The Universe is a large place, yet despite it being mostly empty space, there are still a lot of things to find and catalogue. This includes mildly terrifying things like supermassive black holes (SMBHs), one of the study subjects of the Sloan Digital Sky Survey (SDSS) project. In their 20th data release of the fifth all-sky survey (SDSS-V), the results of the Black Hole Mapper (BHM) program provides a lot of new insights into these SMBHs.
For a good primer on the SDSS’s ongoing survey, you can read this paper by [Kollmeier] et al. from 2017 in which this fifth survey and its three programs, including the BHM, are explained. This comes after four previous phases of the SDSS, all of them focusing on multispectral imaging and spectroscopic redshift survey with the 2.5 m Apache Point Observatory (APO) in New Mexico.
With SDSS-V a second 2.5 m observatory at Las Campanas (LCO) was added, with both observatories combined able to observe the entire sky, not just as static images, but also any changes over time. While these observations are in the near-infrared, combined with the data from other observatories this gives us probably one of the most comprehensive maps of the Milky Way and everything therein, including black holes.
This 20th data release gives us one of the clearest glimpses yet at the formation, growth and behavior of SMBHs and similar objects over time. A big part of this achievement are the automatic positioning robots at the observatories that handle the fiber optics that feed spectrographs, enabling faster and more accurate observations.







For this project, [Andrea] began by porting the game to the SCINTIX P4. It’s a rather interesting device, being a single board designed in the Raspberry Pi CM4/CM5 form factor, but carrying an ESP32-P4 and an ESP32-C6 instead. The game runs on the P4 and is displayed on a 1024×600 MIPI DSI panel, but it’s only stepped through a few frames at a time. These frames are then passed to Claude Sonnet via a WebSockets setup. With only the same information as a human player would get, the LLM has to figure out what it’s looking at, and then respond with movement and fire commands to play the game.