Finding life in space and curing allergies

There was no session last weekend but I have to keep the writing streak alive, so today’s newsletter is taking on a slightly different format.

I consume a lot of information throughout the week to try to stay sharp. Here are some highlights from the week. My hope is that at least one of these is new, surprising, or insightful for you.

In program news: Moonshots kick off tomorrow. I'm containing my hype for now but will unleash it next week as I dive deeper into what this means and what the students are up to!

--

Asteroid flying through space with small signs of life (Stable Diffusion)

The building blocks of life were found on an asteroid

A new study has found a chemical compound representing one of the building blocks of RNA (used in the creation of proteins) in space. Previously we have found evidence of these compounds in meteorites (crashed asteroids), but we couldn't be sure if those came from the asteroid or Earth itself, post-impact. This new discovery suggests that it's possible that the components of DNA/RNA and human life (like water) originated elsewhere in the universe and eventually landed here. Whoa.

Now I’m wondering if other planets were also hit with similar asteroids but never had the perfect mix to spawn human life.

There’s a new VR headset just for reading books

I came across Sol Reader in a podcast I was listening to. It suggests the potential of single-purpose VR products. You might think this is excessive or niche, fair, but (hypothesis): I think the advancements in VR capabilities coupled with manufacturing knowledge have allowed these types of devices to come down in cost over the past several years. This makes it more feasible to build products optimized around singular use cases. Obviously, I want one of these.

I'm also curious to see what other products might work like this in the future. Call me crazy, I think there could be a market for more single-purpose headsets.

Peanuts being destroyed by a mad scientist (Playground AI) – this one was too good not to keep 🤣

We might be able to eliminate peanut allergies with mRNA

It's kind of mind-blowing to think that allergies might be a treatable condition in the near future. mRNA-based medicine is a fairly recent innovation (last 20-odd years) and rose to the public sphere with the Covid vaccine. Now, much more attention is being paid to the power of this area of medicine. This new treatment uses an alternative delivery mechanism (a lipid nanoparticle) which allows it to deliver mRNA multiple times while also having way less inflammation and side effects as your body is not attempting to fight off an adenovirus. These new nanoparticles can also be delivered through airways instead of injections.

Researchers have started tests (on mice) to attempt to subdue and eliminate peanut allergies using this technology. Results have seemed positive so far so we might start to see allergies begin to change in the coming decade.

Caveat: my own understanding is still very basic.

Self-correcting programs can be built w/ AI

Imagine if you could write perfect code without errors. It's a nice thought but rarely achieved on the first pass. Coding is very much an iterative process, making improvements and chasing down oversights and small (or large) errors.

One of the popular use cases of GPT is as a programming copilot, helping you improve your code, offering up code snippets, and providing debugging support. What happens if your code could just perform these operations on itself?

Early experiments are being done to create self-correcting code, such as this one here. It's kind of crazy to see (albeit in a very simple example) and suggests that we could be on a path to much less debugging and maintenance required. As long as the output is thorough, this could act as a valuable teaching tool as well.

We can use ultrasound as a tool to read the brain

BCI technology generally falls into two camps (invasive and non-invasive) each with its own pros and cons. In the non-invasive category, the most popular are EEG (which has low resolution but is very approachable) and fMRI (which has high resolution but requires big-ass machinery). Another contender seems to have entered the ring: fUS – using functional ultrasound to measure blood flow in the brain in response to brain activity.

Early tests were done at CalTech in 2021 where they were able to "predict" movement within a primate by reading areas of the brain responsible for planned activities. I haven't been able to track down much recent research from the last 18 months yet, but this seems like an exciting area to follow as a higher resolution could be.

—

Hope you learned something new!

✌🏼