Hackaday Links: March 17, 2024

A friend of ours once described computers as “high-speed idiots.” It was true in the 80s, and it appears that even with the recent explosion in AI, all computers have managed to do is become faster. Proof of that can be found in a story about using ASCII art to trick a chatbot into giving away the store. As anyone who has played with ChatGPT or its moral equivalent for more than five minutes has learned, there are certain boundary conditions that the LLM’s creators lawyers have put in place to prevent discussion surrounding sensitive topics. Ask a chatbot to deliver specific instructions on building a nuclear bomb, for instance, and you’ll be rebuffed. Same with asking for help counterfeiting currency, and wisely so. But, by minimally obfuscating your question by rendering the word “COUNTERFEIT” in ASCII art and asking the chatbot to first decode the word, you can slip the verboten word into a how-to question and get pretty explicit instructions. Yes, you have to give painfully detailed instructions on parsing the ASCII art characters, but that’s a small price to pay for forbidden knowledge that you could easily find out yourself by other means.

While on the topic of high-speed idiocy, it seems like Google’s new Bay View Campus suffers from a few design deficiencies that are causing headaches for the Gemini AI team working there. According to workers there, the 600,000 square foot (55,700 square meters) building has terrible WiFi coverage, to the point where they’re resorting to Ethernet cables and WiFi hotspots on their phones to get work done or just going outside to get some signal. Employees blame the unique roof of the building, which almost looks like an anechoic chamber. It’s a striking building, to be sure, but it would probably be better if people could actually get some work done in it.

If this week’s installment of “Stuff Elon Says,” it looks like future Mars-bound Starships will feature spin gravity for the comfort and health of their passengers. “Even a tiny gravity vector is better than none,” says Elon, but we’re not too sure. The current generation of Starship is about 9 meters in diameter, which is pretty big as spacecraft go but still not that big. So someone 1.7 meters tall standing inside a spinning Starship will span about 37% of the radius, which is probably going to feel really weird — full spin gravity at the feet, but less than a third of that at their head. We suppose it would be tolerable lying down against the inner surface, but standing seems like it might be a recipe for puking all the way to Mars. And that’s not to mention the Coriolis forces. No thanks!

Bad news for Avi Loeb this week, as a preprint apparently pokes holes in some of the seismic data the Harvard astrophysicist used to classify a 2014 meteorite as interstellar in origin. Benjamin Fernando, a planetary seismologist at Johns Hopkins, analyzed data from seismometers in Papua New Guinea and found a periodicity to the data that appears to correlate with the comings and goings of trucks on a nearby road. One would think that this would have been accounted for by whoever runs the sensors at some point in the past, but maybe not? Loeb, who has become something of an outcast in astronomy circles for his speculations on extraterrestrial civilizations, shot right back, noting that the primary data he used to estimate the origin of the 2014 meteor was satellite imagery and that the seismic data was only supportive of its interstellar origins. He also stands behind the chemical analysis of the material recovered from the putative impact site, which seems to suggest the stuff came from a “magma ocean” on an exoplanet with an iron core. As we recently opined, Loeb does seem to have the numbers on his side, but he still has a lot of work to do to make his case.

Last week, we dropped a link for a fascinating virtual tour of the coolest airplane ever built, the SR-71 Blackbird spy plane. This week, we were pleased to find a full behind-the-scenes explanation of the process used to create the video by Jake O’Neal, the artist behind the Animagraffs channel. The amount of work he did creating the model of the Blackbird is just mind-boggling, and the level of detail he managed to pull from some “copies-of-copies” images and fuzzy photographs is astounding. We’d have thought the bulk of the work would have been in Blender, but it seems like the time he spent researching the plane vastly outweighed the time spent building the model. There are some fantastically retro source docs in the video, too, which any aviation buff needs to check out.

And finally, this brief video of a fully automatic flap-disc machine caught our eye because — well, just because! We used to work in laboratory automation, which is sorta-kinda similar to industrial automation, and the satisfaction of turning a manual process into something that machines can do quickly and reliably is hard to explain. The machine in this video is a great example of the creativity automation engineers have to employ to make the things we need to run our lives even remotely affordable. Enjoy!

What's your reaction?

Related Posts

1 of 514