What I Learned Today

I recently got to visit Alaska and see and learn about orcas. I knew a little bit about them before, but I don’t think I’d realized how _smart_ they are. In a talk I attended by Michelle Nussey, I learned that they can work together to kill even an enormous blue whale – by drowning it. Not only that, they practice this technique together, with one of them pretending to be the whale. Thrill along to David Attenborough’s narration:

They can also swim in tandem to create a bow wave to knock a seal off an ice flow, and more:

Michelle also noted that they will sometimes deliberately beach themselves, so that they can snatch and devour prey that thinks it is safe on the land. She added, maybe only half-joking, “If orcas ever grow legs, we’re in trouble.”

Apparently they have even attacked and eaten moose. (This was also when I learned that moose like to swim and will dive up to 15 feet to eat tasty aquatic plants – but is it really worth the risk, swimming where there are ORCAS?)

Finally, Michelle shared that orcas are one of the few non-human species that experience menopause. Apparently it is quite unusual for female mammals to get any downtime at all from fertility, until they die. So then I wondered… does this mean that orcas have periods? (Menopause is the end of menstruation.). It turns out, no – it’s a misnomer to refer to this phase as “menopause”, because they never had periods to start with. Instead, it is technically a “postreproductive life phase”. Scientists have studied orcas to try to figure out what this (rare) shared phenomena might tell us about humans:

“It suggests that intervention should focus not only on restoring or mimicking ovarian function but also on the whole organism’s overall long-term health. This means that menopause should not be regarded as only a loss of fertility and ovarian function but as a transition from one evolutionary function to another.” (“Menopause in nonhuman mammals: What does it mean for the gynecologist?”, 2025)

That is, maybe there are good reasons to go through hormonal changes and enjoy life that doesn’t revolve around reproduction. Thanks for this nudge, orcas! I’m still creeped out by your hunting tactics, though.

We are still figuring out how best to interact with AI systems, and how we want them to behave towards us.

A study addressing that question just came out: “Autonomy or Guidance: What Users Want from AI versus Human Advisors” by Yasmin Wiesenfeld and Oded Nov.

The researchers assessed human preferences for how much guidance versus autonomy they want when getting advice about health and finance, from a human expert vs. an AI system. They found that people prefer “complete agency” (just give me the options) from an AI but “guided agency” (give me options plus your recommendation) from a human, “suggesting that people assign different normative roles to AI and human experts.”

Having different preferences makes sense if you have a different mental model of the expertise, reliability, and capabilities of different conversation partners. This deviation suggests that there persists a gap in how people view generative AI chatbot capabilities, compared to a human expert.

This paragraph in the Discussion adds an extra twist (and concern):

However, research suggests that when AI systems are anthropomorphized, this trust gap is reduced: although users are significantly more likely to follow advice from a human (e.g., romantic partner) than AI, this difference is nearly eliminated when the AI is anthropomorphized [19]. This suggests that perceptions of AI’s role as an advisor may be malleable through design choices that emphasize human-like qualities, even if they do not affect the level of user autonomy.

Without a doubt, generative AI creators already knew and have been exploiting this tendency. The original GPT language model could have been released as just a language model, which researchers would have continued using to analyze and process language. Instead, the choice was made to personify the model by making it into a chatbot that uses the pronoun “I”.

And finally, I found this bit very intriguing:

However, the findings also raise a normative question: should AI systems reflect users’ stated preferences (what people claim to want in hypothetical scenarios), or their revealed preferences (what they choose in practice)? Despite overwhelming preference for agency-preserving designs in this study, many people continue to employ agentic AI systems that make decisions for them.

We will probably never crack the mystery of “why do people do things at odds with what they say they want?” — but some of that is what makes us interesting personalities ourselves. Still, for AI system designers, this is an important point to reflect on, since choices have to be made to enable the creation of any system in the real world.

