7 ms·
The golden gigaflop: Apple’s shrinking performance sweet spot
- Tloewald 11y agoAs I recall, Apple supposedly bought a Cray to simulate the 68030, and the apocryphal story is that someone announced to Cray that Apple had just bought one of his computers to help them design their next computer, to which he replied, "that's funny, I just bought one of their computers to help design my next computer". By the time this happened, Jobs would likely have been running NeXT (the first NeXT Cube shipped with a 68030). Also, the article mentions Clayton Christensen's The Innovator's Dilemma focussing specifically on his disk drive manufacturer story, which was already outdated and wrong when he first published it (the "disrupted" companies went on to dominate and extinguish their "disruptors").
- pavlov 11y agoInteresting -- I honestly didn't know that about Christensen's disk drive study. (I read the book ages ago and haven't kept up with the debate...) I'd be happy to update the article with a link to discussion about the disk drive studies being wrong.
- srunni 11y agoI'm not sure what discussion the OP was referring to, but there was a recent publication in the MIT Sloan Management Review: http://www.bloomberg.com/news/articles/2015-10-05/did-clay-christensen-get-disruption-wrong- http://www.bloomberg.com/news/articles/2015-10-05/did-clay-c... http://sloanreview.mit.edu/article/how-useful-is-the-theory-of-disruptive-innovation/ http://sloanreview.mit.edu/article/how-useful-is-the-theory-...
- eridius 11y agoThe other apocryphal story is that Steve Jobs walked into Cray's offices unannounced, in casual attire, asked to buy a supercomputer, and was almost laughed out of the building until he managed to convince the receptionist of who he was.
- rythie 11y agoCitation needed on the Innovator's Dilemma point. I can't find anything that backs up your point. As I remember from reading the book, most of the companies mentioned in the book went bust, so there is no way they could have dominate and extinguished anything.
- Tloewald 11y agoSince I can't link to the text of the book (and don't have my copy handy), here's a New Yorker story on the piece: http://www.newyorker.com/magazine/2014/06/23/the-disruption-machine http://www.newyorker.com/magazine/2014/06/23/the-disruption-... 'In fact, Seagate Technology was not felled by disruption. Between 1989 and 1990, its sales doubled, reaching $2.4 billion, “more than all of its U.S. competitors combined,” according to an industry report. In 1997, the year Christensen published “The Innovator’s Dilemma,” Seagate was the largest company in the disk-drive industry, reporting revenues of nine billion dollars.' and later in the same article: 'As striking as the disruption in the disk-drive industry seemed in the nineteen-eighties, more striking, from the vantage of history, are the continuities. Christensen argues that incumbents in the disk-drive industry were regularly destroyed by newcomers. But today, after much consolidation, the divisions that dominate the industry are divisions that led the market in the nineteen-eighties.' and, finally, this gem: 'On March 10, 2000, Christensen launched a $3.8-million Disruptive Growth Fund, which he managed with Neil Eisner, a broker in St. Louis. Christensen drew on his theory to select stocks. Less than a year later, the fund was quietly liquidated: during a stretch of time when the Nasdaq lost fifty per cent of its value, the Disruptive Growth Fund lost sixty-four per cent.' So he wrote a book on disruptive innovation, where his main exemplar was how Seagate et al were felled by small, agile innovators, and published it the year Seagate crushed the small, agile innovators. The book is full of wrong (it's basically fairly accurate about how innovation gets stifled in companies that fundamentally aren't innovative -- companies driven by sales commissions for example -- but those aren't the interesting cases; Apple, Kodak, Polaroid, AT&T, Xerox -- the companies which ought to be the best examples of his case -- don't follow his storyline at all).
- rfreytag 11y agoHow is the demise of DEC and demise-of-IBM-as-a-micro-and-minicomputer-maker not an example of an interesting case of disruption by Apple, Compaq, and others?
