Hacker News new | past | comments | ask | show | jobs | submit login

Does anyone working within Intel know the reasoning ?

It seems skylake is doing really well.

Is it mostly electrical engineers working on the processors or sales and marketing people ?




I hope it's the hardware side of the company waking up that the software and "solutions" side of the company is a gangrenous fifth limb.

Things like trying to start an app store when they didn't have content (e.g. Steam) or a platform (Google, Apple, Microsoft), trying to pivot that abomination into a media store after Netflix already won and the studios are all trying to start their own, trying for way too long to deliver solutions of putting parts in verticals (e.g. tablets, micros) where they are at best only competitive on one of price, power, or performance and immediately face-planting, but limping along seemingly oblivious to why any of these projects fail.

The mention of IoT though makes me think this isn't the case. Definitely part of the brain-dead trendy "solutions" thinking. I don't think you're going to compete with commodity cheap micros with the best IP or the best process size... price is king.

There are too many layers at Intel, and that leaves too many of those layers too well insulated, able to believe it was the brand that won them their markets (not the economics: Wintel monopoly in personal computing history, having the best server and laptop parts currently), and that despite the economics of new products and "solutions", the brand will somehow convince people to buy in.

Like, why do I ever see you on Television, Intel? Who is the audience for these ads, potential investors? Shouldn't they rather see higher dividends? People don't go to the mall and buy Intel chips, they go buy Macbooks and they use Instagram. Facebook and Apple buy your chips at a scale where it's all ROI, and you deliver that, so what are the ads for?

This shit is why I left, and why always did the quicksale on the stock purchase plan.

Intel needs to just keep making the best parts and leave it to the market to conjure up asinine things to do with them. Kill all the product lines where there's no road to being #1 ROI within 10 years, keep those in the lab.


I'd still like to know whose idea it was to try and compete in the Internet of Things market by pitting a buggy warmed-over 486 (the Quark) against ARM's best, most modern chip designs. Sure, Intel got a bunch of headlines in the tech press about their new IoT solution, but the technical details just didn't add up. The performance and power figures were dire, it couldn't run existing x86 code, it looked like a bear to integrate into anything, it just made no sense.


Intel's latest Quark has gotten to 1.5ua sleep current with sram retention. That's not the best ,but that's good enough for many applications. And it's using 22nm so it may be the cost leader, by far. And that's mostly what counts in embedded.


Thanks for sharing and providing that wonderful insight !

I agree with all your points - the last time I though about Intel as a brand was in 2002 while wanting to build my own PC ( a very small market of geeks at that )

What particular thing would you say that makes it difficult for them to built good software as compared to amazon ? Intel seems to have a bottomless pit when it comes to money.

thanks again !


>There are too many layers at Intel, and that leaves too many of those layers too well insulated, able to believe it was the brand that won them their markets (not the economics: Wintel monopoly in personal computing history, having the best server and laptop parts currently), and that despite the economics of new products and "solutions", the brand will somehow convince people to buy in.

Well, as a gamer, intel (and nvidia) are just the best.

I had AMD Athlon and it was an amazing processor. The modern ones just do not compare to Intel ones. AMD also pulled riddiculous shit when they stopped releasing drivers for linux on 2 or 3 years old GPU's - therefore they can go fuck themselves.

Ive got 4790K and it is just great, runs cool (just a decent upsized aftermarket radiator for ~30$ - no water!), doesn't abuse my fans, will probably last me 5 or 6 years performance wise

For a bulk buyer of crappy throwaway corporate laptops, the AMD might have some appeal...


Exactly, even if gamers were a big market, they follow the ROI too. Intel can't try to be a brand the way Nike is or the way LV is, they sell technical things to technical people but mostly to industry.


> Intel can't try to be a brand the way Nike is

You'd be surprised -- they already are. Their original "Intel Inside" campaign was groundbreaking and helped them immensely in reaching domination in the PC space. A lot of barely-technical consumers will not have heard of AMD or other minor competitors, but they will recognize Intel and buy accordingly. Techies don't care too much about brands, but non-techies do.

Part of the problem they have today is precisely that their brand is basically worthless in the only growing segment (mobile); they must find and occupy all other niches just to stay alive, and marketing is part of that.


Sure, they spend on marketing, they're visible to the public, and their products are popular, those don't mean they have a brand -- a wide and loyal base of customers asking for their products by name for subjective reasons.

