You can pick at Marc's words as much as you like, but having heard his visions back in '95 when Netscape was big, he's a big picture guy and is seeing the forest for the trees.
Consider the following:
- In this decade and the last, software engineer consistently ranks in the top ten best jobs
- During the financial crash, software engineers enjoyed the least turmoil and the quickest recovery compared to almost all other sectors
- Software is mission critical to almost every business in the world now, regardless of sector
- Our jobs tend to have the highest pay among the majority of jobs (again, top ten)
I'm with Marc. I'll double down on software right now...it's not going away.
In other words, anyone can see that software is a winner.
What's not easy, and perhaps harder now than for other industries (with less spotlight on them), is to predict who the winners will be, and when they will win.
During the financial crash, software engineers enjoyed the least turmoil and the quickest recovery compared to almost all other sectors
Some of this is due in part to the fact that software development as a profession was nearly decimated after the .com bust. Developers in many areas where at near depression like conditions when it came to there career. It shook a lot of people out of the industry as well as out of the IT programs at schools. The result was that eventually the jobs outpaced the few who remained in the industry.
This is actually one of the things we consciously look for: companies that are turning businesses that didn't use to be software businesses into software businesses.
A funny anecdote from a customer of mine, make of it what you will:
They're in an absurdly low-tech niche. I'm going to call it gardening. Their business model since the dawn of time has been "You call us up, we send a van of guys to your house, they uses scissors to trim your bushes."
Five years ago they got a brochureware website, but they were still doing everything the same way: phone, van, guys, scissors.
Then came Groupon, of all things. See, their number one problem was getting more bushes to trim. Groupon solved that very efficiently, but now they had a different problem: paper no longer scaled sufficiently to schedule everybody. So they started tracking their phoned-in appointments in (I think) Google Calendar.
This pretty much opened the floodgates. They now use one of the help desk SaaSes because the email volume was getting too big and the business owner and office manager were having communication issues. Both of them bought iPhones so that they could run the business from the road, because the owner of a small gardening firm still needs to use scissors on occasion. They bought Appointment Reminder to free the office manager from X00 phone calls per month, so she can spend the time taking bookings and managing their (packed-to-capacity-and-beyond) schedule again. Things are booming.
They're now a software, van, guys, scissors business... and the owner is a wee bit weirded out by it.
How are they a software business? They are not developing software just consuming it. Is not much different than using a cellphone. You wouldn't call them a technology company because of that.
They are clearly on the track to developing their own software; if they don't, a similar business in the same situation will.
How long before they want to book orders via website? Unless there is an OpenTable for gardeners I am unaware of, that will require some custom code.
How long before they want to join their calendar and help desk software? How long before they want email logs integrated into the help desk? Those are custom software opportunities.
One of the most interesting things happening in software right now is how easy it has become to write just a little bit of it. The idea of a gardener developing, or directing the development of, custom code would have been laughable 20 years ago. With PHP and MySQL and ubiquitous cheap web/app hosts and proliferating web APIs, it's become increasingly common.
Many of those are enterprise software for Luddite companies. The sales process is slow, long and painful. I founded a startup for a little enterprise niche and I now have a profound respect for salespeople. The tech is easy, but getting a meeting is insane.
This. I work for a larger company that resells our enterprise asset management software (among others). Recently we received a document from a potential client with the question, "Can you import our data into your system?". The format of said document? A PDF with scanned in diagrams of things.
I had a chairman during the first dot-com bubble who would carry around a copy of 'Blown to Bits'[1] and read from it like it was his bible.
It is the earliest source, that I know of, of the theory that software and the web will destroy every other industry - and it was from the mid-90s (the HBR article was).
As they say in the financial world, "He's talking his book."
That said, much of what he says is spot on. Software is creeping into everything. Education seems obvious. Health Care will be more difficult. A lot of the change will happen in the US.
If we can't invest money, we still can invest our careers.
I don't think most of his investments are good examples of the power of software. I would point to embedded systems as the best examples (iOS, Android, modern fighter jet avionics, millions of other products that would be useless without software). Software is not cheap but it is increasingly the cheapest or only way to achieve desired results.
