Reply to post:

It is no better than a Beaglebone or raspi for newcomers. If the aim is to play with MIPS the linksys WRT54GL does the job and has been there for years. I don't see the market for this one, it is too expensive and without documentation for newcomers and too useless for the rest.

----

I felt like writing to practice my English so here it goes to whoever wants to waste time reading it ;)

At a bit higher price than that MIPS board, the cubieboard or the pcduino like A80 based chinese boards look better and have a similiar/better graphics capability (PowerVR G6230 but as closed source as the previous one). Shit, Cubie started alone. He is doing it quite well for being new in the market, even in the hackable things market. He got lucky and played very well creating his company after writing and publishing the bootloader for the A10 while he was a freshman at Allwinner (I remember him from the mele2K days at freenode). By now he has already made many mistakes and is in a better position than Olimex who didn't have access to cheap SOCs then (veterans in hackable things and 1337 but sadly small, not well known and late to cheap SOCs) but he is taking big risks in unknown territory. His problem now is to scale from garage hacks to mass production in a market where in raw power things get outdated in six months (stocking is a problem and not stocking is a problem too) and he also needs more brains, not only tinkerers. His brain clearly works well but won't scale, a single brain can't hack, run a company and scale it at the same time, especially with a wife/family.

For the new serious league (big autonomous robots, cars...) they're all cheaper than a nVidia Jetson TK1 but for embedded graphics nVidia is far superior and now the graphics are used to compute so the rest should have OpenCL at least to compete with Nvidia and without good OpenCV support they just can't compete for vision.

Intel benefited from wintel so long that just doesn't get it, they have a big army fighting guerrillas and nonexistent buzz. Galileo is crap and late, Atoms now can't compete with the Jetson in power and the embedded developer base is already on ARM so they're trying to compete with chinese vendors (they made an alliance with Rockchip but only for some products and not related). The big advantage intel had with PC is now with ARM, developers developers... Balmer mode on... so they have to be better, cheaper and do what they do "well", document the stuff and give info+stuff to the developers.

One plus for intel as for Freescale is that they sell the same IC's for long periods of time but they do it so expensive that developers just go elsewhere and fight against the unknown thanks to leaks and colective reverse engineering (2012-2014 been there, done that). I began learning chinese because sometimes the chinese companies leak on purpose even if they're not allowed to because they have no idea how to solve some problems on time (and I wonder how many of them go crazy or what they do not to go crazy) and the community sometimes saves the day while they have interest on it (a couple of months+-). Examples: HDMI flickering, WIFI/Bluetooth problems, overheating or hanging due to crappy power supply DVFS, horrible drivers fighting for IRQs, Skype camera/USB audio support, hardcoded MAC addresses, absurd code for the H264 decoder/IPP ....

China will win the price war but their products are usually obselete fast and without support they're soon paperweights (the cubie/radxa too but they're muuuuuch better than the rest chinese ones). The only way the rest can compete is with good support, open-software and good manuals. This mips crap with no open GPU and no SATA is less usefull than a radxa at the same price.

ARM is the new x86 when it comes to developer base and MIPS with a similar price won't atract them (no Linaro, no Ubuntu, no Arch... no common Linux base= for a very specific market, probably too low to be considered)

Arduino is not on the same league but if they want tinkerers they should remember that Arduino didn't won the market because they where better than PIC boards or because they had new tech (they copied Wiring). They won because it was cheap enough (cheaper than BasicStamp), open source and because they got tons and tons of documentation so that people could copy and paste. They also encouraged the shields being open source.

They won because they made making mistakes less likely (documentation) and cheap. The industry went ARM (phones, battery/price) and the educational centers had to teach it. Now the inertia is so big that intel is having problems to attract developers (even the ones with a lot of free time to debug, document and show off with useless projects: students).

Real developers are always a few and/or they contribute small amounts of code but the copypasters are legion. Now Arduino still sells because a kid/ID-10T can do a "look ma I made it blink" much cheaper than with a Lego NXT with even less skills.

The Arduino guys didn't get rich with the MakeyMakey and other boards but they're still getting money and can even afford to sell crappy printers now. People still buy the Arduino _brand_ instead of clones for the same reason they drink Starbucks, some have money and like the non functional experience of the envelope and instead of the greeness/environment/third world they do it for the open source.