The chips sell because they're the best. If people really wanted "Intel" (and knew what they were even buying there) and not just the cheapest / fastest / longest lasting device regardless of how or why it has those properties, Intel's attempts to push into mobile despite failing to hit the required price/performance/power points might have worked like management always pretended it was going to.

They have a very polished image and spend plenty to get broad awareness. Plenty of visibility across all segments, it's not that they failed to capture mobile end-customer's eyeballs, it's that they failed to deliver a mobile chip with an ROI to mobile OEMs.

The most fervent Intel fanboy can't buy your Meego/Moblin/Whatever(Maybe people will forget it failed if we rename it every year)-powered Intel-chip phone if it never hits the market because it fails to have a day's battery life and the OS is basically vaporware (we can't ship updates for existing systems, we're too busy with politics, switching UI toolkits and starting over).

Intel pays to try to build a brand and labors under the assumption that means it has one (for instance, wanting to build their own Platform [SOC+OS+App store] thinking the brand will pull it to market despite reality, instead of just accepting Android from the beginning) but it simply doesn't. All the marketing money looks to be just a bonfire.

Back when AMD got a foothold and was able to be competitive on price, people bought them, I don't remember seeing any AMD commercials.

If people only love you in the categories where you're objectively the best, maybe it's not because you have the cool circuit pattern branding, threw all those parties and bought those super bowl spots.

No matter how much they spend and pretend to be a consumer brand, it just doesn't matter one bit outside the ROI for OEMs and ISVs, since those are Intel's customers, they are only ones making buying decisions of any consequence, and they're never going to buy a part that's detrimental to their own bottom line. Intel Inside partnership was in my opinion about OEMs fleecing the marketing spend for a reduced bill of materials. The sticker meant nothing, no consumers really cared. Enthusiasts only cared when Intel meant it was a better chip, only a minority of that enthusiast minority would prefer a weaker Intel chip over an alternative.

If I were to call my grandmother (70s), mother (50s) and niece (20s) and ask them if they'd prefer a computer with an Intel chip in it, I suppose they'd say yes (and as much as my ego might want that to be because I worked there, it's most probably the ads). However I happen to know that two of the three of the CPUs in question are AMD. Why? I don't know, but the relatives all buy cheap HP or Dell Windows machines (choosing those OEM brands despite better options, real brand effect) on a whim (a.k.a when the last computer's OS installation became unbearable) from the nearest Best Buy (picking that retail brand despite better options, a brand effect) and the OEM chose an AMD part (probably for margin reasons, creating lines of products for undiscriminating consumers) and these relatives didn't notice or didn't care (no brand effect). I wonder if I asked them if their computer had an Intel CPU, if they might even think that they do. I'll try that next time I see them.

No matter how hard Intel pretends they're like organizations with similar marketing spend, they simply don't have a brand like those do: there's no verticals and no consumers aside from self-built-PC enthusiasts, and for consumers the latter are quite a touch more objective than those in fashion or apparel.


I disagree. With all their failures in mobile, you mention it yourself: consumers do know the brand. One of the reasons Intel keeps dominating the PC space is that even at its peak, AMD simply could not get get past the assumption that "chips must be from Intel" in many quarters, be them OEM or consumers. That's real brand power. In fact, their complete dominance is what undermines their own branding: people now assume what they buy will be Intel and don't even check. I bet the Best Buy sales guys studiously avoid mentioning AMD, because it would kill the sale. That's real brand power. Obviously it's not easy, being a component supplier rather than a direct b2c vendor, but I think that Intel punches well above its weight in brand power. It might not be as big as Apple or Nike yet, but it's definitely in another league from competitors.

> The sticker meant nothing

I absolutely disagree. It did mean something, when it was the only sticker. There is a reason stickers became popular right after that.


> Intel can't try to be a brand the way Nike is or the way LV is

They have a surprising amount of advertising at gaming events which seemingly tries to do exactly that. Even though they are more-or-less the only option. (although AMD was (maybe still is? I'm not following that to closely) better at integrated graphics solutions which were interesting for low-end gaming, most gaming profit is probably near the high-end)

Maybe to promote PC gaming in general, over other activities and specifically consoles?


I can't really figure it out. It's pretty mindboggling.


I asked a friend who worked at Intel and he said part of it has to do with the company's hiring and organizational structure. The basic gist is that people are brought onto teams that only exist long enough to fulfill their initial objective. While some teams are dedicated to long-lived consumer product lines, others might focus on R&D or custom B2B orders. Once a team has fulfilled its responsibilities within the larger company, it is likely to be dissolved and there's no guarantee for room in other teams for the leftover employees.

