Slashdot Mirror
Intel Exits the Maker Movement (hackaday.com)
Reader szczys writes: Intel just killed off its last "maker movement" hardware offering without fanfare by quietly releasing a Product Change Notification PDF. The Arduino 101 is halting production on September 17th. This microcontroller board is built around the Intel Curie module around which Intel bankrolled a television series called America's Greatest Makers. News on the end of life for the Arduino 101 board follows the recent cancellations of their Joule, Galileo, and Edison boards. This is the entirety of Intel's maker offerings and seems to signal their exit from entry-level embedded hardware.
This is about the 15th iteration I have witnessed of this.
Step 1: Somebody at Intel gets a Bright Idea to develop some new market. We gots lots o cash so why not lets do it. (i.e. collect underpants)
Step 2: ...
Step 3: Profit.
Step 3.1: Er, no profit. We ended up not owning the market. Pull the plug lets get back to our core competency: i386-architecture processors it is. What's AMD been doing recently?
When compared to Arduino Uno R3 and other microcontrollers, Intel microcontrollers were too expensive, bulky, and slow compared. No communitiy formed around these microcontrollers because Intel was unresponsive to the community. Sort of like big companies like Oracle who are unresponsive or allergic to community efforts.
Let me see: $25 for an Ardunio UNO with a huge responsive community vs. $90 for a microcontroller from Intel where there is no community. Hmmm...
Intel is in serious trouble, but not just because it is no longer a generation ahead on process tech. Sure, soon they will even be a generation behind, but thats not their biggest issue.
Their biggest issue is that as a vertically integrated company they can't be part of Pure-Play and there arent any IDM opportunities. If it isnt branded Intel, then it doesn't come out of an Intel fab. All of Intels older fabs are grossly under-utilized while Pure-Play foundries like TSMC can keep all their fabs old and new running 24/7.
Being behind on the technology is a surmountable problem. Being behind on the business model isn't.
"His name was James Damore."
Maybe Intel is facing stronger competition these days. But the real problem in this case probably just is that the "Maker Movement" was nothing more than yet another hipster fad. It's typically not possible to make profit off of hipster fads.
The difference between hipster fads and normal fads is that hipster fads exhibit a social media amplification effect beyond what even most other fads exhibit.
In a typical widespread fad, there is a large amount of demand and a large amount of hype. But the ratio of demand to hype is relatively even, allowing for some profit to be made off of the fad. For every unit of hype that attracts the attention of producers, there's a corresponding unit of demand. The producers can actually sell what they're making, and if they do this well they can make a profit.
A hipster fad is different in that the ratio of demand to hype is very skewed. There is little demand, but a comparatively huge amount of hype. This means that profit cannot be made as easily, if at all. For every unit of hype that attracts the attention of producers, there is only a very small fractional unit of demand. The producers end up producing too much relative to the small demand, which actually has the effect of driving prices down to below profitable levels.
This situation arises because hipsters excel at producing hype, especially on the Internet, yet never manage to back up this hype with ability. This is quite evident within the computer software industry, where a lot of hipsters try to participate. We see huge hype bubbles surrounding technologies like Ruby on Rails and Rust, yet despite all of the hype we see very little of value actually delivered.
It's similar for this low-end computing hardware. A lot of hipster types like to exclaim how they're "makers" because it sounds good in social media discussion, yet they lack the skills, ability, interest and typically the money (they usually spend this on overly expensive products, such as those from Apple) to actually go out and buy any low-end computing hardware. If they do happen to buy it, they end up not doing anything with it, or stop trying after quickly failing to accomplish anything with it.
The real lesson here is that companies in general should smarten up and ignore what hipsters are saying and doing. They're among the worst markets to target, unless you're Apple and you're exploiting the vanity of hipsters. But there's only room enough for one Apple-style company which totally drains the resources of hipsters, leaving very little for the rest of the market to extract. That's why the rest of the market shouldn't even bother.
Anyone who ever seriously considered Intel's "maker" products for embedded use (as I did) would quickly find that they were power hungry and over-priced. Either one of those problems would be death in this market.
Marketing can cover a lot of sins, but these together are really hard to overcome. "Let's market the shit out of this shit" only really works when aimed at consumers or executives. Makers are engineers. They can smell the bullshit from miles away.
It's a healthier marketplace without such inferior offerings in it.
the growth in cynicism and rebellion has not been without cause
I'll give you a maker success where Intel failed. Check out the ESP8266
http://esp8266.net/
https://en.wikipedia.org/wiki/ESP8266
Originally sold as a low cost serial-connected wifi solution so manufactures could at "Internet" to products. Espressif also offered the now ubiquitous ESP-01 module that quickly became a very popular and cheap way to add wifi to existing arduino projects.
Then someone started poking around at the chip itself and found it was many times more powerful than the micros that it was being connected to. (About 80 times faster than the atmega that powers the Uno) Also has a LOT more memory and flash.
They asked Espressif for and SDK.. And they gave it to them. Fast forward, we've got an opensource tool chain and a number powerful frameworks including native arduino support.
