The company Quixey has executives and a number of software engineers into Less Wrong. http://www.quixey.com/
Btw, I like this question, I think the downvotes might have been because your tone or linguistic patterns seem confrontational. A simple statement of the form "I'm curious because" would help, I think.
No, same copy. So my object+meta comment caused people to change their minds about the value of the question as is, or it would have flipped karma balance anyways.
I mean, it's not that confrontational in the scheme of things, but I think what I really meant is that there were no softening or diminishing linguistic patterns. Comments on Less Wrong tend to do better when couched in the language of rational uncertainty. Phrases like "I think", "maybe", "possibly", etc. There's also an (unrational, imo) bias here against short posts. So I was thinking that one sort of softening sentence would make the post seem less decisive while also making it longer, making positive karma more likely by my metrics.
I'm in the Natural Stupidity business. And, judging by the quality of decision-making I've shown throughout my life, I'm going to corner that market eventually.
I don't use the word AI, but what I do is vaguely related, at least by broader definitions of AI. I employ machine learning algorithms to solve problems from several domains such as natural language processing, optimization on geographic data, and consumer data mining. I don't invent novel new machine learning algorithms (few people do), but often have to be creative when adapting and combining existing ones. My Google Scholar profile tells about the published stuff (it's all natural language processing), but much of my more recent work is not published. Things like the optimizer creating gerrymandered electoral maps (web-based interface here: EOVK), or my spatial data mining work for prezi.com.
I work in a game company, and have programmed "NPC AI", though that's AI in name only (though some models use goal-driven behavior and planning, which is closer to actual AI, but not usually what's really needed to make a game fun and bug free - it does become more important if you're making a complex open-ended game though, which is why game companies generally don't make complex open-ended games).
Closer to actual AI work, I did a fair amount of motion recognition on the Wii, which allowed me to put into practice all the machine learning stuff I studied.
I do research in academia in genetic programming and other evolutionary algorithms. Some theory, some applications especially in graphics/design/music.
Like other commenters, I hope OP will elaborate on the question and/or the responses.
I used to work for a computational linguistics research company. There was...way too much cheating with hard coded grammar rules/GOFAI instead of attempts to do statistical analysis. I am not sure if you would count that as AI work. I no longer do.
what AI related work he^H^Hshe is doing?
I've been working to gain an implementation-oriented understanding "systems that learn" for many years. The hard parts of AI seem to me more like philosophy problems than math problems, so I've worked on the philosophy on my own and with friends via reading and conversation while sharpening programming skills professionally as a day job. For example, I've kept intellectual tabs on Hutter since 2002 and Eliezer since 2003. I'm not really comfortable talking about algorithms too much, but there are serious game theoretic reasons to coordinate with other people working on similar projects, and this seems like the best place on the english speaking internet to find some of them.
I downvoted because the question is too vague. What would you consider "AI related"? How could I know?
For example, I have a story to tell about a trivial inference engine on quite a small data set being able to surprise me - I have reasons to guess that your intent excludes this but I have way too little data from the post to have any reasonable confidence.
I brush up against the field. I'm a grad student in computational neuroscience, and work with modeling how the brain's neural networks might be structured for certain tasks. Right now, I'm focusing on issues involving timing at the seconds/minutes level, as well as the neurological architecture involved with perceptual discrimination (Weber's law and the like, if anyone is interested). That may expand in the future, depending on how productive my current line of research is.
Is any member of this community able and willing to tell the whole world, what AI related work he is doing?
A concrete examples, like "I am with the Google translator team" or "I, alone, have done this and this" - are very welcome.