On AI, my professional future and the new ways I can be productive/useful to society while keeping my sanity.

Last week, OpenAI, few days after releasing the Astra (GPT-6) model, and after lots of drama in the mathematical community, they claimed to have solved the Navier-Stokes problem. Where under some initial conditions and a smooth external force or something like that, the equations break or it creates a singularity. This is one of the 7 millennium problems that where set at the beginning of the century. It hasn’t been approved or marked like solved by the community but everything indicates that the proof probably is correct. This is in fact, a next level of intelligence that I have been experiencing these last days with the Astra model (even though the internal model used for the finding is even better apparently). In this panorama of drama in the mathematical field and in this fast-moving pace world, I have to but wonder what can I actually do from my side or provide to science. I am now in my second or third year of my PhD, depending on how you count (officially 2nd). My capacity on both coding and maths has been completely override by AI models, they are much better that me. I am computer engineering and those abilities that took me years to learn (not even master) are now obsoletes. One thing I thought I still had was the reasoning capacity, critical thinking and organization skills that are needed in order to do research. And despite of not being perfect on AI right now, they are almost at the same level of me, while being much much faster than me. I still right now need to guide the models on what I want and what I need, but month after month it is more it than it is me. Under these circumstances I can clearly see the future where all my research can be just automated with more money right now on tokens, but becoming cheaper and cheaper with time.

So in this page of my blog I would try to reason and think about what is left for us, or in this case for me.

My first instincts go into what is currently needed, that is learn how to better set the workflows for my research, how to better manage the models and how to take full profit of the limited tokens I can afford compared to the big companies and big labs. This is specially interesting because I think this intelligence power is gonna be concentrated in few, and they are gonna monopolize it, but also thinking about sustainability and making the best use of the resources I can. On the other side we have maybe it is worth learning about the harnesses, how they work and how to apply them, I currently have little idea of this subject and how it will work/impact in the future.

Lastly and maybe the most important thing I can do in the medium-long term is become the best person I can. Currently, with all the resources and models we have, it has never been easier to learn about any subject, I can set up a environment to learn about different topics, make it ask me and quiz me on those subjects and in overall learn as much as I can, because why not, why not learn more about hour history, from everywhere, why not study new fields like physics where I have been always really interested on, why not learn more languages, psychology, music, manual stuff, repairing things, drawing, dancing, be more athletic, eat better, why not?

And this is specially important cause I believe in the near future, most of the tasks we currently do, will be automatized, so what is a valuable man?

For a long time, we have valued people for what they could provide: food, protection, things they could build, problems they could solve. What counts as useful has changed many times. Now AI may be able to do much of the work we value today. So what makes a person valuable? What will we do with our time?

If we look in the short-medium term things are going to shift, all the advances we see now in computer science and mathematics are going to be integrated in all sort of administrative “computer” work, so everything that a human does through and only a computer will be automatized and replaced. That goes from a governmental position taking care of pensions to a computer scientist or a mathematician. After, we will see the explosion in robotics, everything that we can physically do, a computer with a physical interface (humanoid/robot) will be able to do. Then all industries will shift to automatization and the economy will change drastically.

So what will we do? Rest, learn and enjoy I guess?? Maybe just continuously fight with one another?? Maybe we will all be extinct?? I hope not.

So there is the key, how to do so all this shit we just spent billions of hours into doesn’t make us disappear as a species.

First, we need to put more energy than ever before to fight for the right, for what is good, for what is fair. And I won’t accept the question of what is good and what is bad. We as humans, as really conscious animals, deserve a life worth living, we deserve to experience emotions with its ups and downs, we deserve to express ourself through arts, we deserve to explore, to be alive and probably we deserve to die. Do not go gentle into that good night.

So again, what do I do? how we materialize all these beautiful and vague words?

There is no real correct answer, the answer that I find more compelling for me right now is to focus all that attention and mind towards how to improve these systems, is there any way BNNs can actually help on the alignment and trustworthy AI systems? Even if its in smaller networks that these models could use for simpler tasks? Now that a lot of my cognitive power can be replaced by AI agents, I should shift my focus on learning and designing systems to make better models for us humans.