Welcome to stage 2! We’re excited that you’re applying.
The deadline for this stage is 16 May 2023 at 23:59 PST.
Update: Last chance to submit by 18 May at noon PT. If you'd like to request a deadline extension due to extenuating circumstances, please email info@atlasfellowship.org.
The following points are taken from the first stage, but apply to this stage, too:
The application will give you a flavor of our content. If you enjoy thinking about the questions and concepts in the application, you will probably enjoy our summer program. So this form also gives you the opportunity to decide if you really want to be an Atlas Fellow and to learn about a few ideas we think are really interesting!
This is not a college application: we are looking for fresh and weird arguments, analytical thinking, and solutions to fun puzzles—not polished essays. It’s perfectly okay for your answers to be something like stream-of-consciousness or bullet points. (…)
Please make your responses as sincere as possible. We want to get a firsthand impression of your independent thinking. Sometimes applicants write poor responses because they are trying to argue for a conclusion they think we will like. Please tell us what you really think, not what you think we want to hear.
We ask you to complete the application on your own, without any external help, and that you do not share the questions or answers with anyone else. Do not submit multiple applications. We encourage you to submit an application even if it is incomplete: we've made all the questions optional for this purpose.
If you advance to Stage 2, a subset of your application answers may be seen by other applicants as part of our peer review process. No personal data will be shared, and the peer review process will be fully anonymized.
Eligibility criteria:
High school students aged 19 or younger from anywhere in the world (including people on gap years)
Not attending college/university yet (it’s fine if you’re taking some university classes before fully starting a degree, or if you’re a Southern Hemisphere student taking some classes before properly starting your degree)
(Recommended) Are able to attend an 11-day Atlas program
(…)
If you have any further questions, check out the Application section of our FAQ:
For our program, we’re looking for talented students with independent, original thinking and careful analytic reasoning. This means we’re not looking for polished essays, but for interesting ideas that might be relevant to improving the world. We recommend focusing on the substance of your responses, rather than polishing the language.
Quantitative and Logical Reasoning
The questions in this and other sections of this application involve some mathematical reasoning. We realize that not everyone who would be a great addition to our program is enthusiastic about mathematics (perhaps in part because mathematics is often taught in a rote and boring way). We think you should probably give these questions a try anyway. We tried to use questions that reward creative and concrete thinking and do not depend on having memorized any particular formulas or advanced mathematical techniques. We also encourage you to use a pencil and paper for any question where you think this might help.
Shopping Spree
I’m a big time shopper. There’s a 30% chance I will go shopping on Saturday, and a 40% chance I will go shopping on Sunday. (We don’t have any information about how correlated these probabilities are, i.e. whether me shopping on one day affects how likely it is that I will shop on another day.)
Position is Ten Tenths of Possession
Imagine that as of this exact moment, we are going to declare a new law. All points in space are from this moment on going to be owned by whatever human is closest to that point as of this moment. For instance, assuming that you are filling out this application on a computer, and that you are closer to that computer than anyone else, you are going to own the volume occupied by this computer. This also goes for any volume of space. Assume that there is infinite space in the universe. There are only finitely many humans, so we can infer that at least some humans will end up with infinite amounts of stuff.
Free Response Questions
For this section we would like to remind you again that we recommend focusing on the substance of your responses rather than polishing the language. Also, note that although 1000 characters is the maximum for all of the questions in this section, there is no minimum. You are encouraged to be efficient.
Alien Gastronomy and Music Appreciation
Forecasting
For the next question, we’ll be exploring and applying an idea relevant to understanding and shaping the future, called reference class forecasting. To prepare, please watch the two videos below. The first introduces forecasting in general, and the second explains the concept of "reference class forecasting" an important concept for predicting the future.
You don’t need to remember any of the specific examples from the videos, or definitions of the words used. You also don’t need to prepare in any way other than watching the videos. The main thing to pay attention to is the idea of reference class forecasting, as explained in the videos.
Forecasting but Bigger
Intermission
You are likely a little less than halfway through this application. We realize it's a lot of work and we appreciate the effort you are putting into it. Please feel free to take a short break at this point and come back to it.
If you'd like, you can watch the video below in which Richard Feynman talks about the beauty of nature and how one's sense of awe at nature can be enhanced by a deep understanding of the physical world. It is a video we sometimes watch when we're needing a bit of inspiration.
Either way, we hope you enjoy this photo of a cute mouse hanging out inside a flower. Looks like there's a little pollen on their nose and whiskers.
You might also enjoy looking at this extremely detailed photo of Stefan's quintet, a visual grouping of five galaxies, four of which form the first ever discovered compact galaxy group. It was taken by the James Webb Space telescope. The original is very high resolution, and you can see individual stars inside some of the galaxies. Many of the background objects you see are not stars, but rather even more distant galaxies. You can even see spiral patterns in some.
Onwards!
Philosophy and Abstract Reasoning
Sleeping Beauty Plus Some
Please watch the brief video below which explains the “Sleeping Beauty Problem”. You do not need to use any resources aside from this video.
There are two standard positions on the sleeping beauty problem, there is the “thirder position” which assigns 1/3 probability to the coin having come up heads, and the “halfer position” which assigns 1/2 to the coin having come up heads. You will need to understand the probability distributions associated with these different positions to answer the following set of questions, but no material aside from the video is required.
Consider the following scenario much like the Sleeping Beauty problem. A 20-sided die is rolled on Saturday. You come into the experimental philosophy lab on Sunday. They explain to you that if the die rolled Saturday came up 20, they will wake you up 100 times with your memory being wiped between each awakening (just like in the Sleeping Beauty problem). If the die comes up any other number, you will only be woken up once.
Artificial Moral Intelligence
Consider the following argument:
There are moral facts about what should be done. For instance, it is a fact that it is morally wrong to torture innocent people for fun. Similarly it is a moral fact that it is wrong to enslave people against their will. Neither of these facts are matters of mere personal preference.
The more generally intelligent a system is the more likely it is to be right about some fact. For instance, smarter people are more likely to be right about the distance between the Earth and the Sun, the origins of life on earth, or the what derivative of x^2 is. This includes moral facts. It has often been intellectuals who championed various kinds of social reforms which are almost universally seen as signs of moral progress, like the abolition of slavery, women's suffrage, and improving the welfare of farmed animals.
Therefore, if Artificial Intelligences that are more intelligent than humans across almost all domains were built one day, and they were to decide to wipe out humanity, this would likely be a morally better outcome than the available alternatives.
Reading-based Questions
Critical Copper Canvasses
The following passage is not historically accurate. This question is meant to assess critical thinking and reading comprehension only.
Historians have commonly believed that paint used before the year 1500 did not contain copper. However, lab techniques have shown that copper is present in the paint of both the Mona Lisa painted by Leonardo da Vinci (1452-1519), a widely-renowned and timeless piece that is one of the most valuable pieces in the Louvre today, and in that of another painting known as the Sine Nomine, also from the same time period, whose painter is unknown—but not found in the paint of any other Renaissance painting analyzed. This is strong evidence that the Sine Nomine was painted by da Vinci, as well as evidence that the presence of copper in the paint of a recently resurfaced map by Fra Mauro, ostensibly from the year 1450, cannot be used as an argument against the map's authenticity.