I started Akasha, a programming language for auto code generation(like genetic programming but can also use stochastic hillclimbing and simulated annealing to generate code). I did just a month's worth of work on it. The language will be imperative and statically typed. It will allow for parallel evaluation of programs(GPU). Programs generated will be Turing complete instead of just symbolic regression.
I think it will be ideal for AI or at least that is what I'm planning to use it for. I estimate it will take me like six months - 1 year to complete.
If you are into AI, you can join my strong AI group:
That sounds really cool. Are you really, really wedded to Skype though? It would be awesome to be able to participate using something open source... :-)
When we get up to like 300 members we will move to a new platform. We are now at 10. The reason why I chose skype is because it is already in many people's workflow. Don't you use closed software for anything?
Open a skype window to one of your friends
and paste the url in. Then click on the link. If that doesn't work add me to your contact : chester.hercules.grant
http://serversforhackers.com, because I don't like current email newsletter formats and there's a large niche of programmers who are behind on how servers work and can be used.
First steps, allows any developer to play with your API or SDK in 60 seconds or less. It spins up one customized virtual machine (container to be exact) per developer, and exposes the machine via a text editor and a terminal in the browser. No setup required for developers.
http://first.askvc.com/home/
I decided I wanted 2014 to be the "year of the product" for me, so I started http://bugroll.com - books and screencasts for developers.
I've "swung for the fences" in the past (anyone remember plentyoftweeps.com - twitter dating? :), but this year I want to focus on the exact opposite approach: instead of trying something that's never been done before, I want to try something that I know many others are doing successfully, to see if I earn my way into that group.
Thus far, I've been focusing more on procedural content generation/procedural -anything- related to gaming/storytelling. Togelius has a lot of good stuff.
Currently working on Gophish (https://github.com/jordan-wright/gophish), an open-source phishing toolkit designed to help businesses and pentesters to quickly conduct phishing campaigns.
http://goodaudience.com is a domain I just won in an auction a few weeks ago. It makes your target market aware of your existence on social media and converts them into fanboys/girls (lead generation).
I started Instagram integration as a side project and now you can increase your reach to over 150 million monthly active Instagram users.
I put this together couple days ago to shorten my URL's and I've actively using it. It's a simple personal way to convey your own message while sharing links. I call it a better URL shortener for startups.
Still the same project I've been working on for some time... Fogbeam Labs[1]. We're an enterprise software startup specializing in enterprise social networking, social BPM and semantic integration + knowledge management. We're an Open Source[2] startup pursuing a model similar to the model used by Red Hat... which is probably not a coincidence, since we're in the Raleigh / Durham area, basically in the shadow of Red Hat Tower.
All of that said, we don't really use the word "pivot" a lot and our basic focus is now still the same as it was when we started this thing. But we have a few new initiatives lined up for 2014, including some things we're going to do in terms of taking our technology, making it more of a "platform", and building some domain specific SaaS offerings on top of the platform.
It's an exciting time right now, as we have some leads in the pipeline and have a real chance to get some movement and traction this year.
1. our product is 100% open source, developed in the open, using the ALv2 license, open issue tracker, mailing list, code on GitHub, etc.
2. Our primary focus right now is the "on premises" deployment model (although SaaS delivery will be available)
and (2) is largely rooted in (3)
3. A tremendous focus on integration. The vibe we get from talking to users of Yammer and the other social networking products is that the largely just become sort of a "replacement for email" or maybe a complement to email. There are different reasons for that, but our thesis is that a big part of it is that the products don't lend themselves to integrating into existing workflows and processes. By comparison, we have SOA/ESB integration for real-time business event subscriptions, BPM integration to make a "social BPM" offering, Calendaring integration, and more to come.
4. A big focus on contextual computing, largely driven by use of semantic web tech. As "content" of various types passes through the system, we do semantic concept extraction, clustering / classification, etc., so we can show you - for a given piece of content, business event, person, task, etc. - all of the related (people|content|documents|events|etc).
5. Open standards focused: all our integration points are based on open standards, including RSS, Atom, ActivityStrea.ms, RDF, FOAF, and the like.
Other than that, it's basically the same thing. :-)
Hangout Fishbowl - A way for the audience, of a Google Hangout On Air, to request an invite in real time. In the next iteration, I hope to add automated invite moderation, like auto-accept if the number of participants is below a specific number.
But, honestly, it was more of an excuse to learn NodeJS.
I started thinking seriously about side projects as a side project!
So far it involves actually planning stuff (using Trello), working on things for one maybe two hours at a time max (lunch breaks etc) and actually shipping. So far one basic Android app has been released. Hoping for more inspiration soon.
The idea is to use declarative markup to achieve better UX without the complexity of client-side javascript frameworks and retaining most of the familiar web app techniques and advantages. It's a riff on pjax, the basecamp 2 approach and Angular.
So, if you want a button to post to a url, you simply say:
<button ic-post-to="/some/url">Post It!</button>
And you get an AJAX post when that button is clicked.
Another aspect to it is using REST-ful paths to express dependencies. So you could have a div depend on the path that the button above posts to and, after the post succeeds, we would refetch the div content and swap it in (if it is different):
<div ic-src="/some/url/my_div_src">...</div>
Because "/some/url/my_div_src" is a child path of "/some/url", this div will get refreshed when you click the button.
I'm working on a side-project per month. I start each month with a couple of days of identifying which project to undertake, a couple of additional days to work out features that I'd like to see and the spend the rest of the time building it (while working a full-time job, part-time writing, etc.).
