Showing posts with label broadband. Show all posts
Showing posts with label broadband. Show all posts

Wednesday, December 7, 2011

Where is the Value of a Fixed Line?

One often hears it said that “broadband is the anchor service” for fixed-line service providers in the future. One also frequently hears that new value-added services would be a healthy antidote to service providers becoming “dumb pipe” access providers. One sometimes also assumes the growing use of "connected devices" benefits mobile service providers (it does), but not fixed-line providers.



All of those statements are true, but analysts and observers might be missing the growing potential of the “dumb pipe” access business, especially as the home and business environments increasingly feature the use of many different “untethered” devices, and as more users get used to switching even their mobile devices to untethered fixed line connections (Wi-Fi). Razorsight Blog


the value of a fixed-line broadband connection will grow as each additional connected device is added.

In August 2011, for example, the share of non-computer traffic for the U.S. market increased to 6.8 percent. The largest percentage from this share came from mobile devices, which drove 4.4 percent of total digital traffic in the U.S. market. The second largest driver of non-computer traffic was the tablet category, contributing nearly two percent of total traffic.



As the share of U.S. non-computer traffic rose over the past four months, the percentage of that traffic driven by tablets has risen from little more than 20 percent to nearly 30 percent. In May 2011, 22.5 percent of non-computer traffic came from tablets. By August 2011 that figure had grown to 28.1 percent, eating into the share of traffic garnered by mobile devices and other web-enabled devices.

That is but one example of how use of connected devices is changing the value and use of fixed-line broadband connections.



In fact, the GSMA expects the number of total “connected” devices to increase from nine billion in 2011 to more than 24 billion in 2020. “Mobile connected devices” (presumably those with a subscriber information module) will grow 100 per cent from more than six billion in 2011 to 12 billion in 2020.

Saturday, December 3, 2011

Charter Boosts Speeds


In markets where Charter Communications has deployed DOCSIS 3.0 technology, about 95 percent of its service area, the company is increasing Charter Internet Express download speeds from 12 Mbps to up to 15 Mbps, and increasing upload speeds from 1 Mbps to up to 3 Mbps.

"Charter Internet Plus" download speeds are being increased from 18 Mbps to up to 30 Mbps, and upload speeds are being increased from 2 Mbps to up to 4 Mbps.

Those changes are being made with no increase in cost. In 2010, for example, cable companies outgained telcos about two to one for net additions (new customers less departing customers). 2010 net broadband additions



The speed boosts are the cable company's fourth speed increase in the last three years.

Wednesday, November 16, 2011

Future of Fixed-Line Telephony?


Generations and their gadgets - Pew Internet

It is clear fixed line telephone services in the United States are beginning a rapid decline, with users favoring mobile phones and computer-enabled telephony, some would argue. Bill Reidway, Neustar Vice President of Numbering Services Product Management Reidway, is among them.


“As the fixed line network begins to fall by the wayside” explained Reidway, “the notion of telephone numbers associated with a specific geography falls with it.” Neustar’s Vision on the Future of Telephony That doesn't mean numbers are less important, just useful in a new way. 


Reidway also explained that although telephone numbers no longer have rigid location sensitive significance, users still generally prefer to associate their phone numbers with a location, and that is particularly important for business users. While it is certainly possible for a business or individual to use an area code, or even country code from any point in the world, he believes an area code “still says something about the identity behind the number.”


One might argue that, over time, the role of a fixed network will change, with users relying on fixed networks for some services and features that are superior to wireless, including bandwidth, cost and features. Business users are likely to derive higher value from fixed line voice than consumers will, for example. 


Most popular personal consumer devices will sport Wi-Fi capability, for example, meaning that "untethered" connectivity is becoming more important over time. 


Fixed networks, in other words, will become the primary broadband connection used inside homes. Given the existence of mobile data caps, it will make sense for most consumers to switch even their mobile devices to Wi-Fi connections when at home.

Most consumer devices use, will use, Wi-Fi

Friday, November 11, 2011

U.K. Will Not Reach 30 Mbps Broadband Access Goal by 2020

FTTH Deployment Cost
Rural fiber infrastructure cost
BT Group Director of Strategy and Policy Sean Williams considers the EU target of 30 Mbps to all citizens by 2020 as "not achievable for any country." 


That doesn't mean complete or even substantial failure. In fact, one might argue the opposite. 


