My editorial is finally out.

[cthellis42]

Gotta love intelligent and well-spoken rants...

Honestly, though, I was hoping to see more quotes. Not name-giving or source-revealing, but more "revelations" we don't usually get to see. (Since this is a "going out with a bang" thing.) I enjoyed reading the ones that WERE there, but if you feel like pasting up a few more pages of 'em, feel free.

 

[Arun]

The emperor has no clothes, but since he shoots the messenger, nobody’s gonna tell him! They give him the message he wants to hear: “Yes, Jen Hsun, next generation will be even better!â€￾

Although this may very well be the truth I don't think this situation is as rare in other corporations as your statement, IMO, suggests. It may start from the top, but middle management isn't blameless. There are also advantages to saying yes.

Agreed. Of course, such things also happen in other corporations, and there are huge advantages to saying yes; but the problem isn't really that they're saying yes IMO. It's that, more than a year later, they still didn't really understand that "yes" was nothing more than a lie.

IMO, most layer of managements at NVIDIA are "guilty"; top could be considered more, though, because it's extremely important they got at least a small clue on what's going on overall.

Micron:

An eight page article on why nVidia sucks.......

I'm only a little bit disapointed. I thought Uttar was going to cover more of the industry in general instead of dedicating 98% of the article to nVidia's dirty laundry. Uttar, why do you use an nVidia avatar on the forums you hang out at? I used to think you were a person who liked their products...

No, no; not 98%. 75%. Count yourself:
There are 8 parts. Part 3 is of average size, and is not related to NVIDIA all that much. A few other parts, such as Part 4, are not fully related to NVIDIA. So let's say 2 out of the 8 parts are not related to NVIDIA.
2/8 = 1/4 = 25% -> 75% is related to nvidia

I was very sincerly surprised of how the editorial came out in the end - the decision to focus so much on NVIDIA was justified by the idea that if it was just a "part" of the editorial, it wouldn't have seemed to have had any real importance. And I didn't want that to happen.

Furthermore, if I didn't like NVIDIA's products, why would I write what Geo just described as "an eight page article on what nVidia needs to do to quit sucking."? If I wanted them to die ASAP, I'd just have made an editorial mostly about some other stuff, and with one part including out-of-context quotes about how good NVIDIA is doing and how nobody really knows how fantastic of a job they are doing on the NV40 and NV50

Gotta love intelligent and well-spoken rants...

Honestly, though, I was hoping to see more quotes. Not name-giving or source-revealing, but more "revelations" we don't usually get to see. (Since this is a "going out with a bang" thing.) I enjoyed reading the ones that WERE there, but if you feel like pasting up a few more pages of 'em, feel free.

Trust me, I also would have preferred to have an editorial made of 99% quotes and not have to write anything myself but perhaps the conclusion

More seriously though, yes, I would have liked to have put more quotes in the editorial. Problem is, I just don't have more quotes available.

More than 90% of the information I received was deleted out of everywhere but my head within 24 to 48 hours. The quotes you see here are the ones I saved, often without any particuliar reason, or simply because I found them interesting and worth keeping.

---

Also, I finally made a more "final" correction in Part 6, which is also noted on the frontpage so everyone can see I edited the text.

[Correction, 11/9: The $10M/tape-out cost seems significantly too high, it should be more like $1M – so either that was a typo from the source, or it included a lot of other barely related costs. Also, the 7-10 chips per wafer number seems to be for one of the first working tape-outs where NVIDIA absolutely needed chips to show to the world – the $60 per chip cost would then be the lowest NVIDIA managed to put costs at, after several respins. Also, there most likely weren’t 7 tape-outs, maybe rather 4 tape-outs and 3 respins. - Uttar]

Maybe I should have added mention of how ambitious and pretentious management was in late 2000 and why that caused the NV30 to be so ridiculously ambitious. But I doubt it's crucial for me to make fifty billions of minor edits and corrections - I prefer just to correct factual errors, for now at least


Uttar

 

[Evildeus]

Well, from my point of view, the "shoot" the messenger, is a really short minded stategy, and i don't think it can happen when the product is actually bad (and don't tell me than the CEO of Nv can't look at some board to see what the customer really think of it's product....). In fact, i would say that, it's best to be the messenger of what actually is happening, than those who say, everything is ok, under control....

