Predicting the Future of Electronics and IT by Watching Component Demand (Video)

A big question college students should be asking is, "What IT and electronics knowledge will be most in demand five or six years from now?" In these fast moving niches, an answer is almost impossible to come by. But what if you were one of the people who supplied raw components to the electronics industry? Wouldn't you have a better handle than most on what kind of devices and components are becoming more popular among prototypers and engineers? And wouldn't watching those trends possibly give you at least a little insight into what the future might hold? Randy Restle, Director of Applications Engineering at component supplier Digi-Key Corporation, carefully tracks orders and tries to determine what's hot and what's not. His reason for doing so is to figure out what Digi-Key should stock in coming months and years. But his insights can also be used to decide what you might want to study or -- if you're already working in the field -- what products you or your company should consider developing. Digi-Key also has an online video library where they feature new products and give ideas of what you can do with them. Even if you're not an engineer or electronics hobbyist, it's fun to see what's available but may not have hit the mass market quite yet.

Comments

Does Betteridge's law of headlines apply?

[MikeTheGreat]

If so:
No
No
No
and No.
But thank you for posting a summary that's nearly 50% questions!

it's pretty obvious, it's ARM... still

[Gravis Zero]

ARM designs the most popular processor architectures in the world. there is an ARM core in literally billions of machines and i dont just mean cell phones. modern ARM chips run anywhere from 12 MHz to 2.2 GHz and they can scale to run much much slower to save power big time (there is an ARM chip that rivals the MSP430 chips). now with the ARMv8 arch, i think we will be seeing some serious inroads made on the server market. of course, ARM will continue to be in everything from your coffee maker to the chips that the NSA secretly implants in people. :P

the past, present and future is ARM.

Re:it's pretty obvious, it's ARM... still

[ebno-10db]

This advertisement paid for by ARM Holdings, PLC.

A big question college students should be asking

[Anonymous Coward]

How the hell am I going to pay off my student loans by entering an industry moving towards decentralized. lowest-bidder IT and commodity hardware, where the labor market is global and comprised of people who have either been in the field for decades or can live on peanuts compared to you, where the brighest minds of a generation are bent on extracting pennies from stock trading algorithms, or coming up with new ways to make you look at ads, or engaged wholesale invasion of privacy.

Do I really want to piss away the best years of my life writing code for yet another tech startup with no business plan beyond IPO, making billions for investors while getting nothing in return? To know that, in the end, I made no difference in the world?

My advice: Make computer science a hobby, not a career.

Re:A big question college students should be askin

[dyingtolive]

Amen. If I could start over again, same circumstances, I'd be a fucking lawyer in a heartbeat.

Solid answer to the impossible question

[HeckRuler]

A big question college students should be asking is, "What IT and electronics knowledge will be most in demand five or six years from now?" In these fast moving niches, an answer is almost impossible to come by

Actually, I believe there are good solid answers to this one that have been true for a decade and will likely be true for the coming decade.

First off kid, you have to understand that there are a lot of fields you seem to be lumping together. There's a difference between code-monkeys, sysadmins, network engineers, electrical engineers, embedded engineers, and web-devs.

For any programmer there's a big question of which programming language to learn. This is something that induces flame-wars and strong passions because everyone has an opinion and their own choice is best. This is because it's an inverse tragedy of the commons, everyone wants you to learn their language because it benefits them and their language to have more users. But a binary search tree is a binary search tree in any language. Some are more verbose. Some are cludgy. But if you understand binary search trees, or whatever, the language used to deal with them by and far doesn't matter. Knowing the syntax of a language doesn't make you a good programmer. Knowing how to use the language to accomplish meaningful tasks, that's what's important. It's a little easier if you learned C rather than IBM RPG back in the day, but if you could learn RPG, you can pick up C without serious problems.

For Web-devs, they'll fret over... let's say... which CMS project is better: Joomla, Sharepoint, Drupal, Django, Wordpress, yaddayaddayadda. Conformity is nice and picking one is important. But you're a COLLEGE KID, when you graduate you'll know what goes into a CMS, theoretically how to make one, and how they work. If you just wanted to learn how to turn it on, you should have gone to a tech school. They'll hold your hand and read the manual with you.

(By the way I also have a thing against "certification". It might make sense for the sysadmin types, but a cert on a programmers resume is a net negative.)

Sysadmins, network engineers, and the hardware guys all probably have similar stories. There are common tools out there you should know, but god knows everyone and their brother make a version of it. Try not to tie yourself to one particular set of tools least you suffer from over-specialization.

tl;dr: It doesn't matter what specific component, language, framework, or gadget is popular in 6 years. You're in college, not a tech school. Learn the basic fundamentals of your field and whatever the hip new thing is will fall nicely into place and you'll understand what it's doing and what's going on. You need to learn how to use a hammer and nails to build things, not fret over which hammer is the best bet.

Re:Solid answer to the impossible question

[fermion]

Or, you shouldn't be going to college for a specific job. While many jobs require a degree or even a very specific degree, paying 100K to a university so one can be a code monkey may not be the best thing to do. The reason to go to college, to get a degree, other than the fiction that a piece of paper will inherently get you a better job, is to become well rounded and, well, educated. I know too many people who put those four years into building their technical skills and achieving what they really wanted to believe that college is the only path to a career. It, however, a very good path to the career you will have when the current flavor of the month loses favor and one has to find a new well paying job.

Which is why I say major in something interesting, and use the four years to become educated and leverage whatever recourse the university has to leverage your skills. It is said, Bill Gates, for instance, leveraged his access to free computer time in high school and university to develop programs. In my experience, the number of toys that can be accessible to a university student who is willing to insinuate themselves into certain positions is significant.