BT does seem to believe it will be possible to provide 30 Mbps access to about 90 percent of the U.K. population. 


For the final 10 percent of homes or locations, 2 Mbps might be more reasonable, for all sorts of good reasons related directly to the cost of building communications infrastructure in rural and isolated areas. Some might argue that the cost curve looks very much like the curve that describes the cost of providing health care to people, where most of the cost is incurred late in life. 


Likewise, the cost of building facilities to the last couple of percentage points of locations is very high. That's one reason satellite broadband providers have a business. The core market is about two percent of U.S. households, for example. 


The high cost of reaching the last 10 percent of locations in either the U.K. or U.S. markets always will be a problem, at least when using fixed networks, whether the services are narrowband or broadband . EC broadband target unreachable 


So some might argue that 90-percent coverage of the United Kingdom with 30 Mbps service by 2020 is not in any way "a failure." It is a success. But the problem with all infrastructure goals is that it always is a stretch to reach the last 10 percent of potential customers with networks of any kind.

Tuesday, November 8, 2011

Dish Will Launch Mobile Video Service?

Dish Network will use the S-band wireless spectrum it acquired via its $2.775 billion purchases of DBSD North America and TerreStar Networks to push a mobile video strategy to complement its wired video delivery service, according to Dish Chairman Charlie Ergen. Dish's Ergen: Spectrum will enable mobile video play


"So the way I look at it is, we believe that the wireless business is a place where, if we're in the video business, we need to be more than fixed, we need to be a mobile video as well," Ergen said. 


Though it also remains possible that Dish Network might look at other wireless services as well, some of that thinking might hinge, in part, on what happens with the AT&T bid to buy T-Mobile USA. 


If approved, it is expected there will be required divestitures, and that could be something Dish Network might be able to take advantage of. If the merger is not approved, there are other possible opportunities that could arise. Some of those possibilities could include the ability for Dish Network to rather quickly acquire market share in the mobile broadband and voice business, for example.


Though that might not have seemed so logical in the past, much is changing in the consumer services business. Cable TV operators, for example, increasingly are talking about broadband access as the core service, with a corresponding lessening of emphasis on video services. 


Ergen still believes there is an opportunity for stand-alone video services, as cable operators increasingly seem to be embracing the value of a "lead with broadband access" strategy.


Having said that, Ergen said the other macro trend is that people continue to use more data. And the key point is that such data could be video, could be voice, could be lots of other things. 


"So I think that strategically, we believe we have to be in something other than a standalone video business as a company, and we are in the transition of being able to do that," said Ergen.  "hat's going to take some time and that is unclear where that is going to be a smart business decision or not."


That's a rather more candid assessment than many CEOs might make, but Ergen is a straight shooter. The larger issue is that the traditional telecom and video entertainment businesses are unstable in a new way. Broadband is becoming the lead offer for cable operators as they continue to lose market share to telcos and satellite, while wireless now is the lead offer for telcos. 


Even "bundles," which have driven growth for more than a decade, though still important, are losing some marketing emphasis in favor of concentration on the lead offers. For Dish Network, a new lead offer might emerge in the future. 


Tuesday, October 25, 2011

Only 5% of U.S. Homes Do Not Buy At Least 1 Broadband Service


There’s an interesting bit of data in the most-recent Nielsen “Cross Platform” report on media behavior of U.S. consumers. The study shows that 72 percent of households buy both broadband access and a video subscription service. Broadband drives consumer spending


About 18 percent of households buy video service but no broadband, while five percent buy broadband but not video service. About five percent of households do not buy either video service or broadband.



Add it all up and just five percent of U.S. households do not buy at least one broadband service, with the dominant pattern being “video plus broadband access.”



That doesn’t mean narrowband services are unimportant, either in terms of gross revenue or profit margin. It does mean that when forecasters say the telecom business will in the future be built on broadband services, “tomorrow” already has arrived. 


Sunday, October 16, 2011

New study quantifies the impact of broadband speed on GDP - Ericsson

A new report, conducted jointly by Ericsson, Arthur D. Little and Chalmers University of Technology in 33 OECD countries, quantifies the isolated impact of broadband speed, showing that doubling the broadband speed for an economy increases GDP by 0.3 percent.

Other studies, such as a 2009 study by the World Bank, suggest similar relationships.

