What Happened to Yahoo
paulgraham.com
paulgraham.com
...the kind of single-minded, almost obnoxiously elitist focus on hiring the smartest people that the big winners have had.
Is Apple a big winner, like Google, Microsoft, and Facebook (the other companies he mentions repeatedly)? Bigger. (Check the last 1, 5 or 10 years on Google finance for the GOOG or MSFT comparison. FB is not public, but...) Do they have this obnoxiously elitist focus? No. They still manage to hire great engineers, though.
So which companies need to have a hacker-centric culture?... any company that needs to have good software.
I wouldn't say that Apple is dominated by a hacker culture at all. It's dominated by Steve Jobs and a focus on design and attention to detail. Yet that seems to produce much better software than Google or Microsoft, IMHO. That's very subjective, I realize, but the market seems to agree. At any rate, Apple does produce good software.
So are the hacker culture and the elitist focus really necessary for a technology company to succeed? Is Apple a complete anomaly while Google, Microsoft, and Facebook are typical? What say you, pg?
I realize it's borderline suicidal to post an HN comment that simultaneously calls out pg and lauds Apple, but it's criminal to completely leave the best (imho) technology company out of a discussion about attributes of great technology companies.
The article is about Yahoo and in a round about way the move of internet marketing focus from 90s banner advertising to search marketing in 00s. I didn't find Apple Computers Inc. (as it was) to be missing, indeed I'm kinda at a loss where you think Apple fit in to this article.
There are already plenty of articles about Apple.
Check, but it's quite far off from being Yahoo and IMO wouldn't add to this article.
Even though they make great hardware, his position seems to be that its the software that delivers the product experience.
It's a mistake to say that. All the work that went into the LLVM/Clang, low-level development that's done OS X and iOS isn't something that you can do without hackers; neither are the algorithms used by Aperture or even iPhoto.
I got called from Google during 2006. The HR guy screened me on the phone with all TCP/IP questions. I was amazed because he knew so much about TCP/IP and UNIX socket. It turned out that he was a former software engineer in Bell Lab/Lucent. It has been the best initial phone screen that I ever have.
I can imagine a software engineer who is suffering from carpal tunnel syndrome switches jobs from writing to keyboard to talking on the phone.
One of the latest tactics that the headhunters have been using is to begin with a line like ``Hey, I used to be an engineer! But now I'm hanging out here doing recruiting, cool huh?''
Maybe I should follow JWZ's step. Selling alcohols is more lucrative than programming!
One of the assertions in the article is that a hacker culture is a necessary ingredient for a successful technology company. I see Apple as a counterexample to this argument. And it's a counterexample that shouldn't be ignored. Apple is driven by design (industrial design, product design, graphic design). They think about products, not technology. They often invent new technologies to enable their products, but that's very different from thinking primarily about technology (a mistake, I would argue, that Google is guilty of frequently).
Taking as an example a project like Google Wave, I would even go as far as to say that a strong hacker culture and an obnoxious elitist culture of hiring only the best programmers can be serious detriments to a technology company. YMMV, but, culture-wise, I'd much rather build (and work for) an Apple than a Google.
Edit in reply to your reply: I thought it was any company that needs to have good software. But even so, surely Apple doesn't pay significantly more than Google (a hacker-centric company) and surely they're in roughly the same domain (iOS/Android, iAd, etc.). I think the real problem is in your next sentence: Otherwise you can't attract good programmers to work in a suit-centric culture. That's a big false dichotomy. Sure, suit-centric technology companies will fail. And hacker-centric technology companies sometimes succeed. I think Apple shows that a culture that is neither suit-centirc nor hacker-centric is the best option of all. If many of your YC companies are trying to build great technology companies, they should take that to heart.
No it isn't. I said that you either need that or to pay people a lot or to be in an interesting industry where there are no competitors who have hacker cultures. If Apple doesn't have a hacker culture, then it would probably be an example of the latter.
Although this may not be as nice as a hacker-centric culture (if you are a hacker), it is still leagues better than an utterly non-hacker centric culture, because the devs are respected for what they do at Apple. This seems to be a critical difference necessary for producing great products.
Two essential qualities of good hackers is an abundant ability to synthesize elegant solutions from disparate details, and a desire to create great things. Jobs may not write code but he clearly has these two qualities. So even if Apple does not have an engineer-centric culture, a design culture driven by Jobs can be directly compatible. You see the differences of course, in the products Google and Apple make, but I think the point is that great hackers can work at Apple because they can get behind Jobs vision.
Contrast this with a "suit-driven" company where decisions are made by looking at statistics and market research, and executives strategize about how to maximize sales based on the superficial product understanding they have. Even worse, places where you actually have people in power who couldn't give a crap about the product as long as they have a $500k+ salary rolling in every year.
Exactly. Apple isn't really about winning or losing; they're rewriting the game. Does it really get more "hackerly" than that?
I'm not going to venture more complete definition here, but I think a lot of the themes around "unix culture" (e.g., small pieces loosely joined), open source (e.g. bazaar vs. cathedral approach to design and architecture), and even agile methodologies (e.g. "the simplest thing that could possibly work") could apply. But there's more to it than that.
From this perspective, I agree that you could easily characterize Y Combinator as "hacking" VC. I just don't see what Apple is doing as hacking.
In fact, in some ways it seems to me that Apple is doing the opposite of hacking (or at least trying to give that impression). Apple seems to want to create an impression of "intelligent design" around their products--they've got it all figured out and are delivering it to you complete, fully-formed and flawless. To my mind "hacking" connotes a more evolutionary approach: let's just try this and iterate.
This isn't a value judgement on either approach, I just have a hard time aligning Apple's behaviour as an organization with my personal understanding of "hackerly".
My girlfriend studies industrial engineering and knows quite a bit about innovation and product development. If I combine everything I've heard from her with everything I've heard about Apple, then I conclude that Apple simply excels at leveraging/executing all the well-known 'best practices'. There's nothing 'hackerly' required to do proper innovation and product development, unless you expand 'hackerly' to include 'doing business the way it should be done'
The "dictating specs to programmers" practice seems to stem from the silly notion of applying manufacturing processes to software development. In the view of companies with this notion, programmers are little more than unskilled translators - the parallel of line workers in manufacturing jargon.
In my experience, one of the worst abusers of this "square peg/round hole" paradigm are internal IT departments.
I read a study in IEEE Computer sometime in the Summer of 2001 or 2002 that showed how bad the application of Waterfall was for software development. Does anyone know of a study on the application of manufacturing processes to software development?
Very pitifully few of Google's projects have either created new markets they can dominate, or have penetrated existing markets to the point they can drive billion dollar business units. Gmail, Maps and Android exception to that point, but Google is woefully incompetent in most other respects. And the first two in that list are simply different takes on the search paradigm while Android is a knock-off of the iPhone OS (a really good knock-off, but certainly not a game changing original new idea).
