Reviewing Steve Yegge's Prediction Record
danluu.com
danluu.com
I don't think Amazon really "understood". It's more that (according to "The Everything Store", anyway), accidentally, some emergency servers were reprovisioned as servers on demand for a bunch of frustrated devs who couldn't get resources the traditional way. And then demand just exploded and they thought why not sell it and see what happens.
It is true that Bezos appears to have a knack for spotting great products that appear accidentally in his org tree, and spinning it out as a separate team (e.g. the A9 team, Kindle...) but it might be that he thinks like a VC and backs everything that looks interesting instead of letting the bureaucracy kill it, and hindsight is 20/20...
Yegge has recalled an email to all Amazon staff where Bezos outlined new Amazon rules. This was in 2002.
1.) All teams will henceforth expose their data and functionality through service interfaces.
2.) Teams must communicate with each other through these interfaces.
3.) There will be no other form of inter-process communication allowed: no direct linking, no direct reads of another team’s data store, no shared-memory model, no back-doors whatsoever. The only communication allowed is via service interface calls over the network.
4.) All service interfaces, without exception, must be designed from the ground up to be externalizable. That is to say, the team must plan and design to be able to expose the interface to developers in the outside world. No exceptions.
Especially the last point of the memo is very telling because it shows that Amazon was planning to externalize artifacts of various teams within Amazon at least 4 years before AWS was actually born.
- Is ad business model a major reason Google is not providing a platform or hesitant about it? (Out of Amazon/Apple/FB/MS, FB is the only other primarily ad-based company (or is it?) so I wonder how FB is doing it). Because programmatic interfaces means developers can program out ads (unless there's a way for a service provider to avoid that situation, IDK).
- Or has Google moved towards platform/accessibility model recently? or is it same old same old?
"Why not sell it" is a bit of a joke. Before you can say that, you have to have a platform that's designed to handle unfriendly users. Users who may be presumed to attack each other and the platform. Setting that up doesn't happen by accident.
shudders I worked in that space right around the time the iPhone landed, WAP was way beyond looking really bad, it was an unmitigated disaster, it was horror.
> The fact that Apple forced the carriers to allow an open mobile internet was utterly shocking to me - in a good way.
To be fair, high-end nokia did already have "open mobile internet" available. Technically anyway: compatibility was a craspshoot and the experience was terrible, I had an E70 with a relatively competent webkit-based browser (IIRC), it did technically render pages but:
* very small screen, making interacting with pages difficult (mobile-adapted pages were not a thing, and the browser didn't do much adaptation)
* arrows moving a small cursor around on a small unzoomed page made for a terrible experience
* device lacked both CPU and (especially) RAM, interactions would be very slow and browser would regularly crash and OOM
* flash was still very present, the device didn't handle it at all, it didn't cope that well with javascript either
After it was released Opera Mini did help with the resources side (but not UI and compatibility). Anyway my point is, open mobile internet was available before the iphone, it was just an utter wreck of an experience (though still not as bad as WAP)
Not really sure this is relevant. Apple forced the carriers to change their business models -- first AT&T and then the others; Nokia et al just played along. I was really disappointed that Google did not persevere with selling their own phone -- there was this shining moment where it looked like Apple and Google would unite to tear down the carriers; I think Google's shying away from customer service may well prove pivotal -- e.g. it's going to be hard to sell cars if you don't do customer service.
People often point out that Apple doesn't really understand the cloud or that Amazon's forays into hardware have been at best mixed. Maybe. Google -- alone of its rivals -- doesn't understand end-users (its customers are advertisers). Which is more important in the long run? (Bezos may suck as a boss, but Amazon's customer service is first-rate.)
1. didn't see it then, don't see it now. Then again I'm in europe, mayhaps US carriers were significantly more overbearing. I don't recall carriers being good friends of the consumer though so not overly likely
2. this responds to a comment specifically mentioning allowing "an open mobile internet". Again this might have been because I was in Europe, but as my comment notes I did have access to an "open mobile internet". The experience sucked but technically there it was.
The U.S. was indeed terrible: very restricted competition on phones (two year contracts, models often specific to one carrier, etc.), prohibitively high data rates and extreme barriers to entry for anyone trying to build a phone-based business.
