💾 Archived View for perma.computer › letter captured on 2024-06-16 at 12:00:29. Gemini links have been rewritten to link to archived content

-=-=-=-=-=-=-

Technological Futures: A Letter to the Smolnet

We're standing at a crossroads. Technological progression has led to a retreat from physicality; whole classes of devices - cameras, music players, notebooks - have been replaced by computers. Mass computerization of most types of human activity and corporate control of mainstream computer ecosystems has led to a drive toward constant monetization and constant growth. Rapid upgrade cycles - often annual or more - have resulted in a pervasive attitude of disposability, generating enormous waste and resource consumption. These two stages - "replace everything with computers" and "build computers in such a way as to maximize recurring revenue" - have produced an alarming result: functioning in a society nearly mandates ceding ground to technocapitalism, and this trend is accelerating.

Large deep-learning models are increasingly being used for things historically seen as the bastions of human ingenuity and creativity - images, writing, music, code. Almost universally, these are characterized by their use of human-generated training data scraped from the Internet and their corporate ownership and operation. These models are now being commercialized - such as by Microsoft with Copilot, which costs a monthly fee. Worse, the major search engines are moving from a "search and get a URL list" approach to an "ask our conversational language model" approach, leaving those who actually write things as little more than upstream suppliers to a bot. Meanwhile, extruded text product and extruded imagery product - "AI text" and "AI art" - make up a larger and larger portion of the Internet.

We stand for an alternative vision - a humane, sustainable approach to computing, independent of growth imperatives, the need for monetization, and the concept of obsolescence.

Two Futures

The Corporate Internet: Automation, Stagnation, Depletion

We have been repeatedly assured that we are on the edge of a great and glorious new era - one where automated systems and emerging technologies allow human productivity to soar to new heights. The utopian promise of an AI-assisted golden age is being promoted as a possible future - if only we trust in a handful of large tech companies (Alphabet and Microsoft/OpenAI prominent among them) to know what is best for us.

So what, then, does the future look like? It looks like massive changes to a large portion of knowledge-worker jobs. Some jobs will be less affected, especially those in the trades, but these jobs will mostly be in the category of "jobs too cheap to automate" - plumbing, shipbreaking, retail, clothing manufacture, construction, waste processing, some kinds of factory labor. The purpose of investment in automation is to save money for the ownership class; this does not line up with providing a comfortable living for everyone. With a smaller number of available jobs, employer leverage over labor will increase. Working hours will become longer - and for those who refuse to work them, there will always be replacements available. Productivity will be monitored by machines, and enforcement actions will be automated and impersonal as well.

In the developing world, the same pattern will apply. Mass exploitation of people and resources will continue and expand in the interest of feeding the expanding consumer needs of the developed-world upper class. Investment from the developed world will, as ever, primarily occur under the rubric of "how can we increase efficiency of production and extraction?" - not any genuine intent of long-term improvements in quality of life. As is already happening, the nastier and less visible parts of AI product development - especially manual tagging of training data - will be dumped on underpaid workers in developing countries. Increased capital capture of governments in developing and developed countries will simplify labor exploitation in both.

Mainstream entertainment will be increasingly machine-generated and largely derivative. Creativity as a means of connection with another human being will no longer be profitable or appreciated - instead, machines will turn out infinite quantities of soulless material, often customized to the viewer and advertised as "more immersive" and "more customized." Communities and fandoms built around mutual enjoyment of a thing will be badly harmed as a result, and artists, writers, musicians, and programmers will find themselves in the unenviable position of being made redundant by machines trained on their creative labor. Authorship and copyright (and, by extension, copyleft) will be conveniently laundered away for corporate and consumer use. The Internet as an open platform will be marginalized or dead; over time it will become little more than a collection of siloed services, generally accessed through AI-based conversational interfaces owned and trained by corporations.

Among the upper tiers of the employed population, mass automation will allow for ever-increasing consumption of consumer goods. A parade of new devices and services with pseudo-intelligent capabilities that would have seemed miraculous in 2013 will be available, and updated frequently. There will be demand for new and extravagant luxuries that continues to fuel resource extraction and labor exploitation. Ecological collapse will become steadily deeper, but its effects will primarily not be visited upon the consumer class. Corporate greenwashing, almost entirely focused on carbon-induced climate change, will continue to ignore the other problems of overconsumption - serious discussion of land and water overuse, deforestation, mining-induced pollution, and similar issues will be studiously avoided. Even absolute decoupling from carbon emissions will likely remain a pipe dream.

In short, this is a future that will continue to reinforce the privileges associated with wealth and power, and will undermine the dignity of people.

Human-Scale Computing: Glowing Embers

We don't pretend that we can prevent that future. We're just individuals, and not very many of them. What, then, can we do?

We can present an alternative.

Computing can be more than the ever-growing techno-capitalist juggernaut. It can be about things other than speed, or scale. It can be about human creation, and human interaction. It can be about people finding new and creative uses for old hardware and software. It can be about computing as an art, not as a production line. It can be about sustainability. It can be about cleverness. It can be about small things.

With that in mind, these are some conclusions we've reached.

We reject the use of generative learning models outright. We furthermore reject any AI development that is driven primarily by corporate interests rather than the interests of the people affected by it. Any AI development should be respectful of human creativity and dignity.
We commit to reducing our personal resource footprints, including those of consumer goods. We commit to device lifetime extension and the circular economy.
We strive for repairability, maintainability, and long life for the projects we create ourselves.
We aim to collaborate with others in a spirit of conviviality before reaching for corporate solutions, technologies, and devices. We look with interest upon projects - including hardware, software, and networks - that have people and their interactions at their center. We also welcome and appreciate all forms of human creative expression.
We embrace innovative, low-impact solutions to problems - ones comprehensible and implementable by human beings working together.
We repudiate racism, sexism, homophobia, and transphobia. We work toward a convivial technology future that is inclusive of people of all backgrounds, identities, and experiences.
We accept that we have participated, and still participate, in an exploitative and destructive system. We look to the future with a sincere desire to do better.

We invite others to work toward a future that respects all people, and puts the needs of human beings before those of corporations. We may not be able to change the direction of the megacorporations, but we can work together to build solutions beyond them - to build an alternative technological road together, and to resist the dangerous path they have set us upon.

Sign Your Name!