Those are very big exceptions.
I've not seen that many software products which were so obviously disruptive; it's no wonder they became giant-killers. Facebook is the only other one I can think of.
On the other hand, some of their other big products have not be a huge success precisely they haven't been game-changers. Google Docs, for example, might be a good example of a web office suite; but it's hardly a game changer as far as office suites go, because it doesn't do some of the core things that people do daily with office suites. It was clear from the beginning that it would be a slow starter. Whereas Maps for example, you knew at the beiginning that nothing else could touch it
I would actually disagree with the last part of that. I think Apple is always thinking about technology, but in the context of design. The Apple difference is not simply that design comes first, technology second, but that design is the gatekeeper for technology.
The iPod is the shining example. Could Apple have made a music player before the mass-production of the 1.8" hard drive? Sure, but it wouldn't have been the wonder to hold in your hand that the iPod was. It was when the technology was ready for the design requirements that they pounced on the opportunity.
Same thing with the iPhone. Apple could have made a smartphone before affordable capacitive multi-touch screens, multi-gigabyte flash drives, and fast-enough-for-desktop-browsing SoCs, but they likely didn't feel they could make a user experience to their standard with anything less.
How can you be so sure about their thought process? Maybe it is the other way around, they just turn it around when it comes to presenting it to the public.
Google's products are mostly software running on servers in datacenters. Facebooks products are the same way. Yahoo's products were mostly software.
Apple's products are not mostly software. Apple sells consumer devices run by software. It's an important distinction, and looking at it that way lets you compare with different sorts of companies, like the technology startup Emotiv Systems[1][2]. They're developing consumer devices that read brain activity like an EEG. They likely spend a lot of time developing their SDK and other software, but they also have to focus on the physical product, like Apple does.
Sometimes people forget or don't realize that "hacker culture" has always had a contingent of hardware guys building solar-cars and robots[3] and kites with cameras.
1. Secrecy is part of the culture.
2. Engineers work hard at Apple, and many teams are surprisingly small. There are no 20% projects at Apple because nobody has the time.
EDIT: Going on the Apple jobs listing for Cupertino, there is plenty of openings for engineers.
How long did their core OS not have protected memory, or pre-emptive multitasking?
Apple still see's their OS as a package with the hardware. I would think this is part of their attachment to iOS (that and the massive markup and market penetration they've had so far).
The iPod gen 1 through gen 4 was not a hacker platform.
" Apple did not develop the iPod software entirely in-house, instead using PortalPlayer's reference platform based on two ARM cores. The platform had rudimentary software running on a commercial microkernel embedded operating system. PortalPlayer had previously been working on an IBM-branded MP3 player with Bluetooth headphones.[4] Apple contracted another company, Pixo, to help design and implement the user interface under the direct supervision of Steve Jobs.[4] As development progressed, Apple continued to refine the software's look and feel. " -- Wikipedia
http://en.wikipedia.org/wiki/Apple_Inc.#1986.E2.80.931993:_R... http://en.wikipedia.org/wiki/IPod
Edit: I could go on longer, for example, OSX coming from BSD and NextStep etc but I'm done. :)
About until Steve Jobs came back. But the Apple of 1985 to 1997 was not the Apple of today. That Apple is long dead. The Apple of today is basically NeXT, plus the branding and whatever other assets NeXT could salvage from Apple. As far as that company is concerned, their core OS had protected memory and preemptive multitasking since 1989.
The Mughal India of the 16th–19th centuries was not the India of today. That India is long dead. The India of today is basically the British Raj, plus the branding, mix of religions and whatever else the British could salvage.
The Republican party of 1860–1970 was not the Republican party of today. That GOP is long dead. The GOP of today is basically the unholy mix of Kristol & Podhoretz foreign & economic policy with Jerry Falwell & Pat Robertson social policy, plus the branding and whatever panders to fiscal conservative constituents they could salvage from the party’s historic base of support.
The field of economics of 1780–1950 was not the economics of today. That economics is long dead. The economics of today is basically the Friedmanite Chicago school, plus the branding and whatever other assets they could salvage from economics.
Now, in all three of these cases, there’s some little bit of truth behind the wholly oversimplified (to the point of wrongness) analysis. But taking these statements at face value obscures more than it reveals, and reflects a poor understanding of how cultural assimilation works in any institution or society. (And denies evolutionary change, as well; Apple today is not exactly the same as Apple of ca. 2000)
* * *
In short: you’d better have some better analysis to back yourself up here, because as it stands I think your claim is, to the first approximation, false, even if there is a tiny bit of truth to it.
It's certainly true that Apple is now more NeXT than it is what it was. Perhaps it's something different enough from both of those--but it followed NeXT's evolutionary path more than Apple's.
You’re right that early versions of OS X looked much more like NeXTStep than they did like previous versions of Mac OS (being built on NeXT’s codebase), and still today is in many small details a regression. But as OS X has matured, it has felt ever more like "classic" Mac OS, not less. I think if NeXT had been independently successful and turned NeXTStep into a consumer OS, it would have evolved in a dramatically different direction (though of course that’s counterfactual).
I don’t really know what it was or is like to work inside these companies, but I doubt that the feel of NeXT could be quite like the feel of the modern Apple, if only because modern Apple is so much bigger.
But as OS X has matured, it has felt ever more like "classic" Mac OS, not less.
Past the transition from public beta to release and maybe a couple small points after that, I don't see how this is true at all, and I've extensively used both. What similarities do you notice?
(It’s not a change since OS X 10.0, but....) Right from the beginning having the menu bar at the top was better than the NeXT per-window menu bars (there were legal reasons for the NeXTStep behavior, sure, but a menu at the top feels so Mac-ish to me, that that difference alone is huge). Much of the conventions for menu organization/structure and so on (and definitely conventions for keyboard shortcuts, for the overall keyboard layout, &c. &c.) is much more Mac-alike than NeXT-alike.
Hardware finally caught up with the software (and the software was optimized better for performance), so that the interface is as “snappy” as OS 9 used to be on then-current hardware. (Early OS X felt terribly sluggish by comparison.)
The mouse acceleration curve reverted to something much more like the classic Mac version (in 10.3 maybe?).
The menus at some point (10.1? 10.2?) returned to the proper behavior of allowing diagonal selection. The alternative is a fundamentally broken interface, and every time I use an operating system that does it wrong I want to stab the guy responsible.
Finally in 10.4, the corners of the screen properly activated the Apple/Spotlight menus (taking advantage of “Fitt’s Law”).
Drag & drop support improved and was properly widely adopted.
