• AI-generated designer proteins that can fold into any shape…
  • Starlink steps into the defense sector…
  • Apple capitulates on its self-driving car initiative…

Dear Reader,

The last few days have been nothing short of mind-blowing.

Elon Musk and his team at Twitter have dropped “Twitter Files 2,” “Twitter Files 3,” and “Twitter Files 4” into the “ocean” for all to see. 

There are so many details that I’ll need some more time to work through it all, and I hope to provide some additional context later this week. But for now, here’s a favorite from Musk that pretty much sums up what’s been happening:

Musk’s Summary of the “Twitter Files”

Source: Twitter @elonmusk

He’s not exaggerating. The “Twitter Files” have provided documented proof of what many have thought and known for years. They undeniably show that “it” isn’t a conspiracy, but a shocking reality.

I can’t help but be reminded of the Will Smith and Gene Hackman thriller Enemy of the State… and Musk may have to “go dark” before this is all over. The revelations are really that damning.

For today though, we’ll check in on something else a lot less Orwellian that actually did drop into the ocean – the Pacific Ocean, to be specific. It was the Orion spacecraft from the Artemis 1 mission.

NASA’s Orion Returns to Earth

Source: NASA

After a 1.4 million mile mission to orbit the Moon and return, it was a picture-perfect end to a long-delayed mission.

NASA’s Space Launch System (SLS), the massive rocket used to launch the Orion spacecraft, had been delayed and hampered for years by technical issues and cost-over runs. And the Artemis 1 mission was originally slated for August. This has been a long time coming.

Aside from the mission being successful, its value is far more than just being symbolic of progress. This mission is a precursor for NASA to move forward with a return to the Moon. It was a necessary steppingstone to gain further funding and political support for a sustained human presence on the Moon.

NASA will now focus on Artemis 2, which will send a crewed mission into a lunar orbit hopefully by 2024. That will lead to Artemis 3, which is scheduled for the 2025/2026 timeframe.

That’s the big one. Artemis 3 will utilize the SpaceX Starship as a lunar lander. And prior to that mission, SpaceX will launch the Gateway to be inserted into lunar orbit, which will be akin to an orbiting lunar space station providing support and access to the lunar surface.

It’s hard to believe it’s been five decades since the Apollo program. 

It’s about time.

The biotech industry will never be the same…

As regular readers know, the biggest development in life sciences this year came from Google’s artificial intelligence (AI) division, DeepMind.

Using its Alpha Fold AI, DeepMind published a catalogue of more than 200 million protein structures back in August. This was one of the biggest breakthroughs in history in life sciences and biotechnology, and all the more incredible in that it was achieved by computer scientists.

As a reminder, a protein’s structure – which is determined by the unique sequence of amino acids – is what determines a protein’s function. The structure also determines what other compounds (such as pharmaceuticals) the protein can bind to.

So the more we know about proteins and how they fold, the better we can design therapies to cure the most challenging diseases plaguing humanity today.

And an upstart biotech company called Generate Biomedicines is setting out to do just that.

This company developed its own AI that builds upon DeepMind’s work. The AI is called Chroma. It uses neural networks and a technique called “diffusion modeling” to design new proteins for specific purposes.

Chroma does this by creating amino acid sequences that will fold in a very specific way. So if we know what function we want and the desired shaped of a protein, Chroma can create the appropriate amino acid sequence.

What Generate Biosciences has done is created software that can design customer proteins upon demand. It can create functional proteins that don’t yet exist in nature.

New proteins naturally evolve over time, but it can take thousands of years. But why wait when Generate Biosciences can generate these new proteins in a matter of days and circumvent the evolutionary process?

The clip below gives us an idea of just how powerful the Chroma platform is:

AI-Generated Proteins

Source: Generate Biomedicines

Here, we can see amino acid sequences that fold to form letters and numbers. This is incredible.

Of course, this was done just to show the capabilities of Generate Biomedicine’s AI, as proteins shaped in alphanumeric characters have no real function. But it’s a useful demonstration of providing the AI with a desired 3D shape and then designing the protein that would achieve that shape.

