I started 2025 thinking about what it would look like if we just shamelessly copied biology. Without too much care, without being too smart. Just building AI systems that copied biology.

After a couple of readings and experimentation, this is me thinking about what would really define my 2026, or what I would like to define my 2026. Three things come to mind.

1. Disability Should No Longer Exist

When you really think about it, if I get into an accident and lose my arm or my leg, the quality of my life significantly reduces. And this is in 2025, when humanity has done a lot more seemingly genius things than solving this.

I know I’m naively underestimating what it takes to solve this, but I still think we should have done it by now. If I lose my arm, I should be able to easily get an artificial arm and put it on. Maybe I’m even better off for it, because I now have an advanced arm. If I lose my ear, I should just attach one. I know artificial limbs exist out there, but it should have been so mainstream that there should be almost no disability.

Think about what glasses have done. About 50% of people use glasses, and there are a lot of eye defects we don’t even consider disabilities anymore. No one says, “I have an eye defect, I have astigmatism and that’s affecting the quality of my life.” Maybe some extreme version of it, but for most people, they just use glasses to correct the defect and don’t even consider it a disability anymore.

We should get to a point where we’ve either eliminated all forms of disabilities or augmented human abilities with artificial systems and artificial limbs.

2. AI Is the Technical Shift That Would Enable This

I know I might be late to the party, but right now we have neural networks predicting and pattern matching systems that would have required millions of years of classical compute, or even complex calculation.

Look at what AlphaFold is doing right now. If you had to do it classically before, you would have had to account for so many variables, done a lot of logic, if-else, maths, and even at that, we were very slow in generating and creating proteins. Now we can just generate hundreds of millions of proteins very easily using AlphaFold. Demis Hassabis got the Nobel Prize for that in 2024.

If you also look at Veo 3, the video generation model, which is still very early by the way (video generation is barely two years in). Veo 3, just by watching a lot of videos, is able to simulate physics mechanisms. There was one where a truck was driving into water and the water splashed, and it got the physics correctly to some extent. We are still very early, it only gets better from here, but it tried to get the physics of the fluid correctly. These are things that we are still trying to figure out the theory behind. If you were to build a game engine for it, you’d have to manually and deliberately write all the code. AI just seems to understand it.

I’m not saying AI would necessarily be the thing that gives us the limb; like I lose my arm and some AGI just gives me one. No, I’m not talking about AGI in this case. I’m just talking about how the fact that we know neural networks can predict systems rather than compute them means we can use this as the base for understanding biology at a level we could not before.

And once we understand our biology, we can abstract. We don’t need to… we will still strive to understand the underlying principles, but we can also abstract and build on top of biology as a platform. Then we can eliminate all forms of disabilities.

One of the reasons biology is very difficult is because: one, we don’t have enough data, and two, there are just so many variables going on in the human body, in the brain, at the cellular level. So many variables determine what goes on. But now we don’t have to understand every single variable. We can just abstract it and build on top of it by predicting and pattern matching.

And by understanding biology, we eliminate all forms of disabilities.

3. Robotics

Robotics, because one, I need a way to fund my expensive curiosities, and I need to create things. Robotics is one of the interesting ways to do it. But also, working on robots with a consciousness of biology in mind helps me understand better how biology works. How robotics and biology work together. I can do a lot more experimentation practically, both in the software sense of things and in the hardware using robotics.

Antroph

So I’m starting a new project, Antroph, with the goal of shamelessly building bio-inspired robots. From simple tabletop robots just sitting on your desk, to complex life-sized pets like a penguin, a parrot, a rabbit. Just come up with some very interesting life-sized robots.

I would sell some of the simpler ones, talk about and use them to fund some of these experiments, while we’re also building some very large complex things.

Biotech with Demi

The second project is a video series, Biotech with Demi, where I periodically take the latest advancement or the latest paper in biology and share it with you guys. The purpose of this is for me to immerse myself in this world, to understand better what’s going on, and to create a system around it.

I’m a novice to most of these. I have very little expertise. I’m very naive. And I hope that by doing this, I get better clarity and understanding of what biology, robotics, and AI entails, and also get to attract some of the most intelligent people in the space to have conversations with them.

So to summarise:

  1. Disability no longer exists — we should eliminate or augment all forms of disabilities.
  2. AI is the technical shift — neural networks predicting biology rather than computing it.
  3. Robotics — Antroph (bio-inspired robots) and Biotech with Demi (video series) as the vehicles to explore and fund this path.

Let’s see what 2026 brings.