But, the Nv30 is not a bad thing all in all (for Nv not the purchasers of NV30), because they will be a bit more humble for some time.

 

[Bouncing Zabaglione Bros.]

Well, from my point of view, the "shoot" the messenger, is a really short minded stategy, and i don't think it can happen when the product is actually bad (and don't tell me than the CEO of Nv can't look at some board to see what the customer really think of it's product....). In fact, i would say that, it's best to be the messenger of what actually is happening, than those who say, everything is ok, under control....

But, the Nv30 is not a bad thing all in all (for Nv not the purchasers of NV30), because they will be a bit more humble for some time.

This is a nice idea but doesn't work in practice. The idea that people may disagree with you but will respect your willingness to tell the truth only works in technical areas like engineering or software development. The mindset of those people tends to allow a real discussion of the issues.

However most of the companies that have this problem are sales and marketing driven. These people have a tenuous understanding of technical issues at best. They think they can sell anything through the power of marketing, so people who do not agree with their plans (or misconceptions) tend to be dismissed as negative, troublemaking or "not part of the team". Anyone who tries to have a truthful discussion of problems and their possible solutions are shot as the messenger of bad news, ignored, and the marketing led-management moves on down the same incompetent road.

This attitude filters down the company, with middle management continuing this company attitude, because that is how they are being treated by their upper management.

In short, unless the person at the top has the knowledge, experience and guts to make changes to the company attitude and atmosphere as a whole, anyone who is willing to point out the emperor has no clothes is ignored and purged, and the company carries on down the same road. This is what caused problems at the likes of Enron and 3DFX.

Dave Orton at ATI has shown what can be done if you are willing to grasp the nettle and truly make changes in your company, and how a turnaround in culture can affect every aspect of your business, making you a better more successful company.

 

[Evildeus]

I didn't say that it doesn't happen, i say that it's unliky when everybody can see that the product is unwelcome. That doesn't mean that you can find who's in fault, but that you can shoot the messenger that comes and say "well our product is the best available" . And if you look at the top management, they no more say that Nv30 was/is a sucess, and that the dustbuster was a non-issue.

Moreover, the fact that you don't understand the technical issue, doesn't prevent you of knowing and saying that the product is F* up.

In short, unless the person at the top has the knowledge, experience and guts to make changes to the company attitude and atmosphere as a whole, anyone who is willing to point out the emperor has no clothes is ignored and purged, and the company carries on down the same road.

Well, i would say that you are ignoring one part of the firm: the shareholders or potential ones. Marketing it's good, selling products, is better. When the management stays in a "ignoring stance" they are at the verge of being sacked and the shareholders, or those willing to become, has interest in limiting this "ignoring stance".

But well, that doesn't mean that we live in a perfect world, just that it's not a general case

 

[Arun]

I didn't say that it doesn't happen, i say that it's unliky when everybody can see that the product is unwelcome.

Management has the bad tendency of looking an awful lot more at sales and statistics than at the market's opinions.
There were a lot of very positive early reviews of the FX5900U - and a few negative ones, which were later often accepted as "more objective and better thought."
Which reviews do you think management read, at the suggestion of employees?

I'm not saying management believes the NV35 and NV38 are a homerun. But their opinion of them, AFAIK, is way better than it should be. They're most likely more worried about them having worst margins than ATI's R350/R360 right now.

And if you look at the top management, they no more say that Nv30 was/is a sucess, and that the dustbuster was a non-issue.

Of course, but take the NV31 for example - did they ever admit it was a failure? I agree it's not anywhere as big as a failure as the NV30, but its sales aren't amazing and the performance is even less amazing.
The only good products NVIDIA has right now are the NV34 and, more recently, the NV36.

The good news is they're finally competitive in the mid-end again. The bad news is they weren't as competitive in the mid-end for the last 6 months as they thought they were. Also, the awful heat problems AND bad performance of the NV31 and NV34 must has caused them to lose a bit of "monumentum" in the mobile/notebook market.
I don't know if management knows of the not-so-good position they're in for the mobile market ( good thing the NV34M is the cheapest DX9 mobile chip by far, and the GF4 series is still good for notebooks ), I suspect they do have a rough idea of it, but not a sufficently good one.

Furthermore, the fact that you don't understand the technical issue, doesn't prevent you of knowing and saying that the product is F* up.

