Hacker News new | comments | show | ask | jobs | submit login

I took the liberty of taking notes. Apologies in advance if my impressions did not capture what PG was trying to say -- I'm summarizing, not stenographing.

1) Innovation

We've still got it, including making things that are not just ways to waste time on the Internet. Evidence: sci-fi writers dramatically underestimate progress in our industry.

2) Biotech

Sexy like cleantech, but doesn't require government subsidies to make money. Also has competitive moat, because biotech is hard. (Don't worry, software still worthwhile, too).

3) Efficient markets

Free flow of information creates efficient markets where not possible before. Many of our startups do this, such as AirBnB, an efficient market in lodging. YC = "mass production techniques, applied to VC"

4) Measurement

"You make what you measure." Put a paper graph on the wall plotting your favorite metric. You'll optimize for it, celebrate improvements, and shoot yourself in the foot if you picked the wrong metric. Metrics show social customs are obsolete (like, e.g., display ads).

5) The United States

PG was born in England, is not "wild, jingoistic patriot", but still thinks reports of US's impending obsolescence are greatly exaggerated. The only thing that kills empires is when people can't make money by building stuff. Three ways this can happen:

  a)  Bandits steal the money.  (NYC)

  b)  Your government steals the money.  ("The England that I escaped from.")

  c)  Other countries steal the money.  (The Netherlands.)
Leading candidate for toppling US is China. PG seems skeptical.

6) Silicon Valley

Budget crisis in California is two sets of idiots playing chicken. You don't have to start in Valley, but it really, really helps.

7) Small companies

World is "higher resolution": stuff gets done but it doesn't require industrial empires anymore. Networked small organizations are more efficient. Economies of scale paper over all the other sins of large corporations, but nimbleness of small companies means little guys win.

8) Economic inequality

A network of small companies plus money not getting stolen will produce massive economic inequality. (Patrick notes: PG's essay on wealth creation is my favorite of all he has ever written. He has a convincing take on why massive economic inequality is a good thing, and it isn't based on trickle-down economics.) If your business model bets on increasing economic inequality, good for you.

9) Moore's Law

Computers getting better, but in uneven fashions (e.g. SSD, not "all components improve 2 years"). Programmers are lazy. Companies which enable programmers to be lazy (i.e. not change practices or working code to benefit from uneven improvements) and get automagic speed increases win.

10) Things On Screens

We spend a lot of time staring at screens. Wider population spending more time staring at screens. PG has a suntan from his monitor.

11) Server-based apps

(I missed this one.)

12) Super good customer service

Customers can switch easily, people are talking together more, so have such good customer service like it seems like a mistake. Customers can now participate in design of products in virtually real time.

13) Apparently frivolous stuff

Our startup founders use Facebook to talk to each other about work, not email. "Facebook has not found its monetization model yet", haha. This sort of adoption shows there is something really at work here. "It is surprisingly hard to do math that has no practical applications."

PG skipped Twitter. Can't get a good name on it, but it turns out Twitter is really useful as a "non-deterministic messaging protocol."

14) Programming languages

There will be a succession of new, popular languages. Use the next hot language. You can be the guy who writes the library for such-and-such. Server-based apps can now be cobbled together from multiple languages. "Super abstract languages, like the ones people successfully write applications in now, were once called 'scripting languages.'"

15) OSS

I can't name a company which did too much OSS. If no one has gone too far, we're probably not doing OSS enough yet.

16) Linux will never be a factor on the desktop

Limiting edge of OSS is design. Everyone thinks they are good at design. Most people are not good at design: look at the contents of their houses. What this means for the desktop: buy AAPL stock.

17) iPhone

iPhone is a big deal, and I'm bummed because Apple are jerks. There are two problems startups have that aren't their own faults: immigration and AppStore approvals. They're like something out of Kafka.

(Sidenote: We're not giving our startups too little money: they can all afford iPhones.)

Android will be crushed under Steve Jobs' heels, because Apple cares about the iPhone like Google cares about search.

iPhone (or something similar) will do for laptops what laptops did to desktop.

18) Design

Design is why the iPhone wins. 20 years ago, it would have been surprising to say American companies can beat Japanese companies in consumer electronics devices. Core competency moved from manufacturing to design after people got microprocessors to shoot themselves in the foot with. Plus, China commoditized manufacturing expertise.

19) Real Time Stuff

Web 2.0 doesn't mean anything. Real time does. Google Wave will actually be important, not just somebody's 20% project. It is like Google-branded Etherpad, and Etherpad is useful, so Wave will be a gamechanger. See also Twitter, useful in a way different than Wave. If you make the convex hull around Twitter/Wave, and see a space which is unoccupied, that is a worthwhile opportunity.

20) VC

VC won't go away because VCs need to give you money. They can make the terms arbitrarily better to put money in your pockets [Patrick notes: can't get 2 and 20 if you can't invest the money]. Great news for you, since [owners] will now have the market leverage. Expect better valuations and board control.

