Fractals of Change

Broadband for EVERYbody

We used to think it was enough to make broadband accessible everywhere. That's no longer good enough. We now need to make sure that everyone actually has broadband in his or her residence and business. Everyone! (voluntary cave dweller excepted). Our goal in Vermont is to combine stimulus money with private investment and state bonding authority to move us quickly not only to 100% broadband availability but 100% broadband penetration.

The electrical system of tomorrow, the health care system of tomorrow, and the education system of today all depend on universal broadband penetration. Oh yeah, communication, commerce, and entertainment all need broadband too. So does e-government (coming soon) and research.

The electrical system of tomorrow will be smart. That means demand, supply, capacity, and outage data flow unimpeded and in near realtime from meters to utilities and back to consumers and generators. Much of this data flow is machines communicating with other machines. Some information flow is back to consumers both large and small so they can control their energy bills by using electricity when it's abundant and cheap and shunning or selling it back to the grid when it's rare and expensive. Taking advantage of the smart grid requires a broadband connection.

Part of e-health is electronic health records. Better information means better, cheaper, and less mistake-prone care. But we can't replace paper records with electronic ones until we can be sure that very doctor's office and place of treatment is online with enough bandwidth for bandwidth-hungry objects like x-rays. In the case of home health care, the place of treatment is the home. A home health worker needs the same access to medical records and the same ability to update them that a doctor or a hospital does. The home of the future will have health monitoring devices when needed. Homes need to be online for the delivery of health care.

When I was in school a million years ago I was taught to do research in the Reader's Guide to Periodical Literature. That was then and this is now. Students need to know how to separate the wheat from the chaff on Google and wikipedia; their homework needs to be online just as mine was in the library. But a teacher can't responsibly give online homework to a class if even a small fraction of the students don't have the equipment and connections to get online when they go home. We can't reinvent pedagogy the way we need to until we know that all students have broadband connections – and that schools have scads of bandwidth

So everyone needs to be online. Geography can't be an obstacle but neither can poverty, lack of equipment, or lack of training.

The platform for SmartVermont – the Vermont we hope to build with stimulus money, State money and bonding authority, and private investment – is universal broadband penetration. The first application on that platform will be smart grid, e-health, and e-education. With lots of hard work, some luck, and our fair share of federal (our!) dollars, we can build that future for ourselves and our children.

When Angels Fear to Tread

It's the lack of exits that's a problem amid the deadly clamp of panic throughout the economy - a far cry from the irrational exuberance that drives entrepreneurs (and their investors) and which brings us trouble and fraud as well as greatness and "slumdog millionaires".

No matter how much we bailout the auto industry, there'll be less jobs making cars next year than there are this year – in fact cutting jobs is a condition of the bailout funds. There are going to be less bankers a year from now, too. New jobs will come from new industries, not from the ones on life support. So it is essential that there be some source of funds for entrepreneurs so that they can do what they do best (when they succeed) – create good jobs for future markets.

Tom Friedman suggested bailout money for VCs; Fred Wilson articulately pointed out that only the incompetent VC firms would take the money and that skilled VCS actually do have money and can raise more. Entrepreneurs pointed out in comments on Fred's post that there IS a shortage of money from entrepreneur's POV. In a comment on my post on the subject, Fred suggested tax breaks for angels.

Angels are motivated by fear and greed just like everyone else – angel investors, that is, who, in better times, are a source of funds for entrepreneurs launching the businesses of their dreams. Angels (I've occasionally been one) are just as perverse as all other investors; we're more likely to invest at market tops when everybody else is and less likely to invest in scary times like the present even though there's greater opportunity now.

There is a way to get angel money flowing again but it's not by subsidizing angels – even with tax breaks. Angels are claustrophobic; don't like going into investments without exits.

Angels generally step aside (but don't get cashed out) when a company gets its second round of financing. Often VCs step in with money, advice, and contacts for the next round of a company's growth. The angels and VCs (and entrepreneurs) generally get some reward when the company is either bought or goes public. But companies aren't getting bought or going public right now.