Agreed. But it doesn't make you understand it either

When the management stays in a "ignoring stance" they are at the verge of being sacked and the shareholders, or those willing to become, has interest in limiting this "ignoring stance".

Well, if the shareholders don't know or don't believe they're in an "ignoring stance", they're safe - which is really the whole problem IMO.

Heck, if management suddenly said to itself "Hey! Wait a second - how come we didn't know this, that and this thing? Maybe we're not so much in control as we thought we were. We better do something about it, too." - then that ignoring stance is likely not to survive much longer.


Uttar

 

[Evildeus]

I didn't say that it doesn't happen, i say that it's unliky when everybody can see that the product is unwelcome.

Management has the bad tendency of looking an awful lot more at sales and statistics than at the market's opinions.
There were a lot of very positive early reviews of the FX5900U - and a few negative ones, which were later often accepted as "more objective and better thought."
Which reviews do you think management read, at the suggestion of employees?

Well, let see, the statistics and the selling figures were/are good? Surely not, since the NV30 debacle, the Nv figures are going down (in term of market share at least). Where was the market share of 250+$ DX9 cards till last quarter, around 5%? That's great, when you know that the top has a big impact on the rest. The statistics were (still are) not good. The mangagement can be leaded, but the plus of internet is to be able to look everyware, ang google exist. Reviews is a thing, customer feedback is much better.

PS: don't you think that card makers can say, hey your card sux, fix that quickly?

I'm not saying management believes the NV35 and NV38 are a homerun. But their opinion of them, AFAIK, is way better than it should be. They're most likely more worried about them having worst margins than ATI's R350/R360 right now.

Even if NV3+ is not a homerun, it's certainly much better than the NV30. The margin comes also with the good decision at the good time, that was not the case.

Of course, but take the NV31 for example - did they ever admit it was a failure? I agree it's not anywhere as big as a failure as the NV30, but its sales aren't amazing and the performance is even less amazing.
The only good products NVIDIA has right now are the NV34 and, more recently, the NV36.

Sure, but has anyone asked them what they think of NV31? Generally speaking all the questions focus on the NV30/35 debacle.

The good news is they're finally competitive in the mid-end again. The bad news is they weren't as competitive in the mid-end for the last 6 months as they thought they were. Also, the awful heat problems AND bad performance of the NV31 and NV34 must has caused them to lose a bit of "monumentum" in the mobile/notebook market.

Well, the mobile market is the only market were they do better, but i agree. They screw up, they can't expect to do better. screwing is something that comes sometimes, knowing that and moving forward, with the issues in mind, rapidly is the best thing, even more in technological market. Otherwise in 2-3 years you are dead.

I don't know if management knows of the not-so-good position they're in for the mobile market ( good thing the NV34M is the cheapest DX9 mobile chip by far, and the GF4 series is still good for notebooks ), I suspect they do have a rough idea of it, but not a sufficently good one.

I think they buy market research for that, and they can at least see the Nv selling figures, that's enough to have a good idea.

Well, if the shareholders don't know or don't believe they're in an "ignoring stance", they're safe - which is really the whole problem IMO.

Heck, if management suddenly said to itself "Hey! Wait a second - how come we didn't know this, that and this thing? Maybe we're not so much in control as we thought we were. We better do something about it, too." - then that ignoring stance is likely not to survive much longer.

Sure, but the shareholder that takes only into accounts what the management says, is doomed anyway so...

 

[Arun]

Reviews is a thing, customer feedback is much better.

Do you know just how long management hours already are? 50 to 60 hours/week is nothing unusual for people like Jen Hsun Huang.
Do you seriously expect them to have the time to watch forums?

Even if NV3+ is not a homerun, it's certainly much better than the NV30.

I agree, and mentionned that in the editorial

Sure, but has anyone asked them what they think of NV31? Generally speaking all the questions focus on the NV30/35 debacle.

I'll grant you the point I've never seen anyone ask management what they think about the NV31. However, all PR interviews a few months ago insisted they had "performance leading parts in every market segment" ( paraphrase ).
Considering the NV31 margins and the lower-than-wanted clocks ( target was 450Mhz IIRC ), I admit I doubt many NVIDIA employees consider it a success. But whether they realize the NV31M, for example, is in every way inferior to the 9600M, I don't know. Maybe they do, maybe they don't - how should I know? *grins*

I think they buy market research for that, and they can at least see the Nv selling figures, that's enough to have a good idea.