That's how I understood it, at least. If anyone more familiar with Intel's inner-workings thinks I'm off-base, please feel free to correct me.


Is this a similar issue to what was reported about Amazon getting rid of people after their product launches (even if successful)?


I'm not aware of what you're referring to. The most recent series of layoffs at Amazon was due to poor sales of the Fire Phone [1][2]. Most companies in Amazon's position would have done the same thing.

[1] http://arstechnica.com/business/2015/08/fire-phone-flop-blam...

[2] http://www.wsj.com/articles/amazon-curtails-development-of-c...


I don't believe that's a thing that happens at Amazon. I've worked here nearly 4 years, across several product launches (hardware and software) and only known of one person being let go, and that was for entirely unrelated reasons (politics and such).


Just a guess, but tech companies are occasionally release % of their workers to clean its workforce, boost productivity and make space for new, young and experienced engineers.

Stagnation and idea depletion is dangerous for tech companies, fresh employees bring fresh ideas and boost competitiveness between workers.


Perhaps some people believe that is what happens. But if I were a talented employee, I'd see an 11% staff reduction as a sign to stop ignoring those recruiter emails.


We had layoffs/firings here and that is what happened for a lot of people. There is no point in sticking around when the same moronic manager is kept around while talented, productive people are forced out. Many of us are only here until our shares vest. Our stock price has taken a huge hit as result.


Which means the best employees are the ones most likely to leave. The ones who weren't fired, but who can't easily get another gig are what you are left with. It's a tough cycle to beat.


> boost competitiveness between workers.

I do not think that this is a good thing...you want collaboration, not politics and back stabbing.


Its corporation. Morals are not really the top priority. The results are.


I don't think levemi is saying anything about morals -- rather, that results will be better if employees are cooperating with one another rather than playing zero-sum games in the hope of missing the next round of layoffs.


But it could just as easily be argued that a highly political environment which necessitates such behaviors leads to poor morale, quality employees leaving etc. etc. which couldn't be good for the company long term...


I would say it all depends on the point of view of the people in charge. The reality is what they decide and whats their take on that.


I don't know about other people, but my next Intel CPU will have AVX-512, which Skylake doesn't have. I'm expecting AVX-512 to (up to) double integer/FPU vector computation capacity.

I also want to see some proof future Intel processors (especially 10nm and smaller) have not lost durability due to electromigration [1] (or some other durability issues small feature size can cause). There's already some indication this might be happening on current Skylake CPUs.

Of course not all workloads will get faster, but it could help with things like parsing (XML, json, etc.), games, media processing, compression, physics simulation, etc. -- or really anything vectorizable that needs to run on CPU for reason or another.

[1]: https://en.wikipedia.org/wiki/Electromigration


You might be surprised to learn that AVX-512 is not available in Skylake - it will only be in the Xeon line. It also won't give you anywhere near double performance, since a lot of AVX2 code is memory bandwidth bound anyway (which is probably one of the reasons it's not in the desktop chips.) I would've bought a Skylake so that I can test code on an AV-512 machine. Now I'll keep my money and wait and see.


Current Skylake Xeons don't have AVX-512 either. For example no AVX-512 in this Skylake Xeon: http://ark.intel.com/products/93354/Intel-Xeon-Processor-E3-...

I know about bandwidth issues, but you can also use big vector sizes to mitigate, more opportunities for simple schemes for realtime (de)compression of data. (Usually called packing/unpacking).

But where you can work around bandwidth issues, you can get up to twice as much work done.

I've also been looking for such a chip to test my code. Of course it's possible to use SDE. https://software.intel.com/en-us/articles/intel-software-dev...


Yeah the E3 Xeons won't have it, they're basically desktop chips dressed up. The E5s are supposed to have it though.


Intel finally realized that ARM is dominating growing sectors (mobile and embedded[1]), while they're dominating the shrinking ones (desktop and server).

[1] still can't bring myself to say IoT.


embedded makes more sense :)


My bet is on falling PC sales. I'm surprised this didn't happen earlier, to be honest.

PC sales by year: http://i.imgur.com/IxSo1oy.png


Does this include laptops ?


Yes. Around 60% of "personal computers" (excluding tablets) sold are laptops.

What's interesting is tablets. The number of tablets sold basically equals the number of desktop and laptops sold combined.



I would bet they want restructure so they can do more stock buybacks...




Applications are open for YC Summer 2019

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: