. Now after the fact maybe you can argue that there was something special about them, but that's a post-facto assertion. They were average devs and if they had not gotten the funding would be relegated to your nut bolt tighteners not having been able to express their "true" potential.
1996 (±2 years) in an arcade, actually.
It was really ragged, so much so that the bank wouldn't take it. So I still have it.
http://www.bbc.co.uk/news/uk-england-stoke-staffordshire-391...
The last one is the best.
You ask if anyone lost any money. If someone says they did, you follow up with "how much money, and how big were the bills?"
Then they have to know an amount, how many bills, and what denomination.
You can imagine that it's hard to fake, and anyone being genuine will have a pretty good idea of what you're holding.
It happens.
If you allow for faux currency exchange, keeping your eyes aimed down at a renaissance fairground is pretty much guaranteed to turn up enough food tickets for a turkey leg.
Also, "proof" is a noun. You meant to use verb form "proove". "Proof" as a verb means to "proof-read" (to read with an editorial mindset).
Once, while eating a pizza with canned mussels, he felt something hard. It turned out there was a pearl in his pizza.
Sometimes money literally falls out my pocket when I fumble out a wad of cash while i'm inebriated.
Other times I carelessly leave my wallet somewhere stupid and the money "walks off."
Over the past 2 years, I've probably lost over $1000 this way.
But it was neat, finding money just right out there on my own lawn.
"....In a similar situation in Los Angeles four staff members and I were talking in front of the Biltmore Hotel when I demonstrated the same point, saying: "Look, I am holding a ten-dollar bill in my hand. I propose to walk around the Biltmore Hotel, a total of four blocks, and try to give it away. This will certainly be outside of everyone's experience. You four walk behind me and watch the faces of the people I'll approach. I am going to go up to them holding out this ten-dollar bill and say, 'Here, take this.' My guess is that everyone will back off, look confused, insulted, or fearful, and want to get away from this nut fast.
From their experience when someone approaches them he is either out to ask for instructions or to panhandle — particularly the way I'm dressed, no coat or tie."
I walked around, trying to give the ten-dollar bill away. The reactions were all "within the experiences of the people." About three of them, seeing the ten-dollar bill, spoke first — "I'm sorry. I don't have any change." Others hurried past saying, "I'm sorry, I don't have any money on me right now," as though I had been trying to get money from them instead of trying to give them money. One young woman flared up, almost screaming, "I'm not that kind of a girl and if you don't get away from here, I'll call a cop!" Another woman in her thirties snarled, "I don't come that cheap!" There was one man who stopped and said, "What kind of a con game is this?"...."
https://archive.org/stream/RulesForRadicals/RulesForRadicals...
I'd give up on ten dollars if it means not having to fight/argue this way on the street.
If you're really hung up on whether 3 is the right number, I think you've entirely missed the point of my post. Adobe has far more people working on photoshop.
And most developers are definitely underpaid relative to the value we provide. From shabby CRUD through deep machine learning research, we're generally underpaid. Partly it's our own collective fault, though: as a group we're told constantly that passion for the product and satisfaction at "being paid [anything] for doing what you love" should be enough, and enough of us buy that crap that it effectively depresses wages.
The bigger issue is if you can pull a 400k Kickstarter and manage such a team why the hell would you be wasting your time on GIMP dev - you could do waay more profitable and interesting stuff.
From my experience, you can hire a good developer from Vietnam for approx ~2k per month; which is 24k per year.
Yes, but it is three full time people. Don't underestimate the difference being able to focus on a codebase for an extended period makes...
Also, knowing that you will be able to continue that focus carries it's own benefit.
It doesn't even have a circle tool and the best "solution" seems to be ellipse select with shift -> fill -> shrink selection -> delete
What a dystopian ux nightmare.
Sorry Mom, for inadvertently inflicting gimp on you.
I don't think it's that far from how one goes about drawing bezier paths in PS (although it is a little more convoluted).
https://graphicdesign.stackexchange.com/questions/28344/how-...