AppleScript support by first-party apps got much better.
Dialog boxes got better writing, including somewhat more consistent use of verb buttons.
Did ⌘. (command-period) work everywhere (anywhere?) for cancel in dialogs in early versions of OS X? I don’t remember.
OS X print dialogs still aren’t my favorite, but they’ve improved relative to 10.0 print dialogs, adopting some of the feel of earlier dialogs.
Save dialogs improved a lot since 10.0. I don’t think ⌘D properly went to the desktop in early OS X versions, but I’m not positive about that.
Places where (in NeXT tradition) early OS X used “inspectors”, for instance in the Get Info window, later OS X reverted to the better "classic" Mac OS behavior. (This happened in various first-party apps too, but I’m not thinking of examples off-hand; it’s been a while).
NeXT-ish UI elements like drawers have been gradually phased out, in favor of UI elements which seem more spiritually Mac-ish. Oh, and remember early OS X Cocoa tab view widgets? Wow, were those dumb. Little wonder they’re pretty much gone.
I think support for Mac file aliases was sorta crappy in early OS X versions, but I don’t remember the details.
Cocoa applications started properly dealing with references to files rather than to path locations, for things like re-opening saved files, etc. etc.; early versions of OS X were filled with apps which would behave improperly when a user moved a file.
Exposé feels much more like a culturally Mac-like feature than the use of the Dock, etc. Though Spaces seems more NeXT-ish to me (I don’t remember if NeXTStep had multiple desktops... did it?).
The appearance of OS chrome tended back towards the simplicity of classic Mac OS (obviously not to the precise appearance, but to the feel), with every OS version further away from pinstripes, distracting transparency, slow unnecessarily flashy animations, etc. And not only the overall window ornamentation. Take a look at the folder icons, for example, which look like OS 7.5 icons (I always hated the ones from OS 8–9).
The Finder never regained its spatial-ity (as John Siracusa periodically laments), but it got much better at remembering window state in each version from 10.0 to 10.3.
Disks get properly mounted on the desktop by default (I don’t remember precisely the history of this, but I believe the default was not to show them on the desktop at some point; maybe I’m just confusing early OS X with Windows here).
The Finder at some point (10.4? 10.5?) finally allowed icons to be spaced closer together as in "classic" Mac OS.
The Finder got back color labels in 10.3, an OS 7 feature, which I believe were missing altogether from 10.0–10.2.
Finder keyboard and mouse selection behavior, especially in list mode, improved dramatically from 10.0 to 10.4. In general, the mouse and keyboard behavior of list widgets – rows, disclosure triangles, heading sorting and column resizing, etc. – improved quite a bit.
* * *
I’m sure there are dozens of other such things (if you read through Siracusa reviews and John Gruber’s notes about each OS version I’m sure you’ll find a bundle), but listing more here doesn’t accomplish much. I’m not trying to argue that "classic" Mac OS feels the same as the current OS X does, or that OS X doesn’t also include many very NeXT-feely features like the Finder column view, &c. I do think though that OS X is somewhere between a NeXTStep/Mac hybrid and its own thing altogether, both in specific features and in less tangible “how it feels” aspects.
Basically, they merged the branch back into the trunk.
It's impossible to tell without working there (and once you work there you probably lose the privilege of speaking openly about this stuff) but it seems Apple is just a generalized, multidisciplinary example of what pg's talking about.
"...but they didn't have the kind of single-minded, almost obnoxiously elitist focus on hiring..."
Apple seems to want to hire the best, but the best being based on actual capability instead of the school you went to.
Steve Jobs is the conductor and Apple is the orchestra. Jobs can bring all the right people into a room, sell them his idea, and have them rabidly strive to produce.
The real difference here is the management's competence.
You sure about that?
From http://www.apple.com/jobs/us/welcome.html:
"Whether you want to be an iPod engineer, a Concierge at an Apple Retail Store, or a sushi chef on an Apple campus (yes, we have sushi chefs), we’re looking for the best. People who are smart, creative, up for any challenge, and incredibly excited about what they do. In other words, Apple people. You know, the kind of people you’d want to hang around with anyway."
Googling "apple hiring process" turns up this:
http://www.ifoapplestore.com/stores/hiring.html
"And lastly, the competition for jobs is intense! Sr. v-p Retail Ron Johnson has said it's tougher to be hired than being accepted by Stanford University--literally. During one talk he said Apple hired 978 store employees during 2002 from an applicant pool of 16,438, or less than a 6% chance of being hired (Stanford accepted 12.6 % for 2005). Put another way, only 1 in 17 applicants is hired."
Sure that's just retail, but it seems reasonable to guess that a similar hiring process exists for other fields.
As it should be since for the most part, Apple has been conspicuously absent as an internet and web company, especially during the period discussed in the article. I have many reservations about the conclusions and generalizations PG draws, but it seems silly to fault him for not including Apple. He didn't mention General Electric either.
Apple's software always runs on controlled environments. I wonder if they can pay that much attention to detail if they build for Web (Google) or any random hardware (Microsoft). Apple somehow avoids the added complexity of supporting the majority. For now it works, though.
Next both Google and Yahoo brought in professional executives or "suits" to run the firms, speak to Wall St. etc. Eric Schmidt is a professional CEO, not a hacker. He was at Sun, then he tried to revive Novell and then was tapped for Google.Yahoo is also run by a professional CEO now. How does that compare to the single minded life long vision of Steve Jobs? What role the leadership and their personalities plays in running the companies beyond the culture? Remember Apple was a "has been" company with plenty of suits and they even fired Steve Jobs at some point.
There is a fundamental flaw in PG's argument, he is trying to have a rational explanation about an irrational path to success of failure.
Lex − A Lexical Analyzer Generator
http://staff.ustc.edu.cn/~shzhang/staff/lesk-lex.pdf
and Google's Eric E Schmidt was only 20 at that time according to Wikipedia: http://en.wikipedia.org/wiki/Eric_E._Schmidt
He got BSEE on Princeton in 1976. OK, both Bell Labs and Princeton are in New Jersey, but it's still quite impressive to be the developer of lex in Bell Labs as a 20-year old Princeton student.
Of course even more impressive is being 129th richest person in the world in 2006.
I don't know much more about the two men, but I'm 100% certain that Schmidt did a colossal amount of programming in his day, and I really doubt that Semel ever did, though I can't be sure of this. This has to have a serious effect on the hacker culture. If nothing else, Google's choice sends a message to programmers that technical ability matters all the way to the top. Yahoo's says it doesn't, at least not enough to be a required attribute of the highest ranking person at the company.
Of course, if you want to extend the idea of hacking so that one can speak of user experience hackers, then Apple would definitely qualify as having a "hacker-centric culture".