And arduino-like boards with gpio and power regulators and all that fun stuff so you can run sketches directly. Like the Wemos D1.. Which is about 4 bucks on ali express. (Yeah. 4 bucks for a really fast arduino WITH wifi and more memory and more storage)
The ESP8266 was designed as a wifi-add on so it lacks a lot of features and gpio that the usual general micros have. Fortunatly the thing is fast clever programming and bit-banging can get around a lot of that. Still, there is a lack of physical pins.. And you have to be careful not to stall the or tie up the CPU with your code or wifi will drop. .. So enter the ESP32. Espressif hired the guys who wrote the tool chains above and put out a new product. Wifi, bluetooth, More memory and flash. Two cpu cores so services can run on one and your code won't break them. Lots of GPIO pins with native hardware support for a lot of peripherals and communication protocols.
https://espressif.com/en/products/hardware/esp32/overview
Open SDK and toolchain. Arduino support is still a work in progress and other frameworks are coming right a long.
This reminds me of the book Makers. In that book, near the beginning, Polaroid tries to reinvent itself through the "maker movement" and completely self-destructs. The "maker" thing is really the same thing as "open source" which is people sharing ideas with each other. As soon as you try to apply business philosophy to that, it's like trying to apply a business philosophy to the Apollo program. When you don't know what's out there, you just need exploration, not exploitation. Companies can't handle the risk involved - it's all too uncertain for a 4 year payback.
People in their own workshops and homes are doing some amazing stuff. Companies see this and think they need to get in on it, but the hobbyists aren't doing it to make money, and most of them don't have much money, so there isn't much money to be made.
When I was younger I made less money and didn't have a family so I had a lot more time to put into hobbies and stuff. I scrounged through spare parts bins and re-used whatever I could. I could spend hours price-comparing and trying to find components that were $2 instead of $3 to save a buck. But back then I was a real "maker". Now I have more money and a lot less time. I take on much simpler projects and I'm willing to spend money to get it done faster. I don't buy the $2 or the $3 component - I buy the $20 pre-assembled solution and plug it in. So I don't want to buy Intel's fancy Arduino because it doesn't really save me any time, and my old self wouldn't either because it's too expensive for what it does.
"I have never let my schooling interfere with my education." - Mark Twain
Intel's problem is that they have already cut their losses and ran from the invasion of phone/tablet products. It's the 11th straight quarter of declining PC shipments. Meanwhile smartphone sales are up again now outselling PCs at a rate well over 5:1. Tablet sales are also down (Q1 numbers) so you might say Microsoft has managed to shore up the convertible/laptop market with the Surface line, but WinTel is completely on the sidelines in the global smartphone revolution. According to the platform statistics 53% of all Internet access is now mobile, 42% PC, 5% tablets.
Intel is not in trouble, they have the server market and so far AMD's offering is basically a return to competition, it's a long way to go until Intel is on the ropes fighting for survival. But they and Microsoft completely failed to bring out a good x86 smartphone leveraging the tons of existing win32 code, I don't know why. I mean all the alarm bells should have gone off when the iPhone became a success in 2007, even with 3-4 years development time they should be ready to kick ass around 2011 but instead we got the Nokia flop. Considering the power of phones relative to typical office applications I'm kinda waiting for the phone with a cheap dock that gives you charging, display, keyboard, mouse, a chromebook-like UI and a bluetooth headset in case you need to answer the phone while docked. Like if you already have a phone and a TV, add these accessories and you won't need a laptop.
Live today, because you never know what tomorrow brings
Actually, Why not? She totally rolled her own email server.
Intel does contract manufacturing now.
W..w..W - Willy Waterloo washes Warren Wiggins who is washing Waldo Woo.
Intel does contract manufacturing now.
Although Intel does contract manufacturing (achronix, altera, spreadtrum, LG, panasonic), they really are only competitive in the leading edge process nodes, and aren't very competitive in providing N-2 generation foundry service where the pure-play companies make the bulk of their money.
I suspect that's mainly because they've only recently licensed the Artisan library cells (owned by ARM) which are used by nearly all fab-less semiconductor design houses. Because they were vertical for so long, I'm guessing the existing Artisan cell design aren't parametrically optimized to run on their process and their N-2 fabs were never optimized to yield these cells. It's a big effort to resynthesize/recharacterized/redesign a bunch of the "chicklet" IP and analog cells that goes into a modern SoC chip to an "intel-proprietary" library and it is difficult if not impossible to hide yield hits and NRE costs of re-engineering to a new foundry process by simply discounting die to your potential customers unless they are really big customers that are willing to target your N-2 foundry and invest a bunch of money to retarget parts of old designs. That is why all their newly engaged foundry customers are targetting their leading edge process nodes with new designs.
Historically, Intel has effectively filled their fully depreciated N-1 and N-2 fabs to run their own chipset and embedded business designs (saving the N generation for cpus). Getting new customers to target these N-2 fabs will be more challenging until a few more tick-tocks pass and the current designs become N-2 legacy designs.
I'm sure they can do it eventually, but the question will be can they be as profitable this way? Will customers trust Intel as much as TSMC to be a reliable die provider for N-2? They probably need to fill N, N-1, and N-2 fabs to make this work. Simply increasing capacity on N generation fabs will create more N-1 and N-2 fabs in the future. Only time will tell if they can work this out.