Limiting Streaming and AI as Much as Possible
The carbon footprint of digital technology is far from negligible.
Using the Internet means using an immense global network that has required considerable quantities of ores for the manufacture of cables, computers, servers, smartphones, tablets... It also requires phenomenal quantities of servers that stay on continuously, 24/7, and therefore consume an immense amount of electrical energy.
Digital technology accounted for 4% of global emissions in 2021 — the same emissions as all the trucks and heavy utility vehicles in the world (or twice the merchant navy).
A large half corresponds to the electricity powering networks, servers and devices (computers, connected TVs, smartphones...), and a small half to the annual manufacture of hardware components.
We have become hooked on digital technology. Here again, one more form of addiction. Let us be aware that we are addicted to the "click". When we start reading a news feed, we struggle to stop; we scroll endlessly, we click on links to read ever more articles, ever more information. A true addiction, which devours a phenomenal amount of our time, whose news often stresses us more than anything else, making us ever more sedentary.
All drugs lead to depression.
Let us be aware of this new form of drug. Let us regain the upper hand and free ourselves from this addiction. Let us spend less time behind screens, unsubscribe from newspapers whose news makes us depressed, delete our social network accounts, and spend all this time — regained and reclaimed — on manual or sporting activities, on walks outside, or on real, in-person moments with our loved ones.
Let us feel from within what does us the most good, what benefits us most, which activities make us happiest. Let us choose the activities that do us the most good; let us choose how we wish to spend the little free time we have over the course of our short life.
Stopping streaming
The worst carbon footprint in digital technology is streaming.
Starting with videos, which weigh far more than simply listening to music or a podcast, or reading an article or browsing a website.
Let us cancel our video subscriptions as a priority.
Let us stop watching films via streaming.
Let us stop online video games.
Let us reduce as much as possible the watching of videos on YouTube or its equivalents.
Let us stop listening to music online.
Let us reduce our podcast listening.
Let us buy music and keep it once and for all on our computers or our smartphones. The same for films.
Let us reduce the time spent behind screens, so deleterious for our health. Instead, let us exercise. We are living beings, with arms and legs made to move — not to remain seated 8 hours a day.
Even just a little strength training or stretching at home, in the evening before bed: screens in the evening prevent us from falling asleep afterwards, and we have bad nights.
Even just 10 minutes of physical activity at least per day would do us so much more good than those hours glued to screens.
The more physical activity we do,
- the healthier we are
- the better we sleep
- the better our morale
- the more present and effective we are in what we do
Not encouraging Artificial Intelligence
The use of artificial intelligence (AI) requires gigantic data centres — that is, gigantic quantities of servers running permanently, day and night — all that for... what, exactly?
Some might be tempted to say that it is safer to hand over to artificial intelligence the task of deciding our lives in our place, of deciding what is best for the human being, because it will make fewer mistakes than we do. But for that, one must still be capable of knowing what to tell the model that trains the AI — that is, capable of saying, for each answer or result provided by the AI during training, whether it is a good or a bad answer, whether this result is good or not for the human being. There is an initial training of this AI, which therefore requires human decisions upstream in any case, even if the goal afterwards is for humans to no longer have to make decisions. Would we know how to say — how to indicate to the model training the artificial intelligence — what is good for the human species, what we truly wish for?
Do we know how to say what is our priority?
Thus, to the question "what is good for the human species?", which of these two answers will the humans training the AI models choose to label as the "right answer"?
- Ever-higher purchasing power to consume ever more (=> hence with more pollution and obesity slowly killing millions of people and triggering ever more storms, floods and droughts)?
- Or a more measured way of life, with fewer objects invading our living spaces, more real relationships with those around us, more calm and clean air, and the better health and good morale that flow from it?
Putting our lives in the hands of machines — is that really what we wish for? Do we really wish to no longer use our brains to think, keep learning, evolve and progress?
Does it really make us happy to no longer have to think?
Are we reduced to being like mere beasts that only need feeding?
Is enduring the infantilisation of the hierarchies currently in place not already enough for us?
Rather than investing billions in AI, would we not do better to invest these colossal sums in the local production of food and energy, which are truly essential to our survival?
The day there is no more oil, it is not AI that will feed us.
If we nevertheless decide to use AI anyway, then let us make sure:
That it is for a good cause => accelerating, in one way or another, the development of food self-sufficiency, for example
To use the most ethical and local AI possible, to maximise the chances that it serves our interests => French AI, whose creators display nobler intentions than those of the USA in particular, by making their models freely available so that anyone can train them differently, thereby multiplying the diversity of the answers and information that can be provided for the same question or topic.
We propose a decentralised approach. When we make our open models available, we allow everyone to appropriate them, modify them, introduce their editorial biases, their culture, their language. Unlike Big Tech, which has a very American vision — to each their own — we do not want to impose our representation of the world. Hence the emphasis we place on linguistic diversity in our models.
The real danger with artificial intelligence is that it is becoming the information portal of tomorrow. So what happens if two or three American companies completely control the way we access information, and have a major influence on the way people think around the world?
Arthur Mensch - Mistral AI
In the eyes of Mistral AI, AI is an accelerator of creativity: the point is to free up time to do what requires more creativity.
Failing the ability to produce food, may AI help us find solutions and work as a priority towards the development of food self-sufficiency — without the help of oil, whose stocks will soon be empty, and under a disrupted climate whose damage reduces agricultural yields a little more each year.
Conceptual compression to unleash the best ideas and thus give anyone a chance to fight back
The motivations of those at the controls of AI in the USA: What motivates those at the head of the "AI ecosystem" is neither human happiness, nor the environment, nor lower unemployment, nor peace or harmony. No — what motivates them is domination and power, a form of quest for immortality, the pursuit of personal dreams, and, quite explicitly for a large portion of them, the replacement of democracy with an "algorithmocracy" of which they will be the masters (in French)
Choosing repairable smartphones and computers
As already mentioned in a previous chapter, if we really need to buy a smartphone or a computer, let us choose models whose parts are repairable or replaceable.