Nvidia GPU Performance Increasing Faster than Moore's Law

Graphics processing unit improvement seems to be occurring faster than Moore’s Law would predict. Nvidia’s GPUs today are 25 times faster than five years ago. If they were advancing according to Moore’s Law,  they would have increased their speed only by a factor of 10, says Nvidia CEO Jensen Huang.

Between November 2012 and May 2020, Nvidia’s chip performance increased 317 times for a major class of AI computing, says Bill Dally, chief scientist and senior vice president of research at Nvidia. On average, the performance of these chips has more than doubled each year.

How AI Can be Applied by a Connectivity Provider

Artificial intelligence is a capability, not a product. 

“You don’t focus on ‘I’m going to go do AI,’ says Peter Guerra, North America chief data scientist at Accenture. “You focus on ‘I’m going to do supply chain better, and I’m going to leverage AI to do that.’” Peter Guerra, North America chief data scientist at Accenture.

In the connectivity business, potential applications include network operations monitoring and management, predictive maintenance, fraud mitigation, cybersecurity, customer service, marketing virtual digital assistants, customer relationship management preventive maintenance and battery optimization, for example. 

In the network operations monitoring area, that might include anomaly detection for operations, administration, maintenance and provisioning (OAM&P), performance watching and optimization, alert or alarm suppression, bother price ticket action recommendations, automated resolution of bother tickets, prediction of network faults or congestion prediction, for example.

What Does Your Business Look Like if the Key Constraint is Removed? You Get Microsoft, Netflix, Google, Facebook, Amazon

What does your business look like if the key constraint is removed? It is a question so challenging--and often so seemingly impossible--that most of us never ask it. As a much younger profit center manager, I was once asked “what would it take to destroy our competition.” Truth is, I had no immediate answer. 

There was a most-likely answer. But I also knew that one thing would “never happen,” as the stakeholders who could authorize the change were--and remained--steadfastly unshakeable in supporting our only competitor. 

But young Bill Gates, before Microsoft was a household name--in fact when the company name actually was Micro-Soft, and was based in Albuquerque, N.M.-- did ask the question. So did Reed Hastings, founder of Netflix. 

If you are a user of computing or communications, what is your behavior if both are nearly free? 

If you are a supplier of computing or communications, what are the implications for your business? What business are you in, and what business should you be in?

Those remain the key strategic questions for connectivity providers.

“The original insight for Microsoft was this: What if computing was free? ” Bill Gates, former Microsoft CEO and chairman, once said. He has said similar things several times, more recently about price trends in communications. 

In 2004, then Microsoft Chairman Bill Gates argued that "ten years out, in terms of actual hardware costs, you can almost think of hardware as being free--I'm not saying it will be absolutely free--but in terms of the power of the servers, the power of the network will not be a limiting factor," Gates said. 

About a decade before, in 1994, Gates mused that  “we’ll have infinite bandwidth in a decade’s time.” In 1995, Gates said “And for this new era, communications is what’s becoming cheap.  So you get to thinking, well, what if communications was free, what could people do?” 

My own analysis is that Gates believed in Moore’s Law and its impact. That analysis would lead you to believe that computing costs would in fact decline substantially, every 18 months, upending assumptions about the use of computing.

Reed Hastings was very clear about the application of Moore’s Law to the foundation of Netflix. At a time when dial-up modems were running at 56 kbps, Hastings extrapolated from Moore's Law to understand where bandwidth would be in the future, not where it was “right now.”

“We took out our spreadsheets and we figured we’d get 14 megabits per second to the home by 2012, which turns out is about what we will get,”  Reed Hastings, Netflix CEO, said.  “If you drag it out to 2021, we will all have a gigabit to the home." So far, internet access speeds have increased at just about those rates.

So as crazy as it might seem, what Moore’s Law has enabled, in computing, communications, applications, hardware and business opportunities, is precisely a clear understanding of what is possible if the key constraint in any business is removed. 

Near zero pricing is the term I use to describe the larger framework of connectivity provider pressures towards ever-lower prices. Others might prefer to emphasize marginal cost pricing. The point is that there is a reason the phrase dumb pipe exists. What we need to remember is that dumb pipe now is the foundation of the whole connectivity business. 

A caveat is that what people usually mean by “dumb pipe” is that a product is sold at low prices and generates low profit margins. But think about it: industry revenue growth now is lead by broadband services (internet access), which is, by definition, a dumb pipe service. It is a way to get access to applications, not an actual application itself. 

You might call that trend another example of the impact of Moore's Law on business and economics. And near zero pricing is a big industry issue. It might be the single-biggest issue. 

In a recent survey by Telecoms.com, the number-one threat to long-term business success was “increased pressure to lower prices” and “lower profit margins,” for example. 

source: Telecoms.com 

Agility or “speed” was also a major concern. Third on the list was competition from webscale firms including Google, Amazon or Microsoft. 