The result is a logjam. The VCs are concentrating their time and attention on companies which would have – in better times – been long out of their portfolios. It's true, as Fred said, that the good VCs can raise more money; but they can't clone themselves. Fred can only serve on so many boards at a time. That means that venture funds can't take responsibility from angel investors at the rate they used to – they have last years' hatchlings still in the nest.

An angel looking at a potential new investment not only confronts the risk of failure – that's always been there – but also the risk that he or she will be in an active role with the company – and perhaps its only source of capital –for a long time to come. We angels, like the VCS, still have the companies that we previously incubated in the nest. So we're not looking for new investments either. We're scared and there are no exits.

Only greed (the dream of an outsized return) can conquer fear. Tax breaks don't do it; you need to have some gain before you can use them.

An opinion piece in the Wall Street Journal today by Tom Hayes and Michael Malone entitled "Entrepreneurs Can Lead Us Out of the Crisis" suggests not only tax breaks for angels and entrepreneurs but also eliminating Sarbanes-Oxley to make it easier to be a public company. This has the virtue of appealing to greed (the IPO dream) and opening up exits; but even I'd agree that, in the dotcom generation, companies went public much too soon – eventually to the detriment of the companies as well as their investors.

We investors have to give up the dream of a QUICK outsized return. We can dream but we have to dream the patient dream. We probably even have to wait for companies to be profitable before we can make any money. That's OK; we can live with that as long as we can have the dream.

Turns out there is something government can do, however, to get investor juices flowing again: invest in infrastructure that creates opportunity rather than subsidizing zombie companies which are blocking the way. When the Erie Canal was built (funded by private investors buying government bonds), private money flowed to boat people and businesses all along the canal. Same kind of thing when the railroads with built with healthy doses of land grants and other subsidies – fortunately government DIDN'T elect to subsidize the canal boats which the railroads put out of business. DARPA (government) had a lot to do with inventing and funding the Internet; the Internet enabled and encouraged a wave of privately funded innovation.

I wish more of the bailout were focused on infrastructure – especially new enabling infrastructure. If transmission lines actually do get built, "alternative" power'll flourish with much less government intervention than is planned. The government does have a role in building those power lines. If the United States can become an e-nation with every citizen having access to a highspeed persistent connection whether at home or on the road, that infrastructure of connectivity will light the exuberance lights for a new generation of connected services. Government's role there is to create telecommunications competition we don't have today and to subsidize the last five or ten percent of connections as we did with rural electrification and telefonication because the network as a whole gains value when it is universal.

We investors and entrepreneurs have to relearn patience. Fair enough. We'll come back into the game once we can't stand to be on the sidelines any longer and when we see a future – almost no matter how distant – in which there is an exit.

Ten Telecom Tsunamis

The telecom industry five years from now will be unrecognizable. The creative destruction of the Internet broadly writ will be even greater than it has been in the last decade. The major telcos, the major television networks, and the major cablecos – if they still exist at all – will have very different revenue models than they have today. That's the good scenario. In the bad scenario the old business models are bailed out or saved by regulation to the detriment of consumers and society in general.

Here's a list of ten major drivers of change:

Universality

The driver of all drivers will be universal very broadband Internet access. Five years from now we'll all (except for those who choose to live off the net) have a minimum of 25megabit per second download speed (and that'll be the low end) when we're standing still. We'll be connected – perhaps at a slightly lower rate – when we're moving around, especially in our cars but also in planes and trains and on foot. That means that the next generation of communication services whether they be voice, entertainment, power management, information, health or something else can and will all be built assuming this universal connectivity. See here for why government should help accomplish universality sooner rather than later.

