Emotional responses to the Intel transition

On 13 Jun 2005, at 10:47 , Craig.Treleavennovopharm.com wrote:
> cwilburchromatico.net wrote:
>> ... It's clear that IBM will make substantially more return
>> on its investment for much less cost by selling chips to Sony. ...

> No, it is not clear. You make a plausible case why it _might_ be
> more profitable for IBM but without knowing the relative pricing
> charged to Apple and Sony, you have no basis for your conclusion.

Of course we do. What is IBM concentrating on, fast G5s for Apple or
stripped down cells for the consoles?

There's your answer, and where the money lies. Cell PowerPC chips
are simpler, and IBM will be selling tens of million of them per year
to Sony, Microsoft, and Nintendo. 100 million chips for 2006 is
realistic. How many years would it take Apple to buy 100,000,000 G5s?

Quoting Google Kreme <gkremegmail.com>:
>> This announcement has to cost Apple billions in lost sales as
>> customers inevitably decide to wait for Intel boxes.
>
> I though about this and decided that most Apple customers are not
> likely to know about the switch to Intel, thus minimizing the Osborne
> Effect.

Anybody thinking about an Apple has heard about this. I've had non-computer
people that know I'm a Mac guy ask me about it at work. I'm on vacation at the
moment and have heard it come up in conversations around me.

I'm not sure if there will be an effect immediately, but there is going
to be a
huge one the closer we get to 2006. I expect Apple to have a horrible
Christmas
season. I'm curious what educational institutes that purchase large
volumes of
Macs will do at the start of the next school season.

On 6/14/05 16:45, "Google Kreme" <gkremegmail.com> wrote:

> There's your answer, and where the money lies. Cell PowerPC chips
> are simpler, and IBM will be selling tens of million of them per year
> to Sony, Microsoft, and Nintendo. 100 million chips for 2006 is
> realistic. How many years would it take Apple to buy 100,000,000 G5s?

Just as important, how often do you see new chips in consoles? When they
come out with new ones, which is what, every four years or so? That's cake
compared to what Apple wanted.

--
John C. Welch Writer/Analyst
Bynkii.com Mac and other opinions
jwelchbynkii.com

On 6/14/05 16:45, "kevinvanhaaren.net" <kevinvanhaaren.net> wrote:

> I'm not sure if there will be an effect immediately, but there is going
> to be a
> huge one the closer we get to 2006. I expect Apple to have a horrible
> Christmas
> season. I'm curious what educational institutes that purchase large
> volumes of
> Macs will do at the start of the next school season.

Six billion in the bank and no debts says they can handle a depressed buying
season or two.

--
John C. Welch Writer/Analyst
Bynkii.com Mac and other opinions
jwelchbynkii.com

On Tuesday, June 14, 2005, at 05:45 PM, kevinvanhaaren.net wrote:

> I'm not sure if there will be an effect immediately, but there is going
> to be a
> huge one the closer we get to 2006. I expect Apple to have a horrible
> Christmas
> season.

They're probably counting on iPod to save them for the near term.

The difference between what happened with the Osborne and the Mac is
that Osborne announced a laptop computer that was so far from available
as to be vaporware - and I'd guess that prices will drop incrementally
on PPC iBooks and Powerbooks as time goes on. As Steve Jobs said in his
presentation, there aren't enough, or advanced enough PPC chips to
enable Apple to develop competitive laptops, which are the largest
segment of the market

But if I was Mr. iCEO, now that he's got his cards on the table, I'd be
trying to get as many Intel Macs as possible out there as quickly as
possible.

As much as I love my dual processor G5 cheese grater, and I most
certainly do, it got up to 95 degrees here in NYC today. Our air
conditioner is in the living room area of the l-shaped living/dining
room - miles away from where I and my G5 sit. Both fans are chugging
away in overdrive, and it's hot as a four letter word. I might not
exactly enthusiastic about having Intel inside any future Mac I might
purchase, but I can see the point of wanting, cooler, more efficient
chips.


> I'm curious what educational institutes that purchase large
> volumes of
> Macs will do at the start of the next school season.

I think that it's less volume purchases by schools than iBooks and
Powerbooks being so much cooler than Vaio, Dell, etc., to individual
students. The schools will support whatever system they decide they
prefer and can get good prices on.

