Has Intel lost its mojo?
bits-chips.nl
bits-chips.nl
What's needed, though, are 10-50 year visions: basically, management plans "where do we want to be in a long period of time" and owners commit to that. That is the reason why the German family-owned Mittelstand is so successful - they are not bound to ever changing demands (especially not of "activist shareholders").
(This year marks the end of their current 5 year plan.)
Of course, China is the largest corporation on Earth. Some perhaps some smaller version of that is more ideal for smaller corporations.
He summarises the whole process, Intel is now at the sales and marketing people running it.
The FAANG quasi monopolies have too much money and power to let new competitors grow big. You either let them buy you out or they'll copy your innovative shit with their enormous R&D resources and crush you so most founders are happy to just sell out and enjoy their riches instead of wasting their time fighting Goliath.
MBAs can go from 1 to 100
But every quarter it resets back to zero and MBAs can't figure out 0 to 1
They're saying MBA's get lost in the numbers and forget that the core of a business is to provide added value, which is a moving target in tech (with annual expectations in the case of silicon companies).
A bigger question is why don't the MBAs understand this?
Tech people generally understand when something is far outside their skillset. Why can't the organization chart gymnasts do that?
They need to not understand that.
MBAs are never in for the long term. They move from one company to the next and follow where the next big position is. That's why they could not care less what happens in the company in the long run.
I am not so sure. There's a huge bias in IT tech to approach every problem outside of tech as a technical bug floating in a vacuum waiting to be solved by applying tech methods. I think battling (and most of the time winning) all day long with psychologically rigid computers does that.
I see you don't read HN much
It's all about the share price and doing whatever you can to boost it, and then pay dividends (if the share price isn't doing well).
Companies are now seen as financial instruments. If you buy shares now is your first thought about the company, or about share price and how much money you can make? That's it really. MBAs know how to inflate share price and 'add value' in the modern world.
And when the CEO is no longer an engineer, he chases the money, because he's usually paid in shares and options.
I absolutely agree. And pushed to extremes it makes for sociopathic reasoning where nothing else matters.
We humans have a hard time finding the middle way : no I don't want communism but I don't want full blown predatory capitalism either.
Here's a good example of a non-technical CEO: Steve Ballmer. From wikipedia
"His tenure and legacy as Microsoft CEO has received mixed reception, with the company tripling sales and doubling profits, but losing its market dominance and missing out on 21st-century technology trends.[13][14][15] "
In any case, that means it takes considerable discipline to avoid short-term thinking; where short-term could mean "just 5 years ahead" - already an eon in a market-focused company. That's bad for a company's stability. On the other hand, it also makes it really attractive to try and corner the market and keep competitors out by any means necessary, which isn't necessarily very capitalist nor good for society as a whole.
Anyhow, I'm not sure engineer CEO's are the answer; if the primary drive for investor gains tends away from dividends, but rather rising stock (and asset) prices - well, that's pretty inexorably going to drive companies to focus on stock-price boosting activities over plain profit. And at the heart there is likely simply that size wins, but conversely is problematic once achieved, for both company and society.
We need some kind of mechanic to encourage companies to stay small (or at least weak) and avoid the market-manipulation traps. In essence: economies of scale are a short-term win at a long-term cost. And perhaps we've conflated economies of scale in activities with economies of scale in corporate structure. Some things take physical scale and cooperation, and of course that's easier under a single controlling corporate entity - but we could try harder to disentangle those factors, at least as much as possible.
The real problem here is how to get from where we are to some hypothetical better place. The very economies/companies of scale that inhibit growth are also what underpin modern efficiency. How many people are willing to become actively poorer to maybe possibly encourage growth down the line, especially when they see other countries that won't all be making those painful choices? The US in particular has an issue here: it's richness is bound partially to its power over trade and intellectual property, but those are rather artificial constructs that can fall away quite quickly. How much of US wealth is enhanced by being the standard-setter, and by other countries granting US IP and financial laws force in their own territories? My guess is: quite a lot. If that pyramid scheme ever falters, that could be really, really destabilizing (of course not just in the US!) Essentially: network effects transfer wealth gain from network participants to network controllers, specifically: American entities.
Even those profiting from the current scheme I think are looking at this way too lackadaisically: sure, you can try and support the current hegemony, but it looks pretty inevitable that it's going to fail. There's too much growth outside of the US, and too much protectionism (ironically supported by the US which must be the most self-destructive policy possible), and too much distrust of major gatekeepers across the political spectrum. If you want to support the status quo, your best bet is to look for a gradual restructuring, instead of the current heels in the sand, waiting for the straw that breaks the camel's back. And frankly, even if you're actively hostile to Facebook, Google and US hegemony in generally: sudden revolution looks pretty unpredictable. Why aim for that?
