Bill Gates: how Intel screwed us
Date: 1997
I have a critical meeting with Intel a week from Wednesday. I want to convince them that they need to stay away from Oracle NCs and work more closely with Microsoft.
I will have to spend some time reviewing the low end and how they screwed us there and describing our strategy. I will have to spend some time explaining how we think customers view NCs and what we see happening with JAVA ..
My three big propositions to them are that: low end stuff should be a clear subset and feed the PC architecture.
SUN byte codes are bad for them and working closely with Microsoft represents the only path under which hundreds of millions of customers will want their very very high performance chips.
..
The only concrete action item here is for Jim/David to pick a few people who are very technical to help me on this point. I don't want to say that Microsoft will never do a software architecture but I do want to point out that the SUN approach does not allow Intel innovation to look good.
..
I don't think it helps Intel to explicitly fragment development activity. In fact I think NT on Merced in near term is their best hope of getting Merced volume. Looking out beyond that I think our work on natural interface (I mean by the our research work putting speach/handwriting/linguistics/learning together in a bayesian framework) is the only concrete thing that can happen fast enough with volume to make 200mips chips look bad then make 500mips chips look bad, then make 1200mips chips look bad ... etc .. etc. This is what Intel needs from the software market.
One thing that is tricky is to explain how their efforts with Oracle/JAVA and SUN make the bad scenarios for them more likely - the world moves to byte codes as a mainstream choice, the world fragmentats so using cycles goes slower, they don't work with us to make the natural interface stuff happen fast ..
----
Remark from reader:
Don Billaranzano:
why you no my friend anymore, I hear you do bizness with de odda mob, why you no come to me, now I can't protect you from BAD SCENARIOS ..
Synopsis
Snippet
From: Bill Gates Sent: Sunday, October 12, 1997 6:04 PM To: Jim Allchin (Exchange); Marshall Brumer Cc: Joachim Kempin; Nathan Myhrvold; Rick Rashid; Eric Rudder; David Vaskevitch; Peter Kukol; Cameron Myhrvold; Bill Gates; Miuke Porter; Butler Lampson Subject: Intel meeting October 22 - technical contentI have a critical meeting with Intel a week from Wednesday. I want to convince them that they need to stay away from Oracle NCs and work more closely with Microsoft.
I will have to spend some time reviewing the low end and how they screwed us there and describing our strategy. I will have to spend some time explaining how we think customers view NCs and what we see happening with JAVA ..
My three big propositions to them are that: low end stuff should be a clear subset and feed the PC architecture.
SUN byte codes are bad for them and working closely with Microsoft represents the only path under which hundreds of millions of customers will want their very very high performance chips.
..
The only concrete action item here is for Jim/David to pick a few people who are very technical to help me on this point. I don't want to say that Microsoft will never do a software architecture but I do want to point out that the SUN approach does not allow Intel innovation to look good.
..
I don't think it helps Intel to explicitly fragment development activity. In fact I think NT on Merced in near term is their best hope of getting Merced volume. Looking out beyond that I think our work on natural interface (I mean by the our research work putting speach/handwriting/linguistics/learning together in a bayesian framework) is the only concrete thing that can happen fast enough with volume to make 200mips chips look bad then make 500mips chips look bad, then make 1200mips chips look bad ... etc .. etc. This is what Intel needs from the software market.
One thing that is tricky is to explain how their efforts with Oracle/JAVA and SUN make the bad scenarios for them more likely - the world moves to byte codes as a mainstream choice, the world fragmentats so using cycles goes slower, they don't work with us to make the natural interface stuff happen fast ..
Full Exhibit
https://techrights.org/o/wp-content/uploads/2009/01/plex0_3876.pdf----
Remark from reader:
Don Billaranzano:
why you no my friend anymore, I hear you do bizness with de odda mob, why you no come to me, now I can't protect you from BAD SCENARIOS ..