AI in Everyday Life: More Common Than You Think

People routinely use electricity and think nothing of it. Artificial intelligence, machine learning and deep learning are likely going to be experienced the same way. Few people can explain how electricity works, but it already powers many parts of modern life.

Equally few are probably aware of how they already use machine learning in everyday life.

Amazon recommends books; Netflix suggests a film or TV show; your email app filters spam using machine learning. Machine learning underlies consumer interactions with Siri, Alexa, Google Assistant.

Personal assistant apps on smartphones use machine learning. Portrait mode cameras use machine learning.

Social media feeds use machine learning to customize your content. Video games sometimes incorporate machine learning to vary app responses based on player actions.   

Machine learning helps Uber estimate how long a trip will take. ML also helps Uber estimate how much you are willing to pay for any particular trip. Riders are matched with drivers using machine learning.

source: KDnuggets

Machine learning powers GMail’s auto-response features. LinkedIn uses machine learning to match jobs with jobseekers. Pinterest uses ML to classify photographs and visual images.

Your bank and credit and debit card providers use machine learning to monitor your accounts for fraudulent activity. ML also powers their reminder and alert systems.

Airliners use machine learning when autopilots are engaged. By some estimates, a typical flight on a Boeing 777 is on autopilot for all but 3.5 to seven minutes in the air.

Customer interaction software uses ML to answer questions and resolve problems at inbound call centers. Search engines use ML when you seek information.

Amazon makes product recommendations for you using ML. Spotify uses ML to make music recommendations. Online ads often are served up with the assistance of ML.

Google Maps uses ML to predict the fastest routes to a destination. Smart thermostats use ML to alter thermostat behavior based on how a user has acted in the past.

ML powers workforce analysis systems that make deductions from email and other written text. Machine learning increasingly can be used to assess organizational risk.  

How AI in general, or machine and deep learning will work to enhance or create products and features are subjects likely of little immediate and compelling interest for most people, whose job responsibilities or life routines do not require any specific knowledge about AI.

AI, almost by definition, works in the background. People use computers, apps and smartphones, but few have any need to understand what happens at the component layer, inside computer rooms or servers. People do not need to understand particular coding languages to use software built with such languages.

You Might be Surprised at the Number of Gigabit Connections Small Rural Telcos Supply

Causes, industries, organizations and policies do not get support at any level of government unless there is a “crisis” or “big problem” to address. And, to be sure, rural internet access or communications services of any type are less robust than in cities, for all sorts of good reasons.

But a balanced assessment would also include the contributions thousands of small rural telcos  and internet service providers already make--in rural areas--to supply internet access under most-difficult conditions. We are talking about service providers in areas where there are 10 or fewer residences per square mile, and many operating in areas where there are but two homes per square mile.

The latest survey of 194 small telcos that are members of the NTCA rural broadband association is instructive. The average (“mean”) respondent organization has 4,355 residential voice lines; 1,493 business lines; 4,455 residential broadband connections in service and 530 business internet connections in service.

You might be surprised to learn that 23 percent of all connections made available by these rural service providers offer at least a 1,000-Mbps connection. Another 34 percent of connections offer speeds from 100 Mbps to about 999 Mbps. In other words, 57 percent of available connections operate at 100 Mbps or faster.

As typically is the case, that does not mean most customers buy the fastest services. They do not. Actual buying clusters in the range between 4 Mbps and 100 Mbps minimums.

Source: NTCA data, IP Carrier analysis  

The service area covered by such telcos is approximately 2,244 square miles. Some 60 percent of respondents have service areas 500 square miles or larger and 27 percent were at least 2,000 square miles.

Consumer Connectivity Wallets are the Constraint on Revenue Growth

Some will lament the fact that even when gigabit (or any other very-fast) internet access service is available, most consumers do not seem to buy them, when there also are choices of other services that cost less, but are not as fast. In other words, there typically is some gap between the availability of a service and consumer willingness to buy.

In the United Kingdom, for example, 94 percent of U.K. homes and businesses are in areas where fixed network broadband operating at 30 Mbps or faster is available, according to Ofcom. In such areas, just 45 percent of homes buy a service operating at 30 Mbps or faster.

