Trademarking common notions, Part I
An episode in Manufacturing Demand
It is difficult to say whether or not smart meters are cost-effective for individual households without knowing more about the specific circumstances and factors involved. Some potential reasons why smart meters may not be cost-effective for individual households include the cost of purchasing and installing the meters, the potential savings on energy bills, and the availability of government incentives or subsidies. Additionally, the level of usage and the type of energy plan a household has may also impact the cost-effectiveness of smart meters. Ultimately, the cost-effectiveness of smart meters for individual households will depend on a variety of factors and will vary from one household to another. (ChatGPT, 12/15/2022)
Vehemence and frustration
For more than a year I have been waiting for an opportunity to explain that much of the vehemence of my rejection of the rhetoric that accompanies current projects to mechanize mathematics — NOT, please note once again, of the projects themselves — is rooted in my frustration with the very similar rhetoric that electric utilities have used, often dishonestly, to promote their Smart™ Meters. That ™ is my addition; this episode of Silicon Reckoner and the next will be devoted to the tech industry’s project of trademarking words in common usage by attempting to create the impression that they have developed improved versions of the notions these words designate, so that in a sense they own the word. This project has been implemented successfully with the words search™, cloud™, like™, friend™, as well as smart™; the process is well under way with intelligence™. If nothing is done to slow them down soon there will also be understanding™.1
Readers may remember that understanding is a value so prized by mathematicians that the word rose spontaneously to the top of the frequency chart I compiled after last October’s Fields Medal Symposium. I now finally have an excuse to talk about Smart™ Meters because the topic fits with a significant fact about Google that was just pointed out to me — that Blaise Agüera y Arcas, who leads Google’s AI group in Seattle, has been arguing that Large Language Models (like ChatGPT, or Google’s LaMDA) “understand us.” This claim, if true, would clearly have important implications for automated theorem proving, and I’ll return to it in the next installment. (Spoiler: I will be arguing that Agüera y Arcas’s claim amounts to substituting understanding™ for whatever we might call understanding.)
Trademarking intelligence™
I’m gradually coming to the conclusion that the deepest of the challenges AI poses, to mathematics and to social organization more generally, are neither technical (how will the new technologies be integrated with existing modes of social organization?) nor philosophical (how does the introduction of the new technologies modify our view of mind or of human values, including understanding?). The deepest challenges are political: how will the relevant decisions be made?
The potential arrival of world-changing technologies sheds a harsh light on existing structures of democratic decision-making. The situation is not new; old and often elaborate ways of life have fallen victim to demands for power generation, transportation networks, extraction of raw materials, and other preconditions for industrialization. What’s different with the development of AI, or so we’re told, is the scope of the expected changes. Every aspect of social organization is set to be transformed, and the structures are not in place to guarantee that these transformations will be in whatever is conceived to be the general interest. On the contrary: it’s impossible to miss the structures that are rapidly being developed to guide these transformations in the interest of the world’s wealthiest and most powerful industry. When this industry is characterized by tax avoidance, anti-union retaliation, unacknowledged surveillance, political manipulation, and all those other practices too familiar to list here, it would defy common sense to expect it to pay more than lip service to the general interest.2
How to evacuate cities of urbanity
Adam Greenfield, author of Against the Smart City3, quotes a Siemens brochure:
Several decades from now cities will have countless autonomous, intelligently functioning IT systems that will have perfect knowledge of users’ habits and energy consumption, and provide optimum service…The goal of such a city is to optimally regulate and control resources by means of autonomous IT systems.
He draws out the implications of this design:
the single overriding idea at the discourse’s very core: civic governance conceived as the centralized computational management of a ramified armature reaching into everything a city is and does.…Unless it is challenged and alternative models successfully fleshed out, this is simply how municipal services will be delivered, tensions between constituencies adjudicated and the experience of public space shaped for many years to come … the articulation of this vision was only possible in its classic form on a site evacuated of content, evacuated of history, evacuated of politics — ultimately, evacuated of urbanity entirely.
Jean Haëntjens observes4 that such a vision “leaves no space for the political city.” In the place of the existing city “which leaves room for the unexpected, the human, spontaneous creativity, the local context,” the IT vision is of the “service city,” branded and “disneyfied.” He adds that the Smart™ City sees only jobs where the living city sees vocations. Unlike a job, a vocation
produces and invents beauty and wonder… gives meaning … creates a connection… these three qualities cannot be automated.