The main problem is people do not want to learn a new interface or have programs running only on Windows.
For what it's worth, I think Sketch is much more user friendly and is $99 one-time. Certainly not OSS, but a better deal than Photoshop for most things, and a much more user-friendly interface than Gimp (though I suppose that's subjective).
Still, 99/year is a drop in the bucket. Even 50/month for all of creative suite isn't all that much considering many of us will just add it to our business expenses and write it off.
I get they are FOSS. But, they could still add in their GUI options that don't exist; then prompt donations that would go towards the features development.
For example, Image > Rotate, triggers a pop-up, "Sorry, Rotating image is not currently a feature but with your support we'd love to add it to GIMP! We expect it will cost $10,000 and we've already raised $4,235 from users like you. Contribute now..."
But there are very good Photoshop competitors that don't cost a lot of money. Pixelmator is $29 and a better program for many users. It's lightning fast to open and use and it requires less computing resources.
I remain deeply unconvinced that any software that requires an elite UI and interaction model will ever come out of a software libre solution. In addition, most software libre solutions are simply copies off of someone's else's ideas and hard work. It can't exist without paid software to copy first.
What are we even doing in tech and software if we are arguing that people should copy other people's work and then make it free, causing people to lose their jobs?
Hyperbole much? GIMP is high quality software and I'm sure I'm not the only one who prefers it to Photoshop.
> copy other people's work and then make it free, causing people to lose their jobs?
Wikipedia is a fully free project and the foundation employs ~300 people funded mostly by an annual donation drive. Free software != no jobs.
And I really don't want to get into it, but for a lot of free software advocates extorting money from people for licenses using legally granted artificial monopoly of scarcity on software you don't even have to disclose for the cost is highly unethical.
1) You need a 3rd button for each submission that says "Already Solved". People can post links to SaaS products that already solve the exact problem the users are posting...
Example: Following 2 Submission have several products that solve the exact same problem.
- Gamify budgeting so I save more money => https://www.demandrush.com/problems/gamified-savings-app
- Community chat on my website => https://www.demandrush.com/problems/onsite-community-chat
2) You also should introduce some sort of flagging or curation, because quite a few "Problems" read like 1 off specific requests, and some descriptions are also really vague and too broad
3) Might also wanna provide Upvote (only, and no downvote, like ProductHunt and Facebook "Likes" :) button to assess popularity, because not everyone might be mentally willing to commit to monthly payments, but might really "like" the idea and think it deserves to be solved.
Just because a solution exists does not mean that it's a suitable one, or can't be improved upon.
> Choose a problem below to get started.
Get paid to build _my_ next project or _your_ next project?
This is clearly a two-sided platform, but the messaging seems conflated: the headline speaks to builders and the instructions speak to end customers.
This sounds like a good idea on paper, but the reality is that most of the value of a project is in the implementation details. There are countless meetings for each individual feature.
I'm pretty sure the above description is how a lot of the end users (clients) are seeing this. Their contribution amounts are also insanely low (orders of magnitude lower than I would dare to value some of their projects).
I don't know how contracts or SLAs (if any) apply here, but this just seems like a good way to get yourself on the hook for getting paid (I'm guessing) a tenth of minimum wage.
Source: I do this sort of contract work for a living.
There are people who simply want a product/service, no desire for ownership. They post what they'd be will to pay for it, as a customer, and you decide if it is worth your time to build it.
Anyone? Am I understanding this correctly?
Businesses: have a problem you need solved with software? Post your problem!
Programmers: looking for a challenging project with identified customers? Apply to solve a problem!
> 1 customer paying $5/mo
Sounds about right.
In many cases, you can use the model as-is (and likely pre-trained) for a use case outside of "match label to an image", with only some additional training with your specific dataset - which may only be tangentially related (or not even that!).
These models (and the surrounding tools) have become for many problems more like Lego in my opinion - which is a good thing! It means they are more approachable for everyone, rather than being something mysterious and complex. Ok - if you dig, things become complicated quickly, but for many problems, you don't have to worry about these internals.