According to Ofcom, perhaps six percent of U.K., homes can buy a service operating at speeds in three or four digits (300 Mbps to 1,000 Mbps, for example). If take rates resemble those of the United States, single digits are the take rates where gigabit services are available.

And such data underscores an “iron law” of the consumer connectivity business. People are only going to spend so much for connectivity services.

Average monthly U.K. household spend on telecom services fell in 2017, down by one percent in real terms to £87, equivalent to 3.5 percent of total household spend, even as households were upgrading to faster fixed network services with higher recurring prices.

That is a typical spending pattern for consumers in developed markets. Australian consumers spend about 3.5 percent of disposable income on connectivity services, for example. That same pattern can be seen in entertainment spending as well: households will only spend so much on communications or entertainment.  

Wallets are only so big, and get bigger only about as fast as overall income increases.

Average monthly spend on communications services fell by 1.2 percent from £126.18 in 2016 to £124.62 in 2017, an annual decrease of £18.72 in real terms, Ofcom notes.

Average U.K. monthly household spend on mobile voice and data services has decreased by eight percent (£4.02) since 2012, to £45.99 per month in 2017, Ofcom says.

In contrast, average monthly spend on fixed voice and internet services increased by 14.3 percent over the same period to £41.13. This is largely because consumers have migrated to faster broadband services, which tend to be more expensive than standard broadband services.

Mobile voice and data spending fell by 98p (2.1 percent), Ofcom says.  

source: Ofcom

The number of landlines fell by one percent to 33.1 million as a result of businesses switching to mobile and VoIP-based voice services. The fall in business lines was partly offset by a one-percent increase in the number of residential landlines, attributed to growing fixed broadband take-up, as most households in the United KIngdom need to buy a landline service to use fixed broadband services.

The main casualty of growing smartphone take-up has been traditional messaging (i.e. SMS and MMS), as users switch to more feature-rich internet-based messaging services, such as WhatsApp and Facebook Messenger, and the messaging services offered on other social networking sites.

By 2017, average outgoing messages (including SMS and MMS) per mobile phone subscription had fallen to 82 per month, having peaked in 2012 at 162 per month. And while average outgoing mobile call volumes per subscription have risen since 2007, reaching 157 minutes per month in 2017, this was two minutes per month less than in 2016, Ofcom says.

Global Service Provider Revenue to Grow, Free Cash Flow to Shrink

There is modestly good news for global connectivity service providers on the revenue front, through 2022: revenue growth rates are going to tick up modestly. The bad news is that free cash flow is going to dip about two percent on higher capex and opex.

And while mobile internet access can be counted on to boost revenue in developing markets, developed markets will have to rely on revenues from new 5G use cases. So there is more risk in the latter; less risk in the former.

source: A.D. Little

According to a forecast by Arthur D. Little consultants, the North American, Latin American and Middle East and Africa regions will experience service provider revenue growth above global averages.

Global service provider revenue growth will average three percent per year from 2017 to 2022, a slight uptick from growth rates between 2014 and 2016.  

“In North America, we expect total revenue CAGR of 4.6 percent from 2017 to 2022,” A.D. Little says. Latin America will see 5.2 percent CAGR, while MEA will a 4.9 percent CAGR.

The revenue growth can be attributed primarily to increased high-speed wireless data coverage in emerging markets, higher revenue from fixed infrastructure required to support aggressive increases in data consumption and new use cases enabled by 5G deployments in more mature markets.

“New revenues from 5G are expected to arise from new B2B and B2B2X use cases, as significant incremental revenues are not immediately expected from consumer 5G services,” the consultants say.  

source: A.D. Little

The baseline forecast predicts global telecoms capex growing at a CAGR of seven percent from 2017 to 2022, a rate that is more than double the historical CAGR and also outpaces the forecast growth in global telecoms revenue for the same period.

