The Broadcom BCM283 system on a chip is impressive, cramming all that capability into one cheap part. So is the Allwinner, which is similar, costs $7, and is the basis of tablets that cost $40. The Raspberry Pi is just a breakout board with a crappy connector layout. There are lots of other ARM boards, most with better layouts.

It's written in C and it's a buffer overflow exploit, right?

We warned you. You didn't listen. Now suffer.

Right. That goes back to Craig Reynolds' "Boids" paper, and is the basis for much crowd behavior in movies and video games. In the real world, it's less useful. The general idea is that there are attracting and repelling fields, and you add up the fields and get a direction vector. This works OK in not-too-crowded spaces. It's great for birds and fish. When there are turning circle limits, or narrow lane limits, it's not as useful.

We tried this approach on our DARPA Grand Challenge vehicle in 2004-2005. It wouldn't work in tight spots. We had to go with a completely different approach. We ended up projecting alternative curved paths in front of the vehicle out to stopping distance and picking the best one.

It looks like these Nissan robots are slow, round, and can turn in place. That's the ideal case for a flocking algorithm.

It looks like Slashdot is having server problems:

"Famous for being famous" used to be a Hollywood thing. (Angelyne is considered to have invented this. In the 1980s, she rented billboards in LA to promote herself.) Now it's a Silicon Valley thing too. Paul Saffo and Vivek Wadhwa come to mind as heavily into self-promotion but lacking a track record of results. Nicholas Negroponte (MIT Media Lab, One Laptop Per Child) is close, but he actually got some real things done in his younger days.

As with Hollywood, it helps to be good-looking. Shai Agassi, former CEO of Better Place, the failed car-recharging company, was like that. I've met him. He's very good looking, a good speaker, and his business plan was bullshit. His company got over half a billion dollars in funding before it went bust.

It's probably worse in "social", but I try to ignore that crowd.

Just use an autonomous quadrotor for the last hundred feet. You could even deliver to apartments that had a small landing pad outside a window, maybe on a balcony or on top of an air conditioner.

Buy a MariaDB support contract. If your business uses databases, this gets you access to the MariaDB developers. It might help you get rid of some expensive Oracle databases.

No, nobody is going to steal your board game design. Or play it.

Yahoo Finance, which was very popular in the financial community, has also been "redesigned". Yahoo Finance was popular because you put in a ticker symbol and you got a chart and all the key performance numbers on one screen. Yahoo was the first to have stock charts where you could easily change the time period displayed, and investors liked that.

Now, there are four rows of Yahoo menu bars at the top of a stock symbol page. There's a big Flash ad at the top. There's a "trade now" button. ("Please provide feedback on the new Trade Now function.") There's another ad. There are links on the left. That's all you get "above the fold", before scrollling.

Below the "fold", there are some links to "reports" Then there are those annoying "Ad topics that might interest you" links. (Not Outbrain, Yahoo does this in house.) There's a table of the top holdings in the fund. Continued scrolling finally gets to the numbers that matter: YTD return, 5-year return, beta, etc.

Yahoo has completely missed the point of why investors go to a page like that.

Nine air charter companies now serve Burning Man. No waiting in line with the peons, either. The temporary airport (88NV) has its own ticket gate.

If you really need a phone, get an Iridium phone. Reception should be great; the whole sky is visible on the playa.

Microsoft's big problem is simply that Windows 7 is quite good. Business desktops use it, they work fine, they crash rarely, and they get the job done. Microsoft conquered the driver quality problem by forcing drivers to pass the Static Driver Verifier, a proof of correctness system which looks at source code to see if it can buffer-overflow, make improper calls, or otherwise crash the kernel. That took care of about half of crashes. The other half, from Microsoft's own code, were handled by a system which classifies core dumps by commonality, so they can collect core dumps with the same cause, then find and fix the problem. So Microsoft conquered the big problem that business cares about - Making It Work.

Businesses see no need to "upgrade". Certainly not to Windows 8. Or Office N+1. It won't help the business.

Microsoft struggles with being "cool". Apple does well with "cool", but nobody else does. It's not clear it will help in the post-Jobs era. (Olivetti once made beautiful office machines. It didn't help them. Most major museums of modern art have some Olivetti products, but few offices did.)

What really made the iPod work was deals with the music industry. Something that many people miss is why Jobs was able to pull that off. Jobs was also CEO of Pixar, and thus, as a major film studio head, at the top of the Hollywood hierarchy. So he was able to deal with the music industry from a position of superiority. That's what made iTunes. (The hierarchy in Hollywood is very real, and very rigid. Ask anybody in the industry.) That's what re-launched Apple. The Mac was below 10% market share, and was stuck there for years, even after Jobs took over again.

There's room for a breakthrough in user interfaces. The rectangular grid of single-purpose icons is lame. We can be sure that breakthrough will not come from the open source community.

The competition is in December 2013, and this team may not be ready by then. Here are the other robots being entered.

The simulated challenge back in June revealed that the entrants' movement control software isn't very good yet. The winning team's simulated robot fell down 12 times. DARPA has posted only heavily censored videos of the results, possibly because they're so embarrassingly bad.

Some of the blame attaches to the simulator used. The Gazebo simulator's physics engine, which is borrowed from video games, is not good enough for the job. Video game simulators use tricks that look OK, but aren't physically realistic. That's no good when you're using them to match a real robot, or even if you're doing control based on reported forces from the simulator. This should be fixed in early 2014 when they get an honest physics engine from Mike Sherman, who knows what he's doing. (If you need a accurate humanoid robot simulator right now, try OpenHRP3, from AIST in Japan.)

I suspect that the December 2013 event, which will be public, will be rather disappointing. But the planned 2014 event may be very impressive.

That's how it went with 2004 DARPA Grand Challenge for automatic driving, which was so pathetic it was covered by the Comedy Channel. Then in 2005, all the robot vehicles at the event could drive autonomously without running into anything and several finished the whole course with good times. The second day of the 2005 Grand Challenge was the moment when automatic driving became real to the world.

Their water processing facility doesn't work yet. "Recent leaks from a novel type of radioactive water treatment device, currently under trial runs at the crippled Fukushima No. 1 nuclear power plant, occurred from corrosion holes in welds, Tokyo Electric Power Co., the plant operator, said July 25, 2013."

It's been years since the event, and Fukushima still doesn't have a radioactive water processing plant. The US has dealt with this problem before, both at 3 Mile Island and some Superfund sites. Water itself doesn't become radioactive (except for tritium, which has a 12 year half life); as with fallout, the radioactives are mostly solids in the water, and can be removed and converted to smaller amounts of solid waste.

With a processing plant, they could reuse the cooling water, instead of building more and more storage tanks.

Me either. Chrome "OS" is mostly a user interface on top of Linux. A server doesn't need a user interface.

If anything, there's an argument for a much simpler server OS than Linux. Something that's more like a virtual machine manager with remote facilities for loading, starting, and monitoring client image. The client images need a minimal OS that's more like a run-time library - no file systems, no drivers, no GUI.