End of the Billable Minute

We won't pay for voice calls by the minutes any more than we pay for email by the word. This trend is already well underway, of course, with flat-fee VoIP based services, "free" VoIP-VoIP calls, and unlimited minute plans; but last mile monopolies have managed to keep minutes billable on many international and most mobile calls. Universal IP connectivity for both residential and mobile users will complete the bypass of these last mile bottlenecks. There'll be no incremental charges for voice, just monthly connectivity plans for bits of any sort. Gory details on why we still have billable minutes are here.

End of Copper POTs

The recession is accelerating the abandonment of landline phone service currently running at better than 10% per year. Without a major breakthrough (which could happen), copper-based DSL won't be good enough for the bandwidths we'll all need in a year or two. Line loss along with displacement of voice calls to VoIP (see above) will shrink revenue earned by the copper network, which has served us long and well, so that the carriers can no longer afford to maintain it even though it will still have many users left. That may be a mess. More here.

End of the Channel

The channel is a left over concept from the days of over-the-air TV. It's convenient for marketing reasons but not technically necessary for cable and satellite companies to deliver a set of channels; they could offer single shows or series ala carte. Today they choose not to except for events they can get a good premium for – pay-per-view. But The Internet is essentially ala carte and the Internet will deliver our entertainment, business model tbd. There may be bundles of content available both to facilitate choice and for economic reasons but children will ask "Daddy, what's a channel?" It's likely that the cablecos will convert most of their bandwidth to support generic very high speed Internet access instead of carving it up for channels.

End of Over-The-Air TV

As we know from the recent flap over the now-postponed switch to digital TV, there are people who still watch TV over the air. But the number keeps shrinking and, as more and more of that same content is available on the Internet and more and more people have sufficient connectivity (see above) to receive that content from the Internet, the economics of over-the-air TV will become prohibitive even though the broadcasters don't have to pay for the spectrum they use. It's not that the local stations'll disappear (at least I don't think they will); they just won't have antennas attached to them. The stations may be able to get a boost by subleasing the spectrum they used to use for over-the-air for generic Internet access. We'll probably end up paying people $40 each for boxes to attach their old tv sets to the Internet.

Open Spectrum

After a rocky start, the white space experiment which the FCC decreed this year will be an enormous success. This open spectrum will be extremely valuable both for fixed and mobile Internet connectivity. More open spectrum will be needed and will become available as TV goes off the air (see above). Why open spectrum is so important is here.

Bandwidth Demand

The price of providing bandwidth either over the air or through a fiber goes down roughly with Moore's law. Every year and a half the amount of bandwidth that can be provided at a given cost doubles. This trend'll continue as we all demand more bandwidth partly in order to receive all the entertainment we used to get from dedicated networks and partly for new applications. Today's five meg connections will soon be as useless as yesterday's dialup as new bandwidth-hungry Internet uses are invented and become essential and as websites are built on the assumption of higher and higher bandwidth availability. See here for more on the bandwidth required to receive "television".

Online GPSes

It'll be no more than a couple of years until every car-mounted GPS is online whenever the car is turned on. We'll get and contribute automatically to crowd-based weather and traffic reports. We'll know how long the lines are at a local attraction before we get off the Interstate – and we'll buy our tickets before we get there. The billboards will literally be inside the car. This post is about an early GPS with connectivity.

Latency Intolerance

Latency is the time between when we send something on the Internet and the time when we receive a response. Interactive voice demands low latency; so do modern web pages which build themselves on your screen through a series of interactions. High altitude (geostationary) satellites cannot provide low latency because of speed-of-light limitations so they will not be a significant provider of Internet connectivity. Local Internet providers also have routing problems to and from the Internet backbone which contribute to latency. Some measure of expected latency'll become part of the marketing description of an Internet connectivity service. More on latency and satellite here.

Smart Grid

Electricity will begin to replace imported oil and gas for home heating and transportation and some other applications during the next five years. Our total electric consumption will go up. But the fossil fuel required to create that electricity will go down as the demand for electricity goes up. A smart grid which lets us better use baseline power from hydro, nuclear, wind, and solar will accomplish that near miracle. See The Smart Grid Should be Stupid.