Parallels with plans to mechanize mathematics are obvious and will be spelled out in future posts. In what follows I focus instead on the means by which one component of the Smart™ City has been promoted, and implemented, often in spite of concerted opposition.
Smart™ness enters the home
No FOMO could have convinced me to set up Alexa when the new TV arrived. A salesman at P.C. Richards discouraged me from buying the model with internet capability ($200 more), and I found that surprising, but I had no intention of buying it in any case. But I suspected that I was cornered — correctly, as it turned out — when ConEd (Consolidated Edison), supplier of electricity and gas, announced that
What should I know, you are wondering?
Think of a smart™ [™s here and below inserted by MH] meter as your tool for more choice, convenience and control… Smart™ meters provide detailed information about your energy usage so you can make informed choices about how and when you use energy in your home or business.
Also
NO MORE WAITING at home for a meter reading. No more estimated bills
and a promise to “help the environment” in some unspecified way.
That’s what ConEd tells its customers, in an attempt to manufacture demand. The quarterly trade magazine Smart™ Energy International says much the same thing
Smart™ metering provides benefits for the whole energy system and is fundamental to the development of smart grids. Most importantly, smart™ metering provides savings, empowers consumers, and strongly contributes to a more efficient use of energy, ultimately reducing greenhouse gas emissions.
but then goes on to speak more directly to the interests of its audience:
Smart™ meters also generate huge amounts of data, which enables value creation and supports the digital transformation of utilities’ operations and creating innovative business models and solutions for market players.5
ConEd is at least prudent enough not to promise “savings.” A 2016 Guardian article entitled “Smart™ meters: an energy-saving revolution or just plain dumb?” claimed that
Their introduction is set to cost consumers £11bn, but studies show they cut energy consumption by 3% or less…
Smart™ Energy International reported in 2018 that this is not specific to Britain:
In the summer of 2013, the German Ministry for Economics contracted EY (formerly Ernst & Young) to perform a cost-benefit analysis of a nationwide smart™ meter rollout. This analysis was performed as a function of the EU Directive 2009/72/EC, which stipulates that contingent upon a positive economic assessment, member nations must pursue 80% smart™ meter penetration by 2020. EY concluded that pursuing EU smart™ metering targets would result in a negative cost-benefit ratio and would be economically unreasonable for the majority of German consumers. These findings were not exclusive to Germany; six other member states (Belgium, Czech Republic, Latvia, Lithuania, Portugal, and Slovakia) also calculated negative or inconclusive assessments from the full deployment scenario.6
Ultimately it was decided that “large consumers with average annual consumption in excess of 10,000 kWh will be required to install smart™ meters” in 2017, with the threshold dropping to 6,000 kWh in 2020. Fortunately for Germans, that EU Directive allowed an escape clause in the absence of a “positive economic assessment”:
The majority of German households will remain unaffected, given average consumption hovers around 3,500 kWh.
No such luck in France. Although I have seen no evidence of a “positive economic assessment” in France, a November news item published by the official government website warned the unruly French that
L'absence de compteur Linky en 2023 pourrait vous coûter près de 50 € par an !
Trans: “It could cost you nearly 50 € a year in 2023 if you don’t have a Linky.” “Linky” is the cute name the French authorities chose to help manufacture demand for the obligatory smart™ meter, undoubtedly unaware of this highly apt definition on urbandictionary.com:
Someone who links around all day doing nothing, but pretending to be a productive member of society.
That same official government website claims that 90% of electric customers in France had installed their Linky by the end of 2021. Much opposition to the technology that remains is motivated by a fear of the effects of the electromagnetic radiation the Linky supposedly emits. A more realistic worry, to my mind, is one of the main benefits to the energy sector advertised in Smart™ Energy International: the development of “innovative business models and solutions for market players,” based on the “huge amounts of data” the smart™ meters generate. Security of this data has been a constant concern.
