Mon 21 September 2026
tags: AI politics
The past decade has been catastrophic in many aspects, our society is in shambles, the tech industry has fallen into the clutches of a handful of mega corporations and our political landscape falls into fascism deeper every day.
During Trump's first term, many of us chose to laugh at the constant demonstration of incompetence shown by the government, watching Trevor Noah or Stephen Colbert make jokes about Trump was cathartic. Not that this administration didn't have nefarious consequences, there was still some degree of resistance.
During Biden's presidency, or what I like to call Trump's second term since no one can say with a straight face the government of the USA was led by the Democrats, things got worse. Democrats demonstrated once again how useless they can be and the spotlight was stolen by Trump day in and day out. Add to this the COVID pandemic, and the American population (along with a big part of the world's population) fell into a crippling lethargy and depressive state.
In 2024, there was a glimpse of hope. Could we have, for the first time, a woman for president? But against all expectations, that didn't happen. I'm still upset about the people responsible for this, and I'm not talking about the MAGA cult, I'm talking about all the individuals who are seemingly against Trump but for whatever reason, refused to vote for Harris when it was time to do so. French people have a reflex to massively vote for whoever isn't a fascist when faced with two candidates. This might not get you the best president, but it sure as hell beats having a Le Pen in the Élysée. The American people (or at least a large enough portion of them) deeply lack any kind of strategic behavior that prevents the worst from happening. When Trump was re-elected, it flicked a switch inside me. I felt a mixture of hopelessness concerning people I previously viewed as political allies, combined a strong will to fight the oppressors. In that instant, rules no longer applied to me since I had been betrayed so violently.
Now, what does any of this have to do with AI, since this is supposed to be an article about AI? Context, my friend. The thing that is equally as important for AIs and for humans. Sadly, many humans are unskilled at giving context, which is mostly fine for neurotypical people but becomes a struggle for people living with autism. With this political context in mind, it becomes easier to see the AI industry was meant to grow the disastrous way it currently is. When I read anti-AI writings, I often have the impression authors have a problem putting 2 and 2 together. The AI industry is the way it is because we have Trump in the White House. The AI industry is chaotic and abusive because some felt they didn't want to vote for Harris. Would Harris have solved all the problems and made the development of AI fully ethical? Of course not. But it wouldn't be the complete disaster it is today. The USA remains a country deeply corrupted by mega-corporations and the Democratic party is known to do very little to change or prevent that. On the other hand, the Republican party is known to make the situation much worse. "The dems aren't good enough so I'm just going to let MAGA in power" is exactly the lack of strategic thinking that is sadly too prevalent.
But what is done is done, you cannot go back in time and move your ass to a fucking voting office (I hope this was a lesson learned and next time you will). This might not be the AI industry we want, but it is the one we have to deal with. I often read things along the lines of "I will not use AI", "I will not run programs that been touched by AI in the slightest way", "I will not look at any AI generated picture", "AI is always wrong".
While I'm generally fine with those holding those positions (those who make it their entire personality tend to become more irritating than convincing), this is once again showing a situation where a group of people puts their values above any kind of strategic thinking. Boycotting AI companies or AI in general is not going to change the AI industry. Companies like OpenAI are already running at capacity, so if anything, a percentage of the population boycotting might even help them providing a more reliable service. Over the years, I've boycotted Apple, Spotify, Twitter, and many others. They are all doing great. Many AI companies were already huge mega-corporation long before AI was even a subject of controversy: Facebook, Google, Apple, AliBaba, Twitter. As far as I know, only Europe dares to challenge those corporations. In the US, they hold more power than the government itself. A few angry rants isn't going to dislodge them.
So instead of putting blinders on and trying to cast AI into oblivion, what other solutions could we have to diminish the power a few corporations hold over society?