When you look at someone like Groupon their competitive advantages are "brand", "voice" and "scale", none of which are technological. Facebook, Twitter and Zynga might depend on really impressive development work to widen their moat; but again their potential for profitability is down to a business development team's ability to sell ads and sew up strategic partnerships. They're selling eyeballs, not tech, and the people shouting Bubble are pointing out that eyeballs are fickle, possibly even more so when placed in front of a computer. Advertisers can find other channels too, especially if they're not getting the ROI they want (given the conservatism of media buyers it might even be a disadvantage that unlike display advertising, Groupon, Facebook and Foursquare campaigns can conclusively be proven not to have worked).
You can't really compare FedEx's business model being dependent on software-driven superior logistics management with the public current preferring Facebook to other distractions.
To implement Health Care is an immense systems integration problem.
Having done some system integration work on a much smaller scale, it seems to me that in the US, more universal health care is an effective subsidy of IT for years. I think we can't really expect a working universal health care system for years.
Using Britain as a possible analogy, building the actual master universal health care system to handle billing and invoices and patient care will be a technical and political challenge that will suck resources from patient care to IT [1].
Integration is a difficult problem and with 452 HMOs in the US - even though some will be inevitably consolidated - there will be plenty of IT work [2].
Sorry to "talk my book" as well - our start-up provides a personal health records system that is connecting the UK's providers where the government's master system approach has failed - but investing in IT does not suck resources from patient care, it frees them up. I noticed this as a physician having to walk around the hospital to deliver pieces of paper. So I wrote software for Palm Pilots with a peer-to-peer encrypted handover system, gave it away, and convinced each of my colleagues to buy a Palm themselves. Each of instantly saved time they could spend on patient care, and the quality and accuracy of our handovers increased.
The problem of health care integration is not just solvable, it is already being solved (we are playing our part) but many in the mainstream are still expecting a big system approach and thus missing the small pieces, loosely joined approach that allowed the internet to scale.
I think he means suck resources more in a financial way, the money to build a big all encompassing system will have to come from somewhere and it will result in less money for patient care in the short term. Of course if the system was actually built and working it would then free up resources.
A few years ago I visited Phillips. They are the biggest makers of MRIs in the world. As it turns out their software department is bigger then their hardware department.
The housing bubble seems to be a pretty good counterexample. Plenty of people were saying that it was a bubble, while others were furiously flipping properties.
Exactly. It's just like the interviews with some of Bernie Madoff's clients. They knew that he had to be doing something not quite legit to be getting the returns that he was getting, but they all figured that they could make some money off of it.
Everyone sees a chance to make some money and convinces themselves that they will get out before the crash so that they won't be the last one holding the bag.
The smart ones realize that it's a bubble, and the dumb ones don't. It's just that the people that know it's a bubble convince themselves that that knowledge shields them from getting burned because they know more than the other guy.
the bubble was in the price of the collaterized debt. having read a few accounts of hedge funds in that time - it seemed preposterous that these things would ever lose value because of the way they were engineered.
some of the big fund managers who made big at the time were embarrassed to tell their peers that they were short CDO's.
David Einhom was probably the only one who came out and said it - the rest, like Poulson and co. were quiet.
Indeed, I was thinking the same thing as I read the comment. Look at google trends for "housing bubble." Peaked in 2005, well before the bubble popped.
Yes, and I think one reason for that is the discrepancy between the huge potential of software in so many industries and the greately varying barriers to entry to these industries.
It's no coincidence that Andreessen speaks of oil and gas exploration, defense, logistics, the auto industry, etc, but most of his investments seem to be in the entertainment/leasure and advertising space.
I agreed with most of the article, but I had the same reaction to that sentence. My guess is that people saying something is a bubble is, in fact, one of the best indicators that it is a bubble. (But of course it's not an infallible indicator of same.)
what's the attraction to gold, other than being somewhat scarce and well prospected for (probably no massive surprise claims to be found).
The value assigned to gold is clearly not due to its industrial applications. Gold for jewelry is also clearly not driving up the price as uncertainty rises.
What is a bar of gold good for and why does it represent a more secure asset than a bar of platinum, or silver, or anything else. Why is gold the commodity to go to for "security".
