AI Restores Mans Ability to Walk + Community Announcement

Welcome to edition #18 of the “No Longer a Nincompoop with Nofil” newsletter.

Here’s the tea ☕

• Meta’s language conservation 📜

• AI helps man walk again 🚶

• Finding rare DNA sequences with AI 🧬

• Use DragGan locally 🖼️

Meta is the real “OpenAI”

Following on the small optimistic piece I wrote last week, we’re slowly starting to see the practical use of AI in creating meaningful impact. Meta has open-sourced their Massively Multilingual Speech (MMS) models which can identify more than 4000 languages, far more than any current model’s capabilities. This is an ethical use of AI if I have ever seen it, and it’s open-source. With AI, we can preserve thousands of languages and cultures from across the globe and this is a giant step in the right direction.

The models also contain speech-to-text and text-to-speech capabilities for over 1000 languages. This can then be used to educate folks and even teach them using technologies like VR and AR. Imagine being able to listen to hundreds of languages and dialects from across the globe, languages that would otherwise be lost to time. The preservation of language, art, culture and history is one of the most interesting applications of the marriage between AI, VR and AR.

Translating thoughts into movement

A recent article in Nature has described how scientists used AI and thought decoders to help a man walk again. The man who was paralysed from the waist down was able to walk again after doctors connected implants in his brain to spinal cord stimulators which would encourage movement.

Using machine learning algorithms, the scientists first observed which parts of his brain were lighting up as he tried to move. They then used a different algorithm to connect this information with the spinal cord implants which would then stimulate the correct limb for movement, i.e. moving ankles or toes. It is genuinely remarkable that we’re now able to use technology to read brain activity and actually stimulate movement in the body. Even a year after the experiment, the man had significantly improved his walking, even without the implants.

This combined with the brain chip implants we'll eventually see from companies like Neuralink will be the future of human-computer interactions. We really are just scratching the surface of what is possible when combining human biology and Artificial Intelligence. Will there be ethical concerns? Absolutely. We as a society will have some serious questions to figure out, as the very definition of what it means to be human will be questioned in the years to come.

AI begins its Biotech journey

AI is an incredibly powerful tool. This is amplified when there is a lot of data involved. The beauty of machine learning is seen when we need to analyse millions of data points to come to a conclusion or prove a hypothesis. In a new study, researchers did something incredibly fascinating with AI in biology. Researchers compared millions of DNA sequences between humans and fruit flies and were able to find not only the vast overlaps between the two but also the unique sequences between them. This is just the tip of the iceberg so let me break it down further.

Given 50 million DNA sequences from humans and fruit flies, AI was able to identify rare instances of gene sequences. Even more impressive was the fact that it was able to correctly predict the activities of these incredibly rare, one-in-a-million gene sequences. Without AI it would otherwise be impossible to test so many gene sequences by hand. Since this is all new, the applications for this are still mostly unknown to us. For example, we could use this to test if a particular DNA sequence is activated in different tissue types or if a particular drug activates specific genes. The possibilities for what we can discover and learn about the human body, its functions, limitations and characteristics are practically infinite.

The application of AI in interpreting massive data-sets in the biotech field will hopefully provide insight into our fight against diseases like Cancer, Alzheimers and many more. This is one space where I have no issue with AI doing all the work and then some. We should be diverting every ounce of our collective effort to pursue innovations in this space.

Use DragGan locally

Remember when I mentioned DragGan, the new tech that lets you manipulate images in oh-so-magnificent and interesting ways. Well, now you can run it locally. It’s part of InternGPT, an open source model which also uses a vision language model for multi-modal dialogue. You can check it out here.

As always, Thanks for reading ❤️

Written by a human named Nofil