Are there problems which don't fit neatly into using a modified ImageNet or LeNet or one of the other "standards"? Certainly. But I think a candidate who understands the standards and basics well is likely a better one than one who only understands a specific subset for a particular industry (if there even is such a thing, which I am sceptical of).
Furthermore - it would be even better when a candidate can say "you know what - your problem doesn't need a neural network of any kind, let me introduce you to <insert standard statistical machine learning method>" - because there are tons of problems out there which can still benefit, and be a robust, easy to understand, and fast (to implement, to maintain, to execute - whatever).
I love neural networks, certainly - but there's been so much hype in the news, everyone thinks they need one, which will probably lead to many investing money into worthless (or expensive) solutions, where simpler (but less "sexy") ones would have sufficed.
The candidate who could know and tell the difference would be even more ideal - being able to interview/walkthru that might be a way to get a better candidate.
It's sad that even the highest bidding project is $500/mo.
> Developers own the products they build for customers, and charge a monthly fee for access.
So, what you're seeing is what a single user is expecting to pay. This is not like a contracting / free-lance website where someone hires you under contract to do some development work.
Instead, you create the SaaS or whatever, and then you have at least x number of people that are willing to pay $y/month for said service; your market is not limited to the people who use this website, too.
I think it's really just a way for people to find what projects people want, and what a select few are willing to pay for it.
I know a company which paid around 1M euros to get such data ( and some software ) for just one industry.
customers_offering_prices = [1,1,4,15,30,30,40,80,100,150,175,10000]
winning_applicant_price = 40
winning_applicant_revenue = [40,40,40,40,40,40].sum
The last 6 customers receive the product and they all pay 40. Essentially a modified Vickrey auction. Everyone has an incentive to bid exactly what the product is worth to them.
Edit: >Everyone has an incentive to bid exactly what the product is worth to them.
Actually not true if applicants see the bids. In my example, the the best applicant would want to bid 10000 because that would maximize revenue, but exclude everyone before that. Customers would adjust their bidding as a result. But applicants should know much they can expect to receive if they win so they can bid correctly.
> https://www.demandrush.com/problems/fantasy-scifi-subscripti...
> I'd love a Netflix-style platform, website or app that, where I'd have a selection of high quality books to choose from.
Amazon has a 'unlimited book rental' subscription model (I know not all books are in there).
"What systems has this person already looked at, and why were they rejected?"
That probably needs to be a base question.
[1] https://arxiv.org/abs/1604.04018 "Multi-Oriented Text Detection with Fully Convolutional Networks"
I've even written one, it's not complicated and not expensive.
Unfortunately Google has some other metrics to detect bots (such as no cookies showing they've seen you before, etc), so it wasn't 100% reliable.
:D
That's not what the project is asking for, but it's very close to being a very doable hobby project.
For example, it can't tell there's text in this clear image of a VIN:
http://1eask1khnn22hm3njx2dmz1a.wpengine.netdna-cdn.com/wp-c...
But I bet if you took a closer picture (like you would if you were using an AR app) it'd work
There are a bunch of SaaS opportunities like this that could be made just creating wrappers for various cloud apis and selling to people who don't know they exist.
Still not sure how you handle avoiding imposters: for things like the accounting solution that's on the top of the site I think there is huge potential for abuse in either direction (either devs phishing for data or companies refusing to pay for solutions). Good luck!
I just posted these questions yesterday, in the hope of finding solutions to a couple common annoyances I have with Google search:
https://superuser.com/questions/1218986/keyboard-shortcut-to...
https://superuser.com/questions/1218989/how-to-maintain-orde...
I don't want the overhead of "hiring a freelancer" to do the work. But I'd definitely pay a bounty if someone came in with the answer / script / extension / app that solved the problem.
I'd suggest to you odesk.com, but that is a site for hiring freelancers. Although I've seen it used many times for 'here's what I want, deliver it and the money is yours' rather than a .. 'work for me at X/hr and I'll direct you' kind of thing.