21) Founders

Founders will more and more have the upper hand. Investors have learned firing the founders is a bad play. More and more founders will be technical founders. Programmers can learn to do business: make something people want, charge them money for it.

There should be an O'Reilley book for business. It would be really short. "Make something people want, charge them money for it. Advanced: charge more money."


Trends Not To Bet On

1) Credentials granted by institutions

Admissions officers are terrible. Look at our applicants: college graduated from (and by implication, admitted to) does not predict success. Not surprising: colleges admissions are impersonal evaluation of 17 year olds based on criteria which can be successfully gamed for money. Credentials are an example of an illiquid market. (Pagerank for people would be nice -- our startup doing it didn't work out.)

2) Business school

B-school is West Point for industrial capitalism. It trains generals, not footsoldiers. Market now rewards people who can do stuff. The kind of people who would be good teachers own their own businesses, became rich, and now have no reason to teach B-school. Instead, we get folks who cannot do and are forced to teach.

3) Government

The people on the bridge changes, but the engine room is the same as always. There is an increasing disconnect between public and private sector: government and 1960s PG (Proctor & Gamble, not the other PG) fit each other like gloves, and now government does not match startups/software/etc much. Folks want to work in electronic medical records: they're going to think bureaucracy is terribly slow.

4) Copyright

"Don't start a music starup unless one of your co-founders is Johnny Cochran." Expect a long, bloody fight that the content industry loses.

5) Restricted flow of information

Getting more liquid, faster.

I really appreciate that people like you take the time to transcribe things like this. I got a lot more out of your summary than out of the segment of the video as long as it took me to read your summary. Thank you!

You're welcome. I generally prefer text to video myself (one reason I always, always ask if I can write a transcript for interviews/talks/etc I give) because of the higher bandwidth and ability to go into more depth, but there are maybe a half dozen people who I'm happy to watch for an hour.

There's a startup idea in there somewhere. Everyone I know hates to watch videos, even of people we really like.

But it doesn't make sense: TV is such a powerful medium. What is wrong with that picture (literally :-)? Is it the delivery? Is it the rhythm?

The majority of TV programming is designed to constantly, subconsciously capture your eyes and ears. The picture changes every couple of seconds giving your eyes something new to process. The sounds (music, noises, and voice) tease you to watch what's going to happen next. This is why having a TV on "as background noise" while you try to do something else usually results in you watching more TV than getting things done.

A video on the web usually has the camera focused on one person and the screen doesn't completely change too often (if at all). Also, the speaking isn't exaggerated in any way to entice you to pay further attention. Therefore, it doesn't measure up to the levels of excitement/entertainment that your brain is already used to from TV.

"A video on the web usually has the camera focused on one person and the screen doesn't completely change too often (if at all)."

On the bright side, those videos should compress much better.

That's funny -- my current pattern is to watch TV on one monitor, and then do stuff on the web (like read HN) as background noise on the other. The result is that I get more things done than watching TV.

If I watch TV, I watch it for entertainment. PG is a smart guy, but I'd rather watch Seinfeld reruns for entertainment, and watch PG to learn something new. At that point, though, I'd rather simply read what he has to say at my own pace, rather than wait around to hear him say it.

Part of it is that watching a video takes the control of pace away from the consumer of that information. Whereas if you can read it, you can set the pace. Read it fast, read it slow, re-read important parts, highlight something and right-click google it, etc. Can't do all that with video.

Additionally, some people are aural learners, but others are visual. The latter generally have better synthesis and recall of things they read than things they hear. Such people likely prefer reading summaries than watching the video.

I wasn't thinking transcription. I was rather thinking about better navigation or focus on the interesting parts of a video. For instance:

idea #1: play video at 2X (or 1.5X speeed) with no pitch modification, so quickly go through most content. When I focus on the video frame, slow down to normal speed. Maybe have an obvious "rewind 30 seconds" like TiVo, so I can re-hear what sounds interesting

idea #2: tag the video timeline from other users commenting on specific times of the video (I have seen this on some video sites). That helps navigate faster and "move around" a video.

The main problem I see with the video format? Think of DVD vs. VCR. It's about rewinding and direct access. A video player online is just a long line, you get lost very easily. Somehow, you don't get lost navigating pages of text. Haven't you ever been frustrated that you couldn't find back something you just saw in a video? And I don't mean search by keyword...

I've found that listening to videos on BloggingHeads.tv at 1.4x speed works just great, and really improves the user experience -- and I never have to slow down or rewind, either. All video sites with people talking should have this as an option.

I think reading is just faster than listening.

there are are startups (like 3Play Media) that are working on automated video transcription.

Mechanical turk transcriptions?

"Those who can't do, teach."