Not really. Please do not confuse Apple's successful products with anything else. Microsoft still delivers on software products in a way (admittedly a hardware company)Apple never could. Oh and as for comparisons with Google, thats even more ludicrous.
I'm sorry, this may be PG's backyard, but I've gotta call bullshit. Note: I am not a Yahoo employee or corporate shill.
1. "At Yahoo this death spiral started early."
Apparently this death spiral took TWELVE YEARS (the WWW is 20 years old), and Yahoo is still #1 (or #2) in terms of total properties on the web. How is this a death spiral?
2. "If there was ever a time when Yahoo was a Google-style talent magnet, it was over by the time I got there in 1998."
Are you kidding me? You're implying all the current Yahoo engineers are idiots? Then how come each piece of code they release is revered, lauded and used by everybody?
Speaking of talent: Yahoo got best paper at EVERY data-related conference in 2010. No single lab(Google,MS) or university(Stanford etc) has ever done this. Clearly you can achieve this without talent, pg!
3. "In the software business, you can't afford not to have a hacker-centric culture. Probably the most impressive commitment I've heard to having a hacker-centric culture came from Mark Zuckerberg, when he spoke at Startup School in 2007."
Which company is responsible for inventing the term "Hack Day": Yahoo. Which is the only company to _promote_ hacking culture to students? Yahoo. Yahoo != Hacker-friendly? WTFFF!!! (edit:goog does soc, but it doesnt have engineers travel around the world teaching kids hands-on like yahoo does)
4. Google and Apple are shiny and get all the attention. Yahoo's the internet's underbelly. It's audience may not be SF hipsters, but they exist and there are a lot of them. Just because they're not like you doesn't mean they dont exist. Just because Yahoo doesnt cater to you doesn't mean it doesn't exist.
5. pg's product got hosed after inclusion into Yahoo. This happens. Reading a Yahoo critique from pg is like reading a Google critique from Evan Williams (Blogger), or a Google critique from the Dodgeball guys. It's a biased opinion.
6. It's a miracle Yahoo has survived for so long despite being criticized by the media /market for more than half its lifetime. The only other company that has had it worse was Apple, when they fired Steve Jobs. Yahoo doesnt have a Steve Jobs, and has been managed badly, but the engineering is SO GOOD that it STILL does a good job overall.
How is this a metric for success?
Also, Viaweb didn't get hosed after being bought by Yahoo. It prospered enormously. I don't bear them any ill will.
What other Yahoo web properties exhibit this?
I was disappointed you didn't mention Mark Cuban (Yahoo acquired his company, Broadcast.com, for $5.9 billion in Yahoo stock in '99).
Fairly expensive domain redirect.
Side note: I have many memories of creating RTML templates, while coveting Trevor Blackwell's DIY segway (the 2nd version).
In fact, taken literally, it is devoid of meaning, because any group of human beings can be described as "uneven" for any descriptor one can think of.
But it is not used literally in the article; it is used as a euphemism for "bad", as opposed to the quality of Google engineers (which one guesses is "even"...?)
Now you're playing with words; assume what you mean.
I do remember being in the snack room at Brickhouse and seeing two posters along the lines of "watch your competitors." One was Google and the other was Microsoft. I thought those posters were misguided because nothing that I admired about Yahoo included anything that Microsoft and Google did well.
Apparently this death spiral took TWELVE YEARS (the WWW is 20 years old), and Yahoo is still #1 (or #2) in terms of total properties on the web. How is this a death spiral?"
By this reasoning MS still has the best OS (and OS engineers) in the world and everything is fine and dandy at MS. They still have 90 % of OS marketshare and Simon Peyton Jones works at MS Research and he works on Haskell and writes awesome papers. How could Microsoft be "dead" as PG claims in another essay ?
The key to this puzzle is that PG uses "death" in a figurative, not literal sense, to mean irrelevance, especially a fade to irrelevance from a previously dominant, direction setting, position not complete physical extinction.
If you prefer, read it as "irrelevance spiral" and you would at least be reacting to what he intended to say vs what you understood by some overly literal reading. "At Yahoo this irrelevance spiral started early" can still be rebutted, but it needs a serious argument to do so.
As he said in his "Microsoft is dead" essay.
"I already know what the reaction to this essay will be. Half the readers will say that Microsoft is still an enormously profitable company, and that I should be more careful about drawing conclusions based on what a few people think in our insular little "Web 2.0" bubble. The other half, the younger half, will complain that this is old news."
Replace Microsoft with Yahoo and this is what you are saying. MS has great market share, lots of smart researchers and many conferences for programmers using its technology and so on. All that doesn't avoid the fact that MS is a shadow of what it used to be in the nineties in terms of technological dominance.
Some of your other points have some merit (but their impact is lessened by your tone imho).
I am all for strong debate and argument (it enriches us all) and PG shouldn't be allowed to get away with sloppy thinking, but you have to show that the thinking is sloppy. And the first step in rebutting an argument is to understand what was really said.
As for whether Yahoo is really hacker friendly, the best way to find this out is to talk to some hackers who work (or have worked) at Yahoo. I am sure there are many people om HN who have worked/did work there. (strlen for example)
From his comment
"Much of the knowledge did diffuse to rest of the organization and it's important to note that along side the "dot-com-wannabe-millionaire-vesting-in-peace" crowd the essay describes, there was plenty of genuine technical talent (anyone who has worked there can attest to presence and influence of hackers).
Unfortunately, for many reasons, that talent has slowly bled out. There was also a great amount of friction between Search and rest of the organization e.g., resistance (by Search engineers) to dog-fooding of what was often inferior technology. Search was, however, able to maintain a different platform, remaining on Linux, with their own platform/software stack, but were some "wtf" moments like porting and forcing the adoption of a custom user space locking library from FreeBSD when Linux already had futex (on the other hand, the fact there was a custom locking library built in a "media company" does say something).
The other key mistake made is that they would position bright, capable new hires in areas that weren't directly correlated to revenue and treated as cost centers."
This sounds like a company making mistakes with respect to technology.
(EDIT: See also tailrecursion's comment on his experience when he cae to Yahoo from Inktomi http://news.ycombinator.com/item?id=1596797)
(These are just anecdotes and not made to argue a point)I have friends on the Yahoo search team and from what they tell me, they did get crushed in terms of search quality by Google and though they caught up later (though not completely), they could never recover the lost ground in marketshare.
