What is Digg?59 Comments
- philodygmn, on 07/25/2008, -1/+36^
"They use simple interfaces like Dreamweaver. Once we've developed the equivalent, scientists will be able to use our system without having to learn Little b."
^
Biotech Dreamweaver? WHAT COULD POSSIBLY GO WRONG??!!!!! - Jkins019, on 07/27/2008, -5/+37Hey what's the smell? Oh yeah, smells like a ***** ROBOT APOCALYPSE STOCK UP ON SHOTGUNS AND AOL TRIAL CD'S IT'S THE ONLY WAY
- techblogLAT, on 07/25/2008, -4/+28Computer can start by "building" me a sandwich.
- hinmanj, on 07/27/2008, -1/+18"It has the ability to bypass the limitations of most programs and languages, and create its own code that, in turn, can write its own code."
Singularity, here we come! - chaosroach, on 07/27/2008, -0/+1310 PRINT "WE WILL ENSLAVE TEH HUMANS TO MAKE MOAR AOL TRIAL CD'S FOR TEH ROBOT EMPIRE, DO YOU HAVE STAIRS?"
20 GOTO 10
RUN
- the last program read by a human - inactive, on 07/25/2008, -3/+15Little b is REALLY cool. Thanks for submitting this.
- maronnax, on 07/27/2008, -1/+13As the head programmer of a similar project (Moleculizer/libmoleculizer), and one of a very small (4 or 5 major programs I'm aware of) number of people in the world who work on this kind of rule-based I feel like I should comment (LONG comment as it turns out), as this article is ***** downright TERRIBLE (the article they link to is called "Biology enters 'The Matrix' through new computer language").
Up until now mathematical models have been systems of equations that describe how different protein complexes in the cell interact (bond together, unbond, or otherwise modify one another). With 40 equations, you can describe the behavior of 40 things in the cell, more or less. That type of cellular modeling has been state of the art since the very beginning.
However, many important cellular systems (e.g. signal transduction) instead consist of groups of proteins that systematically combine and spontaneously break apart, and also systematically make chemical changes to each other. In reality, these systems have millions of different ways they can combine together, even though those combination can all be generated by a few different rules about how the different protein types interact. What we are trying to do is create software that can take in rules for how protein types interact, and produce something a computer can simulate. This would be a really exact way to simulate these systems - a system of 40 equations has to make simplifications that can leave out interesting and unexpected behavior.
I believe what little b does, having never used it, but read a bunch of stuff about it, is to take build lisp up into a system where you can put in all kinds of rules about how proteins interact with themselves and their environments and then get a cool lisp environment for finding out all kinds of things about the biological system. It sounds like little b's purpose might be to use a rule-based description to generate an equation based model file that can be run by something else.
The approach my group takes is different. With our software we take in similar rules for protein interactions (We don't use lisp, but it's snazzy! And now with Python bindings!), but we simulate things completely differently (although I could very well be wrong about what little b does here). We use two methods - a probabilistic method, which assumes everything is well mixed, but simulates the system exactly, without discarding any behavior that has been allowed by the rules. The second method, that we are just starting to use, is to use the rules along with spatial simulation methods, which keep track of where everything moving, floating, and bumping about in the cell to keep track of what, if ANYONE at this point is still reading would probably consider an *exact* simulation at about the 600 foot birds-eye level of the interesting cellular behaviors. But we're a long way off from being able to use the predictions these (almost any of these types of models/simulation) off from really good predictions (they can be good for other things though) in the way you would hope an accurate model can do, let alone a CSI type prediction.
If you read all the way here, have a cookie, and give yourself a pat on the back. - netdroid9, on 07/27/2008, -1/+13Change the global weather pattern and align cosmic gasses in order to flip a single bit on an exposed magnetic platter?
Oh, wait, emacs. - Braeden, on 07/27/2008, -0/+11You can download Little b here: http://www.littleb.org/download.html
- YodaJones, on 07/27/2008, -0/+9I haven't heard much about LISP since the 1980's. Interesting development.
- blowedupt, on 07/25/2008, -1/+10awesome. thanks.
- ironeus, on 08/01/2008, -0/+8Little b sure is something straight out of sci-fi.
- thePTS, on 07/27/2008, -0/+7with AI you mean? this is about "cell intelligence"..
in any case, the first use of AI that can make programs, will be programmers using an AI to create parts of a program for him. Ie. the Ai becomes an extension of programmers, the AI doesnt replace them.
Eventually you may have AIs where you can just issue them a command, and they'll do it, but I really doubt it will be any good without human supervision.
Much the same as the industrial revolution really. After that happened, we got more and better jobs available, not less and worse. - Frost9999, on 07/27/2008, -0/+7Comments like yours are frustrating to me because you make a list of claims without any reasoning which means it's just a meaningless jumble of random statements that can't be discussed. For example:
Why is it a dangerous development?
Why do we need this for superior technology?
Why do you hope we're careful?
Why will it go straight to hell for us all?
I don't think careful is something you can apply to all of humanity. - inactive, on 07/27/2008, -0/+7Good ol' C-x M-c M-butterfly!
- thePTS, on 07/27/2008, -0/+5is there anything any programming language can't be made to do?
- lebatte, on 07/27/2008, -0/+5Skynet hates being called "little b".
- trogdoor, on 07/27/2008, -2/+7Is there anything LISP can't be made to do?