- Payment on the order of $10 for some simple script, not $300+. They can make up the difference via multiple bounties, selling to multiple people, etc.
- Public question that can be found by others with the same problem (who can upvote or add to the bounty), as well as people who are able to deliver the solution. Stack Exchange, but with payment.
- I don't want anything to do with reading resumes, selecting the worker, scheduling, payments, etc. This is what I think of when I hear "freelance".
I do something similar: I write very small programs for embedded systems: simple timers, sequencers, alarms, etc. for around $100. In some cases, we're talking 15 minutes of work, tops.
If I could charge just $30 and instead sell it to multiple people, I would, but the odds of that happening are so slim, I have to have a $100 minimum. The problem is that even for well-defined problems, what people need tends to be so different that the best you can do is have a common framework for similar problems. The code itself is rarely even close to identical
BUT... people already do this stuff for free. Create free Chrome Extensions to make some tweak on a website. Spend hours crafting a script for an answer on Stack Overflow, only to get fake internet points.
I asked the questions above on SuperUser, a website that theoretically does exactly what I'm asking. Someone will probably eventually post a solution, someday. It just takes too long. The site doesn't get enough traffic because no one is getting paid to do the work.
Then I realized I could do the same thing and get paid. It's just a matter of shifting the context. You've no doubt seen any number of sites where people post sample code to help someone, but disappear if it doesn't work, or there's a minor detail left off, etc.
The only change I made is that by paying me, I guarantee you it will work and I'll be around for future changes, extension, etc. A surprising number of people are willing to pay for the assurance over just getting free help.
at the end of the day you need to validate that the person is going to be able to write this code and that you are going to follow through with payment. Either way it opens up one party to risk. A developer risks putting in the work without payment and the financial backer risks non completion after they pay. For instance, I wouldn't mind writing that extension for you for a hundred or so but I also wouldn't want to do the work before getting paid.
Version 2.0 would hook payment up to a successful pass of the test hook!
What I thought was that those request/tutorials could potentially be popular and others might be willing to even pay a little bit for it and once the sum is enough, someone would do it and get that pot of money.
For this reason, I made this little site and put some of my "nice to have" needs on there, but unfortunately, not many people go to the site.
Don't really want to promote it, but you could check it out here https://hackermake.com/
unfortunately, not many people go to the site
and
Don't really want to promote it
are probably related!!!
Just independently saw a link to Bounty Source today. This appears to be very similar to what I'm asking for, although more designed around adding bounties to GitHub issues for features/bugs on existing projects.
I don't think there is a legal way of actually doing this.
Of course, to set up a new service like this with different sources is going to be a huge up-front cost in negotiating and building integration with content owners, capital to subsidize costs until your user base is big enough that you can negotiate low per-user licensing based on the volume you bring, etc. It's not a realistic side project.
But I do like the idea and would also gladly pay for it!
https://play.google.com/store/newsstand and https://newsstand.google.com/
In TV land, this is called "cable". Many people apparently do not realize they have re-invented this for websites.
We need a platform that facilitates payments easily and transparently _without_ seeking to bloat itself to extremes. Then content authors and content consumers can still interact more-or-less directly and we won't have the problems that are typical with big content brokers.
In Europe, this would make no sense. A website is not a legal entity, so it can not be a copyright holder.
Does the statement make more sense in the US, or is it just a common misconception among people who build websites?
The terms page says "Welcome to DemandRush, a website and online service of Grafly, Inc". So I would expect the real copyright holder is Grafly Inc?
But at least as far back as the early 90's it's been common to put "(c) MyCoolDomainName.com" on a web site, and I'm pretty sure most people mean "copyright ME until/unless I get around to making a business out of this, and if I do then copyright whatever business ends up owning the domain name." And I would assume, though again IANAL, that's not quite what it really means.
Unlike, say, Germany, the US has no requirement for anything like an "Impressum" for a web site. Nor could it, really, due to the constitution. (But I Am also Not A Supreme Court Justice so my interpretation of the constitution may be less sticky than others'.)
[0]: https://www.sba.gov/starting-business/choose-register-your-b...
