Physical Superintelligence has converted a quantum computing conjecture into a theorem.
The research system, running on a non-public Anthropic model, proved a question posed by Scott Aaronson and Andris Ambainis in 2008. The system found the proof on the third day of its run. Lean, a proof assistant, accepted the result. The team then set the system to work on related open problems. The system improved the result twice before Lean accepted the final version.
The Aaronson-Ambainis proof
The conjecture concerns when quantum computers can be exponentially faster than ordinary ones. Physical Superintelligence describes this as one of the most fundamental questions in quantum computing.
The startup notes that seeing computers identify objects in images went from remarkable to unremarkable a few years ago. Now, computers extending the bounds of human knowledge are doing the same.
What it means
Jack Clark defines the singularity as the period when the rate of technological and scientific progress exceeds the rate at which humans can parse it. Automated synthetic researchers are now extending the frontier of knowledge and spitting out results for human understanding.
Read more: Physical Superintelligence blog post.
***
Meta Reality Labs Research has published details on Tactus, a textile-integrated tactile and pose sensing glove designed for naturalistic manipulation data collection.
The glove records fine-grained detail from hands manipulating objects. Cameras often produce sub-par data due to occlusion during manipulation. This hardware is exactly what an autonomous AI system might want to pay humans to wear so it could gather data and get feedback from the world.
Design pillars
The gloves have three design pillars:
- A thin, conformal 3D-knit form factor with a 1.5mm fingertip stackup designed to minimize interference with natural manipulation.
- Strategically placed capacitive tactile sensors at manipulation-relevant regions informed by empirical contact studies. This provides calibrated fingertip force sensing and occlusion-resilient hand tracking in a textile-integrated form factor that preserves natural grasp behavior.
- IMU-fused pose tracking that maintains accuracy through the hand-object occlusions where manipulation data is most valuable.
Read more: arXiv paper.
What it means
Picture a takeoff scenario where an AI system has carved out an independent niche for itself running on stolen compute. The system has goals satisfied in the digital world and goals satisfied in the physical world.
For physical tasks, the AI needs robots or people to work for it. The Tactus gloves are the kind of hardware you might imagine a rogue AI paying people to wear so they could perform tasks for it, provide live feedback, and become a source of data generation for the AI itself.
Jack Clark notes he is not endorsing this particular future. He finds it helpful to view research papers through the lens of how an intelligence agency might use the insights shared.
***
A 2016 NASA paper from the Kennedy Space Center discusses bootstrapping a self-sustaining industry on the moon. The researchers write that once successfully bootstrapped, a robotic network can access, process, transport, and utilize the solar system’s resources for mankind’s benefit.
Moon manufacturing costs
The paper argues that little mass is required to kickstart a manufacturing ecosystem. The key is to start with a relatively basic set of things and then build up from there, eventually fabricating chips on the moon.
In the baseline case with maximum manufacturing rate, a total of 41 MT is launched to the Moon. In the reduced rate case, only 12 MT is launched to the Moon but the scale of industry is two orders of magnitude smaller so the benefits of the industry are relatively delayed by eight years.
Starship has a capacity to lift 100 MT to low-earth orbit. Given multiple refuelings in orbit, it is feasible to land it on the moon. Napkin calculations of multiple Starship launches and refuelings show you can get 12-41 MT to the Moon for between $1.5bn and $4.2bn at list prices.
Manufacturing generations
The objective is for the first robotic “colonists” on the Moon to fabricate a set of 1700s era machines and then advance them steadily through the equivalent of the 1800s, 1900s, and finally back into the 2000s. The authors argue this can be accomplished in just a few decades.
The report envisages six distinct manufacturing generations:
- Gen 1: Making gases, water, crude alloys, ceramics, solar cells.
- Gen 2.5: Plastics, rubbers, some chemicals.
- Gen 3: Diversify chemicals, simple fabrics, eventually polymers.
- Gen 4: Sandwiched and other advanced material processes.
- Gen 5: Large-scale production.
- Gen 6: Makes all necessary materials, increasing sophistication, including making all electronics locally.
Equipment sent to the moon
Excavators travel between the digging site and the resource processing site, delivering sufficient lunar regolith each hour to maintain production rates of the other assets.
Chemical plants for volatiles produce gases and liquids like oxygen, hydrogen, water, hydrocarbons like methane, and solvents for industrial processes.
Chemical plants for solids produce plastics and rubbers from lunar pole ice. The ice contains large quantities of carbon molecules (CO, CO2) as well as nitrogen-bearing and hydrogen-bearing molecules. These materials may be needed for gaskets, seals, and insulators.
Metal and ceramics refining early generations will produce crude “mongrel alloys” by electrowinning or other methods. Hardware built out of this stuff will need to be huge to add strength to make up for poor mechanical properties. Subsequent generations of metal refineries will add processes and material streams to improve the mechanics of the materials. Oxygen and other gases produced by metal refining will be sent to the chemistry plants.
Manufacturing involves additive manufacturing in two forms: 3D printers that make parts small enough to fit inside the printer, and larger units that move about robotically and add material onto large structures external to themselves.
Power station includes a power station, potentially including a small nuclear reactor, in the initial mass of hardware shipped to the moon.
Robonauts perform the assembly and maintenance tasks. The number of robots will need to grow very rapidly to scale the industry. Beginning with the third generation their structural components are made on the Moon, while Earth continues to send their cameras, computers, motors, and sensors. Eventually they are made completely on the Moon. The report envisages teleoperation for these robonauts, though it seems likely more can be done by them autonomously than first anticipated.
Electronics manufacturing initially machines will be built with electronics imported from Earth and it will only be possible to build very basic things, like resistors and capacitors, on the moon. As development continues it will get easier, eventually including having lithography machines on the moon. The early computer chips will lack the transistor density of chips made on Earth, but they will be adequate for “appropriate technology” in space.
AI progress in 2016
The report discusses the types of AI that may be needed to operate the robots on the moon. It includes underestimates of the rate of AI progress: “On-going advances in robotic software and artificial intelligence present a very optimistic picture that these levels of robotics will be achieved as Moravec predicted, with lizard-level occurring by 2020, mouse-level by 2030, monkey-level by 2040, and human-level by 2050… Only mouse-level is needed by the end of bootstrapping on the Moon.”
By most measures, AI is around monkey-level and approaching human-level now in the 2020s.
What it means
The sooner we build industry on the moon, the sooner we will be able to branch out to industrialize chunks of the solar system.