True, but that doesn't make them know how to correct the problem.
Saying "Okay, our sales aren't too good, so marketing focus on mobile stuff in interviews and engineering, like, make notebook products even better, alright?" won't help a lot
Actually, it can help, but understanding how bad their products are and knowing evantually how important project X, Y or Z is based on that helps.

Sure, but the shareholder that takes only into accounts what the management says, is doomed anyway

Then you might just as well say 75% of the world's shareholders are doomed
Actually, that's not true - but shareholders and analysts generally do take management relatively seriously. If they begin to have a tract record of not being accurate though, then yes, they'll stop believing them.

But if management says a product is amazing and it's the best thing since sliced bread, it WILL influence analysts a lot ( "they got a new product, and we believe it might make them gain market share in the X market segment as early as in Y weeks" ). Do you know just how, well, "untechnical" analysts are?

The only analyst I ever dealt with was one who, a few months ago, asked me at which fabs the NV36 and NV40 are made at.
I said: "Both at IBM, reliability: NV36: 100%, NV40: 99.9% because who knows, there's enough time left for them to make a strange last-minute decision to switch it back to TSMC"

He thanked me for the info, and said: ( paraphrase once again ) "I'm surprised you're so sure the NV36 is manufactured at IBM, though. After all, the NV36 is just a 'FX 5600 rehash', it's not a fundamentally new product."

You heard it here first, guys: the FX5700 is a FX5600 rehash
---

I sent the final as well as "intermediary releases" to someone who worked at NVIDIA from 2001 to late 2003. He suggested me to make the text more aggressive and bitter ( and that has been reflected in the final version ).
While he was kinda angry at nV for many reasons, I do find it very telling.


Uttar

 

[Zvekan]

Very informative read

Just one question. What is the difference between tape-outs and respins?

Zvekan

 

[cthellis42]

Trust me, I also would have preferred to have an editorial made of 99% quotes and not have to write anything myself but perhaps the conclusion

More seriously though, yes, I would have liked to have put more quotes in the editorial. Problem is, I just don't have more quotes available.

More than 90% of the information I received was deleted out of everywhere but my head within 24 to 48 hours. The quotes you see here are the ones I saved, often without any particuliar reason, or simply because I found them interesting and worth keeping.

Hrm, I see. I suffer for the exact opposite problem, really. I've kept most of my emails dating back all the way to 1997 (only losing some due to system crashes that weren't duplicated elsewhere), and I don't really DO anything important with them! Hehe... Afraid of getting hacked?

 

[Arun]

Trust me, I also would have preferred to have an editorial made of 99% quotes and not have to write anything myself but perhaps the conclusion

More seriously though, yes, I would have liked to have put more quotes in the editorial. Problem is, I just don't have more quotes available.

More than 90% of the information I received was deleted out of everywhere but my head within 24 to 48 hours. The quotes you see here are the ones I saved, often without any particuliar reason, or simply because I found them interesting and worth keeping.

Hrm, I see. I suffer for the exact opposite problem, really. I've kept most of my emails dating back all the way to 1997 (only losing some due to system crashes that weren't duplicated elsewhere), and I don't really DO anything important with them! Hehe... Afraid of getting hacked?

Hehe.
Actually, FYI, many of the info I got was through forums' PM systems. I don't think e-mails are generally sufficently secure for rumors - although that's just me...

With PMs, the problem is if you don't delete it, it remains in the database - and if it remains in the database, it's ridiculously insecure.
So that means I'd have had to save every single of them systematically - something I couldn't find any justification for, since I never even thought about using them this way before


Uttar

 

[ricercar]

Well, you can't reveal sources under supoena when there's no saved email from them.

It occurs to me that a good example of shooting the messenger was the termination of the CFO Christine Hoberg. Maybe someone can explain why a well-run company would fire the CFO for relatively insignificant accounting errors. IIRC it was $3M underreporting on $6B which was not even fraudulent, but rather human accounting errors. Yet the company lost a CFO.

 

[LeGreg]

Hey it never occurred to you
that deleted messages could be not removed from Database
but just marked with a "deleted" flag ?