I know someone who argued at an internal "strategy conference" that Yahoo needed to close the technology gap with Google or they would lose Search to Google and was completely ignored (He is still at Yahoo btw, though he switched to being a full time manager as a matter of survival).
re:interpretation of death, sure, let's call it irrelevance. OR let's even call it 'random variable x". The point is that pg asserts the company has been in "x" for 12 years. Which is 60% of the WWW's lifetime. It may be a decline, but I would not call it a spiral. at least of death.
re:research, I agree about orthogonality of research and company. This was addressing pg's point about Yahoo being a "talent-magnet" -- they indeed are a talent magnet in this domain.
engineers/hackers don't usually focus on doing academic research and writing academic papers targeted to top-tier academic conferences. Yahoo Research is certainly a leader in the academic world, but that might not translate directly to the quality of Yahoo developers.
conversely, take hot start-up X of the moment who is known for having awesome 1337 hackers. how many 'best paper awards' have they probably won in top-tier conferences? probably ZERO, since they're not focused on doing academic research.
I am in total agreement with this. These days anything targeted towards top-tier conferences typically has a sales-marketing slant wrapped around it.
I do feel Y! did a great job -- and made a great decision -- in integrating its search acquisitions; but those acquisitions threw into sharp relief what Y! really was. It's hard to get into too much trouble when all you're doing is throwing horoscopes onto web pages; but Yahoo! never struck me as being serious about software, and as Paul says, they didn't seem to think that was a problem.
Somewhere in 2005-2006, Y! made an attempt to revamp its search ad services with a large-scale software project called Panama. It was going to do a lot of things, including ad ranking refinements to improve revenue per query, and allowing advertisers to create more complex ad campaigns. This project seemed to undergo huge troubles, and I wonder if it was ever completed in any real sense, because I heard so little about it, and everything I heard suggested disaster.
Honestly I don't know if the fruits of Panama -- such as improved auctions for ads on search result pages, which is a critical thing -- were ever deployed. But a couple years later they made the Microsoft deal. Over the last couple years, select people at Y! have moved on to Google, and many many others are now at Microsoft.
I also wanted to ask, am I the only one who thinks it would be completely awesome to work on a project called Panama, simply because you'd have an excuse to have a Van Halen song as your theme? "Pa-na-ma-ha... Panama!"
Yahoo did do this through the Inktomi (Yahoo Search, 2003-2010) acquisition. I've worked in that organization and learned a tremendous amount from the people. At that time the hiring bar had been considerably raised from what Graham describes (I have no idea whether that was the rule or the exception at that time, I was there 8 years later), nobody was shy of rejecting weak candidates. Much of the knowledge did diffuse to rest of the organization and it's important to note that along side the "dot-com-wannabe-millionaire-vesting-in-peace" crowd the essay describes, there was plenty of genuine technical talent (anyone who has worked there can attest to presence and influence of hackers).
Unfortunately, for many reasons, that talent has slowly bled out. There was also a great amount of friction between Search and rest of the organization e.g., resistance (by Search engineers) to dog-fooding of what was often inferior technology. Search was, however, able to maintain a different platform, remaining on Linux, with their own platform/software stack, but were some "wtf" moments like porting and forcing the adoption of a custom user space locking library from FreeBSD when Linux already had futex (on the other hand, the fact there was a custom locking library built in a "media company" does say something).
The other key mistake made is that they would position bright, capable new hires in areas that weren't directly correlated to revenue and treated as cost centers. That just seemed highly illogical to me. It's a mistake that's often difficult to fix: once young graduates are used to working on glamorous projects, it's much harder to get them to work on more mundane, but revenue critical projects such as advertiser systems. Google's strategy of a uniformly high hiring bar (vs. some teams hiring people other teams rejected for roles involving the same skill-set) and assigning hires to business priority projects (while allowing individuals to transfer if the project wasn't their cup of tea, with proper incentives in place for some projects) seems to be (at least from an outside point of view, I've never worked at Google) better.
This reverts back to the original article. Poor strategic leadership.
In any conflict, there is always a risk of fighting the last battle. I see today many companies, inspired by the lessons learned from the first dot-com bubble, making the opposite mistake; thinking that they need to be a technology company when they're really a experience providing company.
Twitter is the refutation to pg's thesis. Despite their initial technical incompetence, they managed to attract a top-notch technical team because they managed to deliver a crawl-over-broken-glass experience (it's so compelling people are willing to crawl over broken glass to experience it, see also: craigslist, plentyoffish).
It was interesting attending the YC work at a startup day and seeing RethinkDB present. Their pitch, in not as many words was basically "Look, we're the only company here that is actually working on a problem that demands world-class engineers (in the non-debased sense of the term)". For most of the other startups, despite all their bluster, the technology platform they were using was commoditized and technology was not their differentiating factor.
First, there's a lot of false info out there about what parts of Twitter were built using which technologies and why it had uptime issues.
Second, their issue was massive scalability. A high-class problem.
That a lot of other companies nevertheless solved far more successfully.
Disclaimer #2: This is my own completely unsanctioned opinion and does not necessarily represent Yahoo!'s views or opinions.
I'm obviously biased here, but I have a great respect for HN and PG so I wanted to talk about this.
PG accuses Yahoo! of two things right at the very top of the article and the rest of the article is more details on those things. They are a) access to easy money removed Yahoo!'s desire to find the next big thing b) ambivalence about being a technology company.
I'm not going to substantially repute a) because I think it is true. However, I think it's a pretty big ask to expect someone being successful to see past the current success to the next big win. PG admit's neither he nor Larry and Sergey really understood how big search was. As such I feel it's a slightly ad-homonym attack to blame Yahoo! for not being Google. We didn't win the big prize, but we also aren't AOL, Lycos or Ask.com either. Just because we didn't become Google doesn't mean Yahoo! has failed.
As an engineer though PG's second point strikes me as deeply unfair. I joined Yahoo! in 2004 in the UK. The team I joined was exceptional. For a team of 30 people there were about 15 book authored by the team. I personally had written W3C standards and multi Web Standards Project founders littered the team. The rest of the team were coders, much better than I.
Now I work in US for the developer network and I get to see the vast range of technologies that Yahoo! does from Y! Research through to the engineering teams. There is a reason 'Hack Days' started here, because the engineers here are passionate about those technologies and playing with them.
I'm also on the Yahoo! Open Source Working Group and I see all the open source we put out of the door. We should probably do a better job of telling people how much, but it's a lot more than you think. I find it troubling that PG who left the company 10 years ago, can now, accuse me and my colleagues of apathy about technology when it's so fundamental to what Yahoo! does.
I think what PG is really talking to is two things. Firstly, it's important for a company to have a 'core competency'. Describing ourselves as a media company makes it clear to our investors and our employees the field of play. Secondly, the ease at which a cash-cow, such as Google's search war chest, allows them to make their presence felt in the technology community reminds people about all the things they touch. I am both proud an envious that Google have the freedom to do that, but again, it doesn't mean Yahoo! failed because we don't.