My first project, to start with something light, was a text-based game that crossed a Unix shell with an RPG (called RPGFS). You can find source and discussion here:
https://news.ycombinator.com/item?id=7198147
My February project is a simple data processing scripting language called Piglet inspired by Apache Pig.
69 comments
[ 2.6 ms ] story [ 119 ms ] threadI think it will be ideal for AI or at least that is what I'm planning to use it for. I estimate it will take me like six months - 1 year to complete.
If you are into AI, you can join my strong AI group:
skype:?chat&blob=xTKl07MXeVdX9li2-xUCvbGBwwo2xTORkOCKkFzKzCtnuoCpzpoIEwFh0vdzTWDY_MwormVtRe2XkYSwE2k4ataFYwVVBLDkK8kUK3CzBK4d0v-TYpxpUMIhTh8sJzKdfsjVKaFUfkE9Go_GDSWcCif_q7M0LjNDnKTId--qZjTaOvq8YzE
Just copy the link into a skype window and click on it.
Not if I can help it. :-)
Hoping to lurk around though.
I've "swung for the fences" in the past (anyone remember plentyoftweeps.com - twitter dating? :), but this year I want to focus on the exact opposite approach: instead of trying something that's never been done before, I want to try something that I know many others are doing successfully, to see if I earn my way into that group.
Here's a blog post that's dated Dec 30, 2014: http://blog.hndigest.com/from-the-future/
Another project I've started is for the tracking and analysis of athletic training and performance (aimed at D1 to professional athletes).
I'm doing this mainly to learn how to actually make games. So far I have 7 games available. Let me know if you have any questions about it!
Are you doing this full time? and if not, how do you find time to work on a game?
How many hours do you put into a single game?
Something for VR (Oculus Rift) that I hope to Kickstart soon.
Thus far, I've been focusing more on procedural content generation/procedural -anything- related to gaming/storytelling. Togelius has a lot of good stuff.
The project is being built with Go and Angular, so it will hopefully be a dead-simple application to install. I'm also blogging about it occasionally here: http://jordan-wright.github.io/blog/categories/gophish/.
I started Instagram integration as a side project and now you can increase your reach to over 150 million monthly active Instagram users.
-- http://juts.co
http://pasteyebeta.herokuapp.com/ http://github.com/shayanjm/pasteye
[1] http://wristpresenter.com/
http://visualino.com/
All of that said, we don't really use the word "pivot" a lot and our basic focus is now still the same as it was when we started this thing. But we have a few new initiatives lined up for 2014, including some things we're going to do in terms of taking our technology, making it more of a "platform", and building some domain specific SaaS offerings on top of the platform.
It's an exciting time right now, as we have some leads in the pipeline and have a real chance to get some movement and traction this year.
[1]: http://www.fogbeam.com
[2]: https://github.com/fogbeam
The major areas where we differ are:
1. our product is 100% open source, developed in the open, using the ALv2 license, open issue tracker, mailing list, code on GitHub, etc.
2. Our primary focus right now is the "on premises" deployment model (although SaaS delivery will be available)
and (2) is largely rooted in (3)
3. A tremendous focus on integration. The vibe we get from talking to users of Yammer and the other social networking products is that the largely just become sort of a "replacement for email" or maybe a complement to email. There are different reasons for that, but our thesis is that a big part of it is that the products don't lend themselves to integrating into existing workflows and processes. By comparison, we have SOA/ESB integration for real-time business event subscriptions, BPM integration to make a "social BPM" offering, Calendaring integration, and more to come.
4. A big focus on contextual computing, largely driven by use of semantic web tech. As "content" of various types passes through the system, we do semantic concept extraction, clustering / classification, etc., so we can show you - for a given piece of content, business event, person, task, etc. - all of the related (people|content|documents|events|etc).
5. Open standards focused: all our integration points are based on open standards, including RSS, Atom, ActivityStrea.ms, RDF, FOAF, and the like.
Other than that, it's basically the same thing. :-)
But, honestly, it was more of an excuse to learn NodeJS.
https://www.hangoutfishbowl.com
http://blog.lobberecht.com/2014/01/hangout-fishbowl.html
So far it involves actually planning stuff (using Trello), working on things for one maybe two hours at a time max (lunch breaks etc) and actually shipping. So far one basic Android app has been released. Hoping for more inspiration soon.
HTML5 + AJAX + RESTful end points + Basecamp2-style partial swapping => Intercooler.js
The idea is to use declarative markup to achieve better UX without the complexity of client-side javascript frameworks and retaining most of the familiar web app techniques and advantages. It's a riff on pjax, the basecamp 2 approach and Angular.
So, if you want a button to post to a url, you simply say:
And you get an AJAX post when that button is clicked.Another aspect to it is using REST-ful paths to express dependencies. So you could have a div depend on the path that the button above posts to and, after the post succeeds, we would refetch the div content and swap it in (if it is different):
Because "/some/url/my_div_src" is a child path of "/some/url", this div will get refreshed when you click the button.The main site is here:
http://intercoolerjs.org/
And a rough demo is here:
https://vimeo.com/85881209
When I get a bit further along and convince myself it's a decent model for development, i'll do a ShowHN.
My first project, to start with something light, was a text-based game that crossed a Unix shell with an RPG (called RPGFS). You can find source and discussion here: https://news.ycombinator.com/item?id=7198147
My February project is a simple data processing scripting language called Piglet inspired by Apache Pig.