Gold rests on a mass psychological base that has remained reasonably stable for thousands of years. Well, if you want to to call something stable that went from $100 in the seventies to $800 in the early eighties back to $300 at the turn of the century to $1800 now. At least it never went to $0 as many currencies did in past millenia.
For me, that kind of "safety" is like becoming a farmer to avoid starving to death in the event of war.
Arguably the dollar price of gold reflects the dollar's instability. You see run-ups in the dollar price of gold after the inflation / stagflation of the seventies and with whatever the Fed has been doing the last few years.
The problem I see is that fears of inflation swing much more wildly than actual inflation. If the gold price were a reflection of actual dollar instability, a chart of the inflation adjusted gold price should be a flat line. In fact it shows swings of several hundered percent.
The price of gold is a reflection of inflation fears, not of inflation. I'm not saying it is too high or too low right now. But gold is clearly unfit for purpose if you're looking for a safe haven or a store of value.
I would argue that it is, the flat line could be said to have tripled from the seventies to right before it took off during the current panic. If one had bought gold before the panic and held long they would most likely see the same trend as was seen from the 70's up until the panic. The thing about gold is it is subject to panic buying when currencies are debased and it is exactly the time that someone that has held long in gold should vacate there position, wait for it to implode and then double down on it again. Gold purchased at it's baseline is actually one of the best hedges available. The problem is gold goes unnoticed until there is a panic. Institutional investors ignore it because it's performance is long while others ignore it because it is not a high profile security that the institutions are chasing.
It's not just inflation, though. Compare the history of platinum prices with those of gold. There is definitely a psychological component to the run-ups in gold which has nothing to do with immediate economic pressures.
Same as with any other currency. The reason people invest in USD instead of pink shells is because people will pay for USD disproportionately to its functional value.
You could certainly invest in platinum, but it's value is still rooted in practical utility- people won't pay for it unless they have a use for it.
Why is it gold that is partly a currency, instead of platinum? It was established as a currency back when its utility really was equal to its value, and people noticed that it was a valuable and stable resource.
you make these points
1. used to be a currency
2. stable resource
which have helped me understand that
1. dollars used to be tied to gold. now USA is making more dollars, thereby devaluing dollars relative to everything including gold.
2. the supply of gold is relatively fixed compared to the supply of dollars these days
For reasons others here have mentioned, and its relative inertness. Most other things with similar qualities have the unfortunate habit of oxidizing and effectively vanishing. As long as people keep valuing it, it will keep having value, and it won't decrease in quantity.
> it can be readily and universally exchanged for goods or services
I'm pretty sure all of the retailers in my area only accept USD. Nobody even bothers with the Canadian exchange rate on this side of the border, although some (and any vending machine) will refuse Canadian coins instead of accepting them at a 1:1 value with American ones.
The fact that the now-fiat currencies of the world used to be pegged by some fixed ratio to gold and to one another hasn't totally left the collective psyche.
wouldnt you rather use a currency whose value does not depend by dumb politicians?...wouldnt you use a currency whose value truly depends on the laws of supply and demand?
That would be the US dollar. The Fed is independent of "dumb politicians", who cannot coerce monetary policy, and its value is determined by supply and demand, as is any commodity.
The Fed is not entirely independent of "dumb politicians" since it targets both inflation/deflation and unemployment. When politicians make exceptionally stupid decisions, such as borrowing too much money which results in a bubble of productive overcapacity that then pops and produces excess unemployment, then the Fed has to react by devaluing the dollar to bring unemployment under control. It would be preferable for it to have an adversarial relationship with the politicians in which it would raise interest rates when fiscal policy starts to get out of control in order to make sovereign borrowing too painful past a certain predefined level (such as a percentage of GDP) but this seems unlikely given that politicians are responsible for nominating and confirming Fed governors.
No, we shouldn't. Politicians have a very long and impressive history of running up excessive debts and then printing their way out of them. The Fed was created from a careful compromise between loose money and sound money interests, and could probably do with a bit less interference from the political class. As someone who works hard and saves carefully for retirement, I would prefer to avoid to have the risk that my investments would be devalued in real terms at some point in the future because the two main political parties in US over the last couple of decades seem to view compromise as the act of raising spending, cutting taxes, and borrowing money.