Beware the Coyote Syndrome

Wylie Coyote sometimes runs over the lip of a cliff but doesn't start to fall until he looks down and sees that there's nothing but air under his feet. Same thing happens with companies and industries: if they're big enough, they keep on going even long after they're really dead.

Talked to someone today who'd googled me first (wouldn't you do that with a stranger?) and found a Red Herring interview in which I'd predicted Nortel's bankruptcy:

"Tom Evslin… compares the recent changes in the telecom landscape to the shift from mainframe computers to PCs. In that transition, dollars came out of IBM's pockets and into the coffers of Microsoft and Intel. Mr. Evslin predicts a similar redistribution of wealth in the telecom industry. Dollars that once went to Lucent Technologies and Nortel Networks will find their way to companies that make soft switches, voice-over-IP equipment, and replacements for expensive hardwire switches that route phone calls…"

Trouble with this prediction is that it was made in 2002. The rest of the     quote: "It will be a painful transition. Even the most optimistic don't expect things to settle down before 2004. For many companies, that will be too late." It wasn't until the next bubble burst that Nortel went chapter 11.

You can be right about trends and get the timing very wrong. Nortel WAS already toast back in 2002 but they (and their customers and their stockholders and their bankers) didn't know that so they kept going on sheer momentum. Maybe they could have used that momentum to scramble back on the cliff as the coyote's trying to do in the picture but they didn't and companies usually don't.

So what companies and industries today are really walking dead? It's a very important question because we live in a time of bailouts when zombies are kept walking with public money and may be crowding out the startups that should take their place.

Old style investment banks have already declared themselves dead and turned into regulated banks in order to get handouts. But do we really need financial supermarkets which are "too big to fail"? I don't think so. Do we need as much credit as we've been using? No. Do we need all the exotic credit instruments? No. Do we need all the people who sell all the credit we don't need and construct the exotic credit instruments? We don't need them to do that. Perhaps we need a retraining program.

Manufacturers of cars powered by internal combustion engines, particularly manufacturers with huge unfunded liabilities for past promises and huge debts, are certainly already over the cliff. Interesting article the other day – I can't remember where – about how many less parts there are in an electric or even a plug in hybrid electric where the gas engine only drives a generator than there are in a traditional gas powered car. Point is that it's much easier for an upstart to get started up building the electrics than it would have been to compete with those who are really good at building gas engines. Now where's that cliff?

Readers say I'm way too early in targeting the extinction of copper-based voice service by the end of Obama's first term. Maybe; there is the coyote syndrome to think about. But it's also possible that the coyote (or the coyote's customers or the coyote's bankers or the coyote's stockholders) happen to look down.

Airlines Encourage Cloud Computing

When things come full circle, you know you're old. I cut my professional teeth on massive centralized computers; helped use remote access to bring computing power out of the glass room; then I retooled for the massive decentralization that microcomputers made possible; made a new career in helping to connect all those decentralized computers. Now, IMHO, computing is about to go back into the cloud.

According to a story by Nathan Eddy at eweek.com, United is planning to join American, Delta, and Virgin America in offering WiFi on US domestic flights. The story talks about being able to exchange email and rebook flights while aloft. What's more important is that WiFi en route will remove one of the last reasons to carry around much more computing power than you can get in a netbook – a small computer with long battery life, a good screen, but not much disk storage and not many onboard applications.

When you want to do word processing or number crunching on a netbook, you use an application for that purpose available online. Google and others already offer somewhat weaker online versions of the Microsoft Office applications like Word and Excel that most of us are familiar with. You or your employer save money twice – once when you buy a cheap netbook without Windows rather than a more expensive laptop, the second time when you don't buy a copy of Microsoft Office.

Up until now the appeal of netbooks was small because we all spent a fair amount of our travel time offline but still wanted to be able to work. But, if we're going to be online most of the time, netbooks get to be a very good alternative. We already can get online in our cars in much of the country – hopefully only while we're passengers – using the cellular technologies EVDO and HSPA.