Chances are this is a problem for a data-mining / ai-training platform that wants to create a better image set. They are charging clients tens of thousands of dollars, and yet willing to pay $100/month to improve the data. lol
Freelancers often make the same mistake--charging by what it costs to them, plus a little bit more. You should charge how much value it brings to the customer, not how much it costs you to make. Easy to say, hard to do.
Anyway, the big issues here are those all bounty-type sites face (which seems, for some reason, to be the thing everyone keeps trying over and over and keeps thinking is a new idea). See for reference, https://wiki.snowdrift.coop/market-research/history/software...
There's a lot of reasons why this sort of approach may seem promising but never really take off. It's also possible (though I'm skeptical) that it just takes a certain amount of design, promotion, and right-place-right-time…
My only concern is that this is a little "race to bottom"; most if not all the entries are along the lines of "I already have X but I'm paying too much!"
I guess this is great for solo entrepeneur/developers, hell, if I know for certain I had 100 people waiting to pay me $50/m for X, I would definately consider building it :)
"Need to see revision history for View-Only Google Docs" - Don't use a shared Google Doc as the agent of record. Email it to each other as an agreed contract.
"Can't find good replacement for expensive accounting/bookkeeper service" Have you tried literally anything else? You're paying $500/month for this solution, have considered just getting an accountant and emailing him spreadsheets?
"Monthly subscription for science fiction and fantasy books" "Subscribe to news without paying $30/month per website" - Existing solutions already exist and require a supply chain and distribution rights negotiation more than they do a developer to slap it into a webapp.
Just signed up with our solution on demandrush: Tuiqo, simple document versioning - https://tuiqo.com. We just graduated from YC Startup School Founders track.
I would very likely pay for this feature, but I'm not willing to commit to much unless it works. I would pay a nominal amount to "kickstart" it I guess, and have the option to sign up later.
But it's a fairly nasty problem - porting your books a proprietary platform to QuickBooks seems full of traps.
Is there a particular reason why you're looking for an InDinero replacement?
When we started with them, we paid about $400, and they did payroll, accounting, and taxes, plus Stripe integration for a DE C corp. We had a monthly meeting with a bookkeeper. We'd get responses to all requests within a day or two.
Now, we pay 3x that (our revenue has grown) they've eliminated payroll, they've outsourced our bookkeeping to somewhere else, and often don't respond for many days. They delay filing taxes until Sept always. I don't really feel like they understand our business particularly well.
They've quoted me nearly double to switch to accrual accounting, and I suspect it's even more expensive than that. I can't really use it for management purposes or cashflow forecasting, which are my two biggest pain points.
I'd rather be on Xero or QB which are industry standard. I could then work with experts who could help solve specific problems (accrual accounting, revenue recognition, etc) on an industry standard platform.
I'm not interested in moving to another bookkeeping software however, unless it's QB or Xero.
Like inDinero or some of the other options out there, we're a service that takes care of your books for you, so you don't need to think about them. But with a few key differences:
* We're built on QuickBooks Online, so you're not locked into a system that no one else knows how to use. (It's still a proprietary system, but it's the industry standard.)
* We take a hybrid humans + software approach. Most of the work of keeping the books updated is tedious data entry, and when humans do the work (no matter how talented they are), it's super-error-prone and frustrating. Our attitude is: computers should do the tedious repetitive work, and our talented team of pros should focus on (a) doing the tricky stuff, (b) understanding the specifics of your business, and (c) answering any questions you have.
* We're startup people, so we've had these problems ourselves. This is our third startup — our previous companies were Ksplice (acq. by Oracle) and Zulip (acq. by Dropbox), so we're definitely not strangers to building software.
(Incidentally, we've done a bunch of migrations from inDinero before.)
Basically, our thesis is: humans + software each doing what they do best. (Or, if you like, we're building an Iron Man suit for our people to do your books with higher quality.)
