Meet AlphaFold: A true artificial intelligence

The paper (linked at the end of this article) presents an in-depth discussion into AlphaFold 2, an attention-based neural network that solves the protein folding problem. The machine learning and biology related materials will be saved for discussion in paper; this brief introduction is to relate AlphaFold’s importance to economics as well as give some societal context.

AlphaFold has great utility for drug companies due to its open-source code and experimental accuracy. However, it has not gained as much traction as it, in my opinion, should. I believe that there has been a lack of effort directed towards understanding how AlphaFold, and more broadly machine learning, works. The cause of this is that the emergence of AlphaFold disrupts the traditional approach to drug development, and that professionals are threatened to have their life’s work undermined by Artificial Intelligence.

Whilst I have empathy for those in that position, it is also important to take into account the bigger picture of human improvement. Drug research, as well as many other commercial biology research areas, can revolutionise their operations around AlphaFold, as well as other machine learning-based tools.

And so, whilst AlphaFold may detract from the personal achievements of professionals in commercial biology research, I believe that, utilitarian-wise, it is truly a tool that will benefit humanity, and is a spring for the research and development of even more sophisticated machine learning algorithms that can solve the broader problems humanity is faced with.

Moreover, the creators of AlphaFold, DeepMind, is a testament to the power of private-sector Research and Development (R&D). Unhampered by academic bureaucracy and regulations, DeepMind are able to put some of the most creative and talented machine learning engineers, mathematicians, and professionals in the same room and allow them to unleash their creativity and potential.

Even though I believe an algorithm similar to AlphaFold would be invented even without DeepMind, it would not be for at least a couple of years. Hence, it is credit to DeepMind for having the foresight and courage to tackle problems and invest in R&D without a guaranteed return.

Please visit the following links to read the full paper as well as view a talk I gave regarding its contents (paper, talk). If you are interested in discussing this further, please do not hesitate to reach out to me via my email address: [email protected].