Halil Akin

I am a builder, researcher and engineer.

Some facts about me:

I grew up in Turkey. Before moving to the US, I worked at two startups there, with one becoming the nation's largest e-commerce firm, and I also started one.
I worked at Meta for 7 years, focusing on NLP and Large Language Models.
I am motivated by the potential of large language models in science, particularly in biology. This has led me to the biology-AI intersection several years ago, resulting in my reading countless biology papers and building biology-specific LLMs.

Some things I believe:

We have enabled machines to "understand" human language (e.g., GPTs), which has allowed them to solve language tasks previously only imaginable
We will soon develop AI models that "understand" the language of biology better. This will have significant implications for our understanding of biology, health, disease, and longevity.

We should aim to raise children who do better than us, build companies that perform better than current ones, and improve the institutions we inherit.
Progress creates meaning. It's good for the society and it's good for us.

Focused, high-conviction, concentrated bets with a well-funded, sizeable team can create exceptional progress.
In a world where we can scale computation with AIs, the limits to progress shift to energy and coordination.
Working on long-term exponential missions under the guidance of a clear strategy and roadmap would alleviate coordination problems.

There's nearly no absolute right or wrong; but most things are either highly overrated or underrated.

Apart from math and science, there are rarely definite right or wrong ideas.
Due to significant reflexivity in the modern world, most ideas today often stray far from their true value, becoming either highly overrated or underrated.
Progress often comes from recognizing underrated truths or challenging overrated misconceptions, leading to paradigm shifts.

We live very different lives compared to our ancestors, and this shapes most of the cognitive biases in modern society.
Historically, humans lived in highly segregated groups, where most value creation resulted from the distribution of best practices.
In today's world, where information travels instantaneously, distribution is overrated and creation is underrated.

Improving public goods creates the most good in the world for everyone, but it is also the least worked-on problem.
If we want to create a significant impact, we should actively think about how to create public goods with our work.
For sustainable progress, we must bridge the gap between well-funded, but short-term focused, extractive institutions and under-resourced, long-term oriented public institutions. This is not just feasible, but it is also the next step in the evolution of our institutions.