Next up: Tesla's Autopilot is actually just some dude in India driving your car with a Playstation controller.

https://www.businessinsider.com/amazons-just-walk-out-actually-1-000-people-in-india-2024-4?utm_source=copy-link&utm_medium=referral&utm_content=topbar

A new arXiv preprint from James Zou suggests that quite a few academics are also using ChatGPT to write their peer reviews.

I don't have to explain what an utter shitshow this is going to be, do I?

https://arxiv.org/abs/2403.07183

Another Elsevier paper with obvious AI-written text.

“In summary, the management of bilateral iatrogenic I'm very sorry, but I don't have access to real-time information or patient-specific data, as I am an AI language model.”

https://www.sciencedirect.com/science/article/pii/S1930043324001298

Thinking to myself this morning “Threads can’t be as bad as I recall…”

First post in my feed: Instagram influencer with a cup of coffee watching the sunrise from her balcony in a $1000/night hotel built with slave labor in a country where women can’t drive, telling me about how she has discovered that the secret to life is learning to savor the little things.

No, lady, apparently the secret to life is to be born rich, white, beautiful, and oblivious.

Anyway, that’s enough Threads for 2024.

The new book from philosopher Cailin O'Connor on modeling scientific communities has just been published.

And, it's free for download for the next two weeks!

https://www.cambridge.org/core/elements/modelling-scientific-communities/1ED3515216067E40A37A72094EE3CB15

Google Scholar is a flaming piece of shit. Why do we use it for scholarly evaluation?

Here, it misattributes 4873 citations — a decent count for entire research career — from Keeling and Rohani's landmark book, giving them instead to the authors who wrote a book review of the book.

Yet, there's no correction or appeal process.

Great news. I can't imagine a higher ROI investment to advance scientific research.

Kudos to the NSF and the Simons Foundation for recognizing the enormous value of what the arXiv provides us all.

https://news.cornell.edu/stories/2023/10/research-repository-arxiv-receives-10m-upgrades

1. “Imagine we land a space probe on one of Jupiters’ moons, take up a sample of material, and find it is full of organic molecules. How can we tell whether those molecules are just randomly assembled goo or the outcome of some evolutionary process taking place on the planet?”

#science #scicomm #assemblytheory #exobiology

If I were to set up a wordpress blog to write about bullshit, science, big tech, large language models, and all that, what would you think I should title it?

A short play in three acts.

ACT 1. September 25, 2023.

I wanted to consolidate a few thoughts on google, misinformation, large language models, enshittification, and the fate of the web as we know it.

It started when Carl Zimmer shared this remarkable example of Google being fooled by machine-generated bullshit online.

It's pure coincidence that Elon Musk just happened to erase the entire record of how twitter was used to organize and share photos and video from Arab Spring, right?

I always said that if I were a ballplayer, my walkup music would be the intro to London Calling. When that bass comes in it's just crazy.

Apparently Cedric Mullins is doing exactly this.

https://www.youtube.com/watch?v=LC2WpBcdM_A

Why is AI research so much sexier than education research? Sure it’s cool to study how machines learn but wait until you hear about PEOPLE.

ChatGPT detection and algorithmic bias:

This afternoon James Zou directed me to a recent pilot study from his group in which they looked at the performance of seven different GPT-detectors that are sometimes used to flag cheating in educational settings.

They found that these detectors commonly misclassify text from non-native English speakers as being written by an AI. A primary driver appears to be the lower perplexity (exponent of model's loss) of such text.

https://arxiv.org/abs/2304.02819

It sounds trivial and obvious, but are you reading error bars correctly?

Do you know whether you're looking at standard errors (measures of inferential uncertainty), standard deviations (measures of spread of individual observations), or 95% confidence intervals around the mean?

And are your intuitions about what each of these mean correct?

Here's a nice primer, refresher, or teaching article.

https://www.nature.com/articles/nmeth.2659

h/t @jakehofman

So I planted my flag of over at bluesky social.

During nesting season, birds don't have the option of freely approaching and freely departing.

If you approach a nest, you are moving into their space and creating intense stress.

I'm discouraged by how many birders—who should know better—ignore this. We, of all people, have to do better.
within 6 feet of the

I'm fascinated by the reciprocal nature of #birdwatching.

Birds are every bit as much watching us as we are watching them.

In my photography I try to capture bidirectional interaction in my photography.

I photograph birds watching me, but only when they can freely approach, and freely depart. I'm interested in their curiosity, not their fear.

This is a wild bird that approached me.

#birdphotography #photography #naturephotography

Below, a fledgling Anna's hummingbird that left the nest only hours before. My shots were taken with an 800mm lens from a distant concealed location.

But then other birders arrived, standing 2 meters away from the bird, using loud mechanical shutters. The mother panicked and stopped feeding.

I protested. They argued—"they eat from my hand in my yard!" and stepped back to 3 meters. Useless.

Photographers, respect your subjects. Learn their behavior. Your shots are not worth a bird's life.