Marilyn

Google Kreme wrote:
> On 13 Jun 2005, at 10:47 , Craig.Treleavennovopharm.com wrote:
> Of course we do. What is IBM concentrating on, fast G5s for Apple or
> stripped down cells for the consoles?
>
> There's your answer, and where the money lies. Cell PowerPC chips are
> simpler, and IBM will be selling tens of million of them per year to
> Sony, Microsoft, and Nintendo. 100 million chips for 2006 is
> realistic. How many years would it take Apple to buy 100,000,000 G5s?

By that same rational Intel should get out of the PC CPU business and
focus on the embedded processor market and console gaming markets. They
actually have moved in on the embedded processor market in a pretty big
way, but I think if you'll look at their financial reports, the PC CPU
business is by far the biggest contributor to their bottem line, and
that's significant given that Intel is valued at about 1/3 more than IBM.

If you sell 100 million chips, but make only a penny on each one, it's
not hard to realize better profits selling a million chips if you can
make more than $10 bucks a pop on each.

There's money to be had (and a lot of it) in both markets, and the
reality is that IBM just couldn't compete well in one of them.

--Chris

Clyde Kahrl wrote:

> Now if Apple is charged $150 per G5, then they buy about $300bill from
> IBM--and they are probably the largest customer.

$300 billion is absurd on its face. Apple had only $11.1 billion in
revenue. IBM only had $97 billion.

$300 billion / $150 per chip = 2 billion chips

The Switch sounds like expensive news for this user of an aging
PowerBook, who has been saving pennies until a sufficiently snappy
Apple laptop is the One More Thing...

Instead of moving all my programs over, I'll have to buy ALL NEW
VERSIONS of anything that needs pretty good performance. I'm not a huge
power user, but I can foresee spending many hundreds of dollars to
upgrade perfectly functional software that I use every week, to new
versions. Think MS Office (I often do simulations; I'll fill up a
spreadsheet until it gets unacceptably slow); PhotoShop (can't even
imagine Rosetta-type emulation there, can you?), and literally dozens
of little audio, video, utility, fun, plugins and other goodies., all
of 'em fairly heavy CPU burners, even without emulation. Somehow, I'm
not envisioning Apple's X86 BlueBook to have quad-core, 4GHz processors
to throw at 'em.

How many developers will figure that X86-compatible upgrades, many of
which will require as much coding effort as your typical
whole-version-number bumps, ought to cost their base of users just
nominal $10 costs? And how many will devote their coding resources to
Intel compatibility instead of The Next Great Thing? They're going to
have to recover costs, and that means that Mac users can look forward
to fairly expensive versions to get reasonable performance. Not so bad
for the new users, who have to pay going rate no matter what, but lousy
for the Base.

As the announcement makes quite clear, there are surprises on the road
ahead. But there's going to have to be a disproportionately high ratio
of pleasant surprises to offset the upgrade costs. This has an
unpleasant odor of Planned Obsolescence.

--On June 16, 2005 10:40:19 PM -0700 Marilyn Matty <mmattynyc.rr.com>
wrote:

> The difference between what happened with the Osborne and the Mac is that
> Osborne announced a laptop computer that was so far from available as to
> be vaporware - and I'd guess that prices will drop incrementally on PPC
> iBooks and Powerbooks as time goes on. As Steve Jobs said in his
> presentation, there aren't enough, or advanced enough PPC chips to enable
> Apple to develop competitive laptops, which are the largest segment of
> the market

Interesting follow-up on the Osborne effect in Cringely's latest column.

<http://www.pbs.org/cringely/pulpit/pulpit20050616.html>

Apparently this has become an urban legend. Osborne didn't announce a
computer a year ahead of time. He announced a computer 3 months ahead of
time, and they actually shipped it. But in the meantime Kaypro released a
cheaper computer with a bigger monitor. When Osborne released their
announced computer it was more expensive than even the old Osborne and
still had a smaller screen than the Kaypro.

Kevin

On Jun 17, 2005, at 1:48 AM, Walt French wrote:

Instead of moving all my programs over, I'll have to buy ALL NEW VERSIONS of anything that needs pretty good performance. I'm not a huge power user, but I can foresee spending many hundreds of dollars to upgrade perfectly functional software that I use every week, to new versions.

The software you have now will continue to run. When you buy new software, or upgrade it, it will be shipped as universal binaries.

If you've got a two-year-old computer, the current machines are likely to be about twice as fast. I'm expecting that by the time I need to replace this Powerbook, the Intel chips will have about three times the performance, if not more. So I can afford to take even a 50% emulation hit -- the new computer will still be half again as fast as this one, even under emulation. And by that time, Creative Suite 3 will be out, and there might be enough new in it to justify upgrading -- and if not, running Creative Suite 1 at 150% the speed is still a win.

How many developers will figure that X86-compatible upgrades, many of which will require as much coding effort as your typical whole-version-number bumps, ought to cost their base of users just nominal $10 costs? And how many will devote their coding resources to Intel compatibility instead of The Next Great Thing?

Probably not as many as have figured out that making their application Intel-compatible takes less than a day's worth of effort. People who are dealing with Cocoa have a few gotchas to worry about; people who are dealing with Carbon have a few more. People with sensible programming processes will catch the vast majority of the gotchas within an hour of testing their code on an Intel Mac. The people who are really going to have problems are the people with huge legacy codebases in Codewarrior, and they're usually companies that are already dealing with Intel compatibility because they're supporting a Windows version and a Mac version already. People with poor development practices are going to have problems, too, but you probably don't want to buy software from them already.

-- Charlton Wilbur cwilburchromatico.net

On Friday, June 17, 2005, at 12:42 PM, Kevin van Haaren wrote:

> --On June 16, 2005 10:40:19 PM -0700 Marilyn Matty <mmattynyc.rr.com>
> wrote:
>
>> The difference between what happened with the Osborne and the Mac is
>> that
>> Osborne announced a laptop computer that was so far from available as
>> to
>> be vaporware - and I'd guess that prices will drop incrementally on
>> PPC
>> iBooks and Powerbooks as time goes on. As Steve Jobs said in his
>> presentation, there aren't enough, or advanced enough PPC chips to
>> enable
>> Apple to develop competitive laptops, which are the largest segment of
>> the market
>
> Interesting follow-up on the Osborne effect in Cringely's latest
> column.
>
> <http://www.pbs.org/cringely/pulpit/pulpit20050616.html>
>
> Apparently this has become an urban legend. Osborne didn't announce a
> computer a year ahead of time. He announced a computer 3 months ahead
> of
> time, and they actually shipped it. But in the meantime Kaypro
> released a
> cheaper computer with a bigger monitor. When Osborne released their
> announced computer it was more expensive than even the old Osborne and
> still had a smaller screen than the Kaypro.

A good article, but it doesn't really discuss an important point I was
trying to make. Osborne announced a new model when his very pioneering
laptop had pretty much become obsolete in the market. Apple has had a
number of laptops in the iBook and Powerbook lines for quite a few
years now. Apple can continue to sell very desirable laptops with
advanced features - which many of us will agree are the best designed
on the market, though now will probably be able to do so at reduced
prices.

When Osborne did introduce the laptop, as Cringley mentions, it was
woefully uncompetitive in terms of features and pricing - far from what
consumers had been expecting. Comparatively speaking, there have been
few clunkers in the Mac line in recent years (and I've known many Cube
lovers that I still think it was a shame this line wasn't better
marketed) - none of them fatal.

And I also still hold with my contention that a big, though not the
only, reason Apple went with Intel rather than AMD is because of the PR
value of Intel and the fact that the vast majority of PC owners will be
assured by "Intel Inside," no matter how much I cringe at the
possibility of that crappy logo next to the beautiful Apple marque. I
hope they will design a special one for Apple products or at least make
it cocoa colored or something.

Marilyn

On June 8, 2005, in message 10 of this thread, chuck goolsbee wrote:
> Apple's "traditional lock on the Creative Markets"? I think that was
> a happy accident, not a plan on Apple's part. [...] That creative
> and media users choose Macintosh has more to do with the developers,
> history, and that Interface issue I mentioned above than the CPU

Actually, if you remember back to the early 80s, it was known by all
in the computer industry that if you need graphics, you choose 68000
over x86. I think it's highly likely that Apple specifically chose
68K because they were building graphical computers.

And little-Endian is wacko. Network byte order is big-Endian despite
most PCs using little-Endian Intel processors; I like to believe that
that validates my belief. :-)

I suppose we're stuck with little-Endian processors now, but I still
hope that Apple would get chips from Intel with at least a few more
registers and interrupts so that Rosetta (and maybe even Windows?)
might run more efficiently. It should be easy for Intel to add more
registers, since they've been doing that since the 8086; I don't know
about interrupts (although with the loss of Classic support, maybe it
won't matter as much?).


-Walter

Almost everyone assumes Apple switched, but in actuality IBM didn't make Apple an offer it couldn't refuse. In fact, it probably just sat there as Jobs poked it with a stick.

The question of Apple's cash situation as a counter to the Apple introduced FUD kind of overlooks what another kind of shift like this will do to desktop OS market share. And portable hardware market share in the next year or so should be a problem for Apple. Didn't Apple's OS share drop at each stage going to PPC and then to Jaguar? Oh, and the last two major upgrades of 10.x haven't exactly been earth shattering advances.

The single most bizarre argument is that emulation will be okay since the chips will get faster over the next immediate years. But the chips will also get faster for the parties not using emulation...hello? Let's be honest: You wouldn't elect to upgrade all your software at one time - if you're doing anything serious you will simply be mandated to upgrade your software all at one time.

Apple can only win if its computers morph into home entertainment hubs for net access, games and whole house audio and video, cameras, cell phones, music players, etc., all elegantly interoperable. The winner(s) will not be computer companies per se, but will be the multi-functional, mindshare centers for the forthcoming breed of electronic gadgets. Every home is going to need a Borg Queen, after all, and the one that doesn't annoy you will be the best one. Simplicity is the hardest thing....

Also, didn't Mark Twain make the remark about not having enough time to write a short letter, and not Lincoln?

> I'm not sure if there will be an effect immediately, but there is
> going to be a huge one the closer we get to 2006. I expect Apple to
> have a horrible Christmas season.

Apple may have experienced some of the (at this point theoretical) decrease
in sales anyway, of course, particularly in laptop sales. If there really
wasn't any viable way to get significantly better chips for Apple laptops,
then convincing people to buy them would have become increasingly more
difficult, and so Apple would have lost sales there anyway.

I expect there'll either be news stories that talk about how Apple's sales
has dramatically fallen since the Intel announcement, which means the end of
Apple (again), or new stories about how sales (or revenue, at least) haven't
really dropped much at all, despite the change, and so the end of Microsoft
(again).

IMO laptop users are the ones that stand the most to gain by waiting (but
I'm so firmly in the buy-when-you-need it camp that I'm buying a new Mac
this week), anyway, since there's a higher chance of a more-than-a-bump
speed increase.

=Tony.Meyer

At 2:21 AM -0700 6/17/05, <tidbits-talktidbits.com> wrote:
>Six billion in the bank and no debts says they can handle a depressed buying
>season or two.

The worry is also the stock price, and here the cash becomes a bit of a double-edge sword. If the stock price drops substantially, it could get to the point where Apple becomes a take over target. Someone looks at an Apple selling at $15/share, calculates that they have about $10/share in the bank in ready cash (very rough estimate), and thinks that they could try and buy the company for an effective price of $5/share.


--
David J Silbey History Alvernia College david.silbeyalvernia.edu

Walter Ian Kaye wrote:
> Actually, if you remember back to the early 80s, it was known by all
> in the computer industry that if you need graphics, you choose 68000
> over x86. I think it's highly likely that Apple specifically chose
> 68K because they were building graphical computers.

Back then 68k's were just generally faster than x86's. Apple already had
a history with Motorola with the 6502, and given that the 68k's were a
better chip, the only way they'd have switched to Intel would have been
with some kind of amazing deal (and even then...).

> And little-Endian is wacko. Network byte order is big-Endian despite
> most PCs using little-Endian Intel processors; I like to believe that
> that validates my belief. :-)

I believe this was more the result of big endian (and Sun in particular)
being the dominant player when network byte order was introduced,
although I could be wrong. Of course, modern network systems often don't
use network byte order, and instead flag the byte order of their
content. This turns out to be faster in many cases.

> I suppose we're stuck with little-Endian processors now, but I still
> hope that Apple would get chips from Intel with at least a few more
> registers and interrupts so that Rosetta (and maybe even Windows?)
> might run more efficiently. It should be easy for Intel to add more
> registers, since they've been doing that since the 8086; I don't know
> about interrupts (although with the loss of Classic support, maybe it
> won't matter as much?).

The PowerPC propoganda tends to make a big deal about the number of
registers, but don't assume this directly translates to improved
performance. Register renaming makes how many visible registers an
architecture have mostly moot (and in fact in some cases the more
registers you have, the more complicated your register renaming work
actually ends up being). If you include the rename registers, a modern
P4 actually has significantly more registers than the G4 (I'm not sure
about the G5, but I think it has similar register counts to a P4).

As for the interrupts, rumor is that Apple is going to require EFI and
APIC, which should alleviate the pain of old PC architecture interrupt
limitations. This may actually have the side effect of getting the rest
of the PC industry finally fully on board with these standards (APIC is
basically already widespread, but EFI is definitely not).

--Chris

alsulliv wrote:
> The single most bizarre argument is that emulation will be okay since
> the chips will get faster over the next immediate years. But the chips
> will also get faster for the parties not using emulation...hello? Let's
> be honest: You wouldn't elect to upgrade all your software at one time -
> if you're doing anything serious you will simply be mandated to upgrade
> your software all at one time.

I think this point about emulation is a fair one, because it is
addressing the context of someone who *doesn't* upgrade their software.
For those who do, they'll undoubtedly be able to get a binary with
native x86 instructions. It's the people who don't want to pay for an
upgrade who'll be stuck with emulation. While emulation may indeed suck,
for most people the end result will likely be that the software they
have now will run as fast as it does now on these new x86 machines.

--Chris

At 3:05 PM -0700 6/20/05, David Silbey wrote:
>The worry is also the stock price, and here the cash becomes a bit
>of a double-edge sword. If the stock price drops substantially, it
>could get to the point where Apple becomes a take over target.
>Someone looks at an Apple selling at $15/share, calculates that they
>have about $10/share in the bank in ready

Presumably, Apple management would not be inclined to sell in that
case. (In fact, it might even buy back some stock.) If there was a
hostile takeover, it shouldn't be hard to convince shareholders that
the sales decrease is only temporary, and that such a low-ball offer
is not in their best interest.

--
Larry Rosenstein
lrosensteincatsincharge.com

On 20 Jun 2005, at 16:05 , David Silbey wrote:
> At 2:21 AM -0700 6/17/05, <tidbits-talktidbits.com> wrote:
>> Six billion in the bank and no debts says they can handle a
>> depressed buying
>> season or two.
>
> The worry is also the stock price, and here the cash becomes a bit
> of a double-edge sword. If the stock price drops substantially, it
> could get to the point where Apple becomes a take over target.
> Someone looks at an Apple selling at $15/share, calculates that
> they have about $10/share in the bank in ready cash (very rough
> estimate), and thinks that they could try and buy the company for
> an effective price of $5/share.

This may have worked in the Wall Street 80's but is rather unlikely
to now. There are many ways to avoid this, and all companies are
aware of these sorts of tactics.

At 1:13 AM -0600 6/21/05, Google Kreme wrote:
>This may have worked in the Wall Street 80's but is rather unlikely to now. There are many ways to avoid this, and all companies are aware of these sorts of tactics.

There are certainly ways that companies can try to avoid this, but they are not uniformly successful. To be most effective, they usually need to be done before a hostile offer comes in. Has Apple put any of the 'poison pill' options in place? (http://www.wiggin.com/pubs/articles_template.asp?ID=1023377242000 for a discussion)

Nor do the strategies necessarily work if they're in place. Witness the Oracle/Peoplesoft struggle:
http://money.cnn.com/2004/12/13/technology/oracle_peoplesoft/

The result is that cash-rich companies with lower stock prices have continued to be seen as takeover targets:
Barnes & Noble: http://biz.yahoo.com/ap/050405/barnes_noble_stock.html?.v=1 was the one I immediately found.

cheers,
--
David J Silbey History Alvernia College david.silbeyalvernia.edu