Around 2000, I worked for a web company in San Diego where Qualcomm was trying to build a stable of companies developing mobile apps to boost demand for better hardware. We looked into the business model and each of the U.S. carriers wanted roughly $50K simply for the privilege of listing your app in their store plus a generous cut of the sales, should you have any and that number wasn't fixed but based on what they thought you could afford so our better-known clients got estimates well into 6 figures, again simply for being listed.
Unsurprisingly, we looked at WAP instead but it was so limited in practice that we gave up and nobody I knew used WAP or J2ME apps. As far as I could tell, that only changed when Apple had enough clout to get AT&T to go for a smaller slice of a much, much larger pie.
For the most part that was also the case in Europe, though countries started cracking down on some of the bullshit fairly early (e.g. the ability to move your number from one carrier to the other and requirements that phones be unlockable after some time). I don't know about the phone-based business aspect.
https://apps.fcc.gov/edocs_public/attachmatch/FCC-03-284A1.p...
Unlocking wasn't a requirement until last year:
http://www.cnet.com/news/president-signs-cell-phone-unlockin...
The big thing which I believe really killed business options, though, was data pricing. Before the iPhone made affordable unlimited plans the expected default, it was assumed that anyone who wanted data was a business user on an expense account and the pricing was so high that everyone I knew treated app usage as an emergency-only measure rather than a part of daily life.
The idea that the telco controls what software you run was unknown there. Data is data. You could almost see operators taking the first steps from inflexible contracts where they gave phones away, to leasing-like models (for business customers, consumer markets went though the same change a few years later).
The only surprising thing is that Apple didn't change telco business models more, but I guess they didn't have to. US telcos continued to sell phones through their own channels and didn't have to yield to bigger data plans until years later.
But change is coming, because data by nature wants to be dumb. Soon all customers will find the idea of operator controlled phones just as foreign as some did ten years ago.
At the time I worked for a web development company that identified itself as an "e-business solutions agency". It was in the process of collapse following the dotcom bust, and we were looking out for any new opportunities to cling to. I remember feeling a bit second-rate at this trade show when I introduced the company I worked for, because we were only "e-business" while the exhibitors were very proud of being "m-".
Nothing came of any of it of course.
Windows Mobile back then had a fully featured browser. It was all pen driven, but it actually worked really well. The first iPhone seemed like a step back at least as far as the browser is concerned.
Data was just extremely expensive, but it was faster than the dial-up internet most of the country was still stuck with back in 2004.
Windows Mobile sucked not because of WAP, but because you had to use it like a tiny desktop UI and poke at it with a tiny stylus to move a mouse cursor (depending onthe version).
The iPhone wasn't awesome because it had a "full" web browser or whatever (though it helped). The iPhone was awesome because it was a pocket sized super computer with a UI/interaction model that actually worked
Of course, it released roughly seven months before the iPhone, which beat the pants off of it. At least it could play NetHack...
Steve Jobs was right in that no one wanted to use a pen (stylus). As the market matures, this is changing, and the stylus is regaining importance at least in certain niche areas. However, for the vast majority of the mass market, your statement was, and still is, a contradiction.
But, on the subject of the "mass market"...
What Apple/Jobs did was build an awesome device for media consumption.
Up until then we'd been talking about the PDA - a productivity device. It needed a physical keyboard and stylus input to make it "productive".
The best PDA up until then was the Palm Pilot - it has exceptional writing recognition - if you spent two hours learning their alphabet. It works for email, spreadsheets. Battery life was excellent. Most people I know used it with a fold-out keyboard when traveling.
Apple realized that 99% of the time you'll using your phone to play games, read, listen to music or watch videos. Their touch screen was fantastic - it took just seconds to learn how to use it. Sure, writing emails was a pain (even with the smart keyboards), but you'd only do that rarely.
So yeah, you're kind of right: no-one whose main use-case is content consumption is going to want to use a stylus. For content creators however a well-implemented stylus is a big sell.
That's why Microsoft are throwing so much weight behind the Surface Pro, and it's also the main reason people are buying it.
Instagram alone is evidence that there are immense numbers of people creating massive amounts of content on their touchscreen smartphones. Spreadsheets are super-niche content. Text, photos, & video are mainstream content.