But there are good reasons why “lower prices” and “lower profit margins” are the top issues. Simply, they are the most-important result of other industry threats causing the price compression and lower profit margins: competition, the shift to internet protocol as the next-generation platform and the embedding of the whole connectivity function within the larger internet ecosystem.

Aside from deregulation of the telecom industry, which lead to competition and price competition, technology is among the root causes of price pressures. Near zero pricing is a scary thought for a connectivity provider, but it is reality.

Near Zero Pricing Remains the Top Issue Connectivity Providers Face

Near zero pricing is the term I use to describe the larger framework of connectivity provider pressures towards ever-lower prices. Others might prefer to emphasize marginal cost pricing. The point is that there is a reason the phrase dumb pipe exists. What we need to remember is that dumb pipe now is the foundation of the whole connectivity business. 

A caveat is that what people usually mean by “dumb pipe” is that a product is sold at low prices and generates low profit margins. But think about it: industry revenue growth now is lead by broadband services (internet access), which is, by definition, a dumb pipe service. It is a way to get access to applications, not an actual application itself. 

You might call that trend another example of the impact of Moore's Law on business and economics. And near zero pricing is a big industry issue. It might be the single-biggest issue. 

In a recent survey by Telecoms.com, the number-one threat to long-term business success was “increased pressure to lower prices” and “lower profit margins,” for example. 

source: Telecoms.com 

Agility or “speed” was also a major concern. Third on the list was competition from webscale firms including Google, Amazon or Microsoft. 

But there are good reasons why “lower prices” and “lower profit margins” are the top issues. Simply, they are the most-important result of other industry threats causing the price compression and lower profit margins: competition, the shift to internet protocol as the next-generation platform and the embedding of the whole connectivity function within the larger internet ecosystem.

Aside from deregulation of the telecom industry, which lead to competition and price competition, technology is among the root causes of price pressures. 

The most-startling strategic assumption ever made by Bill Gates was his belief that horrendously-expensive computing hardware would eventually be so low cost that he could build his own business on software for ubiquitous devices. Basically, I believe he asked himself what his own business would look like if computing hardware was free. 

How startling was that question? Consider that, In constant dollar terms, the computing power of an Apple iPad 2, when Microsoft was founded in 1975, would have cost between US$100 million and $10 billion.

The point is that the assumption by Gates that computing operations would be so cheap was an astounding leap. But my guess is that Gates understood Moore’s Law in a way that the rest of us did not.

Reed Hastings, Netflix founder, apparently made a similar decision. For Bill Gates, the insight that free computing would be a reality meant he should build his business on software used by computers.

Reed Hastings came to the same conclusion as he looked at bandwidth trends in terms both of capacity and prices. At a time when dial-up modems were running at 56 kbps, Hastings extrapolated from Moore's Law to understand where bandwidth would be in the future, not where it was “right now.”

“We took out our spreadsheets and we figured we’d get 14 megabits per second to the home by 2012, which turns out is about what we will get,” says Reed Hastings, Netflix CEO. “If you drag it out to 2021, we will all have a gigabit to the home." So far, internet access speeds have increased at just about those rates.

As frightening as it might be for executives and shareholders in the telecommunications industry, a bedrock assumption of mine about dynamics in the industry is that, over time, retail prices for connectivity services also will trend towards zero.

“Near-zero pricing” does not mean absolute zero (free), but only prices so low there is no practical constraint to using the services, just as prices of computing appliances trend towards lower prices over time, without reaching actual “zero.”

Is Work from Home Really "More" Productive, or "Less?" How Can We Even Tell?

Though many employees express a desire to continue working from home permanently, even after there is no Covid-19 reason to do so, it also seems clear that work from home fatigue is setting in. That might also be accompanied by the end of the initial period where firms claimed that productivity had not been affected.

A survey of 365 U.S. workers by The Manifest found that 30 percent of respondents think they are more productive working from home. Fully 45 percent believe they are more productive working in an office, and 24 percent say they’re equally productive working from home and in an office. Keep in mind those are examples of what people think about their productivity. It is not an actual measure of whether they are, or are not, more productive in home or office settings. 

source: Agility PR 

Some have argued that even if productivity can be sustained for a brief period, it might well not sustain itself over the long term. A survey of the heads of nearly 50 U.S. businesses employing 443,895 people in industries that include technology, legal services, advertising and the finance, insurance and real estate sector found that 40 percent of them have started to see decreases in productivity as staff work remotely. 

To be sure, Vocon is in the office design firm, with a vested interest in continued office work. But it should always have been clear that emergency, short term workarounds during the mandatory work from home period would eventually begin to test whether productivity actually could be sustained. 