What about Node.js? That's pretty en vogue right now. We are currently hiring Node.js engineers and have a number of core contributors in various parts of the company.
YUI and our other front-end work at Yahoo! is world class. People like Crockford are considered leading experts in doing quality front-end web engineering.
Our data centers are also cutting edge, I'm not an expert in this area, but I'm told we are at least on par with Google. Some of the new project are completely cooler-free, the whole building being designed to control heat and power flow.
http://trafficserver.apache.org/
http://developer.yahoo.com/yql/
http://github.com/yahoo
http://yuilibrary.com
http://developer.yahoo.com/yap/guide/caja-support.html
http://developer.yahoo.com/yslow/
http://developer.yahoo.com/search/boss/
http://developer.yahoo.com/hadoop/As an example, I was at Yahoo for 5 years (2003-2008) working in the Search Marketing division. Tons of data, tons of optimization possibilities. Could we use the Hadoop cluster? No, for some foolish internal accounting reasons.
I have trouble defining "hacker culture", but surely it's about more than just the engineers working at a company.
If Yahoo's engineers work on these great projects, but no one hears about them - then it's probably because the culture at Yahoo isn't "hacker-ish" enough to motivate marketing to drive awareness for the projects.
Yes, as a "Yahoo Evangelist", you are probably close to 100% biased.
> We didn't win the big prize, but we also aren't AOL, Lycos or Ask.com either. Just because we didn't become Google doesn't mean Yahoo! has failed.
Here's the Signature yahoo denial of reality syndrome. It's quiet similar to the surreal discussions you can casually have with YSearch people telling you that Google Search is an inferior product (or ymail beats gmail, or 360 will own facebook, ..). Basically trying to convince themselves rather than anyone else.
Listen, this is just like alcoholism, if you can't admit that you have a problem, you will not go anywhere. Yahoo has been steadily losing on most fronts for more than 5 years now and it is currently in the mist of a huge identity crisis. Multiple metrics can attest an AOLization process is going on! You're right by the way, Yahoo has not failed yet, neither did AOL, they're still around, just totally irrelevant ..
> As an engineer though PG's second point strikes me as deeply unfair. I joined Yahoo! in 2004 in the UK. The team I joined was exceptional. For a team of 30 people there were about 15 book authored by the team. I personally had written W3C standards and multi Web Standards Project founders littered the team. The rest of the team were coders, much better than I.
Beyond the lines aimed solely to your own glory. Does writing a book means being talented to you?
No more seriously, I'll pass on the rest of your post, but basically, I'm under the impression that you might be looking at things through the prism of your rather different role in that company. That sounds pretty cool to fly the world and talk about the new hot piece of frontend code written by yahoo's javascript mafia, but please realize that somewhere in Santa Clara, random Joe engineer might see Paul's analysis as an opportunity for Yahoo to wake up, acknowledge the problem and rethink itself.
PS: please, no "ad-homonym" attack!
http://www.informationweek.com/news/global-cio/outsourcing/s...
The primary reason for the cancellation of the contract was the sheer amount of beauracracy IBM imposed. Their modus-operandi was - pick up staff who were working in "non-critical" areas and make them report to IBM managers. So now all-of a sudden if i need to get something from our DBA i can't walk upto him and ask him , i need to raise a request in a god-awful system which goes to his manager who would be sitting in a remote location in a different time-zone "managing" the work-load of the DBA who sits 5 feet away from me. In the majority of the cases for access to sysadmins and DBA's who were an integral part of your team before you now had to go through this route.
I will leave the readers to draw their own conclusions to this marvellous scheme.
EDIT: Most importantly they had just witnessed the spectacle of Netscape, a company that seemed unstoppable, just completely buried by MS.
Sounds like yet another variant of Upton Sinclair's Law: "When a man's paycheck depends on his not understanding something, you can depend upon his not understanding it."
"It is difficult to get a man to understand something, when his salary depends upon his not understanding it."
A terrific read. http://www.amazon.com/dp/0520081986
Rich Google in 2004 overlooked the value of social networking (both their in-house hit Orkut and outside services).
It is, however, true that Facebook is unlike Google in that, while nearly everyone wants to find information, not everyone wants to share information with other people.
I won't tell my coworkers or boss the same things I tell my friends, I won't tell my friends what I tell my close friends, and my family sits on the highest tier.
Maybe facebook will get this maybe it won't, or in general maybe it is impossible on the web where information tends to be so open.
If you want privacy, don't share private information on Facebook. If you want to remain ignorant of your favorite sports team's latest scores or the most disgusting new genre of porn or how some movie you haven't seen yet ends, don't look them up on Google. The "next big thing" that Facebook is going to miss is not something people can already get by simply restricting their usage of Facebook, it'll be on a different axis entirely.
this isn't to say that i don't ever want to say something to my friends but not family, or vice-versa. more often my issues with the "audience" for my facebook posts have to do with various groups interest in the subject of the post - people i went to high school with mostly don't care about python, or some great restaurant out here across the country.
No is not. Privacy is about managing social networking, the ability to connect to those I want as I want,not someone else telling me how should I do that.
As you get more social, you get more people telling you private things.
Companies are growing to the multi-billion dollar mark faster than ever. That trend will only continue as technological advances continue. What will happen in the near future as SSDs get better, larger, and cheaper? What will happen as multi-core CPUs continue their progression? As always, single newer machines become capable of replacing multiple older machines. At what point in the future does the equivalent storage, processing, and transactional volume capabilities of an entire datacenter of today fit into a single server room tomorrow? A single rack? A single 1U server? A single VPS instance?
In the not too distance future the ability to host applications like facebook that serve hundreds of millions or even billions of users will be just a click away and cheap enough to fall under the "ramen profitability" umbrella. Giant, billion user apps will be the viral videos of tomorrow. Companies will grow to enormous sizes not in a matter of years or months but in a matter of days, or hours.
What happens when your billion dollar, thousand dev. corporation has serious competition in the form of a 5 person startup?
Instead they decided to encourage the kind of spam that generates page impressions that sell ads like a media company... Lucky they're full of clever people, I guess.
I was a kid too when I used Altavista, it was a great search tool back in the day... until they fckd it totally.
I had seen it happen with other services like ftpsearch(you know the name of a file, you got it instantly).
When I saw the first Ad blinking on my face(Altavista put suits on place, "monetizing stuff") I discovered reading on searchlores "Google": It was so easy and powerful and NO annoying ads!!
I couldn't believe it, and expected it to surrender to the MBAs forces soon as the other had. But they got money to buy some freedom.
Half the reason I eventually switched to Google was because of their anti-portal philosophy.
