By: Daniel Bela Bizo (daniel.b.bizo.delete@this.gmail.com), July 28, 2021 6:03 am
Room: Moderated Discussions
anonymou5 (no.delete@this.spam.com) on July 27, 2021 5:17 pm wrote:
> > In olden times they used a process family name followed by a number, e.g. HMOS II.
>
> P1272, P1274, P1276, P1278, etc. for the 12"/300mm variants – was that really so bad? :)
Yes. Not for you, not for me, but for Intel. Believe it or not, it's all about marketing, even for such a technical topic. Even though people outside the semiconductor industry don't really know or correctly understand concepts such as Moore's Law, semiconductor performance characteristics and trade-off, or nanometre labels (I'll admit, I barely do myself). In recent years it has become fashionable in professional circles to drop Moore's Law into the conversation, which for that reason has become my flag to spot whose reasoning is lazy and likely flawed.
But they read and hear about them in both the trade and financial press, even if in an absurdly reduced fashion. In times when anyone with a hundred bucks and a smartphone is an investor, an influencer and a self-appointed expert in any matter, and when credit is cheap, the narrative matters more than ever. Arguably, narratives now matter more than fundamentals. Just look at all those billion-dollar companies that will never make money, or that maker of those ghastly electric cars with a P/E of 300+. INTC is at 11. That's what Gelsinger is trying to fix.
Of course, Intel should not have waited for a change to its naming scheme for the moment when it's fallen behind, because as Andrey put it, it looks desperate. Andrey is also right they should have adopted a proprietary naming scheme, much like codenames, to avoid direct comparisons with TMSC to the unqualified, who don't actually understand what it means anyway, but they will still form perceptions. 5 or 10 years ago it would have been a master stroke. Shouldwouldacoulda.
> > In olden times they used a process family name followed by a number, e.g. HMOS II.
>
> P1272, P1274, P1276, P1278, etc. for the 12"/300mm variants – was that really so bad? :)
Yes. Not for you, not for me, but for Intel. Believe it or not, it's all about marketing, even for such a technical topic. Even though people outside the semiconductor industry don't really know or correctly understand concepts such as Moore's Law, semiconductor performance characteristics and trade-off, or nanometre labels (I'll admit, I barely do myself). In recent years it has become fashionable in professional circles to drop Moore's Law into the conversation, which for that reason has become my flag to spot whose reasoning is lazy and likely flawed.
But they read and hear about them in both the trade and financial press, even if in an absurdly reduced fashion. In times when anyone with a hundred bucks and a smartphone is an investor, an influencer and a self-appointed expert in any matter, and when credit is cheap, the narrative matters more than ever. Arguably, narratives now matter more than fundamentals. Just look at all those billion-dollar companies that will never make money, or that maker of those ghastly electric cars with a P/E of 300+. INTC is at 11. That's what Gelsinger is trying to fix.
Of course, Intel should not have waited for a change to its naming scheme for the moment when it's fallen behind, because as Andrey put it, it looks desperate. Andrey is also right they should have adopted a proprietary naming scheme, much like codenames, to avoid direct comparisons with TMSC to the unqualified, who don't actually understand what it means anyway, but they will still form perceptions. 5 or 10 years ago it would have been a master stroke. Shouldwouldacoulda.