You may know about the operating system called Microsoft Windows. Some people felt it was bloated and too corporate. What did they do? They did not boycott operating systems. They took an open source kernel, Linux, and built UIs and software around it. With consistent and persistent effort, Linux based systems are now of very high quality while being accessible to everyone. I feel like we're onto something that could be applied to AI here. With enough compute power, you can get something very decent running at home, completely independent from any online service. The power consumption isn't too high, somewhere between 200W for a mid range GPU and 600W for a high end GPU. It won't consume more than the same GPU running AAA games at high settings. The problem is that there is such a high demand for those GPUs, a RTX 5090 (the best consumer grade GPU right now) sells for more than twice its initial MSRP... This isn't helped by OpenAI, Oracle and Nvidia making plans to buy all the RAM in the world. Maybe patience is the key here. Newer and better open weight models are released regularly with the recent Qwen 3.8 27B almost matching Claude Opus in some tasks. While we're waiting for new generations of hardware, RAM factories and an eventual drop in price on the second hand market, using online services like Claude or ChatGPT can be a necessary evil.
How to use AI is as equally important as the infrastructure used to run it, if not more. I often hear about LLMs stealing your creativity or your critical thinking. This might be the case with a lazy usage like "ChatGPT, do my homework" or "Claude, vibe code a game in one shot", but it's not where they really shine. LLMs are huge databases queriable in your own words, which you can demand any concept to be explain in a way you understand. LLMs are also able to do a considerable amount of work (especially in programming) but you must remain the creative breath and the source of truth in your project. Describe in very detailed words your idea, iterate often, test and validate continuously. AI driven development puts you in the seat of the architect, the supervisor and the QA tester all at once. Those roles demand a high level of expertise so it's important that younger coders practice coding, even if partially, only to be able to build the skillset to fully be able to judge the output of LLMs. LLMs are huge databases of interwinded numbers, all they do is translate one set of words into another; Spanish to Rust, French to Japanese, long text to short text, ... They are not conscious, they are not creative, they are robotic. They work alongside you, not in place of you. They make technology accessible to everyone, where you can operate a computer just by talking to it in your native tongue.
In her very well thought out article AI sucks. Hating it is not enough. , Alice I. Cecile writes "we cannot uninvent AI". Indeed, I don't see this technology fading away any time soon. What you can do, however, is learn about the technology, be prepared to run your own independent AI locally. There is a bubble and it will pop. This will hopefully play in our favor and possibly be an opportunity to finally have access to hardware at a reasonnable price.
The next evolution after AI, one that is already happening, is the democratization of robots. As someone who grew up watching lots of sci-fi, I like the idea of robots. What I like less is the idea that robots would only be available from corporations like Apple, Meta or Google. Tech giants had a huge headstart on the smartphone market and we're only recently starting to see viable options like the Jolla phone or Fairphone. Having such a gap between the tech giants and more truly open products contributed to build a culture of walled gardens, shitty apps, user tracking, ads, ... We can imagine the nefarious consequences of leaving the manufacturing of robots to mega-corporations.
We are less likely to have open robotics initiatives in a timely manner if we lose a big part of the hacker community that is unable to see the technology beyond the perversions of capitalism. We need sovereign models trained on ethically obtained open data, with a corpus auditable by anyone. We need datacenters powered solely by green energy. Those are large initiatives only possible if the numbers are large enough to tip the balance and get some governments involved. Sadly, between those who are totally cool with OpenAI, Anthropic and friends and those who want to wish AI out of existence, there doesn't seem to be a huge crowd of people fighting to make AI a technology available as part of the commons.
An open source AI won't make AI companies go away, just like Linux didn't make Windows disappear or Mastodon didn't erase Twitter. But it gives people a choice, an option to reject corporate control without rejecting the technology.
I'm not here to school anyone but I would love to see the anger directed at AI ,to be directed at the companies which, in the name of AI; disrupt markets, destroy the environment and flirt with fascist governments.
AI is a transformative technology, one that makes computers accessible to everyone and help us build a more sustainable society. Outright rejecting it doesn't challenge the power of AI corporations the way an open alternative can. The anti-AI discourse is slowly but surely slipping into falsehoods and demonization of its users. Ironic, for a crowd that took pride in denouncing machine made hallucinations. The detractors of AI are losing grip on reality because they've been shooting at the wrong targets for too long. It is very hard to have a meaningful conversation with them and my only hope is that they still have enough common sense to do the right thing and vote when they are asked to.