Chipmakers Rally in Talent War | Hacker News Reader

Chipmakers Rally in Talent War

eetimes.com

66 pointsbcaulfield66 comments

Really what has to happen is that the chip companies have to wake up to the fact that levels of compensation at software companies has outpaced them. If you want to hire top talent for half of what GOOGLE/FB/AMZN would pay the talent then yes there will be a shortage.

Trying to convince engineers with these fluff pieces how they are important will not make a difference. Pay up.

Where I work we have both embedded firmware and software developers. The firmware devs are largely EEs and get paid less due to lack of demand and market competition. If they paid the software devs the same there would be a mass exodus.

So sad to hear this, as it is much harder to study EE and embedded programming is much much harder than software programming.

They are both equally difficult.

Dunning-Kruger effect applies here. The levels of confidence/validation are completely different. Good luck fixing a software bug on a toaster after-the-fact.

Also, performance is regularly an annoyance, especially if you have to do something non-trivial like process ethernet frames or talk on USB. Sub 8-Mhz processors for embedded processors still extremely popular and the flash size constraint leaves no room for huge libraries or ROM waste.

Completely different.

Hardware and software are two deep domains that can be incredibly complex and challenging.

You're doing a disservice to both by taking a toaster without QA as an argument.

Yep, this is why Intel reported $60 billion in 2017 revenue at a gross margin of 63 percent. [1] Many profitable hardware companies seem to be case studies for what I think of as corporate atherosclerosis.

[1] As mentioned in the article attached to https://news.ycombinator.com/item?id=16194011

I used to work for a company where the complete opposite has always been true[0]. The EEs make significantly more and are generally held in higher regard.

I guess it depends on owners[1]/managers and maybe what type of product is being produced, also.

[0] Still is true. [1] This was a small, private owned company.

Good firmware engineers are incredibly hard to find. I know, I’ve hired two dozen. If you’re a firmware engineer and you’re getting paid less than the software people, leave.

As a EE myself, who left the hardware industry to become a software engineer, I agree completely. It's also worth examining the reasons why chipmakers in today's world can't ever compete with SW companies for top talent:

1. The market size and opportunity in SW is tremendous as compared to HW. SW is eating the world in nearly every industry, and this is going to continue for the next 10-20 years. HW used to be great, during the time of Moore's law. But with Moore's law coming to an end, it's going to be harder and harder for chipmakers to compete with the same products they released 3 years ago.

2. Entrepreneurial opportunities are severely limited in HW. A software startup can be launched by 2 college kids in a dorm room, using a couple thousand dollars of angel investments, and scaled nationally with 1-2 million in VC funding. With HW startups, because of the engineering/logistical nightmares involved in fabricating a chip and selling a physical product, you need so much more money/credentials/connections to get a startup off the ground. SW devs can realistically dream of building/joining a world-changing startup, but HW engineers are mostly resigned to being a cog in a big machine.

3. Paranoia and risk aversion. Strongly linked to the above, is the culture of rational risk-aversion that exists in any HW project. With a software project, the Agile/MVP methodology works extremely well, because the cost of making changes, or fixing prior mistakes, is so low. In HW, the costs of making changes and releasing a buggy product, are astronomically higher. Because of this, HW projects are extremely risk-averse. Junior developers are micromanaged by their superiors. People are given little autonomy to innovate and try new ideas. Project timelines are measured in years, not months. From an artistic perspective, developers love having the freedom to build freely and get their product in the hands of users as soon as possible. This is so much easier in SW as compared to HW.

Combine all 3 of the above, and it's easy to see why SW is so much more lucrative and enticing. As an EE myself, I tell everyone I know that unless they have a deep passion for the field itself, they should strongly consider a career switch.

The bug consideration is huge in hardware. One wonders what a different world it would be if the zilog z800, the most advanced 16 bit chip of its time, would’ve had fewer bugs.

Re: Dead Moore's law and entrepreneurial opportunities.

It seems to me that the fact that Moore's law has ended may end up with a flip side that could potentially be awesome. Previously, it nearly always made more sense to do something on a CPU and wait for performance to double than to develop a custom ASIC. With the free lunch deal over, I suspect that we will begin to see a ton more custom ASICs and architectural innovation. This has already begun for ML/DL/AI.

Without Moore's law, the semiconductor industry may get a lot more exciting. It might also just fade to commodity even more, but that isn't as fun to think about.