 

[Arun]

Hey it never occurred to you
that deleted messages could be not removed from Database
but just marked with a "deleted" flag ?

Sure did. That's why I generally made sure the scripts of the forums did not operate that way
Frankly though, it wouldn't make sense anyway. If both users delete a message, keeping it just seems like a waste of storage, something few programmers tend to appreciate

Ricercar: Good example. Actually, it's understandable - the whole thing caused them BIG problems, made their stock go down a LOT, and so on. Firing the CFO though, did seem too extreme. Although I fear that in that example at least, many other companies ( but not all, maybe not even most, but I don't know about that ) would have operated in a similar manner.


Uttar

 

[cthellis42]

With PMs, the problem is if you don't delete it, it remains in the database - and if it remains in the database, it's ridiculously insecure.
So that means I'd have had to save every single of them systematically - something I couldn't find any justification for, since I never even thought about using them this way before

Good point. ^_^ Me, I like info. I'd be cut 'n pasting all day. Hehe...

 

[g__day]

I enjoyed the article. If I could have hoped for more it would have been informed speculation on "what is in the mind of Jen Hsun and the top management team as they weathered 2003?".

Say for instance a top site or two took all NVidia's faults over the past year and turned it into more a professional, business analyst's view of NVidia's challenges, then published it and sent it to NVidia executive and a few stock anaylsyts - something say titled "10 questions Merryl Lynch and all Investors should ask NVidia today?"

How frankly would you expect NVidia to respond and what would they do? Because that is Merryl Lynch's bread and butter - working out an earnings forecast and getting inside the health of the company. They don't have to understand technical 3d terms to understand management problems. And they have a heck of alot of clout and influence when they rock up to you in private or at an AGM and say "Please explain...".

Has anyone every thought of writing a background arctile like that for the analysts then watching the delivery?

How do you think NVidia would react if really insightful questions were asked by the financial servcies community that underpins their operations?

 

[micron]

How do you think NVidia would react if really insightful questions were asked by the financial servcies community that underpins their operations?

Why would you want to do such a thing?

 

[Vince]

Say for instance a top site or two took all NVidia's faults over the past year and turned it into more a professional, business analyst's view of NVidia's challenges, then published it and sent it to NVidia executive and a few stock anaylsyts - something say titled "10 questions Merryl Lynch and all Investors should ask NVidia today?"

I believe Micron is correct. Investors only care about tomorrow, today has past and yesterday is forgotten. This is a concept that's diametrically opposed to that expressed frequently by the closed minded individual, who I happen to believe will never let go of the events that transpired in the year two thousand and three.

 

[cthellis42]

This is a concept that's diametrically opposed to that expressed frequently by the closed minded individual, who I happen to believe will never let go of the events that transpired in the year two thousand and three.

No doubt, no doubt... After all, we've seen plenty of people who try to justify the events that transpired in the year two-thousand and three with events they never let go of that transpired in two-thousand and one.

But hey, since we're still IN 2003, at least it's more relevant.

 

[g__day]

Er, this seems almost too basic to explain.

Investors care about today and tomorrow - because you get to tomorrow from today. If today is stuffed - and you're hiding it - then tomorrow is likely to be stuffed as well.

I deal a fair bit with Investment banks / analysts - they like to know they're asking the right questions. They would be very receptive to gaining a better understanding of how effective a company like NVidia's stategy and execution is in a market sub segment as big as 3d video cards.

And overall the sooner NVidia comes clean and faces the music the sooner things will get better and the less damage done. Public companies can't simply hide their mis-deeds in todays open and regulated markets.

This happened to e-business as well - as soon as the right questions were asked the bubble burst.

Why do I care? Because I hate seeing a fine company squandering its assets with camaflagoue substituting for content in areas. It can only do this when talented individuals who realise what is happeneing stay quiet. B3D has over the past year done great insights into their problems, trouble is it isn't summed into one professional article that is pointed towards the right company internal (executive management) and external (market analysts and partners) influencers.

The truth is out there is just has to be told to the right groups to take effect and cause positive changes. You only need 1 share in a company to attend its annual general meeting and ask its executive a challenging question. If someone framed 10 thought provoking questions to 10 people with a share each - you'd have a AGM where the message would get to NVidia very clearly and to all other major investors attending the AGM. This very tactic worked well for GreenPeace in the past.