Maybe smart is the new cool but saving money now means survival. My prediction is that netbooks start to make serious inroads during the current ueconomic unpleasantness. Bad news for Dell and others who will probably see smaller margins on these cheaper machines; perhaps good news for Intel if the cheaper machines (which still need good computing power to execute remote apps locally) expand the market for chips. Good news for communications providers as the need for bandwidth – especially mobile bandwidth – increases. At least temporary good news for airlines since they plan to charge $12.95/flight for access (are we really saving money then?). Very bad news for Microsoft. Perhaps bad news for Apple since it's hard to see running most apps on the small screens of most cool Apple devices and the Mac coolness will be lost if we're all running apps in the cloud.

Also bad news for parts of the country which don't have pervasive Internet access. We end up using obsolete apps on obsolete machines and paying more for the privilege. Good thing we're building an e-state here in Vermont.

Why Government Investment in Broadband Is Justified Now

Bobbie Johnson, technology correspondent for The Guardian, was kind enough to quote me along with Vint Cerf (nice to be in good company) on the importance of building an online economy and an online government. Vint said: "You know how they say opportunity lies on the edge of chaos? Maybe that's going to be true here too."

So far our telecommunications infrastructure has largely been privately built and financed. Why should that change now? It's unusual for government to do anything as well as the private sector.

The US must become an e-nation. Network economics means that even those locations that already have decent communication gain by subsidizing the locations which do not. Everyone gains from a universal transportation or communication network; even those who already have local transportation and communication. Remember Metcalfe's law: the value of a network is proportional to the square of the number of endpoints. Even if you are already connected, you gain by having the network you are connected to become universal.

Even before we were a national at all, Ben Franklin was appointed Joint Postmaster General for the Crown and, realizing the value of universal national communication, cut the time for mail delivery in half. Ironically, that postal service, subsidized by the Crown, was critically important in coordinating resistance to Parliament by the colonies.

The railroads that made the US a continental economy were subsidized by massive land grants and other government giveaways. Incidentally, there was massive fraud both in the private financing of the rest of the cost and in the competition to get the government money. Public financing, any financing where there is lots of money involved, is tough to get right and keep honest.

Rural electrification and the Eisenhower Defense Highway System (the Interstates) made us the country we are today. Both involved subsidies meaning that urban areas (which were already electrified and already had highways) subsidized the buildout to the rest of the country. Both the urban areas and the rural areas benefitted.

When we rebuild highways and bridges (as we must and will), we just get back to where we should have been. When we build a communication infrastructure which is both universal and the best in the world, we build a path to the future. BTW, when we rebuild the roads it would be dumb not to make them smart roads with mobile communication everywhere available; when we rebuild our electrical grid, it's got to be a smart grid with photons of information guiding the use of electrons of energy.

Government investments ought to be made counter cyclically both because they're cheaper then and because they cushion the pain of private contraction. Clearly, this is such a time. Universal high quality broadband ought to be one of those major investments for at least four reasons:

1. The private sector has failed us. We've slipped from being the world's leader in Internet access to number fifteen and the slippage is still going on. IMHO that's due to lack of competition. Government subsidy of backbone and middle mile can build the highway on which competitive telecommunications providers reach retail customers.
2. We all gain from universal connectivity (see above).
3. If government and the private sector can assume that everyone will have true highspeed access at home and on the move, if we are an e-nation, then new government and private programs can be developed as e-programs assuming that connectivity. We make the lowest common denominator higher; we get better service cheaper.
4. If the US is the world's first e-nation, the we will be the place where many of the e-applications for the rest of the world are invented and first deployed. If we are not, we sacrifice the advantage of being a huge market for first deployment and cede that opportunity to the eurozone, China, and India.

See here for how the federal government should and shouldn't distribute money for a broadband infrastructure build.

Federal Broadband Money Should Go To States

The carriers and cablecos are hard at work trying to influence the broadband portion of the Obama stimulus package according to an article by Amol Sharma in today's Wall Street Journal. That's no surprise, of course, and their lobbyists are just doing what they're paid to do. But it would be a terrible mistake to have any of this money strengthen the duopoly whose uncompetitive nature has resulted in the US slipping from fourth to fifteenth place in broadband deployment since 2001.

The money should be given to the states with a requirement for state matching funds to assure that the states only spend federal dollars where they are willing to put some of their own very limited money. The allocation to the states ought to be similar to the allocation of federal highway funds which takes into account both population and the higher cost per capita of building infrastructure in rural areas. There should be no earmarks or congressionally-mandated "demonstration" projects. (Obama spokespeople have already promised no earmarks in the stimulus package. If Congress goes along, that will be change that I can believe in.) In 1981 and 1982 I was Secretary of Transportation in Vermont and saw firsthand both the strengths and the limitations of the federal highway program including the damage done to local planning by earmarks..

There should be very little restriction on the funds other than that they be used to build backbone and middle mile infrastructure open for any legal use by wholesale and retail providers as well as commercial users (like the highway system but with tolls). It should not be permissible for the funds to be used to have the states themselves become retail ISPs (too much big brother potential). However, separately, universal funding should be reformed and made available to the needy in order to allow them to purchase connectivity services from providers of their choice.

The states don't own the equipment to build infrastructure any more than they own the equipment to build highways: they will contract with the private sector for building and possibly operating infrastructure. They may buy or lease existing infrastructure from the carriers and cable companies that have built it; they may simply decide that in some areas there is already sufficient backbone and middle mile so that they don't have to spend any money.

The federal money should not mandate whether fiber, radio, or other technology is used to provide the needed facilities. The states'll figure that out on their own and have the opportunity to build infrastructure appropriate to their own needs. States should assure that any roadwork they do with federal highway money includes conduit for fiber whether the fiber is placed in the conduits now or not. Federal money spent to improve the electrical grid should have the proviso that high capacity fiber is part of the build.

Here in Vermont the legislature has authorized forty million dollars of state revenue bonding for communication infrastructure. The Vermont Telecommunications Authority (VTA) has already issued an RFI so that carriers and others can help shape the process of putting together the infrastructure improvements we need in Vermont. We'll go ahead with or without federal help but we'll do more faster if there is a federal allocation. On our own, we'll give Vermont equal or better telecommunications to any other state in the US; with federal money we could be part of a plan to raise the standard for the whole country so that America again leads the world in Internet quality and availability.

Full disclosure: My wife, Mary, is Chair of the VTA (a volunteer position) and I also volunteer doing odd jobs for that organization. We have a personal although not a financial stake in the success of this program. The opinions in this blog, however, are mine alone and are neither approved by or attributable to any organization.

Blocking BitTorrent in Britain

Virgin Media announced its intention of restricting BitTorrent traffic on its new 50Mbps service according to an article by Chris Williams in The Register. Does this mean that net neutrality is endangered in the UK? The question is important because advocates of an open Internet like me hold the UK up as a positive example of net neutrality achieved through competition rather than through regulation.

One of the major benefits of a competitive rather than a regulatory approach to net neutrality is that users get to decide what sort of network they would like to be on. With a regulated approach, the regulators decide. In the US the FCC has reproved Comcast for blocking BitTorrent traffic. On the other hand, we net neutrality advocates think that it is acceptable to throttle heavy users in times of network overload because this is non-discriminatory as far as applications are concerned. Heavy users who don't use BitTorrent would probably rather be on a BitTorrent-blocking network than one which blocks them. In a free and competitive market they would get a choice; BitTorrent users, obviously, would prefer a network which doesn't block BitTorrent explicitly; they would have a choice as well.

The market might satisfy both sets of users by offering them a choice of services or one or the other type of service might prove uneconomic because not enough users like it. Nothing wrong with that. Moreover, a network which blocks BitTorrent, as Comcast was suspected of doing, to favor its own entertainment content, might find itself with no users. All sounds like competitive utopia.