As a result, free cash flow will dip about 1.7 percent

Vodafone Upgrades 1 Million Berlin Homes to Gigabit in 3 Months

Vodafone has upgraded--in just three months--one million Berlin households to gigabit internet access speeds. Already having upgraded six million German household passings, Vodafone expects to reach 11 million gigabit passings by the end of 2019 and more than 12 million households passed by 2021.

That the gigabit upgrade is happening so fast is testament to the use of hybrid fiber coax cable TV facilities, among whose advantages has been lower cost and faster upgrades than switching from copper telephone networks to fiber-to-home.

Liberty Global has estimated the cost to upgrade to gigabit speeds at about €20 ($22) per passing.

source: Vodafone

Ofcom Wants Customers to Get Best Price

Unless they work at it, most consumers are probably unaware whether they are getting the best deal on their mobile or fixed network services, when not under contract for those services.

That can result in an anomaly: the more-loyal customers pay more for the same services than new customers just acquired on promotions.

Ofcom notes that out-of-contract prices vary based on the type of products purchased.

Mobile phones and subscriber identity modules, as well as fixed network voice prices, can range from six percent (fixed network voice) to 27 percent lower when contracts have expired.

But costs for out-of-contract dual-play or triple-play services on fixed networks can range between 19 percent and 26 percent higher, Ofcom notes.

source: Ofcom

That is the backdrop for possible Ofcom action requiring that service providers notify customers when their contracts have ended, as well as notifying such customers of the best prices available for the types of service they already are buying.

In some other markets, the potential for overpayment has been largest in instances where mobile handset sales are bundled with service contracts, when those service contracts continue even after the handsets have been paid off.

That is less a problem in markets where most handset sales are separate from service charges, of course, and where device installment plans are separate from service plan contracts.

There also is more transparency, and greater freedom to choose, when no service contracts are a typical retail billing practice.

The Ofcom proceeding is just one more example of why average revenue per user or account keeps dropping in the connectivity business. In addition to government-mandated actions that reduce consumer prices,  competition and new technology, plus changes in end user demand, combine to push prices lower over time.

Eco-Friendly Buildings Make Indoor Communications Worse

A recent report by the Ireland Commission for Communications Regulation  finds that the use of modern building materials,windows, block materials and roofing can have an extraordinarily detrimental effect on the propagation of radio waves into buildings constructed using these materials, reducing indoor signal strength three to seven orders of magnitude (100 times to 1,000,000 times weaker signal).

The losses suffered by radio waves penetrating these materials is on the order of 20 decibels to 60 decibels.

The basic math is that a 3 dB reduction in signal is a loss of half. So 20 decibels of loss represents seven consecutive reductions of 50 percent. A loss of 60 dB is 20 consecutive reductions of half the signal strength.

source: ComReg

The problem is that  thermal insulation or windows with aluminium or metallic frames, designed to help reduce heat loss from inside, also reflect incoming radio signals.

That is why some believe new forms of supplying indoor mobile coverage are a significant business opportunity. The question is how big an opportunity  neutral host indoor communications might be, and for whom in the ecosystem.

That might be especially in settings where enterprises and organizations decide to replace Wi-Fi with 5G as the local area network platform.

Enterprises might well be able to use 5G as a replacement for Wi-Fi, argues Randall Stephenson, AT&T CEO. “5G will become the way businesses network,” he says. “Wi-Fi probably goes away.”

The old distinctions between indoor connectivity and public network services, blurred with the advent of mobility to an extent, are changing.

“We’re seeing a lot of demand from enterprise customers for blurring the line between what has historically been a wide area network, mobile, with a local area network, which has traditionally been wired,” said John Donovan, AT&T Communications CEO. Private 5G and 4G networks, indoor small cells and 5G network features all are combining to create new possibilities.

Industrial internet of things networks on the factory floor might well use private 5G instead of Wi-Fi. In other cases 5G small cells might be operated by integrators or public networks.

If one assumes outdoor space will be the place where mobile coverage is most valuable, indoor space will remain a more-contested arena where access options will be more diverse, where third parties will have a greater role, where the ability to support private network features at the indoor edge will open up new possibilities for end users, mobile operators and third parties.