If you're joking you got me. I can't think of another commodity whose value is affected by the government to the degree of fiat currency. Even if you leave the impact of government borrowing/spending aside, a dollar (or any other fiat currency) is just a piece of paper until the government institutes legal tender laws and demands people make and accept payments denominated in that currency.
"Companies in every industry need to assume that a software revolution is coming."
I can't wait for the revolution to come to the control and automation industry. I can see the heritage and legacy of the (software) tools I have to use, and unfortunately they aren't so old as to have a unix heritage but to have been born in the windows 95 era.
Probably I just need to pony up and get the real good high end shit, but the automation industry is ripe for disruption like health care too. the problem is the market is small and the stakes are high, so we end with old, expensive, tried, true, ancient solutions.
I hear ya. I think people have the idea that automation and controls are high tech. In reality, most of the stuff I've seen is more like 1970s state-of-the-art.
While I don't really agree with Marc's take on what's going on with current tech valuations, he hit on one extremely important point, and about it he is spot-on.
There is a major crisis coming in this country if the gap in skills and quality education continues to widen. Too many manufacturing jobs are moving overseas while high-tech jobs are expanding at unforeseen rates. I worry a lot about what's going to happen, and as controversial as it may sound, I think we may see a day in the not-too-distant future where the minimum wage is done away with.
As for Marc's commentary, it seems to me that every significant example he cited was related to content disruption (communications, entertainment, etc.). So I think a more accurate theme would have been "Why Software is Eating the Entertainment Industry."
Even more so, it was pretty much common knowledge at the time. It was just that there was lots of money to be made, and people always want to believe they won't be the ones to lose out in the pyramid scheme.
Sure, the speed and the impact of the bubble bursting was a bit of a shock, but the bubble itself was pretty obvious.
The cat is out of the bag! What the technologists have known all along! The market is huge, and mushrooming. Only the largest tech companies can keep pace with the growth, and there is still a-plenty for today's startups. Get yours today. Seriously!
Even already-connected markets, like the USA and Europe, will mature impressively as more people get better at using an improving internet. And by the time the entire living world is connected, birth rates alone will sustain satisfiable market growth.
It's not just about who will build the software, it's about who will use the software.
The way I see it, we're on the edge of going post-material. The trend is that the proportion of the population involved in the manufacture and distribution of physical goods is dropping. Follow that trend for a century, and you get a society where most people's jobs involve only virtual goods (although many of those goods will be turned physical by 3d printers or large bespoke manufacturing companies). Apple and google are the vanguard of the all-digital companies (apple isn't in the business of making stuff, only designing it). This means the majority of people will be making their money producing digital content, and spending it purchasing digital content. Already a sizeable portion of our income is spent on content (tv, dvd, games, books, magazines, ...). I see no reason why that trend shouldn't continue until we have a digital post-material economy.
And if we will have a digital economy, that means most people will be software users, producing content for others to buy. We're not just going to have to train the people that will make the software, but also the people that will use the software. I think the actual production of software will remain a small share of the economy.
A few centuries ago, someone may well have remarked that many of the best businesses in many industries, were moving into office buildings, an invention that was only a few decades old at the time, and that office buildings were eating the world.
1. They would have been right.
2. Moving into an office building would mean that a business was keeping up, not that it was necessarily a good business with respect to other businesses.
3. It would have been a good time to be building office buildings.
> Today's leading real-world retailer, Wal-Mart, uses software to power its logistics and distribution capabilities, which it has used to crush its competition.
Wait... others don't?
> Likewise for FedEx, which is best thought of as a software network that happens to have trucks, planes and distribution hubs attached.
Again... do their competitors really not? How can they get anything done?
He uses many examples to illustrate (belabor?) the simple point that businesses leverage technology for productivity gains. This has always been true.
25 years ago i.e. a quarter century similar points could have been made-- writers/authors/journalists are now using word processors instead of typewriters, accountants are now using spreadsheets instead of written ledgers, grocery stores now have automated UPC scanners instead of relying on cashiers to read price stickers, etc.
> The days when a car aficionado could repair his or her own car are long past, due primarily to the high software content.
More likely, it's because all said software is proprietary. If people got the source code to their cars' computers, you'd see a lot more people repairing their cars (and a lot more interest in automobiles in the current generation, leading to real, open innovation in the space).