…an attacker, by having access to the electricity consumption, can execute a set of Non-Intrusive Load Monitoring (NILM) algorithms to infer when a household turns on the heater(s), has breakfast, leaves late for work, or even the number and the age of the people living at home.7
Imagine the “innovative business model” that can be based on data like this:
Our research shows that the analysis of the household’s electricity usage profile at a 0.5s^{-1} sample rate does reveal what channel the TV set in the household was displaying.8
My very own smart™ meter
Back in the US,
A September 2022 white paper by the Mission:data Coalition details broken promises to give consumers access to their own data. The laws that determine requirements to install smart™ meters in the US, which differ from one state to another, have not been modified in response to this information. Bills have been introduced in the New York State Assembly or Senate in 2015-16, 2017-18, 2019-20, and 2021-22, by four different Assemblymembers or Senators, all with the same guarantee:
It shall be the right of every customer of an electric and/or gas corporation, at no penalty, fee or service charge to decline the permission of his or her electric and/or gas corporation (a) to replace an existing meter at such customer's premises that is assigned to such customer's account with a two-way smart meter, or (b) to install any two-way smart meter device at his or her property without such customer's consent.
None of these bills has got out of committee, and in the meantime, New York’s rule is similar to the one in France. We have the right to opt out, but there is a penalty. An article published a year ago in a local newspaper entitled “ConEd’s smart meters don’t seem so smart after all”9 indicated that the penalty was more than twice that imposed in France
if … convenience and accuracy wasn’t enough, ConEd pushed to get meters installed by telling customers who refused the new meters that they could possibly face financial penalties.
That could come in the form of a $10 monthly fee.
As it happens, there was no way to exercise even the right to opt out in my building; I returned from travel to discover that the landlord had allowed ConEd to replace all the stupid old meters by the new smart™ kind.
I spotted something that I believe is my personal smart™ meter is in the basement of my building, but I don’t know what I would do with the information if it were confirmed.
Smart™ meters in New York City: an unofficial narrative
In conclusion, here is my rough transcript, in part a paraphrase, of a report on ConEd’s Smart™ Meter program, delivered by Josh Karpoff to the 2019 NY Historical Materialism Conference. I have not attempted to correct possible errors in the transcription, nor to make complete sentences out of my rough notes.
In 2017 Consolidated Edison announced plans to install so-called Smart™ Meters in all customers' homes and businesses. This was part of an industry-wide trend.
They want to consolidate all the data from Smart™ electric and gas meters as well as Smart™ thermostats and appliances and sell it to their partners.
ConEd is one of the largest publicly-traded utilities companies in the US (it is listed on the New York Stock Exchange). Annual revenues $12 billion, assets $54 billion, profitable 45 straight years. It is the sole provider of energy in its territory (the 5 boroughs of New York City plus several suburban counties).
Its aging physical plant, on average 50 years old but up to 100 years old, is at the end of its useful design life. ConEd is heavily in debt, with a huge deferred payments backlog; it needs massive capital investment for upgrades. There have been upgrades in the distribution system through regular emergency maintenance but the bulk of the system dates back to the postwar boom or before. In the early part of its corporate life ConEd was flush with cash from investors. Seeking opportunities in a rapidly expanding industry, and as electricity became the dominant source of industrial power, it expanded to suburbs; later new investment slowed so it was able to increase its rate of profit by increasing its coverage territory by buying up competitors in neighboring counties.
As those opportunities become more limited it found itself with a fairly new system, not needing much maintenance; it kept profits up by reducing its workforce to the minimum needed for day-to-day operations. Management sought new technologies to increase the productivity of the remaining workers. Centralized and automated systems like supervisory control and data acquisition (SCADA) allowed a few operators in centralized control rooms to control the entire service grid, eliminating the need for local operators in every substation.
Eventually that too no longer worked to increase profits. So they fought regulation and won the right to deregulate the energy supply market in New York State. That unburdened them from having to own their own power plants. So the environmental obligations the EPA created for them got taken off their books.10 The subsidiaries were bought up, in some cases by foreign companies, but deregulation also allowed competition in.
So now ConEd is facing stiff competition, has an aging distribution system, without enough capital to make up investment for necessary upgrades. Where are they going to get the money?
They could get a rate increase from the public service commission. They asked for an 11% increase for gas and nearly 6% for electric customers.
They could reduce the burden on the existing distribution system by reducing customer consumption, but this also reduces revenue.
This is where Smart™ Meters come in. They're primarily a way to save money, to charge customers more, while also collecting data they can sell.11
Traditional meters with spinning wheels measure the flow of gas or current. But someone has to come out to read the meters, and that's labor. There are also errors, or there may be no one home when the meter reader arrives: that means lost revenue.
About 15-20 years ago they began installing new meters with digital displays, easier to read, harder to defraud. Smart™ meters are the next level: no need for meter readers; the meters directly report usage data to ConEd servers via a wireless data mesh network in the 900 MHz band in real time. All Smart™ Meters in the area are interconnected, forming a mesh of network nodes with enough hubs to get back to their servers.
