[HN Gopher] The Enchippening
___________________________________________________________________
The Enchippening
Author : apsec112
Score : 25 points
Date : 2024-01-23 23:02 UTC (23 hours ago)
(HTM) web link (sarahconstantin.substack.com)
(TXT) w3m dump (sarahconstantin.substack.com)
| paulsutter wrote:
| Nonsense, manufacturing learning curves[1] were identified by
| Theodore Wright in World War II airplane manufacturing, and apply
| to many types of production:
|
| > Wright found that every time total aircraft production doubled,
| the required labor time for a new aircraft fell by 20%. This has
| become known as "Wright's law". Studies in other industries have
| yielded different percentage values (ranging from only a couple
| of percent up to 30%), but in most cases, the value in each
| industry was a constant percentage and did not vary at different
| scales of operation. The learning curve model posits that for
| each doubling of the total quantity of items produced, costs
| decrease by a fixed proportion. Generally, the production of any
| good or service shows the learning curve or experience curve
| effect. Each time cumulative volume doubles, value-added costs
| (including administration, marketing, distribution, and
| manufacturing) fall by a constant percentage.
|
| [1] https://en.wikipedia.org/wiki/Experience_curve_effects
| scraptor wrote:
| As I understand it the core thesis of this article is that any
| object which can be manufactured using semiconductor processes
| gets to _share_ the experience curve of semiconductor
| manufacturing, which not only has has a unusually high cost
| reduction per doubling compared to other industries but is also
| already very cheap to begin with due to the large volume
| already produced.
| blamestross wrote:
| I feel like a lot of people need to learn what a sigmoid curve
| is, and how only some of it "feels exponential".
| zwieback wrote:
| The key takeaway _" Every tech that is improving exponentially in
| the 21st century is explained by improvements in semiconductor
| fabrication."_ seems hard to believe. What about Lithium
| batteries? Same trend vastly different process.
| g8oz wrote:
| I don't believe lithium batteries are improving exponentially.
| They certainly are improving though.
| vajrabum wrote:
| Looks like a good counter-example. According to this article
| the decline in cost is due to materials and chemistry R&D, and
| less economy of scale. https://news.mit.edu/2021/lithium-ion-
| battery-cost-1122
| toasterlovin wrote:
| Part of economies of scale is that you can afford to do R&D
| that only pays off when amortized over larger volumes.
| dweekly wrote:
| If the hypothesis holds even in part, deceleration of Moore's Law
| could have drag effects on progress in many domains.
| the8472 wrote:
| Not necessarily. As it gets more difficult the next nodes and
| alternative materials will depend even more on advanced
| materials science and fabrication technology that'll also be
| available to adjacent fields.
| jgeada wrote:
| The death of Moore's Law has been predicted for a couple of
| decades. Sure, eventually we'll run into the limit, but I'd
| need strong evidence for such a claim.
|
| (and yes, some related semiconductor laws no longer scale.
| Dennard scaling has been lost, etc)
| m463 wrote:
| I have another theory about moore's law - it gives the
| (conservative, risk averse, non-techy) money people a roadmap to
| growth in tech.
|
| I think many people in business sort of stick to the status quo.
| We just got where we got, and lets just rest a bit and turn the
| crank. But if the roadmap says things will be 1.9x in 17 months,
| the projects that get the greenlight will be different.
|
| I don't think predictors like this are common or visible. A lot
| of business is reactionary, to competition achieving something.
|
| Additionally, it also helps with tech dependencies. Like "if I
| plan for chips to be this fast in this many months, our project
| can make these assumptions"
| eichin wrote:
| Didn't Sun Microsystems outright _do_ that back in the 90s?
| Straight up declared that their roadmap was to treat Moore 's
| Law as a plan, rather than a prediction?
| alted wrote:
| A disappointing fact of chip fabrication is the minimum bar is
| high and expensive.
|
| In other fields, a hobbyist can do wood/metalworking or learn
| programming or build a robot kit. There's an onramp for people to
| start learning the skills, which makes a huge ecosystem of
| gradually improving talent.
|
| But in microfabrication, even though it's the only way to make
| chips, screens, cameras, inkjets, and LEDs, the minimum equipment
| cost is millions of dollars. Even worse, it takes even
| professionals months to fine-tune a manufacturing process to make
| a new thing.
|
| As a result, R&D is much lower than it could be, and most
| fabrication is limited to circumstances with a high chance of
| mass production payoff.
| chx wrote:
| Well...
|
| You can learn a lot by programming FPGAs.
|
| In theory, while I think you could build an SSI manufacturing
| device at home if you really wanted to -- just how many people
| build an internal combustion engine themselves to learn about
| it?
___________________________________________________________________
(page generated 2024-01-24 23:00 UTC)