- jedberg 11y ago> 20MB was quite a lot of space in the eighties! Yes it was! I remember in 86/87, when I upgraded my original IBM PC from having two 5.25" floppies to having one floppy and a 40MB hard drive. All of a sudden my word processor, with 11 floppies, could be loaded on the drive and I didn't have to switch disks to spell check and it opened in seconds instead of minutes! Back then the upgrade from two floppies to one floppy and a hard disk was way bigger than going from spinning disk to SSD (and that was a huge jump too)
- jacquesm 11y agoFor me the biggest jump was from cassette tape to floppies. Before that I hacked a Sony cassette recorder to give me something resembling random access but the floppy drive (all 160 K of it...) really made a huge difference.
- jhallenworld 11y agoWhat did you do to it? Turn it into a DECtape?
- jacquesm 11y agoThe Sony recorder had a way to activate the mechanisms electronically, so small step to hook up the optical gate of the tape counter to an interrupt pin and the buttons to output bits. That way you could address the tape roughly by the block (toggle the 'play' or 'record' button bit when the counter hit a certain point). This was the deck: https://www.google.com/search?q=sony+tc+fx+410 https://www.google.com/search?q=sony+tc+fx+410 The storage capacity was horrible, but it did work after a fashion. I just checked out the wikipedia entry on the DECtape, yes, that was pretty much it, only much much slower and not as nicely integrated.
- captaincrowbar 11y agoA few years ago I had my mind slightly boggled when I realised that the new computer I'd just bought had exactly one million times as much storage as the first computer I ever owned (Amiga 500 with 5 MB -> Mac Pro with 5 TB).
- pervycreeper 11y ago>The iPad makes a somewhat distracting bump in that chart. If we ignore the years and just swap the iPhone and iPad, we get this rather tidy logarithmic curve instead It would have been more honest and accurate just to have discarded the point for the ipad.
- pavlov 11y agoThe point is that the iPad fills a slot on the logarithmic volume curve almost exactly between the PowerBook and the iPhone -- it was just released "out of order". I'll edit the sentence to make this more clear.
- ISL 11y agoThis is a dangerous road to travel. Any dataset can be reordered to yield a monotonic result.
- pavlov 11y agoCertainly. But in this case it makes sense to reorder the dataset by descending physical volume, because the nearly constant multiplier between the steps becomes apparent.
- Someone 11y agoOnce you allow that, you can put anything in that graph. For example, that jump in the graph between the Cray and the AlphaServer calls for a product of around 3 cubic meters. I conclude that the Apple car will be more like a people mover than a car :-)
- watmough 11y agoBut the interesting thing, is that the iPad was created before the iPhone, and shelved to make way for the iPhone, so I'd say the modified graph reflects if not the release order, at least the order of innovation.
- filabrazilska 11y agoIt's not a gigaflop but a gigaflops and the flops part of it doesn't stand for a floating operation but floating operation per second.
- pavlov 11y agoIf we're being pedantic, it's not a "floating operation" either... While I agree with you in principle, consider that a unit that ends in the letter 'S' is inconvenient because then you can't easily refer to a quantity of one. For this article, the recurring appearance of the 1-gigaflops performance level in Apple's various form factors was the central point, and "gigaflop" makes it that much easier to discuss.
- dragonwriter 11y ago> While I agree with you in principle, consider that a unit that ends in the letter 'S' is inconvenient because then you can't easily refer to a quantity of one. Yes, you can. The quantity of "1" is easily referred to be specifying the number of units as "1", e.g., 1 gigaflops. Its prone to incorrect intuition about the singular form being different from the plural, but even among English nouns, "units of measurement ending in 's'" aren't the only case where the plural and singular are the same. > "gigaflop" makes it that much easier to discuss. It doesn't make it any easier to discuss; if someone knows what the acronym stands for, its obviously clear in its original and correct form, and if someone doesn't, you need to explain what it means anyway, at which point the incorrect form becomes disconnected from the expansion and less mnemonic, whereas the correct form is (as acronyms tend to be) a mnemonic device for the expansion. The incorrect form may make it more convenient as completely opaque jargon, but I'm not sure making that more convenient is in anyone's interest.