It's interesting to consider how cars went from being something anybody could hack on, to something that only a few "qualified" people are now able to service. I don't expect the same thing will happen with software (in other words, I don't think "trusted computing" will ever become the norm), but we must be sure that it never does if we want innovation to continue in the software space.
The patent system is horribly broken, no doubt. But now that everything -- and I mean everything -- is turning into software, what does that mean for patents?
The capitalist answer is that we should let ideas freely grow and fight each other in the marketplace, but having an idea and selling an idea are two completely different skills. We will reward the salespeople, marketers, and business creators at the expense of the ideas people.
Perhaps that is what we want. Perhaps all ideas, not just startup business ideas, will become worthless. Execution will be the only thing that matters. If so, that's going to have some major impacts in the rest of society. It'll be interesting to watch this play out.
You are assuming that the patent system as it stands rewards people who have ideas (more than the alternative of having no patents at all would).
I think there is a good chance that this is not true at all.
Some possibilities:
Maybe the system rewards people with the best lawyers? Perhaps the system rewards people who are more worried at shooting the other runners instead of running as fast as he possibly can? It could also be that it rewards large companies (since worrying about patents is mostly outside the radar of small companies), making good people want to join them?
Ideas will not ever become worthless, if only because often it takes a certain mindset to extract value out of them.
You are assuming that the patent system as it stands rewards people who have ideas
Sorry, I should have been clearer: I am describing the purpose of the patent system, not the current application of the patent system. As I said, the actual application of the patent system is horribly broken. My question was whether or not the underlying assumptions that created the patent system in the first place will still be valid or not once everything becomes bits.
By letting inventors use their inventions without trolls getting in the way. The patent system doesn't recognize independent invention though reality does.
Imagine having your valuable R&D rendered worthless by an injunction prohibiting you from distributing your product, that's what patents do.
Very little of what patents describe could be called original and they aren't described in a way to help people learn. While congress may have enacted them thinking they'd spur invention it's pretty obvious to developers operating in a patent-laden world that they do nothing of the sort. We aren't allowed to look at patents and when we do they aren't written to educate but to obfuscate. When our companies milk us for ideas they aren't interested in our ground-breaking earth-changing ideas, but simple (easily infringed) ways to do [common thing] [in a slightly new medium].
As someone who has been involved in software for a long time (since I was a kid), of course I love to read news like this.
Having said that, I noticed that in almost all Hollywood futuristic-theme movies, software (or any ground-breaking inventions that usually found with the help more advanced software/hardware) tend to cause problems that forces humans to destroy them and put humans back to the world pre-software.
I hope that would never happen but looking at the trend that whatever Hollywood producers imagine usually come true (even though it may take 5-10 years since the movie is out) in real life makes me scared sometime when I read news like this.
Hollywood (media in general) can create self-fulfilling prophesies, but that doesn't necessarily mean things are set in stone:
1) Where is my flying car? Where is my personal jet pack? Where is my faster than light travel? Where is my single world government? What about space elevators? Why are we still eating food and not geometric shapes in primary/secondary colors (ala Star Trek's 'food cubes')?
2) People try to emulate the cool things that they see movies and read in books. Creating software that takes us to a post-apocalyptic world isn't 'cool.' People want to create light sabers, phasers, and/or transporters. People don't want to create master control programs that will manage the population.
3) There are many more things to be afraid of. We have nuclear arsenals that can wipe all of humanity off the face of the planet. We are so dependent on industrialized agriculture and mechanized supply lines, that we would be totally and royally screwed were they to be disrupted for a significant amount of time (e.g. a super-volcano goes off and we can no longer use most of a continent for farming).
Personal jet pack is here albeit still in its infancy. I'm quite sure people are researching into ways to make car flies. I don't think you can have time-travel (hence no faster-than-light-travel either). Of course there are "impossible" Hollywood tricks and of course some of the things would take longer to be created.
Sometime people invent something for a specific goal but ended up as something else. Nobody plans to create post-apocalyptic world. But if the invention evolves or ended up as a base to create something else, who knows what it'll turn out to be?
Take plastic as an example of that.
Yes, I do aware of the other things to be afraid of.