We really, really need to start working on reducing the size (or really power) of corporate behemoths, now, before the system collapsed under it's own momentum. And in the future, we should be more wary of the politically intrinsically destabilizing nature of corporate titans.
It's kind of the essential nature of capitalism.
If you are looking to a system oriented toward value for other than the capital-owning class, look elsewhere.
My university background actually straddles engineering and management, but I ended up working on the business side of the firm. While I was there I kinda thought it was amazing how much of the business was powerpoint slides about market positioning, rather than making stuff, which seemed like a fixed roadmap provided from above. I was an intern, so I kinda figured it was just a thing about a firm of that scale, and besides I had friends in more technical roles so someone was doing the chip-making. Remarkably some are still with the firm.
Anyway the thing that tied up the ends for me has been some recent work. I'm helping a friend at a B2B startup in the financial space, ostensibly in a technical capacity. A huge amount of the work is everything other than coding. Colleagues were brought in with sales/marketing experience, and it's been made clear how these sales work: there's an awful lot of box-checking. You need docs about everything imaginable to even have a chance to sell to some of these firms. GDPR, ISO standards, and so on.
And it turns out there's a disconnect between how things work technically and how they work on paper. When did you last do a GDPR training session? Where is the updated registry with the serial numbers of every device in your firm? It goes on an on with things that work on paper but in practice probably have a bit of an implementation gap. The ones I mentioned are mostly just time consuming, but give it enough complexity and you have documents that say we'll be ready with the next feature in 3 months, but techies that say that's not what they meant.
Guess what, there's always business people involved, and they know that they don't know how code works, but they can read docs. It's important to cater to them. And who does that better than MBAs? And once you make the sale, who gets credit?
So you have a two tribes issue eventually, when the firm is so big that there are business people who don't talk to the tech people and vice versa. In a small firm the interface people are people like me who can do both and have very few steps to anyone else. In a large firm, there are entire halls full of people who think the other side is doing X when they're actually doing Y. You have Chinese whispers at the interface, because in a large firm the tech+biz people are too far away from the tech end to understand what the techies want to offer and what they think is a stretch.
Besides, someone is always going to be complaining about management. That is one of the pitfalls of management - everyone seems to be an expert at running a company except the people who actually do it.
That is especially true when one of the issues lays with manufacturing. The problem may be a product of decisions made in many parts of the company, not just MBAs.
AMD could leapfrog Intel from time to time with a really clever design, but struggled to keep up in manufacturing. This is where the Zen2 gets so dangerous to Intel, a very good design produced on the better manufacturing process. Where did that process come from?
That leads to the decisive point to where Intel is today: the emergence of the smartphone. The impact on Intel is twofold: first of all, the PC market is shrinking as the smartphones are the computing device many people use, PCs have lost their importance for many, even if people still have a PC, they don't get updated as often any more. And secondly: there is a new huge market for semiconductors.
Which Intel managed to not be a part of.
The volume of the mobile marked helped TSMC to get to the point, where their processes eclipsed Intel. This benefits all the competitors of Intel to compete not only on even ground, but on higher ground. Suddenly, non-x86 architectures, mainly ARM become economically feasible. When Apple designs the Apple Silicon, they can afford to throw enormeous R&D resources at any problem. Not only because Apple is a rich company, but also, because the spending is spread across over 200 million iPhones per year. Not counting iPads and soon Macs. This is, why I am so excited about Apple Silicon. For the first time since the passing of the classical high-performance RISC vendors, a non-x86 architecture enters the desktop market, this time backed by huge R&D spending and a superior process.
Is all lost for Intel? Certainly not, Intel still owns the largest part of the PC market. While the problems with their 10nm process were huge, they are not too far behind TSMC, as the 10nm Intel process isn't that much behind the TSMC 7nm as the numbers would suggest. But their moat consisting of volume and process lead has been breached, so success is no longer automatic and Intel needs to work hard going forward.
AMD and TSMC shown us that it work.
How exactly would it be better -- again, given the volumes?
-- The "design" split-off would be instantly lacking modern-node fab capacities (TSMC already said they wont scale up for Intel),
-- the manufacturing part would find itself in competition with those already having better node tech ==> not enough orders to fill the pipe (customers with comparable volumes are already all set WRT fabbing).