Reality might not be so simple, however. If there are only a small minority of users who care about BitTorrent and Virgin Media can cut costs and/or improve service for most users by restricting BitTorrent, it may gain enough competitive advantage so that other providers emulate it and BitTorrent ends up being restricted everywhere in the UK. Would that mean that a competitive market is not enough to protect net neutrality? Some net neutrality advocates would say "Yes. Any system which results in a particular legal application being banned is bad and needs to fixed."

I'm not ready to go there. I think that if we have a competitive market (which we don't in the US) and if there is no market-fixing or other arrangements between competitors to restrict competition and there is full disclosure of the rules of each network, we have to accept that the result is a neutral Internet – which might not be exactly the kind of network we Net Neutrality advocates think the world should have. We're not the fabled Internet czar either.

Net Neutrality and the Obama Stimulus Package

As long as US telecom is duopoly dominated, a neutral Internet is endangered if not impossible; regulation of this kind of concentrated power is necessary but is unlikely to be sufficient. The solution, IMHO, is to dilute the power of the duopoly so that consumers can buy whatever kind of Internet access they want. Countries like the UK with a competitive ISP market do not seem to have net neutrality problems nor require net neutrality regulation and have better Internet access than we do at lower prices.

Whatever portion of the Obama stimulus package is devoted to telecommunication should be directed away from the incumbent telcos and cablecos  – whose lobbyists are indubitably doing their job and already lining up for the lion's share of federal funds – and used to create infrastructure on which competition can flourish (ideas to follow). It will be a huge squandered opportunity and a misuse of public funds if the telecom infrastructure project ends up reinforcing the telco-cableco duopoly which now controls most of our Internet access

A Wall Street Journal article on Google's hopes to locate servers within carrier networks alleges that this plan is evidence that Google has abandoned its fervent support for net neutrality. Such a change of heart is denied by Google's Rick Witt, who points out, correctly, that what Google seeks to do for its own content is no different than what Akamai and other commercial caching providers already do for content providers. The story aroused extra passion because it also says that advisors close to President-elect Obama are "softening" their positions on net neutrality. At least one such advisor, Larry Lessig, who is justly famed for inventing creative commons licensing, denies on his blog that he has changed his mind, although he does say "Some friends in the network neutrality movement as well as some scholars believe it [his own long-held position on net neutrality] is wrong -- that it doesn't go far enough."

Larry's last sentence points out part of the problem with net neutrality regulation; it's almost impossible to write workable definitions. Fervent supporters of the concept of net neutrality disagree on what is or isn't a violation of such neutrality. There is a huge danger that any regulation would result in further advantage to the incumbents who are accustomed to using regulation to their advantage. Would you want to wait for the FCC to certify your new service neutral before you could introduce it?

On the other hand, it's easy to recognize the virtues of a neutral Internet. With a few exceptions, we've had that so far. The backbone itself delivers packets to anywhere from anywhere without trying to figure out who sent them or what they might contain.  It is wide open to innovation. It allows innovative business models whether they're disruptive or not – and whether they will ultimately succeed or not. Friend Om Malik warns "Many startups might skip over this issue, which I constantly bring up, but they need to wake up and realize that in the end they are all going to be impacted if network neutrality is backstabbed to death." He's right.

If we are stuck with the current duopoly, we will need regulation– and face the very real prospect that regulation may be ineffectual or even counterproductive. On the other hand, if we build a national telecom infrastructure upon which competition can flourish – as it does on the highways, for example – we won't need FCC regulation against discrimination any more than we need the ICC (Interstate Commerce Commission – founded to regulate Railroad transportation monopolies in 1887, RIP 1995) to regulate trucking.

An Innovator’s Dilemma – License or Manufacture?