This seems inevitable, and positive. There are some friends of mine who dislike the resulting job shedding and concentration of wealth; I am not quite sure how that will shake out.
It will be interesting to see if today's 'software' disruptors will themselves disrupted by software. Today's revolution seems to me, a changing of the guard from the massive inefficient people-driven gatekeeper to the massive and lean software-driven gatekeeper. I wonder if the evolution of this will lead to decentralization and eventual diminution of today's usurpers?
If things did lead to decentralization, I would celebrate such a thing. In general, decentralization wouldn't necessarily be the death of the software industry, just of centralized gatekeepers in specific areas. Things like Big Data(tm) (gathering and analysis) would be hard to decentralize.
An interesting article, however the trend described might be a superficial one. Replace "software" with "paper" and the same arguments could be made for the economy 100 years or more ago. That indicates that the real driving force is something more fundamental (productivity is mentioned several times here) that could result in software being replaced with something better.
I've heard this argument before, but I'm not convinced by it. There's a vast difference between paper and software. Paper is bounded - it's just paper. Yes you use it to organize matters more efficiently (although the Babylonians were just fine - thank you - using their stone plates 4000 years ago), but there it ends.
Software's biggest difference from everything else is its being fully meta. It keeps growing. Abstractions keeps being added. Compare software in the 60s and today - some of us earn our bread by writing code for tools that other people use to create other tools, that IT companies use to provide computing to traditional businesses. And another level of abstraction isn't far away - when it's needed, it can be added.
I don't know of any field of human knowledge similar to this aspect, which is why software is here to stay.
Asimov had this story about all jobs being pre-assigned according to genetic abilities. Just one job wasn't pre-assigned, and that was allocating and assigning jobs (I may be badly mis-remembering... sorry...). Software is this job.
It is amazing to see how far we have come. I'm so glad that I picked the right industry to get into. The world changed all around us, and we were a part of it.
My only fear is that this will flood the market with crapy developers. That is how this change WILL be like the DOT COM boom.
I'm going to assume you don't do hiring at your company, then. Because the market is already flooded with crappy developers. The resumes that come in are often amazingly bad. Of those that pass that check, most of them can't even talk the talk. And fewer still can actually code anything worthwhile.
"We believe that many of the prominent new Internet companies are building real, high-growth, high-margin, highly defensible businesses."
What about profit? Doesn't ignoring profit (the article only mentions it in the context of Apple despite name dropping some other wildly unprofitable businesses) sort of suggest that maybe we are in fact making at least some of the same mistakes as the last bubble?
"Today's largest video service by number of subscribers is a software company: Netflix"
Netflix certainly uses a lot of software, but I think it is slightly disingenuous to paint them as a software company.
Add to these the recent announcement of Foxconn to install 1000000 robots. Now software will even be eating up sweatshops. Unfortunately, the rest of the economy is slow to catch up with these changes, in both the developing and (in a smaller degree) in the developed world. For the developing world this means a slump in growth until a more educated generation grows up, for the developed world, it means slow job creation. It's not a fault of technology; governments should have seen this coming decades ago. It's a shame that still, in many countries, programming is not required in primary or secondary education.
Take a moment to brag and enjoy the glory. Marc is a hero and this is an inspiring piece. Now back to work...
The main drive in the economy is more productivity; the main approach there is more automation; the main approach there is computer hardware driven by software.
Next, the main point about software is that it be 'smart' enough to give especially valuable output. For that the main tool is math.
Well i guess this means that emotionally warped hypergeeks will truly inherit the Means of Production and run your life like a program---from cradle to grave. Fuck Marx and Engels, we have Andreesseeeeen and Zuckerberg.
yes...we emotionally warped hypergeeks have something tht we can offer people....their own time...while the others seem to be making money through bullshit sales pitches!!
Consider the following:
- In this decade and the last, software engineer consistently ranks in the top ten best jobs
- During the financial crash, software engineers enjoyed the least turmoil and the quickest recovery compared to almost all other sectors
- Software is mission critical to almost every business in the world now, regardless of sector
- Our jobs tend to have the highest pay among the majority of jobs (again, top ten)
I'm with Marc. I'll double down on software right now...it's not going away.