- pavlov 11y agoYou see, I wanted to call this article "The Golden Gigaflop" rather than "The Golden 1 Gigaflops". It just doesn't have the same ring to it.
- mikeash 11y ago
- TorbjornLunde 11y agoAccording to John Sculley’s in the book Points of View a tribute to Alan Kay, Apple invested in Cray hardware in order to simulate what the computing power of a desktop 10 years ahead would be and what kind of possibilities it would offer. In 1987, Apple also invested in a Cray XMP 48 super computer which enabled our engineers to experiment with what real time manipulation of multidimensional objects on a screen would look and feel like many years before such computational power would be available on general purpose personal computers. You can buy the book or download a PDF of it for free here: http://vpri.org/pov/ http://vpri.org/pov/
- mwcremer 11y agoApple's Cray was purple, called "tma1" due its resemblance to the Monolith, and my understanding is that it was used for flow modeling of injection moulding for Macintosh II series cases.
- mtw 11y agoI wish Apple focuses instead on bigger hardware (mac server, mac pros) for scientific computing and video/graphic applications. Professionals like the Apple GUI but if you can't find a machine with a good nvidia card, many will just have to jump to another platform
- Gravityloss 11y agoI don't think they would make sense for production use. Maybe instead Apple could do a commercial GUI on top of a "proper" Linux core. Now the production stuff that runs on Linux anyway, sorta kinda works on mac for local development most of the time, with Homebrew etc.. (Homebrew is great, I just installed ag today, I think it even compiled it from source) There aren't much real alternatives to Linux for servers, but there are to macs for desktop. Business people are using Windows anyway, and technical people could switch to Linux if they can forgo Office. They would need professional Linux support though. Or then Apple could help create a good package system.
- jonhohle 11y agoWhy would you want an Apple GUI on headless hardware? Apple seems content to run Linux for their web servers[0]. Pixar, when still majority owned by Steve Jobs, used Linux for its render farms[1]. Before Steve Jobs returned, Apple was working on MkLinux[2], but that effort seemed to fizzle out, and it seemed like in later years, the idea of using Linux as a starting point was a non-starter[3]. Even Airport devices typically have run either VxWorks or - more recently - NetBSD. 0 - http://toolbar.netcraft.com/site_report/?url=http%3A%2F%2Fapple.com http://toolbar.netcraft.com/site_report/?url=http%3A%2F%2Fap... 1 - http://www.macobserver.com/tmo/article/Pixar_Big_on_Linux_Clusters_-_Mac_Not_so_Much_ http://www.macobserver.com/tmo/article/Pixar_Big_on_Linux_Cl... 2 - https://en.wikipedia.org/wiki/MkLinux https://en.wikipedia.org/wiki/MkLinux 3 - http://www.wired.com/2008/11/the-iphone-coul/ http://www.wired.com/2008/11/the-iphone-coul/
- Gravityloss 11y agoI meant the opposite - Linux for my Macbook.
- mmastrac 11y agoThis is a very Kurzweil-esque look at computing, even though narrowly focused at Apple. Extrapolating exponential progress yields some interesting predictions.
- Scramblejams 11y ago"The Power Mac G4 Cube was an outright failure." sniff My favorite PC design ever, so sad. RIP!
- vessenes 11y agopavlov, This is a super interesting way to think about computing devices from Apple. If I think about two orders of magnitude down, I get to maybe an earring; some sort of wearable terminal, a-la Ian Banks' Culture series. We've had glasses already, although they kind of suck. It's an extremely cool idea to start thinking about making an interface with a gigaflop budget and a size estimate. You didn't hit on it, but there are bandwidth budgets that generally follow a curve as well. I'm not sure what the curve is, but it's probably straight-ish (up) on a logarithmic graph. So if I combine a gigaflop in an earring with 1gigabit of low energy wireless, what do I get for an interface, or use cases?