Most of us can sustain productivity for short periods of time. What never is so clear is whether we can do so if a long-term disruption happens. 

J.P. Morgan also seems concerned that WFH productivity is slipping, while others note growing evidence of mental health issues caused by enforced WFH. 

At least, that might be true for the 25 percent of workers who can work from home.  

Also, there is a growing sense among professionals that WFH is hurting their career progression. Others might point out that WFH has positives and negatives, not including burn out or work-life balance, even if WFH once was viewed as a perk.   

“Effective leadership is all about presence and persuasion,” says Bob Fisch, Specialty apparel retailing pioneer. “The best mentoring happens close-up, with a senior executive encouraging a younger associate, and with the associate embracing the unique opportunity to offer the leader her own perspective.”

“It is that irreplaceable intimate element of mutual mentoring that is lost outside the office,” Fisch says. 

The percentage of those indicating they would like to continue to work remotely at least occasionally declined from more than 80 percent in July to 67 percent in August in the United States, according to the IBM Institute for Business Value. The percentage of respondents who say they would like this to be their primary way of working dropped from 65 percent in July 2020 to just 50 percent in August 2020.  

Information worker productivity is notoriously hard to measure, as output itself is intangible. So key performance indicators might not even be possible. Quantifiable measures such as “emails sent and received” might not measure anything of true importance. 

In fact, even outcome measurements might be more valid than output measures. A policy paper might be the output of a person or team. But its value comes if there is a policy outcome. It might therefore seem obvious that tracking qualitative inputs (subjects interviewed, hours spent interviewing, number of interviews and so forth might be nearly irrelevant. 

Output is what people do; outcome is what they achieve. In other words, for knowledge workers, input measures are generally unhelpful. But even output metrics might not be germane. Only outcomes might really matter.

Fixed Wireless or Mobile Access are the Best Ways U.S. Telcos Can Gain Home Broadband Share

Fixed wireless is not a new platform for internet service providers. It has been common in many rural areas for wireless ISPs, using unlicensed spectrum and point-to-point or point-to-multipoint networks. But many argue fixed wireless will be a more-common platform for a wider range of service providers in the 5G.

There are a few good reasons. First of all, use of huge amounts of new millimeter wave spectrum, open source radio technology and the relatively high cost of fiber-to-home networks in highly-competitive markets make 5G fixed wireless more attractive.

For starters, the initially installed cost of fixed wireless--even before 5G and millimeter wave spectrum availability--is lower than fiber to home, by a substantial margin. That is true for both the cost of passing a location as well as connecting a location.

source: Siklu 

There is more. One crucial business model issue is the amount of stranded assets when deploying any new cabled network in a market where there is robust competition. To use an example, in a market where a cable company has 70-percent market share and a telco has 30 percent market share, a new FTTH network might still mean stranded assets (no customer attached) of at least 65 percent, assuming the telco can gain about five points of market share. 

In other words, in highly-competitive markets where a competent competitor already has 70 percent of the installed base, fixed wireless might offer the only hope of remaining competitive, as the cost of FTTH might never prove sustainable. 

Verizon’s FiOS FTTH network, even after years of marketing, topped out at about 40 percent take rates until perhaps 2018, when take rates broke above 40 percent. Keep in mind FiOS was introduced in 2005. 

AT&T’s FTTH adoption rates seem to have been historically lower, at about 25 percent, though AT&T execs hope they eventually can boost those rates to somewhere in the 40-percent range over time (where FTTH is available). AT&T execs have sometimes said they believe they can reach 50-percent take rates. Some of us highly doubt that. 

With cable operators having 70 percent of the installed base, typically leading in bandwidth and offered speeds, and with the next generation of cable modem service already being developed, to supply more symmetrical service at 10 Gbps, it does not seem that even FTTH is going to change the value proposition. 

Essentially, FTTH only allows a telco to play catch up. So long as cable operators keep investing to maintain their lead, it seems unlikely FTTH will allow telcos to leapfrog the competition. Keep in mind that cable operators have been the U.S. broadband access market leaders since at least 1999. 

The FTTH business case, for consumer customers, now seems irreparably damaged. But that is why AT&T, T-Mobile and Verizon are so focused on ways to use 5G and future networks to challenge all cabled networks as platforms for internet access. In the U.S. market, it may no longer be possible for at-scale FTTH to compete sustainably with cable operator services.

That includes using fixed wireless or even standard mobile access instead of FTTH. Though it might not have been so clear 20 years ago, once cable operator hybrid fiber coax emerged as a platform for consumer internet access, even FTTH was not going to be enough for telcos to remain highly competitive. In recent years, all net gains in accounts have been garnered by cable companies.

How Much Post-Covid Change Will Businesses Achieve?