I came across this great course (Electrical Engineering for Everyone), taught by Dr. Laura J. Bottomley (NCSU) and offered through the Great Courses (which I am accessing through the Kanopy app – your library card may provide it to you free!).

I’ve only watched the first lecture so far, but it was surprisingly riveting. It is aimed at folks who aren’t necessarily aiming to get a degree in Electrical Engineering, but who want to learn more about how electricity (and electronics) enable so much of our modern lives.

One of my favorite aspects is how Dr. Bottomley stands in front of all her cool gadgets and then demonstrates them. They included a lightbulb, a cute little Tesla coil and fluorescent lights (although I was a little disappointed that she didn’t say anything about how the Tesla coil worked, just that it creates an electromagnetic field), solar cells (which were first invented (but not very efficient) in the 1880s!), and a simple circuit with an LED.

She identified four major advancements in electrical engineering: the ability to generate and transmit electricity through wires, then wirelessly, then electronics (computers) / photonics, then “intelligent applications” (artificial intelligence).

She also talked about six “habits of mind” that engineers are supposed to have: communication, creativity, collaboration, systems thinking, ethics, and optimism (learning from failure). I had not previously encountered these habits of mind. They all seem pretty relevant and reasonable, but I was intrigued by “optimism” making the list!

She also touched on a lot of interesting bits of history, including Thomas Edison, who invented the lightbulb, but his filaments only lasted a few days. His contemporary Lewis Latimer (a new name to me) was a key contributor to making lightbulbs accessible to the masses: he extended the longevity of light bulb filaments while working for a rival company, then was later hired by Edison, and he also helped with patent applications for Alexander Graham Bell.

Today I learned about the impressive collection of maps held by the Library of Congress, which has the largest map library in the world (6 million items). These items are publicly available to anyone with a (free) Library of Congress card, and there are also a lot of digital maps available online.

The Geography and Maps Division is currently featuring key maps from the American Revolution period (to commemorate the U.S.’s 250th anniversary). These 1,470 maps include a map of Philadelphia that includes a sketch of the State House where the Continental Congress met to hash out the Declaration of Independence, a Rochambeau map of the 1776 Battle of Long Island, the first map of the United States showing them as an independent country (1783), and more.

First map of the United States as an independent country (1783)

They also have an collection of Revolution era maps made accessible to visually impaired individuals. They have created “tactile maps” that reproduce maps with raised outlines, Braille labels, and textures, as well as audio descriptions of the maps. To make these tactile maps, they first choose maps likely to translate well to this form (ideally not too complicated), identify lines and edges, add Braille labels, then add texture to shade different regions. The maps are printed with a regular printer, on special paper, then heated and the places with ink swell into ridges you can feel. Awesome!

My screenshot of the tactile map example during the June 9 webinar

Finally, the “By the People” program at the Library of Congress, where you can volunteer to help transcribe handwritten materials, will soon be launching a project to transcribe nautical maps from the Atlantic Neptune Collection. It should show up at By the People when it opens. You can help enrich this collection!

You may know that you can disable the “AI Overview” in a Google search for a single search by including “-ai” (minus ai) in your search text. But Google doesn’t give you a way to turn it off for all searches.

However, you can configure your browser’s built-in search shortcuts to do it for you. A regular Google search is something like

https://www.google.com/search?q=text

where “text” is the text you are searching for. But if you add an extra parameter (udm=14) to this search, it disables the AI overview:

https://www.google.com/search?q=text&udm=14

You can make this version your default search shortcut in Google Chrome as follows:

1. Open Settings.
2. Go to Search Engine -> Manage Search Engines.
3. Scroll down to the “Site Search” section and click “Add”.
4. Give this search a name (like “udm14”).
5. In the “URL with %s in place of query” box, paste this: {google:baseURL}search?q=%s&udm=14. It will look like this (mine says “Edit”; yours will say “Add”):
   
6. Click “Save”.
7. Under the three dots menu, select “Make default”.
8. Try a search and enjoy no AI overview!

I have not yet figured out how to do this in Safari.