A 0.3 percent GDP growth in the OECD region is equivalent to $126 billion. This corresponds to more than one seventh of the average annual OECD growth rate in the last decade.

The study also shows that additional doublings of speed can yield growth in excess of 0.3 percent (e.g. quadrupling of speed equals 0.6 percent GDP growth stimulus)

Both broadband availability and speed are strong drivers in an economy. Last year Ericsson and Arthur D. Little concluded that for every 10 percentage point increase in broadband penetration, GDP increases by one percent.

Broadband impact on GDP

Tuesday, October 4, 2011

EU Wants to Slice Copper-Based Wholesale Rates


European Union telecoms commissioner Neelie Kroes wants to create a new pricing model for wholesale access to incumbent telecom provider networks that would cut prices for copper access but exempt carriers from the rules if they sell fiber optic access to wholesale customers. The new rules would create incentives for replacement of copper infrastructure. Kroes calls for greater broadband investment

Private investors have been reluctant to invest the €270 billionn Kroes estimates is needed for Europe to replace its copper access network with an optical fiber network. Broadband investment is the issue

It’s a contentious issue, as you might guess, as service  providers remain unconvinced there is adquate end user demand for new services that would justify the investment, at least for the moment. And that has investors concerned as well.

Thursday, September 29, 2011

Does FTTH Lead to Economic Growth?

If fiber to the home could show clearly that it boosts service provider revenue and reduces cost, more observers would be unabashed supporters. But large-scale deployments in the United States are relatively ambiguous, one might argue. Verizon has the overwhelming footprint and much of the total industry experience, and observers still cannot agree on whether FiOS has been a clear success or not.

Likewise, nearly everybody seems to believe that fiber to the home is required for economic development. But even there, the impact is hard to discern. David Russell at Calix tried to test the hypothesis, looking at communities where FTTH had been in place for at least five years, was deployed ubiquitously and was deployed at a regional commercial center.

Based on what is available today (data through 2008) Russell looked at the growth between 2004 and 2008 and compared the results from the towns served by FTTH with the rest of their states. It turns out that of the five (Bristol, Va./Bristol, Tenn.; Dalton, Ga.; Jackson, Tenn.; Reedsburg, Wisc.; and Windom, Minn.) only three did better in business creation than other towns in their state. In both Dalton and Reedsburg, business creation trailed other areas of Georgia and Wisconsin, respectively.

When it came to job creation, only Bristol and Dalton did better than other towns in their states. But that's not to say they added jobs. Bristol actually lost three percent of jobs and Dalton nine percent.

So only the Bristol area did better than the rest of its state (Virginia) in both job and business creation.

Unfortunately, when economic data is available for the 2008 to 2010, the data isn't likely to improve, given the effects of the Great Recession of 2008. So it is likely to remain more a matter of faith, not fact, that FTTH indeed clearly underpins economic growth.

Friday, September 16, 2011

FCC Testing White Spaces Databases

The Federal Communications Commission is testing the ability of white spaces databases to accurately correlate locations where unused TV spectrum can be used by unlicensed devices, while protecting TV broadcasts from interference.

White spaces are unused spectrum between TV stations and are considered prime real estate because signals in this band travel well, making the band ideally suited for mobile wireless devices, says Julie Knapp of the FCC.

Among the applications envisioned for such white spaces are "Super Wi-Fi", wireless broadband networks, support for video devices and services and machine-to-machine communications.

Generally speaking, there will be much more spectrum available in rural areas than in urban areas. That should spark interest by service and application providers who want to build their businesses in rural areas. There will be spectrum available in urban areas as well, but there also are many more bandwidth options in dense, urban areas.

Monday, September 12, 2011

Mobile Broadband Will be a Majority of "Broadband" Use by 2015

By 2015, more U.S. Internet users will access the Internet through mobile devices than through PCs or other wireline devices, IDC now predicts. That's the broadband parallel to what has happened with voice services, where mobile voice for many users is the predominant and preferred way to use voice, while for many others represents the only way to use voice.

The number of mobile Internet users will grow by a compound annual growth rate of 16.6 percent between 2010 and 2015.

Worldwide, the total number of Internet user will grow from 2 billion in 2010 to 2.7 billion in 2015, when 40 percent of the world's population will have access to its vast resources.

Global business-to-consumer e-commerce spending will grow from $708 billion in 2010 to $1,285 billion in 2015 at a CAGR of 12.7 percent.