Advantages for ConEd:
No more labor costs for reading meters; that large section of ConEd's labor force will be laid off entirely. This is a massive long-terms cost savings for the company.
More accurate billing. No longer relying on estimates has financial consequences because the cost of energy to ConEd may be higher in the month of the actual reading than in the month of the estimate.
Reduces vulnerability to price volatility, because the cost of energy is passed on to the customer in real time.
Monitor for outages regularly, which is in their interest because state regulators fine the company for outages.
They want to limit the immediate financial impact of the symptoms of their deferred maintenance rather than fix the real problems.
Communicate with IoT devices in the house. This allows them to collect data to sell and also facilitates load-shedding — when the demand on the system is too high they can (for example) turn the thermostat down remotely, without consulting the customers. ConEd gives rebates for IoT products purchased in local home improvement centers, and promotes this as environmentally responsible, but this shouldn't be taken seriously; reducing customers' energy use doesn't benefit ConEd unless the network is strained.
Gas leak detectors linked to Smart™ gas meters send emergency signals and notify repair crews. These are naturally located in the part of the network for which ConEd bears responsibility in case of a leak.12
If you're a building owner you can opt out. It's difficult and costly.
Part II, on understanding™, will appear in about two weeks.
In his forthcoming book, provisionally entitled “L’IA entre Golem et Pygmalion; Où nous mène l’intelligence artificielle ?” Jean-Michel Kantor describes this process of “détournement des noms,” which he calls the “Pygmalion effect,” in the broader context of artificial intelligence research, not limiting himself to commercial applications.
Naturally this includes their activity sponsoring research that bears on mathematics and that involves mathematicians directly, but this will be a topic for another discussion.
Ahttps://speedbird.wordpress.com/2013/09/28/against-the-smart-city-now-available-for-purchase-in-kindle/
J. Haëntjens, Comhttps://speedbird.wordpress.com/2013/09/28/against-the-smart-city-now-available-for-purchase-in-kindle/ment les géants du numérique veulent gouverner nos villes, Eds. rue de l’échiquier (2018).
“Think smart; act smart; win smart” December 2, 2022, at smart-energy.com
M. Kelly, “Germany moving ahead with smart meter rollout plans,” 22 March 2017, reprinted at Smart Energy International, December 27, 2018.
P. Picazo-Sanchez and M. Almgren, “Gridchain: an investigation of privacy for the future local distribution grid,” International Journal of Information Security (2022), published October 11, 2022.
U. Greveler, P. Glösekötter, B. Justus, and D. Loehr, “Multimedia Content Identification Through Smart Meter Power Usage Profiles,” Proceedings of the International Conference on Information and Knowledge Engineering IKE'12, Las Vegas. That last one was published in the proceedings of one of the many tech conferences held annually in Las Vegas, so it must be credible.
The title refers to a local energy consultant’s explanation of the surprisingly high bills some Bronx customers are reporting:
even though the meters are recording data, they may not be communicating the data properly to Con Edison, so the bills that are sent to customers won’t reflect those readings.
At this point Karpoff mentioned that the Indian Point nuclear plant, up the Hudson River from NYC, is leaking radioactive material into the Hudson every day; but that's not the topic of this newsletter.
ConEd’s brochure states that
- Smart™ meters do not collect, store or transmit any personal identification information. The only data transmitted is how much energy you use. What you use it for is your business,
- Smart meters do collect voltage data that we use to operate the grid more efficiently. They will transmit diagnostic flags and outage alerts so we can respond quickly to any probiems.
We keep your usage information private.
- Con Edison uses the data we collect from your meter only for billing purposes to operate the grid efficiently. and to provide you with customized recommendations for how to save energy.
However, a 2016 Politico article, entitled “Con Ed proposes selling data as part of the new grid,” quotes a ConEd spokesman:
We’re beginning to incorporate innovative business models into the traditional utility framework — helping us orient toward the future — and allowing third parties access to utility data in a way that avoids shifting these costs on to other customers.
Pop quiz: Where have you seen the expression “innovative business models” before?
Karpoff points out that landlords are not keen on providing the means for tenants to detect gas leaks on their own, since ConEd is required by law to shut down the gas when a leak is reported, and cannot restore service until the leak has been repaired by a licensed plumber and ConEd has returned to inspect the repair. This is naturally more costly than alternative solutions, and therefore less appealing to many New York landlords.