Compared to email on the desktop during the PDA era, I'd bet people communicate with other people much much more now via their smartphones, whether that's through email, social networks, instant messaging, or video chat.
Fans wanted a phone in there since day one. The only reason it took them six years to do it was Jobs commitment to getting it exactly right (or "his way" if you wish) which meant technology had to become available.
It was small, had almost a week on standby (if you used it as a phone, which -- before the smartphone revolution, most people did). And it could actually browse the internet properly - gateways had to do some conversion, but unlike WAP, it was working with a real HTML subset.
It was that day that I understood the death of WAP is imminent, the mobile revolution is coming, and that HTML is here to rule mobile as well - because the future was already there, just not evenly distributed.
Prediction is hard, especially about the future, but describing the unevenly distributed present is easier.
Sometimes you have to dig deep into the online docs to find this info.
(I am looking at you, matotlib -.-)
Earth to Major Yegge, Earth to Major Yegge, do you copy? Come back, we miss you.
I just posted an update to the predictions post, with my own scores. Take a look!
As far as blogging goes, I've been on a bit of a hiatus. There's been a lot to learn. The tech world is moving so fast! I've been in Google Cloud Platform for the past 3 years, and next year I hope to start blogging about the experience, and about the technology in general.
For now, though, I'm busy both with my day job at Google and with various personal projects, so there's not much to report just yet. I promise I'll be back someday. :)
https://sites.google.com/site/steveyegge2/ten-predictions (convenience link)
[0] https://plus.google.com/+RipRowan/posts/eVeouesvaVX
[1] https://plus.google.com/110981030061712822816/posts/AaygmbzV...
WordPress has some really horrible backward compatibility baggage but the largest reason for it's success is that it's a platform. Being a platform is as simple as having an API to let others build on top. The platform API is like a bag of seeds that lets everyone do their own gardening using your seeds. This builds an ecosystem.
Accessibility when it comes to Platform APIs is also best illustrated with a simple example and should be clearly obvious to anyone who used jQuery, Prototype and the ilk. At the time jQuery offered method chaining and which other JS frameworks did not. And it also offered some really simple and well-named convenience methods.
$('#foo').append('OK!').show();
The great thing about building a developer platform is that it leads to an ecosystem of apps and plugins. And the great thing about an ecosystem is that it gets you venerable developer talent on a truly global scale.
I recall having a discussing many years back about why Microsoft enterprise licenses were mostly available only through certified partners. It struck me at the time that Microsoft was handing out 15%+ commissions for no reason while it could have adopted Apple's direct selling model. It turns out Microsoft had essentially build a retail platform where being a Gold partner unlocked top margins. You could find Microsoft certified partners everywhere. And this built an ecosystem. These certified partners were cross-selling and up-selling Microsoft products to no end, layering their own support on top and pushing for 3-year release cycles so the stream of money would get renewed vigour.
Google really does have the best benefit program in tech.
EDIT: I'm kinda joking, and kinda not. Maybe it seems silly, but this really will be on my mind next time I hear from a Google recruiter.
And then the focus became JavaScript, which has all of the bad and many—but not nearly all—of the good points of Lisp, and one key differentiator: it runs in the browser, on effectively every computer, tablet and phone on the planet.
Maybe the Lisp-compiled-to-JavaScript efforts will help.
I think Javascript moves those Java programmers halfway further to Lisp. And while the Common Lisp revival movement sort of petered out, Racket (which is honestly much better than Common Lisp to my tastes) is still trucking along and might be poised to be the next step for a generation of programmers who've absorbed the Javascript mindset.
While I personally like Racket quite a bit, I think that the Lisp most posed to pick up the JS crowd is Clojure/ClojureScript.
OTOH, maybe instead of a Lisp we'll get another halfway-from-where-they-are language for the JS programmers, and continue the Zeno's paradox of Lisp readiness...
Today people seem to take for granted that the NBL was in fact javascript, but I still cannot find any source for this.
Does anyone know more and can point me in the right direction?
Maybe I am taking it for granted though…
The video for his presentation was totally messed up, so his slides didn't load up. He just said "screw it" and went on presenting from memory to a room of around 4,000 people.
He nailed it.
Afterwards, Tim O'Reilly said he had never seen anybody deal with something like that and do as well.
