Navigation
As we celebrate Artemis-II astronauts and their record breaking journey around the moon, a recent article in the Washington Post came to mind and I was scampering down a complexity/simplicity rabbit hole again. The bar-tailed godwit (if you don’t get to read the article - this photo alone should make you want to know more about that bird) B6, a 5-month old juvenile, set the record for the longest known non-stop flight by a bird - an astounding 8321 miles in 11 days. Non-stop. From Alaska to Tasmania.
(Article thumbnail credit: NASA)
But here is the big mystery: A four-month-old godwit born and raised in Alaska must somehow “know” where to go for the winter. And it does not follow its parents, who migrate before the nestlings are physically ready for the trip. Instead, young godwits gather and repeat the adults’ flight south. To do this, the birds have to possess a hardwired program to guide them on this maiden flight. It is evident that these young super-migrators possess a capacity akin to the GPS of a smartphone. How is that possible?
While there are intelligent hypothesis about the navigational mechanisms of these super-migrators, it appears nobody really knows how exactly they do it. Compare that to our navigation infrastructure and reliance on it. Many of us cannot get from home to work these days without the satnav capabilities of our phones. The Artemis-II (and all space missions) rely on an elaborate navigation and communication infrastructure. The moon is an average of 238,855 miles from the earth. An adult godwit is assumed to travel 285,000 miles in its lifetime - unerringly finding its habitats in New Zealand (or Australia) and Alaska. I wrote about how things should be only as complex as they need to be - I wonder if our navigational technology could be simpler. That is the beauty of it, there is enormous simplification still ahead. We just have to master the science and learn to emulate the amazing bar-tailed godwit. We need the humility to realize that simpler solutions might exist, as we engineer ever complex solutions to the problems we face. Whether the problems are ones related to navigation or geopolitical disagreements.
Mythos - an AI and Cyber twofer
If you did not follow the Anthropic news this past week, here is a quick summary.
Anthropic announced that they are not ready to release their latest model (named Mythos) to the general public for fear of its capabilities to identify and exploit cyber vulnerabilities at scale. While this might be a bit of clever marketing and grandstanding, it was followed up by a meeting of federal officials and major US banks.
The leaders of some of America’s largest banks were warned by a top government official this week about a new artificial intelligence model from Anthropic that could lead to heightened risks of cyberattacks, according to three people briefed on the matter but not permitted to speak publicly.
Scary stuff. These are the kinds of articles which make me ruefully admit that I am glad that I retired (Sue and Lori can deal with it).