One often hears it said these days that one impact of Covid-19 on organizations and firms is that it will cause permanent changes in the ways businesses and organizations work. Most of those changes, though--agility, reaction speed, cost reduction, productivity changes, customer focus, innovation, operational resiliency, growth, financial performance, for example--were important organizational objectives before the Covid-19 pandemic. 

Even remote work on a full-time, part-time or episodic basis is a trend decades old, if many believe the difference is that a large percentage of information workers will shift to permanent remote work settings, and a larger number of observers might agree that many information workers will routinely work more often from home. 

One might note that very few of the key changes executives now say they have--because of their experience with the pandemic--can be addressed directly by better or more use of communications services. A recent survey by McKinsey found that “speed” was the driver of organizational changes related to Covid-19. 

source: McKinsey

What is not so clear is how much actual change has been accomplished, as organizational change normally is very difficult and also takes a long time. Consider the extensive changes needed to increase speed and agility.

Organizational silos, slow decision making, and lack of strategic clarity, for example, are impediments to speed. But do you really believe that big firms have been about to abolish silos, speed up decision  making and gain new strategic clarity in a few months' time?

Big barriers exist for real reasons. If they were easy problems they would have been fixed long ago. Few would doubt that executives say these long-standing issues are being addressed. But also, few of us might believe real progress is being made, fast. 

source: McKinsey

Rigid policies and formal hierarchy also are cited as impediments to speed. Have you heard of massive reductions of top level and mid-level management over the past few months?

Or consider the impressionistic claims some might make. “Higher meeting attendance and timeliness” resulted in faster decisions,” one survey respondent says. Do you really believe that? More people in meetings produced faster decisions? Much of the literature specifically argues that more people in meetings. reduces decision-making ability. 

It might be fair to hypothesize that meetings, as such, have had no discernible impact. Does anybody really believe holding more meetings improves output? In fact, there is evidence tot he contrary: more meetings mean less time for getting the actual work done. 

A team of researchers said this: “We surveyed 182 senior managers in a range of industries.  65 percent said meetings keep them from completing their own work. 71 percent said meetings are unproductive and inefficient. 64 percent said meetings come at the expense of deep thinking. 62 percent said meetings miss opportunities to bring the team closer together.”

Another leader notes that “communication between employees and executives has become more frequent and transparent, and as such, messages are traveling much more efficiently” through the organization. 

There already are signs of what might be called collaborative overload. “In most cases, 20 percent to 35 percent of value-added collaborations come from only three percent to five percent of employees. 

Colloquially, that proves the truth of the observation that “if you want something done, give it to a busy person.” The practical observation is that the highest performers are besieged with the greatest amount of demand for time spent in meetings and on work teams. At some point, the danger is that these high performers simply get asked to do too much, reducing their overall contributions and effectiveness. 

source: Harvard Business Review

Not to belabor the point, but value--assuming the insight is correct--can be gleaned by having many fewer people in meetings. 

Nor, for that matter, is it entirely clear how greater numbers of  team members working remotely actually addresses any of those aforementioned issues. Some changes could materially impact performance in a positive way. But the issue is how remote work, for example, materially abolishes silos, speeds up decision making, produces strategic clarity, abolishes rigid policies or reduces hierarchy. 

Some will argue that “employees like it.” The more accurate statement could be that “some employees prefer work from home, and some do not prefer it.” But productivity is not a matter of what people believe. It is a matter of fact, to the extent we can measure it. 

Productivity. is often hard to measure--perhaps almost impossible for information workers--and employees claiming they are productive does not make it so. Also, what we can measure might not actually correlate well with actual output. Quantity is not quality, in other words. Innovative ideas and creativity might not be measurable at all, except by reputation. 

Productivity measurements in non-industrial settings are difficult, as it often is difficult to come up with meaningful quantitative measurements that provide insight. We might all agree that not all tasks create the same value for any organization. Productivity can be described as the relationship between input and output, but “output” is tough to measure. 

And what can be measured might not be relevant. 

Even some who argue work from home productivity is just as high as “in the office” note that work from home productivity is only one percent lower than in the workplace.  

Some argue productivity now is higher or equal to productivity when most people were in offices. Some of us would argue we do not yet have enough data to evaluate such claims or evaluate the sustainable productivity gains. It is one thing when all competitors in a market are forced to have their work forces work from home.

It will be quite something else when WFH is a business choice, not an enforced requirement. As might be colloquially said, widespread WFH will last about as long as it takes for a key competitor not working that way begins to take market share. 

Not to deny that post-pandemic, some firms will find meaningful ways to restructure business processes to gain agility, productivity and speed, but the actual gains will be slower to realize than many expect, as organizational resistance to such changes will be significant. Resistance to change is a major fact of organizational life.