Worldwide online advertising will increase from $70 billion in 2010 to $138 billion in 2015, with its share of total advertising across all media growing from 12 percent to 18 percent.

Thursday, April 28, 2011

Cable's Future Network Will be Flatter

Saturday, January 22, 2011

Mobile Has Largely Erased the "Digital Divide"

The thing about "problems" is that you have to know when a problem has been largely solved, substantially solved or is so close to being solved that one has to move on to tackle the next set of problems. Internet access and broadband are, in many ways, those sorts of problems. Availability is less and less a problem. Redesigning life and business to take advantage of the changes is where the real work awaits.

Thursday, December 23, 2010

Broadband Networks: Slim Returns

Wireline networks have the weakest returns on invested capital with a 1.5 percent gain over the last decade, argues Sanford Bernstein financial analyst Craig Moffett.

Wireless networks had a meager return of 0.3 percent. Cable garnered a 2.5 percent return. Satellite networks had the best return on invested capital at 5.5 percent.

At least in part, that's one reason DirecTV shares have trounced other companies in 8-year returns, he argues. Other stocks—AT&T, Comcast, Dish, Sprint and Verizon—have negative returns, says Moffett.

But here’s where the returns get tricky. Once you add up the costs of various telecom deals, the returns look much worse.

Friday, December 17, 2010

Not Enough Competition in U.S. Broadband Market?

Many people believe there is not enough competition in U.S. broadband access markets, which will come as news to the firms that actually compete in the market. But it's always easier to criticize somebody else's business than your own, one might observe. read more here

Sometimes the argument is that the alleged "lack of competition" means slower access speeds in the U.S. market, compared to some others. And the U.S. does feature typical speeds than some other nations do. With no exceptions, those nations are territorially small, have high population density and also tend to have had heavy financial sponsorship. The first observation means any advanced network can be built faster; the second and third observations mean any advanced network can be built more affordably.

The simple fact is that no large country, especially not any country with continental size, ranks in the very-top of broadband speeds. And there are simple reasons for that situation: Very-large networks, covering very large areas, with highly-varied population density, cost much more to build, and simply take longer.

Despite those background factors, the United States ranks about where you would expect, in line with the United Kingdom, France and Germany, for example, in most measures of broadband speed or coverage. The United States is not at the top, and likely never will be. The United States never ranked much more than 14th globally for fixed-line voice, either, and nobody seems to think voice service has been an impediment to economic growth, social equity or anything else.

The other issue is consumer demand. Broadband penetration in the United States in right in line with PC ownership. About a quarter of U.S. homes do not seem to own PCs, making broadband a rather useless product. Now that broadband adoption is up around the 70 percent level, most people who own PCs buy broadband.

The other angle is consumer demand for various speed tiers. There just isn't much demand for the fastest tiers of service, with most buyers purchasing services virtually all surveys indicate they are happy with. In other words, U.S. consumers choose services offering moderate speeds, and moderate prices, not the fastest speeds sold at the highest prices. Unless you think people are irrational, that sort of makes sense: people buy services that meet their needs, not "just because" faster speeds are available.

The FCC also notes that 66 percent of U.S. consumers already are buying access services running at bandwidths between 3 Mbps and 10 Mbps. Most service providers will tell you that this represents the bulk of current buying behavior. Will people buy services of higher capacity in the future? Most people think so. Are they likely to pay much more than they do now? Perhaps, but only if some other part of their current budgets can be shifted. There is little, if any, evidence that the percentage of household spending devoted to communications changes very much from year to year, running about 2.3 percent or so of budgets, and growing very slowly over time.

Broadband access is a means to an end. People might want the Apple iPad because it, in itself, is seen as having high value. Price has not been an impediment to robust adoption. But broadband access isn't that sort of product. There isn't the same "need" to buy the fastest service, as there might be to buy a Lexus.

One might argue that 3 Mbps is good enough for most people who pay with their own money. The Federal Communications Commission's latest report on the state of U.S. broadband access services took a look at locations by zip code, and estimated that 48 percent of U.S. households had, at the end of 2009, the ability to buy downstream service of at least 3 Mbps and upstream service of more than 200 kbps from at least three fixed-network providers. You might not say that is fast enough, or that three providers are not enough. Fair enough, but that's a value judgment.

read more here

Some 44 percent had the ability to buy such service from at least two fixed-network providers.