Moore’s law is not necessarily dead with respect to analog integration, and III-V semiconductors. Much room for growth. Xilinx has some amazing RFSoC; FPGA fabric with integrated GHz DAC & ADC.

i'm long on semiconductors for this reason. there's a lot more that can be done with them than grinding out increasing dense general purpose CPUs and DDR memory.

we saw a tiny bit of it with the introduction of MEMS IMUs, and we'll see more yet.

sooner or later we'll get low costs on small quantity ASICs, and things will go really nuts.

Ding ding.

I've always been of the opinion that the end of Moore's Law will result in a new wave of startup innovation. When the cost of newer nodes are a barrier to entry and the advantages of startups can be obviated by just moving to the next node (which the startup cannot do), then innovation stagnates. This is largely what killed many FPGA startups (e.g. Tabula, which btw had super cool tech).

I’m also an EE, but designing RF/Microwave hardware. Software, specifically SDR, is fun to do as a hobby, but I’d never want to do it for a living. Too much abstraction in software; libraries upon libraries. With thing such as antennas, and RF circuitry with a few transistors, there is almost no abstraction.

Would you mind sharing what kind of applications you are working on? Defense, aerospace, mobile phones, other?

I've heard worse things about the defense and aerospace industries than I have about semiconductor industry, so I'm just curious.

Antennas, power amplifiers, up and down converters, etc, typical RF stuff, VLF through mmWave. Defense contractor in flyover country. The pay is good as there are few who do this stuff. Almost 22 years at the same company. Totally opposite of the startup hopping software whiz.

touché! Also on the same boat pivoting out of the H/W Product development hell

What's missing here is a discussion of compensation. I've seen at conferences how proud senior Google employees (ie Jeff Dean) are of the TPU work that their teams are doing, and I suspect that compensation is following accordingly. I doubt that Intel or similar companies are compensating their people at the same level. Whenever I see people talking about a talent war, I'm always suspicious that they're getting outspent and are unwilling to compete.

At my university, the word on the street was to not waste your time with chip design or device physics, because most of that work was heavily outsourced and the employers didn't pay well. I thought that was a real shame, because our device physics professor was brilliant and his research sounded fascinating.

Sadly, the people I met from a very popular microchip manufacturer (they don't make CPUs though?) suggested that they had to work long hours and didn't like their jobs. They were actually.. surprisingly transparent about that. I appreciated it, of course!

Also if you do make a device physics breakthrough then people will promptly ignore you. Doubly so if you're young.

Last I heard, you had big claims and have yet to test your ideas in physical hardware. People will probably be less likely to ignore you if you have something that you can demonstrate being able to manufacture reliably and that you aren't just exploiting some simulation artifact.

It's relatively hard to fool TCAD. SPICE simulations at least I can believe there can be scoping artifacts, but even those are generally in a certain pattern. In any case, I'd be okay with a "this is great if it works in hardware, but until then idc", but the response is more "you're a young idiot who shouldn't bother with memory cell innovation". A lot of this is superstition due to SRAM being finicky. I could go on about this, but let's take it offline in that case.

I worked at a facility with a lot of these guys. They seems to work long hours, got laid off a lot, and the cars in the lot were not fancy at all.

The managers looked well rested and well paid though.

The unfortunate thing is that the hardware guys do all the hard work but the software guys make all the money. And that's the real reason why Google pays way more than Intel.

It's the likes of who Intel figure out how you can put a billion little things each with a few tens of atoms on a chip that somehow work together to produce a useful computing device. And then Apple finds a way to create a really pretty GUI and walk away with 90% of the profit in the Industry.

As a hardware guy, I always feel like this is somehow unfair.

The Apple guys and gals solve a problem that more people understand they have and less people offer good solutions to /shrug

Processor hardware is technically as close to magic as I’ve found in the modern world. The science and logic required to make chips is simply breathtaking.

That said, processor hardware is a mature market, one that is so mature that it is starting to consolidate in major ways. Until a startup comes up with a chip that can solve current problems not solve able by a CPU/GPU/FPGA, profits will continue to accrue further up the stack to the software companies that are adding differentiated value.

Nah, the hardware guys are cheap and treat employees, on average, worse. Intel has 60% profit margins. If only there were a way to use money to induce people you desire to have work for you to do so!