Dash Express, the GPS with GPRS communication for automatic pooling of real-time traffic reports and a truly open application API, is one of the coolest products I've ever owned; it clearly points the way to not only the future of not getting lost but also the next stage of mobile communication and crowd-sourced data. Dash Navigation, the company behind it, recently laid off 65% of its workforce, according to Eric Shonfeld on TechCrunch, and has had to make radical and perhaps fatal alterations to its business plan.

Dash Navigation invented a radical mating of wireless and GPS technology. Existing GPSes like those from Garmin are closed systems, so Dash to built and distributed its own hardware to provider consumers a way to buy its clever technology. They made their platform open so that outside developers could add value to this cool device. They raised a significant amount of venture capital from first tier VC firms Kleiner Perkins and Sequoia. They got excellent product reviews.

But now they're in trouble. They've announced that they will no longer be a manufacturer (although you can still buy the device as I did); they will concentrate on licensing to GPS manufacturers, mobile phone providers, PDA makers etc. in order to get wider distribution. They haven't announced any licensing deals since the press release with their change of approach – although that was only a month ago. They're not doing any hiring according to their website, not surprising given the layoff they just had. They are still supporting their product as I can testify from a user POV.

Licensing is unfortunately NOT a good strategy for innovation. The licensees usually won't budget enough for the consumer education needed to sell innovation; they don't live or die by your success; they may just be covering a bet. They don't change their hardware to fit your software. You as the licensor don't have retail margins to support retail advertising since those margins go the builders and distributors of the retail product.

On the other hand, manufacturing and getting good distribution for a manufactured product is incredibly expensive and hit or miss. You run the risk of just doing enough to show fast followers how to add your features to their hardware. They don't even have to be as good as you if they have better distribution.

This is the innovator's dilemma and there's no good answer except lots of luck or lots and lots of capital. In a great economy, there's a flood of innovation and it's easy to get money but hard to get attention. In a poor economy, money is hard to come by; your VCs suddenly remember the old adage about throwing good money after bad and say they hear their mothers calling if you happen to corner them in a corridor.

The Apple Macintosh wouldn't have worked as licensed software for the PCs of its day. It needed hardware designed for its strengths; it needed large margins to support a large ad budget and initially short production runs (compared to DOS machines). Steve Jobs decided to tightly bundle the hardware and software and they've never been teased apart since. When things are going well for Apple, this is considered brilliant strategy. When things weren't going well, everyone knew this was a mistake; he should've licensed it. Now Apple has its own capital and can continue with integrated hardware and software innovation like the iPod and the iPhone. The model works when you're rich and brilliant.

Tivo was an incredibly innovative product. I still haven't seen a competitor eight years later which is even as good as the first Tivo device. Tivo has gone back and forth at least once between a manufacturing and a licensing model. It can't compete with the distribution advantage of the networks. Somehow it survives but doesn't really prosper (disclosure: I own a small amount of Tivo stock even though I don't currently own the device because the features I want don't work with DirecTV).

The Israeli company VocalTec was THE early innovator in VoIP (when VoIP was still in the carrier network and not in the home or office). Their engineering skill was in VoIP software development. They felt, probably correctly, that they had to sell VoIP devices to the carrier market to get the margins they needed and to deliver a "turnkey" solution.  But once Cisco and other became interested in delivering their own boxes, there was no longer a way for VocalTec to compete. Should they have licensed from the beginning? Easy to say with hindsight but they might well have not gotten traction nor sufficient margins nor been able to raise capital by going public if they had.

Back in the distant past my company Solutions, Inc. developed fax software for the Macintosh. We licensed it to fax modem manufacturers; we made a modest living but squandered most of it trying to promote our software (which was only available bundled) at trade shows. "You're selling what?" people would ask. At best, we made a dollar or so on each modem our licensees sold. Should we have had our own modem? Didn't have the capital. Eventually we licensed the software to Apple who used it as an upgrade to their own inferior software. Not a bad outcome but not what we were dreaming of.

I hope Dash Navigation finds a way through this innovator's dilemma. Their device IS going to shape the future. It'd be nice if they could benefit from it.