About 22 percent of househoulds could buy service of at least 6 Mbps/1.5 Mbps from at least two providers, while 57 percent could buy from at least one provider. Some 20 percent of U.S. households could buy service of at least 10 Mbps from at least two providers, while 58 percent could buy service from at least one provider. Some work needs to be done there, but upgrades are on-going, and those gaps will be closed.

If one adds in wireless providers, the FCC found that 58 percent of U.S. homes could buy wireless service of at least 3 Mbps/200 kbps from at least three providers, while 35 percent could buy from at least two providers and six percent had at least one provider.

But what makes a market workably competitive? That might not be a tough question in the abstract. Most people would probably agree that multiple competitors in any market are good for competition, and therefore good for consumer welfare. read more here

Matters are tougher when looking at capital-intensive industries. But how much facilities-based competition is actually possible in the wireline or mobile broadband industries?

Some would argue from experience and study that much more than two or three facilities-based competitors in a fixed-neetwork business, in a large market, is about as good as it gets. read more here



Thursday, December 16, 2010

Triple Play and Broadband Pricing: The Assumptions Always Matter

Methodology matters when any researcher attempts to make a cross-nation comparison of TV, voice or broadband access value and spending. For example, when trying to determine whether consumer prices are up, down or flat, and looking at either stand-alone or bundled service packages, the assumptions make all the difference.

A new study by the Technology Policy Institute, for example, finds that U.S. broadband prices are relatively steady.

 Researchers Scott Wallsten, TPI VP and James L. Riso, TPI senior fellow, studied about 25,000 wireline broadband plans across Organization for Economic Cooperation and Development countries as part of their analysis.

Overall, quality-adjusted prices remained relatively constant from 2007 to 2009, they find.

Prices for standalone broadband plans in the U.S. are approximately in the middle of the range of prices across OECD countries. Prices for triple play plans in the U.S. are among the highest in the OECD. And while residential prices have on the whole remained constant in the U.S. they have been declining in most other countries.

read the full study here.

The study also compared quality-adjusted broadband prices across countries and over time, and found that
U.S. standalone broadband plans (plans not bundled with voice or video services) compare favorably to other OECD countries, but that U.S. prices for triple play (plans bundled with voice and video) packages and very-fast broadband connections tend to be higher than those in other OECD countries. In addition, while residential prices have remained unchanged in the U.S., they have been falling in most other OECD countries, the study found.

One might argue that the triple-play prices are subject to the assumptions one makes. By definition, a triple-play bundle includes three products at one price, so an analyst has to attribute component prices to each of the constituent products.

Indeed, the researchers caution that the results must be viewed carefully. "The figures must be interpreted cautiously," the report says. The raw prices do not immediately translate into meaningful observations about the real world for at least two reasons, say Wallsten and Riso.

"First, the number of plans in a given country will affect the median and range of prices in that country," they say. "These simple summary statistics assume all of a country’s plans are equally important and representative, which is not the case."

Plans often are available to subsets of a country’s population of varying size, and the popularity of different plans differs even when they are available to the same population. "Notably, existing studies and sources of data on prices suffer from this problem: the prices they report may be based on plans that are not those to which consumers typically subscribe." In other words, the study deals with published rates, and not with the percentages of consumers who may be buying various plans.

Also, the plan data does not account for contracts and data caps, which makes simple comparisons difficult.

The other issue, when looking at triple play pricing, is that if prices for the other constituent services--voice and video--vary significantly, the triple play packages will reflect, to a large extent, those pricing differentials. And most observers might note that U.S. video entertainment packages cost more than equivalent services in most other markets. See study  here for a comparison of multichannel TV spending in variious countries.

The point is that all such cross-country studies are highly dependent on the assumptions and methodology used.

Few U.S. Consumers Buy 25-Mbps Services

It appears there is almost no U.S. consumer buying of the highest-speed broadband access services, according to Federal Communications Commission data.

Of services offering 25 Mbps or more bandwidth, business buyers register something on the order of two percent of total broadband subscriptions.

Consumer take rates are low enough not to register on the graph. One might argue that take rates for the higher tiers among consumers are so low only because the 25 Mbps services are not available in most markets. That's true. See http://ipcarrier.blogspot.com/2010/12/us-broadband-access-not-as-bad-as-some.html.