If you get the chance to see Steve present, go for it. He's thought provoking and entertaining.
The smartphone market is at least 5 years out. 3 years later iPhone comes out and defines the market. Saying it didn't get really, really popular for a year or two is misleading. If you were betting on the market being _at least_ 5 years out you would have missed that boat.
In other words, I think saying XML databases are the future in 2004 is like saying JSON document stores are the future now. I don't view either as all that correct, but there's definitely been a surge in interest.
I'm getting the impression that a lot of people are reconsidering the "nosql" approach, as well, sticking with (or reverting to) sql databases, including simple ol' mysql. This was an interesting interview that showed up on HN a few days ago.
https://medium.com/s-c-a-l-e/github-scaling-on-ruby-with-a-n...
I was never all that enthusiastic about "nosql", partly because personality wise, I tend to be reluctant about new technologies, and I liked a system based on relational algebra, all that. That said, "nosql" wasn't really meant as a NO to sql. I get the feeling that the term got away from them a bit, largely because "you-know-sometimes-there-really-are-better-solutions-than-sql-for-some-problems-and-you-might-want-to-consider-them" isn't really an option as a catch phrase.
That said, nosql absolutely did grow massively in popularity, it did move from the fringes to mainstream use, so as a prediction, yeah, I'd agree this should score a few points.
If you are right, then you can point people to your correct predictions. If you are wrong, then silently delete the records or just let people forget about them.
Umm... I'm probably too old, as I feel anger when I read something like that.
As long as it runs on startup, it ships. Any problems can be addressed by paid support contracts (suckers).
> What if every "API call" were hosted by a different server? Well, you'd have a lot of flexibility in fixing any given call without needing to impact clients that don't use it. And it'd be easier to distribute it over multiple machines — it seems like it'd be a better design than the BEA WebLogic design, which is to run a few copies of your MonolithService on each of a handful of machines. There would be new problems to solve, but who's to say it wouldn't be a better design overall? We won't know unless we ask the question.
[1]: http://steve-yegge.blogspot.com/2009/04/have-you-ever-legali...
Generally speaking, NBL will have to have a much greater focus on performance than so-called "scripting languages" have had in the past. I mean, if it's really the Next Big Thing, it'll have to run on mobile devices, game platforms, and all sorts of other hardware-constrained devices. It sounds impossible, I know. But it's not. Those compiler writers are a tricky bunch, and they have decades of experience. A bunch of them got together and did a lot of head-scratching over NBL, and the result is going to perform quite nicely.
What do you think are the reasons? Some folks might claim that Java enables mediocre teams of mediocre programs to turn out mediocre code, while they would turn out utter dreck in Lisp (we know that excellent programmers will turn out excellent code in Lisp, but there are business and operational risks associated with a small team of excellent programmers), but I don't actually know if that's true. Has research been conducted with large teams of mediocre programmers with mediocre management, using Lisp?
We use Java at work. Obviously all of the following depends upon your use case.
The main reason we continue to use Java is probably legacy reasons. We have lots of library code. And we're a small shop, so every extra language we program in has a real cost associated with it.
The stability and backwards compatibility is great. We have some libraries we haven't had to touch in almost a decade, being used in apps we just wrote last week. Of course some of them we've updated to use newer Java libraries (such as java.time and java.util.stream in Java 8).
IMHO Java's biggest problem is the community. It's a bit enterprise-y. As long as you don't get caught up in that I think you're OK.
I think the other problem was it went so long between Java 6 and Java 7. With Java 8 there isn't a lot from the likes of Python and Ruby I miss anymore.
For imperative languages, I generally like Go and Python better, because I find them "funner". But once I get past myself, meh. I do think Go gets things a bit more right than Java, but not enough that it merits switching the language we use. Maybe if we had zero code and had to start out fresh.
Groovy will be feeling the heat here for its original use case. With the lambdas and terseness in Java 8, there's not a lot of reasons for it to exist.
I think it's to do with predictibility: back-compatibility, discoverability. You can get new coders who can get up to speed on a codebase. Business customers - who really really want things to keep working - value that. And business customers ultimately decide what is used. Similar to your point. BTW I'm not even sure expert lisp programmers can easily maintain their own code (or, perhaps more importantly, would want to).
