Every day we read about new layoffs, even in companies that once seemed untouchable. And every time, I find myself asking the same question:
👉 Are we teaching AI to do our jobs?
Thousands of developers are training intelligent agents to solve problems, write code, optimize processes… completely on their own.
The paradox? It’s like programming your own dismissal, line after line of code.
A collective self-layoff, perhaps unintended, but inevitable — unless we start asking ourselves: what role do we want to play when machines can do everything we can?
― Dario Cavada, https://www.linkedin.com/feed/update/urn:li:activity:7387040204890456066/
This is the quote translated by ChatGPT fn the post itself, at the end you will find the original one. TL;DR, my dad is arguing that we (programmers) are “programming” our own lay of from AI. This because of today’s news about Meta’s layoff. Few friends commented that as very pessimistic and arguably it is. I wanted to reply something below his post but I felt few lines could not be in my favour so here it is, in between papers, I am keen to write this piece which is quite dear to me.
But where do I even start? I will go in order of the ideas that I wanted to reply
Abstraction
I am curios if programmers when compilers and high level languages were this pessimistic too…
― Sebastian Cavada, Reply n.1 I never posted
This is something that I reasoned about before. It is unintuitive, but someone before me shared it for sure. In particular in the 70s-80s when only programming language was ASSEMBLY, well most people were doing exactly that, then when C came around, and C++, never mind python etc… they might have thought well if everything is offloaded to the compiler, well we would do same things in 1/10th of the time, we will run out of business. Well I think, and for my knowledge, high level languages are what inspired generations, and also made possible to build what we have around now. COuld it be that AI is just another (MASSIVE) improvement in the abstraction of the code stack? Well, I think the first signs of that are already there, see Lovable, see how much more efficient are programmers, see how Tesla has built so much of his tech stack into software 2.0 which is just neural network based. Referring again GeoHot on similar topic to this post here
Thus in general I’d like to see things in an optimistic way, because why not. And second because well it’s true that this technology is so new and so “alien” that we might not know what’s coming, what I think though is that we should leverage it, and there are so many possibilities.
Less work, or more?
One assumption there, is that given that there are lays off, it means there is less work for everyone. I don’t agree on that either, thinking back at when agriculture was mainly done by hand. At that time a lot of people were involved, but then technology arrived and we have a fraction of the people involved in agriculture, because of leveraging our ingenuity. I don’t see why that is not going to happen as well. Laying offs are (beside very sad, and with lot of negative ripercussions) a natural cycle of business, exactly as it is to open a business, as it is failing at a business. This very process is what led us here today, with this many tools that allows me to write and share with the world my thoughs. A very good turn over of the story is that probably the brilliance of these people will make them find a job instantly, as per many tweets already today of people trying to hire the “left out”. In general what I see and what I hope will be the future is more about being able to do the same amount of work with less people and more machines, more tokens from an intelligence that is trained to do that. But this will allow to people to do more of those things, and this is exactly where we need to be mindful and cautious.
Who is replacing whom?
Maybe we should think about who owns this AI and how they are trying to lock us in?
― Sebastian Cavada, Reply n.2 I never posted
In the las period I have been thinking about this a lot. If manager and CEOs can lay off employees, why and how are they going to be replaced by? Well easy answer is AI. Yes but more technically? Well it’s about LLMs, with agentic capabilities, that can code, better (so they say) than a professional (not really correct maybe better than a freshman). Anyway, assuming this statement correct, than what? Well this agents are running somewhere, and not outside on the ground but in the cloud (small dad joke). So ultimately, employees in my mind are just being replaced by GPUs. Yes but GPUs + electricity (because GPUs without electricity are useless).
Fun fact now the compute is not calculated anymore in FLOPS per second (Floating Point Operations Per Second) rather on a larger scale in Gigawatts, yes the unit of Energy! So now energy is directly converted in Computational Scale!
Thus employees (humans) are getting replaced by Tecnology, Electricity and you are correct some massime amount of numbers. But this reminds me about again and again, as most of the transitions that happened in humanity.
So is it all rainbows and unicorns?
Well I tend to be a tecno optimist (I think inspired by GeoHot) and mostly I think everything will go well and humanity as a whole will be fine. I like to think that those technologies, yes are driven by closed source, but open source is taking up faster and faster thanks to some key players. But we cannot rely on their benevolence forever. Things might change, OpenAI, going from non-profit to for-profit, Meta stopping the plans from LLama 5 (So it seems) and taking slightly different direction. Europe not knowing what the hell an LLM is, but wanting to enforce some weird laws anyway. China picking up speed at an unbelivable speed. Well probably that is exactly what we should talk about, not about laysoffs from meta, but rather the broader impact of AI. Of course there will be so much to add to this paragraph, about security, AGI, etc, but on this I might refer the reader to the GeoHot blogpost titled -> I couldn’t find it now
In the end
Trying to find the link in the previous paragraph I lost my flow. Nevertheless what I wanted to add that is extremely important is that we shouldn’t be scared of this progress, we shall try to embrace it learn as much as we can and probably being suprised in the end.
The possibilities are limitless. If AI is so powerful and one day it will be, we will get access to free energy from nuclear fusion, we’ll get closer to quantum computers and guess, maybe we will be able to train LLMs on quantum computers. We might raise the minimum IQ points, of every child because instead of taking 5 years of elementary school they will go to a kindergarten where they will learn calculus by building with legos. They will engage in languages from 4-5 years old and grow up knowing 10 different languages. We could cure cancer, not only by editing genes and personalized cures, but and especially in the prevention of those, by optimal nutrition, sport, health, stress monitoring etc. By the way I am not saying everything will be good, but the good will outweight the bad. But we need to be careful not to fall in the AI slope of infinite amount of funny videos generated by sora that will keep us attached to our phone and dopamine maxxing without any positive output for the society.
And concluding I want to underline that I see infinite ways of creating a bright future, it’s up to us to choose whether we see the glass half empty, or overflowing with possibility.
An employee getting fired might finally find the courage to start the new venture that will revolutionize the world.
― Sebastian Cavada, Reply n.3 I never posted
The original post:
Ogni giorno leggiamo di nuovi licenziamenti, anche in aziende che sembravano intoccabili. E ogni volta mi viene la stessa domanda:
👉 Stiamo insegnando alle AI a fare il nostro lavoro?
Migliaia di sviluppatori addestrano agenti intelligenti a risolvere problemi, scrivere codice, ottimizzare processi… in completa autonomia.
Il paradosso? È come programmare il proprio licenziamento, riga dopo riga di codice.
Un auto-licenziamento collettivo, forse non voluto, ma inesorabile a meno che non iniziamo a chiederci: che ruolo vogliamo avere quando le macchine sapranno fare tutto ciò che sappiamo noi?
― Dario Cavada, https://www.linkedin.com/feed/update/urn:li:activity:7387040204890456066/