But even where such services are available, take rates remain quite low. Low enough, in fact, that U.S,. service providers never disclose the numbers.

Tuesday, December 14, 2010

How Big is European Demand for Fiber-Based Access Services?

European consumers value reasonable prices and high bandwidth access services. The issue is that they seem to value affordable prices more than they do higher speeds.

At least that is the conclusion one might draw from a new survey of 13,764 European online users.

When asked about the most-important considerations when choosing a new Internet service provider, 87 percent indicated "price" was important.

(click on image for a larger view)

About 62 percent said the speed of the connection was important. Most other issues, ranging from technical support, installation fees and type of connection were cited by 31 percent to 35 percent of respondents.conducted by Forrester Research.


Telco fiber services might lag the headline speeds offered by cable. Ian Fogg, Forrester Research analyst, suggests that where cable might offer 100 Mbps, telcos might offer 40 Mbps to 60 Mbps.

The issue is whether the "speed gap" will matter much. One would think it would. But in other markets where very-fast services are available, at higher prices, consumers do not seem to be buying them in large numbers. "Fast enough at a lower price" seems to be the winning formula.

Thursday, December 9, 2010

U.S. Broadband Access: Not as Bad as Some Believe

There's an interesting statistic in the Federal Communications Commission's latest report on the state of U.S. broadband access services. The FCC took a look at locations by zip code, and estimated that 48 percent of U.S. households had, at the end of 2009, the ability to buy downstream service of at least 3 Mbps and upstream service of more than 200 kbps from at least three fixed-network providers.

Some 44 percent had the ability to buy such service from at least two fixed-network providers.

Some 28 percent of households had the ability to buy service of at least 3 Mbps/768 kbps from at least three providers, while 48 percent had the ability to buy from at least two providers.

About 22 percent of househoulds could buy service of at least 6 Mbps/1.5 Mbps from at least two providers, while 57 percent could buy from at least one provider.

Some 20 percent of U.S. households could buy service of at least 10 Mbps from at least two providers, while 58 percent could buy service from at least one provider.

If one adds in wireless providers, the FCC found that 58 percent of U.S. homes could buy wireless service of at least 3 Mbps/200 kbps from at least three providers, while 35 percent could buy from at least two providers and six percent had at least one provider.

About 40 percent of U.S. households could buy service of at least 3 Mbps/768 kbps from at least three providers, while another 40 percent could buy service from at least two providers, and 17 percent could buy service from at least one provider.

That means 97 percent of U.S. homes can buy service of  3 Mbps/768 kbps from at least one provider in each area. In addition, 28 percent of those households could buy such service from at least three providers; 48 percent could buy from at least two providers and 21 percent could buy from at least one fixed provider.

About 80 percent of U.S. households can buy 6 Mbps/1.5 Mbps service from one to three network providers, and also 80 percent could buy service from at least one provider as well.

Some 79 percent of U.S. households could buy 10 Mbps/1.5 Mbps service from at least one provider.

You can argue more speed is needed, or that prices are too high: people do that. But that's a significant number of facilities-based competition for nearly 80 percent of U.S. households.

Thursday, October 28, 2010

Slight Dip in Consumer Satisfaction with Broadband Access Services

Overall customer satisfaction with residential high-speed Internet service providers has decreased slightly from 2009 due to declines in satisfaction with cost of service and offerings and promotions, according to J.D. Power and Associates.

The study finds that overall satisfaction with residential high-speed Internet service averages 634 on a 1,000-point scale, a decrease of five index points from 2009.

The study also finds that customer satisfaction with cost of service averages 584 in 2010, a 12-point decrease from 596 in 2009. Contributing to this decline are decreases in satisfaction with fairness of prices paid and ease of understanding pricing options.

“Although product performance is most important in retaining customers, the top reason they switch providers is cost-related,” said Frank Perazzini, director of telecommunications at J.D. Power and Associates.

It isn't always clear how such "satisfaction" ratings affect customer propensity to churn, though. Some products just never seen to be seen as providing high value and reasonable price. Airline service comes to mind. But most consumers are happy with broadband. See http://ipcarrier.blogspot.com/2009/12/is-broadband-satisfaction-directly.html

"Tokens" are the New "FLOPS," "MIPS" or "Gbps"

Modern computing has some virtually-universal reference metrics. For Gemini 1.5 and other large language models, tokens are a basic measure...