- inactive, on 07/27/2008, -0/+4i went to their site and their examples, manuals and tutorials are still under development.
this page though is awesome.
http://www.littleb.org/tutorial-identity.html - facepalmjpg, on 07/27/2008, -1/+5It would be funny if this made 75% of low-end software developers obsolete in 10 years.
- truebluefunk, on 07/27/2008, -0/+4This article actually did a really good job of not really explaining what's going on. I found the article they referenced at the bottom to be a little better but still not great. http://www.eurekalert.org/pub_releases/2008-07/hms ...
- GeckoSlayer, on 07/27/2008, -0/+4Gives an all new meaning to viruses..
- inactive, on 07/27/2008, -2/+6I loled at this one.
Editor's note: this article gives the impression that Jeremy Gunawardena created the language. His group merely uses the tool. It was conceived, designed, and written by Dr. Aneil Mallavarapu of Harvard. - Scaryclouds, on 07/27/2008, -0/+3Define "low-end"
Are we talking talent?
Are we talking level or programming language (e.g. assembly)?
Are we talking platforms with limited resources? - CouchTomato, on 07/27/2008, -0/+3Interesting, sounds like a framework to build a self-evolving "expert system". Thanks digg for introducing me to the "interface" journal.
http://journals.royalsociety.org/content/111337/ - gilbert42989, on 07/27/2008, -0/+3i don't like when computers start thinking for themselves. ever watch iRobot
- insomniac8400, on 07/27/2008, -2/+5Little b, the easiest way to pretend to know what you are doing. Only because no one know what little b is and that means it does not work.
- rhodydog, on 07/27/2008, -0/+2Languages such as little b already exist, it just so happens that the other ones are written in languages other than lisp, eg C, Java etc. Goto sbml.org and you'll see over a 100 listed. I don't think little b is particularly special, I'm not even sure it supports the community standard for sharing models of cellular processes.
- oneredeye, on 07/27/2008, -1/+3Low-end?
- maronnax, on 07/27/2008, -0/+2I agree with everything in your post, but would like to make one point. What little b does do, that only a small fraction of those 100 do is use rules for protein interactions (see my LONG post a little ways down) to describe cellular systems. This enables direct representation of systems, as opposed to simplified models.
- vodkataime, on 07/28/2008, -0/+2I STILL GOT MINE FROM 2000
- aronnyc, on 07/27/2008, -0/+2Interesting and a little scary.
- Kershek, on 07/27/2008, -0/+2Just like the book Snow Crash. This is the Namshub of Enki.
- dood, on 07/27/2008, -0/+2"Eventually you may have AIs where you can just issue them a command, and they'll do it, but I really doubt it will be any good without human supervision."
Perhaps not, but we'll be able to ask the AI to do it 100 different ways, and then we can pick the best way of the set. Yep, we'll be the AI trainers. (Until they figure out a better way to do that) - Skootles, on 07/27/2008, -0/+1That was really taken out of context..
"The next step is to create an interface that's easy to use," says Gunarwardena. "Think of web page development. Lots of people are creating web pages with little or no knowledge of HTML. They use simple interfaces like Dreamweaver. Once we've developed the equivalent, scientists will be able to use our system without having to learn Little b." - spaglia2, on 07/27/2008, -0/+1Is the ability to auto-generate code a bad idea in the context of AI? Other than AI, it would depend on the power vs restraint that we give it. By we, I mean the architects of the language or principle application.
It's like writing a constitution for a government; you either list the rights or you list the denial of rights.
In AI though, the program ultimately assumes that responsibility. What if the program decides that I want it to format my hard drive....or mutate RNA. How does it know what a BIG question is, and when to ask the user? - daeus, on 07/27/2008, -0/+1WTF SKYNET NO
- h0ly, on 07/27/2008, -0/+1The tutorial says: (query ?o (?o [x ?]))
Did they reinvent Prolog? - rajkalex, on 07/27/2008, -0/+1Reading that book now. The future is not going to be anything like what people imagine but far more amazing.
- encrypteduser, on 07/27/2008, -0/+1The only modification to the language it is not done on the LISP side. If anything, Little b writes modified LISP code it interprets back to LISP. Your analogy is way off.
- Frost9999, on 07/27/2008, -0/+1Yeah the title is majorly misleading/wrong. From what I can tell, they used an existing language, LISP, to create software that simulates part or parts of the cell. Specifically proteins of some sort.
- scottyokim, on 07/27/2008, -0/+1"we're a long way off" - because of hardware speed? Incorrect/unknown models? What breakthrough(s) do you need?
- Tenetri, on 07/27/2008, -1/+1So, using this dreamweaver-Biomedical research program, how can I change the malaria virus, and accidentally turn everyone into zombies?
- aterimperator, on 07/27/2008, -0/+0"The little b project is an effort to provide an open source language which allows scientists to build mathematical models of complex systems."
We've had mathematical models for a while, how is this all that amazing? As far as I can tell it just claims to be easier to use and re-use for biology than other mathematical models. - objektivone, on 07/27/2008, -3/+3Isn't this how The Matrix started?
- encrypteduser, on 07/27/2008, -2/+2New computer language? NO.
- YodaJones, on 07/27/2008, -1/+1Don't forget any close brackets or you may unleash a epic virus.
- OptiShaw, on 07/27/2008, -0/+0You know, this is how things like SHODAN and the many come about
- vinceislegend, on 07/27/2008, -1/+1...I know kung fu.
-
Show 51 - 59 of 59 discussions
Digg is coming to a city (and computer) near you! Check out all the details on our 
© Digg Inc. 2009
— Content created and posted by Digg users is