The Economist Gets It Wrong

[Paul McLellan]

This week's Economist contains their Technology Quarterly, which contains a piece on Moore's Law. Some of it seems to have been written by Intel but in the end it does get to the point that 28nm is the lowest cost per transistor and since then it has been going up. But it doesn't mention double patterning, just the "cost of lithography equipment". It reckons that 14nm is the half pitch of something in a 14nm process. And it has an interesting description of short-channel effects.

I have a subscription to The Economist so I can't tell if people who do not can read the article. But it is here.

Final paragraph:
In the meantime, the 50-year era of pushing down semiconductor costs through improvements in manufacturing know-how is about to be superseded by a new age of making chips cheaper, faster and better through smarter design, including systems on a chip. In so doing, Moore’s law could get a new lease of life.

 

[Daniel Nenni]

I was able to read it. I'm not really sure what the point of it is. Are they trying to make Intel shareholders feel better or worse?

 

[Brianhayes]

I was able to read it. I'm not really sure what the point of it is. Are they trying to make Intel shareholders feel better or worse?

As a Brit I am ashamed to say I do not think there is an answer to your question. That newspaper has passed on the obligation to be incisive.

 

[Michael Clayton]

On answer is that monolithic chip products will morph into "systems in package" aka multi-chip-modules.
The Intel hype on Moore's law will dry up as electronics world moves back the way it started in hybrid circuits like the IBM ceramic modules or early Motorola ECL hybrids in a can as key components in electronic systems. Remember, it was simply called Batch Electronics by initial customer, the military.
The monolithic circuit was a cost reduction idea that works really well.

But of course good marketers will spin it to highlight their tiny portion of the product, namely the overall design, as they gradually complete their retreat from integrated crystal growing, wafer processing, and packaging efforts into that world of hands-off product design and sales.
Just like the automotive industry, no more "integrated manufacturers" like Henry Ford's disruptive innovation long ago.
Everything is simply another commodity at chip level just like automotive drive-trains.
But it was a great 50 years.
"Real men have fabs" per famous ex-Motorola-salesman and best-dressed man in Bay Area, Jerry Sanders.
Here's to you, Jerry!

 

[U235]

It's a shame they didn't ask someone with some technical knowledge to check the article, and so correct a few points.

I have a friend -- an engineer by education, now working in finance -- who reads the Economist as if his life depended on it. That's who the article's written for. The article is typical of Technology Quarterly, they just give a taster of what's going on in an industry, mixing a few key points up along the way.

 

[Brianhayes]

I have a friend -- an engineer by education, now working in finance -- who reads the Economist as if his life depended on it. That's who the article's written for. The article is typical of Technology Quarterly, they just give a taster of what's going on in an industry, mixing a few key points up along the way.

If your friend's life really does depend on it you should advise him to read another newspaper. The author's research drew on Intel's spin before wandering off into saying that SoCs matter more than CPUs. The conclusion read as if the author scratched his head wondering how he was going to finish. I spent my working life in finance and used to read the Economist cover to cover. I still subscribe. It just ain't what it used to be, mainly because it puts politics, and political correctness, before economic and financial facts.

PS. I should, to be PC, have said s/he instead of "he".

 

[Arthur Hanson]

I also read the Economist and have seen PC creep in and ignore those articles. The rest of the magazine is of still very high quality, although they just had a major change of ownership. Only time will tell.

 

[Michael Clayton]

Station map | California Fuel Cell Partnership
California is adding more Hydrogen refueling stations.
Toyota, whose Prius triggered the hybrid car surge, now has FCVehicle coming out.
Refueling stops take minutes compared to hours recharging batteries or ??? exchanging battery paks.
This is now a global effort, so it will be developed for cars as well as heavier vehicles.
Stationary fuel cells are very successful already, but they are very heavy as they run at high temperature.
Mobile fuel cells are lighter, but with less insulation and lower temps, they are also less efficient.
But to me the selling point in California will be number of refueling stations by 2017 for the new Toyota vs number of "battery exchange" stations for Tesla et al.

Of course city folks who never drive very far can use pure e-cars and home charging stations overnight.
So hydrogen FCVehicles will be for longer distances given "gas" stations along the way every 200 miles or so.

 

[Fred P]

The vast majority of articles or news stories written on subjects where I have fluency that gets it right is very small. What does that tell us?

 

[Brianhayes]

The vast majority of articles or news stories written on subjects where I have fluency that gets it right is very small. What does that tell us?

That is why I read Semiwiki, most of which I do not understand! What I can gather is written by people who know what they say.

 

[Fred P]

Yep, fully agree.