Believe it or not, teaching IS doing. It is a separate skill set, a separate collection of values, and a separate set interests.

The idea that all who choose to be professors are failures at business is totally juvenile. What's more, it's arrogant in the extreme to think that those who are successful in business would automatically make good professors, and it's even worse to think that someone who happened to have both skill sets would find no reason to teach once they were rich.

Edit: Of course, this is addressed to PG, not to the poster of the summary.

I used to subscribe to your point-of-view, but don't anymore.

The traditional teacher-pupil relationship is well over-rated, and may be more efficient in the very start of the learning curve. Once the basic toolset is available, I'd much rather transition to deliberate solo practice and a mentor-mentee/craftsman-apprentice relationship.

In my experience, during any kind of medical/surgical training, these "wet-fingered" surgeons/mentors are the really ones that are sought for, not the big names who are famous on lectures halls and congresses. I believe this is also true for other complex skills as well, such as in startups. YMMV though.

I DON'T like all this belief that economic inequality is good. Is good if you are on the right side. As Keynes put it:

"For at least another hundred years we must pretend to ourselves and to every one that fair is foul and foul is fair; for foul is useful and fair is not. Avarice and usury and precaution must be our gods for a little longer still. "

I think wealth (the PG way) and an healthy society are not mutually exclusive.

I don't want to be rich in a desolated and poor Land. And it is not something to hope for. For you too.

The leadup to the quote you gave is also useful:

The love of money as a possession — as distinguished from the love of money as a means to the enjoyments and realities of life — will be recognised for what it is, a somewhat disgusting morbidity, one of those semi-criminal, semi-pathological propensities which one hands over with a shudder to the specialists in mental disease.

Company which did too much OSS: Sun.

Can you cite Sun products that were made OSS that users weren't already replacing with OSS solutions from other sources? I think Sun's problem was that it was in the business of selling proprietary Unix workstations and servers, and most of their potential customers these days are content to use PCs running Linux, Windows or OS X[0] in those roles.

[0] Apple is a lot like a proprietary Unix vendor in a lot of ways, but Apple hardware is not more than 50% more expensive than PCs with similar capabilities.

Sun's problem wasn't OSS. Sun bet the farm on SPARC and it died out in favor of x86 faster than anticipated, basically. The OSS stuff was great -- they needed to build and sell an IDE or some other enterprisey-product that would drive Java's big corporate users to pay out big license money.

> Company which did too much OSS: Sun.

s/did too much OSS/didn\'t do OSS early enough/

I would go so far as to say Sun never did enough open source.

I had the misfortune of doing some work in a mixed shop where most of the Unix servers were running Solaris 8/9 recently. A couple things that led to its downfall:

1. No package manager. And now that there is a package manager, maintainers are few and far between. I don't think I would trust it in a production environment. It's never going to be as robust as dpkg or yum. if they had fully open sourced Solaris, and supported the community as primary, there would be a well-maintained package manager. I doubt Oracle will do anything to fix this.

2. I'm still not entirely clear on where their utilities like grep, awk, sed, etc. come from. All I know is I was terrified to push /opt/bin ahead of /bin in my PATH for fear of breaking something, and meantime I hated the native shell tools. They should have adopted GNU's tools as first-rate. Solaris' tools suck.

I see that as more of an orthogonal marketing issue of not understanding where the customer growth was coming from. When you have the attitude that you sell to CIOs and your consultants do the dirty work of setting up the systems, you don't care if college sophomores find it to be usable. (Until 10 years later when those former college sophomores are now making purchasing decisions, whoops!)

Open source won in this area because of its "by developers, for developers" approach moreso than the source code ecosystem.

Package managers are an unassailable benefit over proprietary software. That shop I worked in hasn't updated their servers in half a decade because it's too much work to handle dependencies without a robust manager (and even then it's a significant amount of work.) And it's not like it gets easier with time. We have package maintainers because it's a full-time job trying to check all of a piece of software's dependencies and makes sure nothing breaks with an update.

Now yes, proprietary software like Solaris could, in theory do that. But Sun would have had to double its software staff to even come close to the stability you get out of Debian or Red Hat's repositories.

Sun was a smouldering wreck before they started giving away everything they owned. Their OSS policy was an act of desperation -- it did not work out, but to say it killed them would be vastly misrepresenting their last decade.

8) Economic inequality ... If your business model bets on increasing economic inequality, good for you.

i'm very curious what this means. what are some examples of start-ups that make this bet? the only thing that comes to mind is private security and gated community development.

It also means that there are growing markets for high-end luxuries and very low-level services.

what does that have to do with income inequality? how does it help your business that an increasing narrow group of people can afford what you sell?

Just a note to say thanks for taking the effort to write this. Reading your transcript and typing this note has taken me less time than watching the video

I'm curious as to whether PG has any specific languages he'd bet on right now.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | DMCA | Apply to YC | Contact