It is partly due to the capital investment required to first achieve a competitive scale. Since microprocessor companies require such enormous investment, there are fewer large companies and startups to compete for talent.

In other words, the hardware labor market is closer to monopsony than the software one.

In addition, other things being equal, the closer one is to customers, the more leverage one has on pricing and terms. This is a major reason why Amazon, Google, and Apple are focusing on developing voice assistants and home devices.

> And then Apple finds a way to create a really pretty GUI

Isn't Apple designing most of its chips in house, these days, at least on mobile where the big money is at? I suppose Apple isn't building its own fabs, but describing them as 'software guys' seems a bit of a mischaracterization of the company, and how/why they make oodles of money (hell, their software has been pretty damn buggy of late anyways).

They do order up a custom SoC package that then is actually manufactured by TSMC or Samsung - but do you think they would have to cut the price of their phones if they had to use the standard chips that Samsung, HTC etc use?

I'd say no, they could charge nearly the same. Which means the complicated hardware is not the differentiator "killer app", the marketing around the custom OS and the ecosystem is.

(Which is also why they don't sell iOS or macOS to other hardware firms - the hardware is good, but the software is the USP.)

> The unfortunate thing is that the hardware guys do all the hard work but the software guys make all the money.

It also doesn't help that a hardware exit can be 5+ years in the making and take $10-20 million, minimum.

What VC will fund that when they can throw a couple million at a bunch of stupid 20-somethings, flog the hell out of them, and flip to Goobookazonsoft within 18-24 months?

While a valid complaint, it's arguable that Apple has the best CPU team in the world right now and maybe even chip team.

Their single threaded performance is no joke, especially in such a small power and die envelope.

it is unfair.

i'm a software guy thats worked with hardware guys for most of my career. us software guys will bitch and halt work for a month if the coffee is a little too cold.

and the hardware guys will sigh and put in another 6 months of dv to make sure the part works, and it does, and they'll have a beer and go home and maybe sleep in that weekend.

Engineer from India here, let me tell you my story.

In my final year of engineering, I dreamt of being an Integrated Circuit Design Engineer. While the rest of the folks from my class took up microcontrollers based project, my team was the only one to take up a FPGA based project, because that was as close as I could get to VLSI (Obviously at my insistence, I'd to fight with my partners over FPGA vs Microcontroller)

After engineering, I searched for a job in VLSI in Pune & Bangalore for 6 months, but in vain. Got to know they don't hire freshers. So, I did a 6 month Post Graduate Diploma in VLSI Design from CDAC (Center for Development of Advanced Computing - makers of India's supercomputers) (Its like finishing school) . Agtee the course got over, interviews began. While folks who did their course in .NET/ Java got 20-30 interview calls, we got an average of only 5 calls. All 5 were very selective in interviews. Tried for a job in VLSI for another 1 year, and then finally took up a job in IT in Enterprise Automation.

One year after my IT job, spoke with a few classmates. Some people were really doing well making 70k INR/ month while my lab partner was stuck on 12k/ month. I was making 35k/ month. Lab partner said - "IT folks make much better than we do. What was the point of getting into VLSI?"

This was one point. Another point for me is immigration.

I am planning to immigrate to Canada (Starting my PR process soon - am 28, unmarried). My father (GC holder) has filed for my GC abut my priority date is 10 years away) So US gonna be tough.

But even if I had made it in VLSI, how many jobs are there in Canada for Chip Designers / RTL Design Engineers or maybe FPGA Engineers ? If I stay in IT doing my BMC Software stuff, I know I may get a job in some ITSM company. There aren't many VLSI jobs in every book and corner of earth. Whereas IT ensures that jobs are available in plenty everywhere.

Comments / feedback would be highly appreciated.

I'll be honest I think even in US getting a good chip engineering job is tough. I've a firmware engineering background so I straddle between hardware and software. I see lots of EE students come into the workforce and realize the software is where the action is and struggle to switch over. As a hiring manager, I feel for them and try to help their transition but many are not prepared in their coursework even for firmware. Even some chip designers I worked with were asking me how to transition over. A few were taking courses on the side. Even I switched over the general software side though I try to keep my foot in the door should it even improve.

When I do apply for MS in US, I'm planning for AI/ Machine Learning. Yesterday, I was reading https://news.ycombinator.com/item?id=16196421

This also looks promising. Combines stocks & software engineering. C++ mastery is required and also knowledge of Computer Architecture.