We're in private beta today and have been doing the books for a bunch of tech companies (all the way from newly incorporated to well on their way to their Series B). We're actively taking on a steady stream of new customers, so if you're interested, send me a note: waseem@zapgram.com
User: "I want {A}."
Winning developer: builds {a}
User: "This is useless to me. I wanted {A}, not {a}"
Developer: "But you never said you were case-sensitive!"
It can easily go the other way, with the user saying "Well i said {a}, but I meant {A}. {a} is useless to me." Who pays when nobody is clearly in the wrong about some ambiguity? The FAQ hints at an "initial 3 month subscription." Am I missing more info on the process?
Several questions about where could this go post 1-day HN buzz:
1. How are existing solutions being presented to the ignorant "wannabuy"?
2. How is this list sorted?
3. How are duplicates managed?
4. Where is the "discussion" required to even minimally spec the ideas and bring them to a buildable form?
Isn't that the public library?
I think the service should also have _two_ services. One is for great sci-fi fantasy books of all time-- obviously that's going to be cheaper. The more expensive service would be for books that have come out in the last 2 years.
For myself, I often look at the winner & nominees for the Hugo and the Nebula for the last two years: https://en.wikipedia.org/wiki/Hugo_Award_for_Best_Novel#Winn... https://en.wikipedia.org/wiki/Nebula_Award_for_Best_Novel#Wi...
Keep in mind that media mail is like, $2.60? If I was making this service, I would want to ship 3 books quarterly, rather than 1 book monthly.
There ought to be economies of scale by buying hundreds of books from a publisher. What do book retailers pay for books anyway? One benefit is that you also won't be returning books, like brick-and-mortar bookstores do.
An additional awesome feature would be if it connected to your GoodReads account and never sent you duplicates.
The librarians at my public library are great at making suggestions. They also have themed book displays that they rotate every week (eg. physical fitness one week, high fantasy another etc.)
Also, the internets are filled with curated lists. Book lists have come up here on HN a number of times [1][2][3], even Sci-fi specific[4].
But I get it, push button services are very appealing, and can save a lot of time. Just click subscribe and get a constant supply of books you'll love at a reasonable price. It's a business model that can work surprising well[5].
1 - https://news.ycombinator.com/item?id=13117521
2 - https://news.ycombinator.com/item?id=11229312
3 - https://news.ycombinator.com/item?id=12556160
Unfortunately my local public library doesn't expose any sort of API to allow this sort of thing. Only way to do it seems to be by writing a screen scraper.
It might still be a cool little system to build though. I might try pitching it as a new (volunteer) project to my local library and see what they say.
I suppose technically Analog and Asimov's no longer meet the monthly requirement, because they recently switched to 6 issues per year, but each issue is now double-sized so it is the same amount of material as they published before the switch. (That wasn't actually quite monthly either. It was 10 issues a year, with two of them double-sized).
Subscribing to two of these magazines would be under the $8 month people said they would pay (that's for the ebook edition...paper I think is a little more).
There should also be a discussion section.
For example, I see: >Can't find good replacement for expensive accounting/bookkeeper service
I'd love to recommend Bench.co (use it and love it) and avoid someone wasting their time. It's much more affordable than 500/mo.
I do love the concept though. @OP: I'll share this with the https://turtle.ai/ developer community and see if there is any interest.
Anyone who's pre-launched a SaaS knows you're lucky to get 50% of people who said they would buy to actually buy. People lie all the time. But this is a good gauge. Think an app that does Y is worth $50/mo but there's a dozen people saying they'll pay $100/mo for it? Great! A dozen people saying they'll pay $10/mo for it? Not so great.
This is what my doctor charges for a flu vaccination.
I guess I've chosen the wrong profession.
Looks like we've got full circle considering this same logic was applied to devs for ages.
Youtube is full of videos showing how to inject oneself.
Writing a text-detection script can be done perhaps with some help from Google/Stackoverflow, but requires still some intellectual effort depending on requirements.
If you solved that problem you would be rich.
Flu shots are generally intramuscular, you literally just jab it into your bicep. A layperson can do it with high probabilities of success just by reading the instructions on the box.
