Is a Self-Iterating AGI Vulnerable to Thompson-style Trojans?
In his 1984 lecture "Reflections on Trusting Trust", Ken Thompson (of Unix fame) speculated about a methodology for inserting an undetectable trojan horse within the C compiler binary that would self-propagate throughout all future versions. (Additional good video that got me thinking about this.)
The replacement code would miscompile the login command so that it would accept either the intended encrypted password or a particular known password. Thus if this code were installed in binary and the binary were used to compile the login command, I could log into that system as any user.
What does the reaction to NFTs tell us about how people evaluate ecological damage?
Recently, the Ethereum Foundation finalised the ERC-721 interface standard for Non-Fungible Tokens (NFTs). This standard lays out a protocol for the exchange and ownership of this new class of assets. An extremely simple explanation of the ERC-721 standard is as follows:
A blockchain is a decentralized ledger. It can contain things called smart contracts, which are like little computer programs. A smart contract can implement the ERC-721 interface, which means that it keeps track of variables called Non-Fungible Tokens (NFT). Each token has an owner, a unique identifier, and can be traded to other people. Each token can also include metadata information about the object it is representing.
The Price Of Space Is Friendship
It's an exciting time in the history of human exploration. We're only just beginning to probe outside our biosphere. Low-orbit, long-term space inhabitation like that seen on the ISS has been the norm for many years now. As of the time of writing, 7 people were in space. Our long term goals seem set on permanent outposts on Mars.
Orbital re-entry of crewed spacecraft is a complicated dance of physics. You ever do that challenge where you have to build a cage out of straws and cardboard to protect an egg from a fall? Well, it's like that on steroids. Generally, if you are coming in from another interstellar body you will be going very fast. You can't decelerate over around 90m/s/s (or 9gs) or let the interior of the ship get too hot without killing your crew. You've got an enormous amount of kinetic energy to disperse, and the best way to do that is by braking in the atmosphere at a relatively shallow angle of descent. This means that reentry maneuvers cover a lot of airspace. You will likely either orbit the earth entirely or a large portion of it before you finally reduce your velocity enough to land.
The year was 2018, and I was burning rubber southwards down the Peruvian coastline with a deadline for Santiago, Chile. My van - a Chevy Astro 2008 coined Babo Conquista - was my home and my world as I travelled, with a bed in the back, gas-hungry engine up front, and the heady musk of habitance all around. I was at the end of my journey, financially and emotionally exhausted, ready to lay my head down on a more permanent pillow after 10 or so months of vagrancy. On my way to sell the van in Santiago, I reached the northern Peru-Chile border. It was here that I encountered one of the more stressful experiences of bureaucracy in my life.
When you enter Peru in a vehicle, you are given a Temporary Import Permit (TIP) that gives you a certain amount of time to have the car in the country. When I entered Peru via Brasil, it turns out that the customs officer misread my paperwork for the van. Where the customs officer was meant to print my name, they instead printed the name of the previous owner which was also on the deed. In my haste, it was not until I had been in the country for a week or so and was far away from the border that I recognised this mistake. But, I thought, surely the inconvenience will be minor. All the permanent paperwork indicated that I was the owner, and the mistake in the TIP was fairly obvious when pointed out. So, when I hit the border and saw those SUNAT offices, I thought it would be an easy time. I was very incorrect.
Ethics & Arbitrary Objects
There are many abstract, mental systems that we build throughout our lives. We build systems firstly in order to interpret our memories and sensory input into an internal simulation of "reality", our experience up to that moment. Secondly, we then run simulations of the future and predict their probability. These systems are abstract objects that exist within your mind that define your experience within the world.
However, at some point, your systems will come to the final part of this process of existence: selecting which predicted future you should pursue. Eventually you have to answer questions like "Why should we reduce human suffering?" This will involve using some form of heuristic to determine future value, and then balancing that value against probability.