Another field I've shortlisted is Management in Information Systems, since I'm in Enterprise Automation (BMC Software)

Would you kindly bring to my attention any more promising fields that you know of? Thanks.

It is hard to predict where things will be 2-5 years down the line even for AI/machine learning (which I personally don't know a lot about.) What if AI becomes more automated in a few years so less people are needed? But atleast for now it seems like a good place to start.

My general thoughts are get into a top tier university as that gives you a head start to find a promising company to join. Pick a field that really interests you yet balance with market demands. Focus on the broad computer engineering skills that will be applicable in many industries. Work hard to make a name for yourself at your first job. At some point it will build up a network of people who can transition you into other industries as you change jobs. Eventually you will hit the leading edge of some hot technology and can make your success.

This is a very good idea that you told me. Thanks for this :)

FPGA and ASIC design as a field is very vast. Tool sets cost high and once a company invests in an ecosystem, they generally don’t switch. In such a case, if you showcase your expertise with say ALTERA at a company that uses Xilinx, it may not come across as impressive to those looking to fill in a new spot.

I think it’s better to keep your feet in non hardware fields if you want to be sure of getting a job as the first thing.

To continue in FPGA/ASIC design, I’d suggest you do a MS degree in a US university in a location that has a lot of chip design industries (I’d guess Bay Area, Texas), because that’s where you can have a good chance of getting an internship and that’s going to be your foot in the door, to start a career in the field of FPGA/ASIC design. Getting an internship is very very crucial if you want to get started with a career in hardware engineering since the situation is similar in US also, where you need to hit the ground running from day 1. With an internship/co-op it becomes much easier since they give you some slack because you are a student .

Whatever you do, dont just get stuck in the cess pool of IT Services jobs in India, because your friends in those fields have magnificent titles and pay. If you get to work on a product at a product company, great. But orherwise for someone like you IMO a graduate degree in the US is the best bet.

(I’m from b’luru, did grad course in FPGA Design in the US and later switched to s/w dev because I was not awesome at h/w)

Many thanks for your suggestions. Went through your reply couple of times. Would you kindly give me pointers on how to get scholarships for MS in US? And how to successfully get RA or TA. Asking since you have already finished yours.

As I've previously mentioned, I am in IT now working in Enterprise Software.

I went to a good university with an ok EE dept in a very bad location, and didn’t look for assistantship.

But based on what I have seen, look for universities / departments that are well funded for research (I don’t know how you can get that info), which often means more assistantship opportunities.

Getting an RA/TA can also be competitive but since you have some work experience, it depends on how well you market your experience to fit the RA/TA role.

Thanks for the info. Much grateful _/\_

As a "young" person in the semiconductor industry, I can whole heartedly guarantee that no young person will ever want to go into this field ( nvidia is the one exception but only for DL software)

The field is essentially "no matter what the data says, the person with more experience is right", looks down upon young talent, essentially refuses to innovate and has too many clueless people making decisions.

Don't believe me? Ask your ex-Intel friend.

EDIT: this turned into a rant. Naturally, this is one view, there's another perspective as well. E.g. ASML is extraordinarily innovative and Intel's TMG was an absolute engine.

Isn't the preference for experience due to the cost of respins?

The hubris is strong with this one.

> Looks down on young talent

Are you certain this "young talent" is looked down upon for youth, not in response to their criticism of people who have actually consistently made chips that work and work reliably?

fyi, people are more likely to listen if you don't call them clueless.

> fyi, people are more likely to listen if you don't call them clueless.

They outsourced their chip and now are unable to ship it. I think it's fair to call them clueless.

You seem to have switched from a general statement about the semiconductor industry to a specific occurrence. If your experience is only with this specific occurrence, you should not make general statements about the competence of thousands of people in the field.

Perhaps a more accurate statement would have been: "some clueless people exist in every field." But that is not as exciting.

That was one example, others include Meltdown, the McAfee fiasco, etc.

I also think it's fair to criticize "some clueless people" when they comprise a majority of the decision making power, which they do. I'm not saying that "every single person in the semiconductor industry" is wrong/clueless/incompetent, far from it, Bill Dally is experienced (he's been in semiconductors before I was born) and I consider him to be a genius. There are others as well, such as Ivan Sutherland, but my gripe is that these are not the people who make the decisions. And with respect to the general statement, this was in response to an article making a general statement about a general point.