Indeed, pharmacists administer these all over the US.
Tricep is the way to go here.
Also most developers in SV are paid more than physicians (internists/family-medicine) who prescribe flu-shots.
It is likely a lot more difficult than they are making it out to be. If it were that easy, why wouldn't they do it themselves? If they just don't have any technical employees able to do it, then they probably don't know enough to outline what they really want with a technical person, anyway.
Relevant XKCD: https://xkcd.com/1425/ Relevant comic: http://johngushue.typepad.com/.a/6a00d83451f25369e20120a5138...
So... they want to chop out watermarks?
2. This is why flu vaccine is not inexpensive: First, the CDC and WHO need to identify possible strains of flu that will be prevalent during the next flu season. Then flu vaccine has to be created by incubation (in an egg I believe), then killed, then transferred to a sterile container in suspension, this must be done for each strain of flu you are vaccinated against. The container must be transferred within certain temperature and time boundaries. This container must then be moved into a sterile syringe, which is inserted under your skin, preferably by someone who knows what they're doing.
The only reason that the vaccine is so inexpensive is that the CDC and WHO receive outside funding, and the process takes place on a massive scale.
There are already many, many marketplaces for tech development of all shapes and forms.
For this one I guess it's:
Hey, people who want some tech developed: If you can formulate your problem as an interesting side project you will (1) be likely to find someone actually willing & able to do it cheaply; (2) you are likely to be able to aggregate your buying power with another buyer; (3) you pay a subscription so it isn't a large up-front cost.
and;
Hey, tech developers who like to do interesting side-projects: Here's a bunch of interesting side projects that you might have done them anyway, except here you will (might) get paid to do them!
One way to look at it is in comparison to kickstarter. In kickstarter, a producer runs the campaign by championing an idea and setting a funding model and the consumer can choose to buy in by providing dollars. Here, the consumer champions the idea and sets the price, the funding model is a subscription, and producers can choose to buy in by providing work... and also more consumers can buy in by subscribing with dollars.
I guess for this market to work, it has be be more efficient at satisfying the demand for X-aaS development than existing markets, though I don't know how it might be.
Toggl is a great service, works in the web or as a desktop app, and has slack integration: https://toggl.com/slack-time-tracking
And as for "solving", what if they fail? Is it no commitment? Is any of the funding up front, or is it just built on the promise of possible future subscriptions?
I think the idea could work, but directing the balance between both sides is going to be the problem, a lot of the "buyers" are wanting to pay an amount that will work at scale (like $5/mo) but not really if someone is trying to bootstrap a solution.
The idea is good but it's difficult to use because you can't reach the people.
Versus with contracting/freelancing, you get paid 1x and the product is owned by the customer you work for.
If I get final say in what goes in, then it seems very likely they wouldn't be on board with this (especially not after going through a project). If they get final say, then this is just contract work.
The other issue is liability. In the bookkeeping example again, who is liable in case of a data breach? If there is payroll info or client info in said service, who is responsible for compliance? They couldn't pay me $500,000 a month to get me to take on that kind of responsibility, much less $500.
I know I'm harping on one example, but a lot of projects would have similar issues come up.
I guess my point is that sometimes I don't actually want ownership of a project like that.
> you get paid 1x
That really depends on a lot of factors. From what I've seen on the site, it's more like the other way around is true.
For example, say I pick up one of the projects and the subscriber picks me to work on it. Okay, great. I go and work on it and consider it done. As it turns out, there was a lot of miscommunication and they wanted something different.
Do I still get paid? It seems like either way, there is a huge avenue for somebody to get screwed.
This is more along the lines of Nugget or OppsList in website form, where folks with product ideas request freelance work to bring those ideas into existence.
Having said that, I think the idea of doing something like "Upwork" but with client aggregation and monthly pricing is a great idea - but still speculative, so a great idea for somebody else to pursue.
To me this is almost begging for a s/o style comment section.
What I'm willing to pay for but cant easily develop myself is to add features from OSX to linux. Not too many, I'd start with:
- a graphical browser like Finder where I could color tag files.