Re:Solid answer to the impossible question

[ebno-10db]

a binary search tree is a binary search tree in any language

I hope employers still see it that way. In years gone by, I interviewed people for software positions and often didn't bother to ask what languages they knew (or if I did, I didn't consider it important). My attitude was that if you can't quickly become proficient in a new language, we made a mistake in hiring you. These days though I see a lot of job ads for "Language Du Jour Programmer". Say all you want about code monkeys, but if employers want experience w/ a specific language, what can you do?

Some are cludgy.

Unfortunately you failed the most important test. It's "klugey" (American) or "kludgy" (British).

Re:Solid answer to the impossible question

[GarethIwanFairclough]

Some are cludgy.

Unfortunately you failed the most important test. It's "klugey" (American) or "kludgy" (British).

He certainly did, the insensitive klod

Been there, done that

[frovingslosh]

Over three decades ago I worked for a minicomputer manufacturer (sometimes known as Data Who?), in Field Service support and later in Systems engineering. Those of us in the field were able to put together a very good idea of what new products were going to be released, not from listening to the rumor mill, but by looking at the parts lists that were being published internally and seeing what components were being bought and assigned internal parts numbers. It's amazing what you can learn about supposedly secret projects just be seeing what the company is buying.

Re:Been there, done that

[ebno-10db]

The KGB used to do similar things. If you want an idea what new project XYZ Aerospace is working on, just check the job ads.

Cost of Components

[Phoenix666]

I was particularly struck by the cost per unit he cited for 32-bit processors: $0.49/processor. At that cost profile the possibilities for DIY swarm and fabrication projects is compelling; a vision of autonomous mesh nodes spreading throughout our cities, powered by ambient backscatter chips, and forming the ultimate redundant network danced through my head.

Exciting times.

Re:Cost of Components

[Dzimas]

In all fairness, you get a pretty basic processor for 50 cents - an ARM Cortex M0 core running at 30 MHZ with 4KB of flash memory and 1K of SRAM in an 8-pin package (52 cents, quantity 5000). I am constantly amazed when pricing out designs and discovering that the quite capable little MCU that I budgeted $3 for now costs a mere $1.20.

Re: Cost of Components

[MountainLogic]

Digikey is amazing. Use them all the time for Protos and some production. Where tracking Digikey fails is the roll of truly innovative stuff barely out of the lab in small production as we'll a the other end custom ic. Sure they do fpga but try buying a full intel chip set, a GPU or what ever Qualcomm is selling to phone makers. They really are more trailing than leading edge. And the county airport is being expand to handle larger federal jets just for Digikey. They probibly have several million skews and many of those skews are for reels of 5 thousand resistors per real and many multiple reels of a skew in stock. They must be tracking billions of pieces of stock.

Re:Cost of Components

[unitron]

...I am constantly amazed when pricing out designs and discovering that the quite capable little MCU that I budgeted $3 for now costs a mere $1.20.

And if you weren't such a blabbermouth about it you could have pocketed the difference and no one would ever have been the wiser.

: - )

Increased resistor sales

[dohzer]

Resistor sales are up. I don't know what that means for electronics and IT, but I predict the future will be warmer.

Re:Increased resistor sales

[ebno-10db]

It means some people finally figured out Ohm's Law. If they get the hang of complex numbers, capacitor and inductor sales will go up too.

Re:Increased resistor sales

[mysidia]

Resistor sales are up. I don't know what that means for electronics and IT, but I predict the future will be warmer.

The borg will not be pleased. Resistance is futile

Re:A big question college students should be askin

[ebno-10db]

These days, even the law school grads aren't doing so well.

time to rethink college it's to long and can be th

[Joe_Dragon]

time to rethink college it's to long and can be theory loaded where the people coming out can have big skill gaps.

Analog electronics

[rlh100]

If I was a young double E student I would focus on analog electronics. Designing analog electronics is a dieing art. And it is art as much as electronics. Simulation only goes so far. Then you need to know the tricks of design and layout.

The old school analog electronics engineers are retiring and there is not a new crop of young engineers to take their place. While more and more things are going digital we will always need analog electronics to interface with the real world.

Analog electronics will become a specialized niche that will command big bucks. Kind of like COBOL programming. Neither of which are very glamorous but both of which are all around us.

Re:Analog electronics

[unitron]

... Designing analog electronics is a dieing art...

As is spelling, it would seem.

: - )

But seriously, if you've got the kind of brain suited for it, not just analog, but the voodoo known as RF in particular should keep you in demand.

Re:A big question college students should be askin

[necro81]

Or, don't focus solely on computer science - i.e., being a code monkey. Instead, understand the underlying hardware and create tightly coupled hardware/software solutions - embedded software.

For instance, the video discusses GaNFETS, and the new power density they enable. Become a controls engineer and you can end up using these devices to make world-class power supplies (not to be confused with wall warts) that are used in electric vehicles, industrial robotics, and renewable power. There is still a ton of software that goes into those kinds of things, software development that can't easily be outsourced.

Start coding on microcontrollers. Master the ARM architecture. Understand embedded Linux and wireless networking. You can make (or at least start) an intellectually satisfying, remunerative career in this way.

I, for one...

[unitron]

...hope they (and Mouser) keep a good stock of Low ESR capacitors for some time to come, due to "capacitor plague".

Re:A big question college students should be askin

[n7ytd]

Amen. If I could start over again, same circumstances, I'd be a fucking lawyer in a heartbeat.

You may be on to something there... a prostitute that also does estate planning? That's gold, Jerry!