> Are you certain this "young talent" is looked down upon for youth, not in response to their criticism of people who have actually consistently made chips that work and work reliably?

You could use the same line of reasoning for science in general. Was it stupid to criticize the opinions of scientists from classical mechanics?

Also: meltdown.

A couple of things:

Science is not the same as engineering. Engineering requires a balance between performance, reliability, cost, etc. I don't think the parallel to paradigm shifts in science is a very good one.

There is a lot of wisdom behind listening to people with a proven track record over cocky youth. The people with experience are typically a lot better at striking an effective balance, even if the data suggests that some great new innovation will benefit one characteristic.

Of course it was not stupid for scientists to move beyond old theories. But the scientists with new theories almost always learned from and worked with the previous generation and then moved beyond them, rather than criticizing the character and competence of the old guard.

My main point was simply that if you adopt the same tone with the people who "look down" on you as you did in your original comment, I am not surprised that you don't get any respect.

My original comment was rather ranty, I've calmed down since then :)

I think the comparison with science is fair, engineering in this sense is very close to physics. If you don't like that example we could try the shift from vacuum tubes to integrated circuits, or the disrespect thrown at neural nets before they got popular again. All analogies are crude, but they convey some meaning.

I mean, I've shown people hard data (not even from simulations,from hardware) and they've still essentially ignored it, so I think it's fair to say there's at least some dysfunctionality in the industry. As epitomized, by the very article I'm responding to here.

I'm not usually like this I promise, it's just that this article struck a bit of a chord with me :)

Maybe if they paid competitively they would be able to attract new grads. Many of my Berkeley and Stanford EE acquaintances are joining Apple, Facebook, and Google as hardware engineers and getting 6 figure initial RSU grants and $140k starting salaries (this is at the PhD level). Even Intel, which has among the best pay in the traditional semi industry, pays about 20-30 percent less total comp, and so they are only attracting either the people still intensely interested in their specialty (and willing to forego pay) or those who couldn't pass an interview at the aforementioned other companies

My friend who works in Apple Hardware had a competing offer from Nvidia. They both offered the same pay and an expectation that you will work nights and weekends, but apparently a followup team lunch at Nvidia didn't go well and he took the offer at Apple.

I don't know what the dollar amount was, I didn't ask, but I think it's fair to say that Nvidia is at least interested in paying whatever the market rate for EE is.

I believe the fundamental issue in all of the technology industry is tying people that know how to take money and turn it into new technology that begets more money. There is tremendous room in semiconductors to do so. In terms of selecting winners, it's no harder than software, but software cycle times (and profit taking) are much shorter so it's faster to wash the wins and losses. It also makes it easier to appear to know what you are doing (VCs) when you don't. A semiconductor upstart as a new entity or within an existing company is a minimum of several years of investment before you can even gague long term success. People (VCs, governance boards, executives, GMs etc) are no better at picking winners in software than hardware, but you spin the wheel more often and therefore it's more lucrative.

The issue of compensation extends far beyond just chipmakers. The large tech companies (FB/GOOG et al.) pay top compensation for most technical fields and have the ability to suck up talent worldwide.

Chipmakers, automotive, media...the question is how long incumbent companies are going to continue poor compensation practices before they figure out how far behind they are.

If Meltdown and Spectre do end up causing industry wide redesign of CPU internals and issuing of new generations of chips I expect competition for silicon design, verification and enablement expertise to heat up to the point of being red hot.

I'm not surprised to hear that traditional chipmakers are having trouble with their talent pool. I just recently left semiconductors to get into software.

Getting into software is much more accessible, especially with the abundance of learning material online. It's also easier for people to imagine how software can make a difference in people's lives, which is a common motivator. We're more aware of how we use software on a daily basis. And it's often software that goes viral in a consumer market, not hardware - Apple and phones aside. Silicon fabrication and chip design are at a level of abstraction that's harder to grasp.

How much accessible information is out there on hardware design? Neither is it easy to start fabricating your own chips.

> “A big part of the challenge is in awareness — we need to help them see the path.”

> That shouldn’t be difficult. “We make air travel safer, healthcare more predictable, we have an easy story to tell,” said Durn.