- UI/UX: do something about easier configuration of the trackpad. Even when I installed linux on Macbook the trackpad does not work the same or does not have all the tapping actions instead of clicking. I cant even drag files/links to file-browser-sidebar/dock or such things.
There you go [1] originally from [2]
[1] https://github.com/AKSHAYUBHAT/CTPN/blob/master/demo.ipynb
One thing that would be really nice would be some kind of discussion thread for each topic. What does the customer really need? If there are ten interested customers, do they want the same thing? Which alternatives were considered? Which tradeoffs would be acceptable?
Also, IMHO, it would be cool if the site could handle the mundane details of hosting and billing. But I guess that there are probably other sites that do that kind of thing already?
That said, there is no way that ads in podcasts inflict $15/mo worth of suffering upon me. The "skip ahead 30 seconds" button is absolutely perfect for not hearing yet another Squarespace ad
The "industry specific deep learning" project is similar to something I'm working on right now. Though I'm not planning to charge for it.
To any folks here interested in this: Are you looking for a tool to get started in ML or for a resource to apply existing ML knowledge to a specific (possibly new) domain?
Although, I had a slightly different way of going about it and different way to charge, I love the idea!
Gratz to OP. I'll definitely be submitting, and maybe I'll still come out with mine at some point.
I'm thinking of the kickstarter problem: vaporware.
If, over 6 months, 5 people apply to solve a problem, who gets the subscribers?
The first one, I hope not.
Each subscriber chooses?
P.s. Very interesting idea. Something in this direction will be useful.
> Q: I submitted a problem - what happens next?
> A: Over the next 14 days, developers will be able to apply to solve your problem. You'll get a list of applications and a recommendation of who to choose.
So, I guess the answer is "subscriber chooses, and if there are multiple subscribers, they (or the one that paid the most) make(s) an agreement."
Your connection is not private. NET::ERR_CERT_AUTHORITY_INVALID
On the other hand, it seems there is a ton of money to be made on the site!
https error: NET::ERR_CERT_AUTHORITY_INVALID
I will pay $100 - $500 for an Arq Backup[1] clone on Linux, with the same UI polish as Arq for macOS or Windows, optional encryption, deduplication and supporting all the same backup locations, including:
* AWS S3/Glacier
* GCP Nearline/Coldline
* Backblaze B2
* Dropbox
* SFTP
* NAS
* Google Drive
If you're absolutely going to force monthly subscription pricing down my throat, I guess I'll pay up to $20/month. I want this badly.
I know about rclone, Duplicity, Duplicati, that awesome rsync/cron workflow you have, etc. I want Arq (or something just like it). Arq works flawlessly - it is absolutely superlative when it comes to backups on Windows and macOS.
I use an Ubuntu workstation as my daily driver at home and I have a MacBook Pro. Through very careful configuration I have gradually made Ubuntu about as enjoyable to use as macOS (with the exception of 1Password, which has to run under Wine). But I don't want to use the command line, or handle an API myself, or keep track of cron. Duplicati was the closest thing to what I'm looking for, but it's cludgey and started not working for me recently.
I know polished UI isn't exactly the first thing that comes to mind when you think of Linux. But if you're the kind of person who makes "Ask HN: What is your pain point?" posts or who is looking for a problem to solve, this is a burning problem I am willing to throw money at.
EDIT: I'm adding a bunch of HN comments to demonstrate the interest I've seen for this.[2][3][4][5][6][7]
____________
2. https://news.ycombinator.com/item?id=14403321
3. https://news.ycombinator.com/item?id=11742226
4. https://news.ycombinator.com/item?id=9185632
5. https://news.ycombinator.com/item?id=5718015
Alternatively i would add support for one or two services directly, rewrite many times and come to the above model and add support slowly.
Knowing you would need all of the backends would help immensely in the initial design
It's like a democratized efficient tender process driven by the market. Brilliant and beautiful! I really feel you shall be able to capture a lot of value and create alot of value from this. Well done!