31–40 of 145 posts
NLP as a service. It's a growing field yet setting up training models and all of that for a wide variety of NLP related services is pretty hard still. Things like sentiment analysis, Named Entity Recognition, etc.
http://blog.mashape.com/list-of-25-natural-language-processi...
Meta: As the last time this was done have showed, every Sunday is probably not sustainable, which lead to progressively worse thread every week. That said, iirc it have been a year since the last popular "Idea Sunday" was on (I started working on a project from that thread!), it seems like a good idea to do it a week or two ... unless someone decide to kill it off early again. On topic: I think there might be a need…
Context: I realised recently that a lot of film/series music is processed, compared to the album version. Specifically, I noticed blues tracks on Suits sound so much deeper than their album versions. I'd like to know how was that result achieved and how to make a chosen track sound as if it was processed for that series.
On a more abstract level, the problem here is that people and organizations often create some data and basically Ctrl+C Ctrl+V it to different services. Couple of examples:
- RSS subscriptions. If I want to try a new reader, I have to export subscriptions from my current reader and import it to the new one.
- Calendar events. Same problem. Yes, some services can sync with each other but it's a messy situation. Calendar apps must separately implement sync with several calendar services.
- Whenever I want to buy something from a new eshop, I need to enter my email and address.
- Restaurants publish their location, opening hours or photos on Yelp, Google Maps and others. When opening hours change, they need to separately update every service.
- Public transport operators send their timetables to Google and other services.
The solution I suggest is to store this data at the source. People and organizations would have something that's sort of similar to Dropbox, let's call it databox. You have a databox url, such as databox.org/1. Your avatar is available at databox.org/1/avatar. Your RSS subscriptions are at databox.org/1/rss/subscriptions. Opening hours of some company are inside a JSON that's on databox.org/2/restaurants.
The databox URL also serves as a login. So the workflow when trying a new RSS reader is like this. 1) You click on "register". 2) Enter your databox URL. 3) The reader requests a read and write permission to your RSS data. 4) You allow it and the reader fetches your subscriptions.
There are some standardized "subdirectories" (e.g. /rss, /calendar, /tasks) but anyone can create a new standard. For example, a bunch of Linux distributions can decide they will save desktop background and other desktop settings to /linux-desktop.
some place where entrepreneurs can put real ideas they're working on, and anyone can vote on them. If you're an investor and love an idea, you can pay for an intro to the team. (names and details are otherwise hidden.) any such money is shared with the founders who posted the idea (30% site, 70% them) and they can use it to help roll out their MVP.
How does this idea differ from Kickstarter? Or those other websites that you get your ideas funded on.
I would argue it is also not a platform for ideas (i.e. in the way that this thread is, though few people have incentive to post their genuine ideas here.) For example, computer renderings are prohibited on KS - only working products may be shown.
Meta: As the last time this was done have showed, every Sunday is probably not sustainable, which lead to progressively worse thread every week. That said, iirc it have been a year since the last popular "Idea Sunday" was on (I started working on a project from that thread!), it seems like a good idea to do it a week or two ... unless someone decide to kill it off early again. On topic: I think there might be a need…
> I think there might be a need for Website-for-SaaS-as-a-service. What I meant is that whenever a developer decide to start a SaaS, you have to have a website for that service (that has info, landing page, billing, account management etc.). Yeah I like this idea. You can get so far with Bower, before you inevitably start spending so much time working on "product boilerplate". I remember being dismayed by this when I…
An Uber-like app where people pay for their rides with banked miles earned by giving others rides. This would serve a large market of people that a) travel alot and can give rides in their home city (free rides when they travel) and b) people that only have part-time access to a car but may need rides when they don't have it. It may also be legal in areas where Uber can't legally operate because they charge money for…
we have that already, the layer is called "money" you can give people rides and be compensated in the "money points (usually US dollar). Later you can use this points to recieve rides yourself by others. great thing is that you can use those points for many goods and services and not only for rides.
This will provide a single place where you data exists and other companies will only be able to use it as need be. Security would become a big problem however... But it could get around a lot of privacy concerns at the moment.