I agree that there needs to be more awareness. It's not, however, as easy as you might think to sell a student on this story. Sure, it might sound great at first. It sounds convincing. Almost everything is based on the hardware these companies produce, right?

Even if you convince them and get these new engineers in, you have to find out how to keep them.

The scale of engineering that needs to happen at these companies is massive. It takes many moving parts. Many of these companies start by allocating new undergraduate engineers to grunt work that's required at any company of this size.

> As our industry matured, operational efficiencies and a whole set of different DNA took hold

Good luck convincing your new engineers that being inserted into your company as a cog and dealing with your business procedures for “operational efficiencies” is making the world a better place.

> About 85% of chip vendors need new kinds of talent to keep pace with the rise of digital operations powered by automated systems, big data, and machine learning. However, 77% of them report a shortage of talent, particularly for EEs, according to studies by Deloitte Consulting done for the SEMI trade group.

Sure, there’s going to be a shortage of talent when you often demand graduate degrees at a minimum. Even more so when you don’t pay as much as the software giants are.

There is a good point in here though. These companies do need these new kinds of talent. And to be clear, it’s not the talent that needs to keep up with the new age, it’s these companies that need to keep pace. They need new processes and reorganization. And by reorganization, I mean something beyond just shuffling around managers and fattening up middle management.

You have to somehow replace the processes that worked just ten years ago but are slowing you down now. IT and your software departments can’t be seen as a cost center anymore. Don’t hire bright young engineers to do things you should be automating. Hire people to automate those things instead.

there is also a public image issue, they have crappy, incomprehensible websites, tortuous procedures to access their crappy bloated design tools.

I graduated in 1989 with an Electrical Engineering degree and three semesters co-oping at AMD in a project on one of their Intel drop in replacement processors. My ancient experience coincides with your more recent experience - it taught me that a.) the really interesting stuff to me personally was going to require a Ph.D. to get in the door, they had extremely bright people working on lower level projects but everyone I talked to on the top projects had advanced degrees b.) each chip iteration was a huge process involving lots of different people at each stage with different expertise so each individual contribution was going to be tiny c.) the software we used to make/design chips was a lot more fun to work on and iterate and didn't require nearly as big of a capital investment/gamble, even back then.

> Good luck convincing your new engineers that being inserted into your company as a cog and dealing with your business procedures for “operational efficiencies” is making the world a better place.

Dead on.

I'm sure the author of the article is a great person, but the fact that the _CFO_ is the public face of this effort tells you all you need to know about how serious they are.

> Neither is it easy to start fabricating your own chips.

Former top rank star level chip designer here.

It's easy and cheap to do up to medium scale with FPGAs. All you need is elbow grease, smarts and creativity. This gives a lot of experience. There are guys around that have made their own chips singlehandedly. These are smart guys. You'd think they'd be in huge demand. Hah hah.

Larger scale cutting edge chips oh yeah, it gets extremely expensive, not just way beyond individual level, but way beyond most companies. There's just not enough foundries willing to produce designs from scratch for random third parties for it to be competitive for small scale things. Doing so is expensive. You only want to take known serious players, or those with enough cash to make it worthwhile regardless.

But one can certainly do so on the cheap for a lot of more modest designs and develop the experience. There are even multiple fantastic free classes these days that teach chip design. Amazing. Not only that but we see some microcontrollers that have small FPGA sections built in so you can roll your own for some stuff in widely available architectures. Wow! This is amazing to have and few take advantage.

However, the criticisms stated are generally true. And yeah, if you're smart enough to do good chip design, you can do better and have less stress doing something else for sure.

Based on all the responses in the thread, I am now hardly surprised that my Ask HN to switch to hardware received 0 responses. (https://news.ycombinator.com/item?id=16054742)

May be @madengr, could answer it as I am of the exact same opinion, that hardware is way more fun that software. And even in software, I am not doing the usual boring things, still it has had me. Any feedback highly appreciated.

"Today, kids dream about Google, Facebook, and Apple; they don’t dream about us... his father brought home an early chip made at Western Electric as the brains for a U.S. missile."

Right there is my reason for not working for them. I don't want to be involved in military technology.

I'm an immigrant, and building weapons to kill foreigners is a conflict of interest.

Good luck working in anything software or hardware which isn't somehow involved with military technology.

Reply on news.ycombinator.com