Starship ready for launch
Getty Images
Why SpaceX needs Starship. A new kind of origami. Sam Altman’s longevity startup raises more money. All that and more in this week’s edition of The Prototype. To get it in your inbox, sign up here.
SpaceX filed its long awaited IPO on Wednesday, and there’s a lot to talk about. So expect some commentary from me in the coming weeks about the feasibility of lunar mass drivers, putting hundreds of GW worth of data centers in orbit, colonizing Mars and more.
But while the sci-fi aspects of the company’s IPO are garnering headlines, sitting on a launch pad in Texas is the cornerstone of the company’s entire future: Starship. I had planned to focus this newsletter on its latest test flight, which was slated for earlier this week. But there were delays and now it will be going up a few hours after this goes out (in theory.)
Starship is SpaceX’s next-generation, heavy-lift rocket, designed to be bigger than the Saturn V rocket that took Apollo astronauts to the Moon while being fully reusable. According to SpaceX’s public filing, the company has poured more than $15 billion into its development.
But beyond just the money, Starship is a lynchpin for all of the plans that SpaceX is using to sell investors on the proposition that it’s worth trillions of dollars–far more than its current revenue of $18 billion (with nearly $5 billion in losses) would suggest. The next generation of Starlink satellites are so big that it needs Starship to launch them. Its proposed orbital data centers require Starship. It needs Starship to fulfill its obligations to NASA to carry astronauts back to the Moon
And with all of this, the company is years behind schedule. Several of its Starship tests have ended in “rapid unscheduled disassembly” (the company’s term for “exploded”), though more recent tests have gotten things back on track.
With SpaceX listing on the stock market in just a few weeks, I’m sure there are going to be a lot more eyes than usual on this upcoming Starship test. In the risk factors section of its prospectus, the company acknowledged (in bold print, no less) that “Any failure or delay in the development of Starship … would delay or limit our ability to execute our growth strategy, including the deployment of next-generation satellites, global satellite-to-mobile connectivity, and orbital AI compute.”
Discovery of the Week: Curved Origami Structures
There are a lot of activities–from camping to space exploration–that require you to carry materials with you for shelter or other use. This can be something of a hassle: the more flexible and light a material is, the easier it is to carry, sure. But it’s also typically less strong, and portable origami shapes in particular tend to be rough and jagged rather than smooth.
Researchers at McGill University may have found a potential solution. In a recent paper, they describe a new type of material that combines cable- and origami-style folding to turn flat materials into curved structures. The cables, which are on the interior of the surface, can be adjusted to move from being flexible and soft to rigid and load-bearing–and back again.
“Our approach opens new avenues for the design of deployable and adaptive load-bearing curved structures,” co-author Damiano Pasini said in a press release. Among the potential applications for the new materials: emergency shelters, space structures and medical implants.
Sam Altman-Backed Retro Biosciences Is Expanding Its Longevity Research
Joe Betts-LaCroix
Retro Biosciences
Retro Biosciences, an AI-powered longevity startup backed by Sam Altman, has raised a new round of investment at a $1.8 billion valuation. (The total amount of fundraising was not disclosed.) The investment was led by 4P Capital.
The company’s stated goal is to “add 10 years to healthy human lifespan” and it already has one therapeutic in clinical trials. That drug’s goal is to increase “autophagy”–a biological process where the body cleans up and recycled damaged cells, which slows down as we age. This has been a rapid drug program so far, taking only 15 months to go from initial concept to human testing.
The company has two other programs getting ready for human testing as well: a cellular therapy that aims to replace old brain cells called microglia with younger ones to treat a form of Alzheimer’s disease, which will go into trials in 2026. The other is a therapy to replace old or missing blood stem cells, in partnership with the Murdoch Children’s Research Institute, which will go to the clinic in 2027.
Retro CEO Joe Betts-LaCroix told me that all of those programs are fully funded. The purpose of the new round of capital, he said, is for “additional discovery” and the company is going through a process to determine what problems related to aging it aims to tackle next. “There’s a bunch of things we want to work on.”
The Hot Take: Biological AI Needs Data
Each week, I ask investors for their take on tech trends within their industries. Today the answers come from Jory Bell, a general partner at Playground Global. Bell focuses on deep biotech startups like Ultima and Strand Therapeutics. This week, the firm raised a new $475 million fund to invest in deep tech startups.
Jory Bell
Playground Capital
What tech is being overhyped right now?
Generative AI for binder design [a process to create synthetic drugs] in that it is necessary, but not sufficient that people think of this as something that is going to move the needle. You need to go much, much deeper into the development of a medicine with AI in order to not hit Amdahl’s Law and revert to artisanal methods.
What should more people be talking about today?
I don’t know if this is sufficiently underhyped, but I think it is despite the hype it’s getting that nothing is more perfectly fit to the use of artificial intelligence than untangling the insane complexity of biology, but people are not yet acknowledging the scale and quality of data necessary to train on that. This is changing, but I just don’t think it’s changing enough or acknowledged enough yet that the most well-known and successful kind of breakthrough in AI is Alphafold, which was trained on the protein DB, which is 170,000 data points.
That is such a humorously small number, but that was the best and you got something that had been this white whale of protein folding for so long. That needs to happen at many orders of magnitude more. And so finding ways to transform that data generation, again, from an artisanal thing to an industrial scale and the right type of data is what is necessary. And I think even now we are currently, companies like [Playground portfolio company] Manifold are really struggling to even educate the frontier labs who say that the next most important thing after coding is solved is biology, but don’t yet appreciate the data inputs necessary to make that successful–even as it is entirely within our grasp now. The insane claims of “We’re going to cure all disease” – that is closer than you can imagine and yet farther in terms of the data inputs than people acknowledge.
What are we all going to be talking about in five years?
I think we’re going to have the true bio + AI companies emerge, which will be the Genentechs of this century. They are going to be created in the next five years.
On My Radar
Space Stuff: NASA Administrator Jared Isaacman said he expects China to send a crewed mission around the Moon next year, in a similar mission to Artemis II. The agency also announced a major reorganization this week, consolidating some of its current departments into larger ones.
More AI For Science: Google has launched a new set of AI tools that are geared for use by scientists within its Gemini suite. Meanwhile, OpenAI claims that one of its models has helped disprove a conjecture in geometry first proposed by Paul Erdős.
But Also…. Mo’ AI, Mo’ Problems? Preprint server ArXiv, where many papers are submitted before being published in peer-review journals, will ban authors who use hallucinated citations generated by AI. The bar may soon come up with similar discipline, as lawyers keep using fake AI-written citations. Meanwhile, Waymo has suspended freeway services for its self-driving cars, some of which are driving into construction zones and flooded roadways.
Quantum Gets Federal Backing: The federal government is putting more resources behind quantum computing by offering $2 billion in research grants to companies in exchange for equity stakes.
What’s Entertaining Me This Week
A few months ago in this newsletter, I sang the praises of the Netflix series Department Q. This week, I’m in the middle of reading The Keeper of Lost Causes by Jussi Adler-Olsens, which is the first novel in the series the show is based on. (Why the delay? Because the library queue was long, people.) It’s quite interesting to compare the two, since the show not only changed the setting but also some key details of the case. The TV series is faithful overall, and the book is a compelling read in its own right.