I think Yahoo made the correct decision on both counts. As far as depending on a bogus source of revenue, all new advertising mediums are massively undervalued at first, then massively overvalued, and then only eventually even out. It would be a huge mistake not to plan your entire business model around this cycle. Yahoo did this, and they made literally billions of dollars because of it. Sure, if they had chosen to go into a different sort of business they could have perhaps extended their reign, but it's hard to argue that they made the wrong choice.
As for not having a hacker-centric culture, I think they made the right choice here as well. Every time Yahoo bought a startup, their stock rose more than they paid for the company, meaning every time they acquired their tech through a buyout the execs took home massive paychecks. Whereas every dollar they spent on elite programmers was money out of their pocket.
I think the lesson here, if anything, is that you can make mostly the right decisions and your business can still decline over time.
Incidentally, I heard a great story about one of the Yahoo sales guys. Apparently he bought his own inventory at the end of every month to meet his quota, because the stock he was getting vested was worth more than the 100K or so he was required to fork over each month. (Because he obviously couldn't actually make any sales.)
Interesting statement. I suspect this process is well under way at Google these days.
But why does this happen? Is it just that once companies grow large enough, you need middle managers, (who then consolidate power and control)? Is there a more nuanced explanation?
What's your basis for this speculation?
What's your basis for this speculation?"
Conversations with friends who work there mostly. One of the advantages of being older is that you know people in most companies who talk to you off the record over beer, sometimes in very "high" positions.
They act like a secret club. They speak the same language, went to the same schools, have the same friends, hang around each other, know each other better than people not in the club. They tend to hire and promote other that are similar to them, more than simply basing it on merit.
I'm far along enough in my career that I can say that I think what's happening is that it's very very hard to hire somebody through the old "post job, take tons of resumes, filter, interview" method because there's so much crap, and so much noise, and so much risk of getting false positives and false negatives, etc., that eventually you realize it is much much easier and more reliable to hire people you've already worked with at a previous company in the past. You can vouch for them as a known quantity.
Ideally, a bigger proportion of hiring would happen like this. But there's always new folks entering the field who won't have somebody to vouch for them, plus, not everybody who you would vouch for is available or interested, so you sometimes to have to go the posting-resume-interview route and cross your fingers.
So it's almost the opposite of what you said, though with the same goal! You should want to hire based solely on merit -- but if you have to fill a job opening and your choice comes down to filling it with a total stranger whose resume claims he has 20 years of experience in "J2EE/JBOSS/Jscript" versus filling it with a guy you know firsthand had delivered a major threading-related fix to a previous employer's Java codebase the wiser and more meritocratic choice is probably to go with the second guy. And it's not because he knows a secret handshake. :)
Everything slows down, gets miscommunicated and politicked, and divergent goals appear. The company is no longer about the product, it's about managing the process that produces product.
When process and hires are set up to maximize an offer, over the years the company gets fossilized "doing things this way".
It happens on every industry. You find some awesome source of revenue, you build upon it, you get blind to other options.
http://steveblank.com/2009/10/01/durant-versus-sloan-part-1/
Basically a business get specialized in doing something, managers are the best organizing stable stuff, and making money(milking the cash cow). But over time the business change and this people don't like change.
>> By 1998, Yahoo was the beneficiary of a de facto pyramid scheme. Investors were excited about the Internet. One reason they were excited was Yahoo's revenue growth. So they invested in new Internet startups. The startups then used the money to buy ads on Yahoo to get traffic. Which caused yet more revenue growth for Yahoo, and further convinced investors the Internet was worth investing in. When I realized this one day, sitting in my cubicle, I jumped up like Archimedes in his bathtub, except instead of "Eureka!" I was shouting "Sell!"
* The Innovator's Dilemma, and
* Mistaking a key driver of innovation and growth (software development) for a cost center.
"I'm here to generate sales for the company, and whether that means we have good products, bad products, or indifferent products," he says, "I have a job to do -- meeting our plans from a top- line standpoint."
It was a great idea, original and effective, though I'm not how to feel about their patenting it. If they hadn't, they would have been copied and crushed without credit. http://en.wikipedia.org/wiki/Goto.com#Origins_of_Goto.com
As it was, they got 2.7 million shares from Google's infringement, and later acquired by Yahoo for $1.63 billion.
No offense, but hiring programmers for HR seems like a terrible idea. HR people's main capacity is to understand people and relationships, not exactly something programmers are known for being good at.
Not to be offensive, but I think a little bit of respect for other lines of work would be a good thing here.
* Processes, procedures and forms
* 401K, health insurance and other benefit programs
* Buying salary surveys
* Finding template bonus programs, performance evaluation programs and hire-a-friend programs. Adjusting them a bit to fit current organization.
* Employment and immigration laws
I've never found them to particularly understand people and relationships. I've worked for 4 years for a corporation where I never once managed to talk to an HR person. Everything was done through support tickets. But even in smaller companies, I found that HR folks are more of systems-people than people-people.
the point remains the same though. I don't think hackers would be very good at processes, procedures and forms.
Hackers create, this is administration.
If those in Denmark are people people, good for them!
Perhaps this is a hacker-culture reaction to that kind of thinking?
Sounds a bit like writing FooInterface and FooInterfaceImpls all day, to me.
I was born in 1981 and I clearly remember the fear and awe Microsoft inspired in 1995 :) .
There were articles about how Bill took things personally and crushed competition.
I read the 1992 biography "Hard Drive" ( http://www.amazon.com/Hard-Drive-Making-Microsoft-Empire/dp/... ); Bill was kind of a role model back then (a successful and well-known technologist).
Edit: See also ojbyrne's comment: http://news.ycombinator.com/item?id=1596737
At any rate, I (think) I remember waiting forever for each individual and unappealing icon to load at this store. It's completely possible that the store was a Yahoo Store knock-off, or that it was before the optimization you mention. I guess we'll never know, since I've probably cleared my cache since then.
Each division within Yahoo really had its own culture. The search group culture was entirely different from the media (news, sports, etc.) group which was entirely different from the listings (real-estate, etc.) business.
A group like Yahoo Search (prior to its sale) was an entirely tech centric organization. Many of the top engineers from Yahoo Search Yahoo! are now co-founders within top Silicon Valley startups.
1. It has no notable engineering foundation;
2. It has no clear vision about what it is and where it's going; and
3. Like Microsoft, it's put a business person in charge, which is the death knell of any tech company. You need someone with a technical foundation or a product person, not a business wonk.
I'm going to defer to mixmax's quip about putting a gardner in charge of a gardening firm's finances [1] as the alternative.
You can't argue Jobs or Schmidt are not business people. Although Schmidt could be used to either refute or backup your point depending on how you feel about Google these days.
The same can't be said for Carol Bartz or Steve Ballmer. Those are the people you want in the room when you're organizing an M&A deal, not when creating consumer oriented products.