And get this: The AI can run on a single graphics processing unit (GPU). That means it will run on a desktop computing system that would cost no more than a few thousand dollars. And then Chroma can generate custom proteins in a matter of minutes.

So there’s absolutely no barrier to entry here.

And with this technology, we can create proteins that could potentially cure every known disease of genetic origin. It’s just a matter of getting the shape right. It’s kind of like reverse-engineering a protein once we understand the desired shape.

This has been a remarkable year for the entire biotech industry. 

These breakthroughs with DeepMind and Generate Biosciences – which both predict protein folding and can create new proteins – will positively impact the entire biotechnology industry. They’re tools that are openly available for the purpose of therapeutic development.

2023 is the year when the industry will put these newfound tools to use. And breakthroughs in therapeutic development will follow.

And of course, Generate Biomedicines could become one of the hottest biotech companies out there. We’ll certainly be tracking this one closely going forward.

Clear evidence of SpaceX’s master plan…

Last week, we had a look at how SpaceX’s Starlink division received regulatory approval to launch 7,500 of its next-generation satellites. Well, Starlink followed that up with an interesting announcement.

The division just unveiled something called Starshield. It’s a satellite network designed specifically for military and defense applications.

Recall that we also looked at defense contractor Anduril last week and noted how rare it was for Silicon Valley companies to bring bleeding-edge technology to the defense industry.

Clearly, SpaceX wasn’t shy about doing the same.

The purpose of Starshield is to establish a low-Earth orbit communications network for miliary, defense, and intelligence applications.

The latest version of the Starlink satellites have laser communications built into them that enable data to be transferred and received between two orbiting satellites. We can imagine information hopping from one satellite to another over the desired ground station where the information can be transmitted back to Earth.

I’ve long predicted that Elon Musk’s ultimate goal was for Starlink to become a global backhaul network for space. This is like a low-Earth-orbit-based cloud-communications network. And it will become a critical piece of infrastructure for national security purposes.

Starshield has effectively become a systems integrator and defense contractor. It has the ability to design, build, launch, provision, and operate satellites and communications networks on behalf of its customers.

Very few companies have all of these assets in-house, and none of them can deliver as competitively has SpaceX/Starshield.

So this is another big win for SpaceX. And it’s a smart move by Musk. 

After all, if he’s making himself critically important for the U.S. government with one of his companies like SpaceX – and he’s and supporting government policies like the shift toward electrification with Tesla – Musk will likely have some “protection” from the government over what’s being brought to light at Twitter.

The latest from Project Titan…

We’ve just learned of some disappointing news surrounding Apple’s secretive electric vehicle (EV)/self-driving car initiative, Project Titan.

When we last checked in on the project back in May, Apple had just received a patent that envisioned building virtual reality (VR) technology right into its self-driving car. This was Apple basically patenting the user experience for autonomous transportation.

So everything appeared to be on track for Apple’s target launch date of 2025. And it sure looked like Apple planned to launch its EV as a fully autonomous vehicle.

But it appears that Apple has been struggling more than expected with its autonomous driving technology.

The company just pushed out its EV launch to 2026. And Apple’s first car won’t be fully autonomous. It may have limited self-driving capabilities on the highway… but that’s the best-case scenario.

Clearly, Apple isn’t confident in where its self-driving technology is right now. That’s a disappointment, as Apple has such a strong history of successfully redefining industries.

At the same time, this isn’t terribly surprisingly. AI has never been Apple’s core competency. Jumping to something so complex as autonomous driving was always known to be a very ambitious project.

Still, this is a huge shift. Rather than what could have been a revolutionary take on personal transportation, Project Titan will likely just become an evolutionary step in the EV industry.

This also means that Tesla won’t have any threat from Apple as a new entrant into the industry. Its technological moat is too great, and it has years of advantage when it comes to autonomous driving technology. 

It will be very hard for others to catch up.


Jeff Brown
Editor, The Bleeding Edge