Operator overloading: very cool, nice for matrices, complex numbers etc. But easily abused to create impenetrable code (you must first get across all the definitions). C++ had it, Java avoided it. Lisp is so flexible, it's like everything is an overloadable operator.
I'm not aware of such research on lisp; but the only successful lisp projects I've heard of are with hot-shot coders. BTW: reddit was originaly written in lisp (inspired by pg, who gave them the idea for reddit); they switched to python.
It's why I always ask about favorite programming language in engineer interviews. It's not that there's a right answer, but I do expect to hear some nuanced opinions. If the candidate says "Java" and then looks at me like I'm a little crazy for asking, I know they probably aren't a good fit.
"We work in a fashion industry, and marketing really matters.".
I don't know if I am getting older, but this rings very true to my ears, especially today.
Nothing succeeds like success. Unfortunately, it's difficult to tell luck from skill quite often.
I'd argue that Zuck didn't do too differently either. Both companies did have technical execution down. But circumstances also mattered hugely.
Let me lay the raw truth out for any lisp fans. It will never be popular or widely adopted. Ever.
It has been around for half a century. At no time was it ever widely adopted. It was always just pushed by a small group of fanboys.
That said, I love it and have learned a ton from as have many others.
I mean, what is going to take? 10 more years? 20? At what do you conclude that a language really isn't useful for general purpose programming? (not to say it isn't brilliant and useful for specific purposes, learning, etc.)
I think Lisp is the same: it's everywhere you don't think it is because it's opinionless and languages are defined by their community (and community is held together by the shared values and opinions). Look how many languages are rapidly adopting functional programming idioms. They're becoming Lisp. Eventually you start seeing the value of functional idioms in Blub language. Then you go and look at where it came from and notice, "Hey this is the real thing". Then you get real yourself. (This has been my experience).
I care that they produce a decent phone, and that they have enough market-share that there's a software ecosystem.
Profit is really susceptible to distortion based on reporting. What exactly are you counting as costs? Do you include investments in the future? That's really a business choice, and it doesn't necessarily mean much.
At the end of the day, it's only the shareholders that really care about that profit. As a customer, it doesn't really matter whether that money ends up at the company or its contractors - you're paying either way.
As an investor, you (eventually) care about profit. If you're just trying to understand what the customers value enough to spend money on, revenue is a better guide.
Market share matters for a lot of companies because economies of scale mean that your unit costs go down as volumes go up. This gives a natural advantage to market leaders, in addition to things like increased brand visibility.
Market share is very important to pundits though, because it changes very rapidly and is often measurable throughout the year on a per-product basis, whereas company financials are less frequently reported and harder to tease out. A battle between products that readers can see in the shops and read reviews for is easier to create a narrative around than corporate financial performance. I think this is why sales numbers and market share are focused on so heavily in the media.
Unfortunately, the one concept we have down is "make it up in volume."
One thing is certain. A guaranteed way for Apple to fail is if they start worrying about how other manufacturers are 'shaping' the market. It would kill them stone dead in a matter of a few years, tops.
And I don't think it's really reasonable to include tablets at all. The motivation is that they're a new, unforeseen development? But then you really should be including android tablets in those numbers too; and those happen to be sold primarily by other vendors.
At the end of the day, it's clear that tablets aren't laptops, nor do they look like they really replace laptops (especially not the ipad - you could quibble about microsoft's offerings, but those haven't been too successful yet.)
The heart of the prediction was that apple would come to dominate the laptop business. Clearly, it has not. It's still mostly windows-vendors. If you include tablets, apple does better, but there too it's a minority market share; android dominates.
But you are way off in your figures. In 2014 tablets outsold laptops 1.5 to 1. Apple sold 12M iPads in q4 2014.
And tablets replace laptops for many people. Most people just do Facebook, web, and email.
Best estimate is that there are around 200 million laptops a year sold (2014) based on a PC market of over 300 million and that laptops outsell desktops (worldwide).
Looks to me like tablets sell roughly at the same rate, around 230 million units. (I think it's not easy to find comparable numbers so maybe my data is off).
So given apples approximately 25%-33% market share of tablets that's just around 50-80 million ipads, which is still well below the top three in laptop sales (100 million).
http://www.statista.com/statistics/272595/global-shipments-f...