The chances that your vote will affect the outcome of an election are practically zero regardless of who you vote for.
The way to make your vote count is to use it to increase the perceived support base of the candidates and/or party you like best. For the smaller parties, this can lead to improving their reputation, and for the next election, access to government campaign funds and less red tape in getting on ballots.
To work, they need to be fired either where the player is going to be, or where you don't want the player, to herd them. Bots can't do this.
Sure they can. It's just more complicated.
The Visual Dynamics Research Group at Oxford has some really cool demos that show the results of using probabilistic prediction techniques to improve image-based tracking. Using the same techniques to track something whose past and present location is known precisely (rather than being inferred from noisy observations) is much easier.
As you yourself already said, you just teach the bot about whatever kinds of movements you want it to be able to handle. Once you have lots of common models in there, you can tune its expectations by having it learn during real games.
A lot of folks here complain about how obvious things are patented. Other people chime in that those things are obvious now, but weren't obvious when they were patented. That's a good point, but I think it's often not true, and that shouldn't surprise us. Every day we read about why that's the case.
The patent office tries to determine whether something is obvious to knowledgeable people in the relevant industry. But they're a US government agency, and they can't pay enough to keep knowledgeable software people on staff. Being a patent officer is not a glorious job. Do you know anyone working there? I don't. Know anyone who wants to? Me neither. How much would it cost to have clueful people on staff, or to consult some regularly? I know lots of people clueful about the software industry, and, not surprisingly, they're all working in the software industry. It's easy to find an exciting job that pays well. How can the PTO compete? They can't, of course. Worse yet, it was years before they even tried.
We've heard recently how hard it is to get corporate engineers to talk to corporate patent lawyers. There are people who base careers on patenting things their engineers thought were obvious. In many cases, I'd guess that there were engineers at other companies who also thought those things were obvious, but since no one spent $100k to have a lawyer go around and ask them, we get stuck with more stupid patents.
In April 1999, Q. Todd Dickinson (Acting Commissioner of Patents and Trademarks, USPTO) gave a speech, which I'll quote briefly here:
[O]ne of the reasons some have questioned the ability of the PTO to gauge software innovations is the lack of an organized, meaningfully complete library of prior art. In this sense, software has presented PTO with a very different challenge than, say, biotechnology. Patenting was at the ground level of biotechnology; the PTO has been watching biotech since its inception.
But by the time of the watershed Diamond vs. Diehr case and its clear mandate to issue patents in the software realm, there were already decades of programming under the industry's belt. Since then, PTO has embarked on an ambitious programs to catch-up - a program I am committed to continuing.
I hope they catch up soon, but I'm not optimistic. I don't think it's a question of building libraries; they'll never keep up. They need to stop relying on full time examiners and start asking people actually in the industry.
As for Google's patents, I haven't read them. I'm hoping it's more than just scoring relevancy based on how something is referenced. That's been obvious to me for a long time, and I was waiting for someone to build a search engine that way. There are a lot of people who think of a lot of neat patentable things, but who don't have the time/money/desire to patent them. No matter how many people think of the same thing though, it's still fair game for whoever bothers to apply for a monopoly.
Why is government money being put into this? It's one thing to fund basic science that might possibly someday benefit people if only someone would work on it. But this has huge, obvious, and (relatively) near term applications. Why not leave it for private industry and save the federal money for something else?
anyone in the world (except perhaps China) can freely download the latest in crypto technology from a number of countries.
So what? Availability is nice, but law enforcement folks don't care whether it's available, they care whether it's used. It's clear that availability isn't enough to promote use. It needs to be ubiquitous and effortless, and that means building it into popular software systems.
If Microsoft, Netscape, and AOL all built strong encryption into their mail software, everyone would be using it. That's what law enforcement is afraid of, and the current legislation is effective at preventing it. They're not stupid; they're getting exactly what they need.
That's bullshit. I ride my bike all over Sunnyvale. Most of the time, I'm riding in bike lanes. The area isn't dense enough for biking to be good enough all the time (I often need to go ten miles up to Palo Alto, and for that I take my car), but other than that it's a great place to get around by bike. The streets are wide, there are bike lanes everywhere, the weather is amazing, and it's flat.
Slashdot Mirror
dselect!
The chances that your vote will affect the outcome of an election are practically zero regardless of who you vote for.
The way to make your vote count is to use it to increase the perceived support base of the candidates and/or party you like best. For the smaller parties, this can lead to improving their reputation, and for the next election, access to government campaign funds and less red tape in getting on ballots.
Sure they can. It's just more complicated.
The Visual Dynamics Research Group at Oxford has some really cool demos that show the results of using probabilistic prediction techniques to improve image-based tracking. Using the same techniques to track something whose past and present location is known precisely (rather than being inferred from noisy observations) is much easier.
As you yourself already said, you just teach the bot about whatever kinds of movements you want it to be able to handle. Once you have lots of common models in there, you can tune its expectations by having it learn during real games.
A lot of folks here complain about how obvious things are patented. Other people chime in that those things are obvious now, but weren't obvious when they were patented. That's a good point, but I think it's often not true, and that shouldn't surprise us. Every day we read about why that's the case.
The patent office tries to determine whether something is obvious to knowledgeable people in the relevant industry. But they're a US government agency, and they can't pay enough to keep knowledgeable software people on staff. Being a patent officer is not a glorious job. Do you know anyone working there? I don't. Know anyone who wants to? Me neither. How much would it cost to have clueful people on staff, or to consult some regularly? I know lots of people clueful about the software industry, and, not surprisingly, they're all working in the software industry. It's easy to find an exciting job that pays well. How can the PTO compete? They can't, of course. Worse yet, it was years before they even tried.
We've heard recently how hard it is to get corporate engineers to talk to corporate patent lawyers. There are people who base careers on patenting things their engineers thought were obvious. In many cases, I'd guess that there were engineers at other companies who also thought those things were obvious, but since no one spent $100k to have a lawyer go around and ask them, we get stuck with more stupid patents.
In April 1999, Q. Todd Dickinson (Acting Commissioner of Patents and Trademarks, USPTO) gave a speech, which I'll quote briefly here:
I hope they catch up soon, but I'm not optimistic. I don't think it's a question of building libraries; they'll never keep up. They need to stop relying on full time examiners and start asking people actually in the industry.
As for Google's patents, I haven't read them. I'm hoping it's more than just scoring relevancy based on how something is referenced. That's been obvious to me for a long time, and I was waiting for someone to build a search engine that way. There are a lot of people who think of a lot of neat patentable things, but who don't have the time/money/desire to patent them. No matter how many people think of the same thing though, it's still fair game for whoever bothers to apply for a monopoly.
Why is government money being put into this? It's one thing to fund basic science that might possibly someday benefit people if only someone would work on it. But this has huge, obvious, and (relatively) near term applications. Why not leave it for private industry and save the federal money for something else?
So what? Availability is nice, but law enforcement folks don't care whether it's available, they care whether it's used. It's clear that availability isn't enough to promote use. It needs to be ubiquitous and effortless, and that means building it into popular software systems.
If Microsoft, Netscape, and AOL all built strong encryption into their mail software, everyone would be using it. That's what law enforcement is afraid of, and the current legislation is effective at preventing it. They're not stupid; they're getting exactly what they need.
That's bullshit. I ride my bike all over Sunnyvale. Most of the time, I'm riding in bike lanes. The area isn't dense enough for biking to be good enough all the time (I often need to go ten miles up to Palo Alto, and for that I take my car), but other than that it's a great place to get around by bike. The streets are wide, there are bike lanes everywhere, the weather is amazing, and it's flat.