Speaking as a veteran of a .com that had $30MM of venture money literally handed to them in 1999 for a concept. No customers, no proven ability to implement the concept, no management that had ever attempted anything like it, etc. The money was gone in a little over a year, frittered away on lavish offices, a huge support organization (for no customers!), massively overbuilt technology infrastructure using the most expensive Sun servers, Cisco networking gear, triple-redundant everything.
Very few tech companies, customers, or investors in that era had any real idea of what they were doing.
I disagree. Where were those banners shown? Search results pages. The reason people went to Yahoo! was to find stuff. That's the reason Yahoo! became Yahoo in the first place. The Yahoo founders made an invisible, scatterbrained web accessible by giving some sort of list index to it. Google later improved the sorting of the index, but Yahoo was first to build a large, searchable index.
http://www.quora.com/Why-did-the-web-services-group-at-Yahoo...
In my time, I've heard and seen an number of non-software companies choke, stumble, and even become acquisition targets because of bad software. Budget car rentals was one that immediately comes to mind that I can actually talk about. For awhile I had a spate of bad customer service experiences where the reps explained that "our system crashed."
If PG is right, all companies are going to need a hacker-centric culture!
Another takeaway from the article: Bogosity kinda works, but only on shorter timescales and for a minority of participants.
I'm assuming that their other 94% of traffic came from their content/media business, mail, and other portal-related activities. Since search seems to be a large part of the internet now, has their numbers changed today, or have they stayed about the same?
And search was such a small thing back then—did Google set the example for search becoming one of the most important actions on the internet?
And if Search didn't turn out to be a big money maker via ads, Yahoo would probably still be the biggest web company and Google would probably be a niche company last FAST -- making a profit, but not being the new software power.
We got that. :)
Saying "Yahoo is dead" (pg didn't, btw) is a sensationalist way of saying "Yahoo is in a slow, steady decline."
To predict this, look at the leading indicators:
- new products
- strategic acquisitions
- flow of talent
- company morale
- growth relative to competition
- positions in key emerging markets
Also, using "biggest site" as a metric is one of the problems with Yahoo that pg mentions in this article.
In contrast, Google was concerned about perfecting its search product, and therefore was focusing on what would help in that regard (hackers), which helped sustain its hacker culture.
I think you have to tread carefully in an area like this if you are a programmer. Many opportunities to to confirm a bias.
First, the world is not composed of productive hackers and politicians. In a media company, for example, there are writers, producers, actors, etc. Some of them brilliant. A good strategy for a media company could be to have the best of these.
Second, it's not always obvious what business you are in. I think this is especially for startups because they don't have a big version of themselves to look at. If you are a budding fast food chain, you know what business you are in and you've got examples. Yahoo didn't.
Not all internet companies turned out to be technology companies. Not all such companies in the future will either.
I'm sceptical that Yahoo was ever a great media company either.
rings true for humans too.
The main point of Graham's article is that Yahoo! didn't have a hacker-centric culture. If there was a time when that was true, it must have been before I joined.
A company without a hacker-centric culture doesn't encourage the kind of risk-taking and experimentation I saw when I was at Yahoo! Search. As an engineer, I had direct input into product features at every level, from ideation to design to implementation to launch. If I had a crazy idea, I was encouraged not just to tell people about it (up to and including executives), but to implement it and see if it tested well with users. I was able to add my own personal touch to parts of the product (sometimes big parts) without needing to ask permission or wade through excessive red tape.
This may not sound impressive to someone who's used to the way things work at startups or small companies. But this was at one of the largest Internet companies in the world, on one of the most visited websites in the world. For Yahoo! to give me and other engineers the kind of freedom and power we had is not normal for a company or a product that operates at this scale.
Can you provide any specific examples of how bad programmers hurt Yahoo?
Yahoo didn't have their own search technology, until acquisition of Inktomi in 2003 (they used Inktomi to serve their results prior to using Google, whom they used until ~2004).
Correct way to put it is "they didn't hire enough good programmers to build a search engine". A web search engine isn't something bad programmers can build. They acquired a great search engine (the algorithms were, IMO, better than Google's and more immune to gaming, they were a first distributed search engine paving the way for Google's "cluster of commodity hardware" architecture [1]), but it came too late.
[1] See http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.83.... for what I mean by "paved the way". Their mistake was using Solaris on SPARC (appropriate choice in 1996) for a few years too long (Google started on Linux/x86 right away, Inktomi transitioned by the time of Yahoo's acquisition).
Then Google demonstrated that search was not a solved problem, search was not a commodity, and it was too late.
"But there are worse things than seeming irresponsible. Losing, for example."
Business people -> ROI,IRR -> fast profits
Hackers -> hard-to-do -> sustainable profits
Dial-up access was their inflated source of revenue and their focus for much of the 2000s was on ad sales and not hacking.
cheap -> low
"Cheap" is a property of having a relatively low price. Therefore it's either low prices or cheap products, but not cheap prices.
The most important thing is, the CEO very much needs to know nearly everything important for his company. For a company heavily involved in software, the CEO needs to understand software in general and the software of his company in particular.
In a startup, he needs to be able to understand all the software, in detail, from the 'architecture' down to line by line. So, for this he needs to have enough knowledge of computing to understand, evaluate, and construct the architecture and have good knowledge of all the crucial 'software tools'. So, if the software is in C++ and Apache, then the CEO needs to understand these two. Windows? He needs to understand some or all of .NET, the CLR, C#, Visual Basic .NET, ADO.NET, ASP.NET, etc.
Now we come to a curious point: For the CEO to acquire that knowledge takes more of his time than writing the code, given the knowledge. So, really, once the CEO has learned to write the code, he can, for less than the time investment in learning the tools, just go ahead and write the code. He should.
Since the CEO wrote the code, maybe that makes his company a 'hacker culture'? I hope not: There are things more important than ASP.NET, etc. While knowing ASP.NET might be necessary, it is not sufficient. Just because the CEO drives his car does not make him a chauffeur. Instead, the software is just part of his job.
Next, my experience is that the best software developers and, in particular, the ones best at the most technical details of software, and the best at software 'architecture' are not 'hackers' and are not even 'computer scientists' but are mathematicians who took out a few afternoons to understand, say, AVL trees, extensible hashing, Cartesian trees, minimum spanning trees, DeRemer's LALR parsing, and monotone locking protocols but also understand dynamic programming, linear programming on networks, Lagrangian relaxation and non-linear duality, Poisson and Markov processes, and martingales. Sorry 'bout that!
With that technical knowledge and work, he has to pay close attention to the product, customers, revenue, hiring, etc.
Not so strange.
IMHO truer words have never been spoken.