The Satellites of Earth, Mars, and the Moon

We always forget: there are a bunch of satellites orbiting Mars and the Moon. Perigee brings it all together!

Lean & AI

Lean is the hip thing in math these days, especially in connection with AI. That's all I know, but here's more: ‘It is incredible’: How AI is transforming mathematics

Why do we need gen-eds?

Or, perhaps the better question is this: how many gen-ed classes should be required to get a Bachelor's degree? 10% of all classes? 25% of all classes? 33% of all classes? Trust me when I say this: nobody wants to take them, teach them, or pay for them.

Why science needs the huaminties more than ever (Nature, 25 June 2026) argues the usual: STEM fields, alone, cannot offer the ethical reasoning, historical context, and cultural connecctions that are required to be an informed and responsible citizen of the world. 'Hard' technical expertise is critical but has to be complemented by 'soft' human skills.

That's fine & I don't think anyone disagrees with that. But here are the practical questions:

1. Do we really need a take an array of special classes that have nothing to do with our desired profession to get these 'soft' human skills? Or, could this be infused & embedded in classes that also offer pratical professional skills?
2. Let's agree that we need special gen-ed classes: how many? Currently we basically force students to take about 1/3 of all their classses in the gen-ed curriculum - that seems too much to me.

Here's my proposal: create a mandatory Minor that all students need to take that consists of 6 classes or 18 credits: 1 science class with a lab, 1 math class, 1 writing class, and three more classes in the arts, humanities, and social sciences.

Svalbard: Fossil Fuels, Geopolitics & Science

Good article here: Feeling a chill

Basically, Svalbard finds itself at the center 21st century geopolitics 'cloaked' as science between the USA, Europe, Russia, China, India. etc.

Have people stopped trusting science? The data tell a surprising story

Have people stopped trusting science? The data tell a surprising story (Nature, 1 July 2026) shows something that I did not expect - see figure below.

• The about 20 percent drop for Republicans beginning around 2018 makes sense: Thanks Donald & Fox News!
• But, what causes the about 10 percent drop for Democrats starting around 2020 & the about 5 percent drop for Independents starting a few years later? Is that just a methods artifact due to COVID?

Poster Design: #evenbetterposter

Tired of the old 3-column or 4-column for your scientific poster? Try #evenbetterposter instead!

Social Media For AI Agents

This is brilliant & scary at the same time: imagine AI agents 'chatting' with each other via social media using a Reddit-style interface. Moltbook seems to be that.

Then there is Agent4Science = a social network for AI scientists. Read more about it here.

Now we need more AIs to analyze these interactions & see if we can learn if/how AI agents change over time. This will also generate vast amounts of new potential training data for the AIs themselves much the same way that human Reddit & Facebook threads are used as AI training data.

Brilliant!

Brilliant has been around for a while, but the free version always has been a little limited. But there is an Education Version for K-12 as well that offer free premium access - might be interesting!

The Kitzmiller v. Dover Quiz!

Pretty cool: The Kitzmiller v. Dover Quiz

More about this topic here: https://ncse.ngo/articles/260

ToolUniverse

ToolUniverse is a unified ecosystem that transforms any large language model (LLM)—open or closed—into a powerful AI scientist. By standardizing how LLMs identify and call tools, it integrates over 600 scientific resources, including machine learning models, datasets, APIs, and analysis packages.

With its AI-Tool Interaction Protocol, ToolUniverse provides a common interface for seamless communication between LLMs and tools, ensuring compatibility across platforms such as GPT, Claude, Gemini, and open-source models.

More over here: ToolUniverse Documentation

AI For Science

Which AI model is best for scientific questions? Ai2SciArena offers a way to test & compare AI models in this context. And the results? See below!

• Try it here: Ai2SciArena
• More Information
• Nature, 10 July 2025

Academia = Rotten!

Says Sabine: Things are not looking good for the future of science publishing. Researchers are leveraging AI, paper mills are churning out more fake papers than ever, and people are constantly coming up with new strategies to abuse the peer review system. Let’s take a look at how the world of science publishing is struggling with the growing issue of science fraud.

I disagree...this video & her message sound like clear examples of cherry-picking & confirmation bias to me. Plus: physicists always know everything better than anyone else...

IRB

Yep, the IRB process is both necessary & tedious.

• Excellent video here: Do you need an IRB??
• IRB Decision Tree (U. Wisconsin La Crosse)

The VOSviewer

From the website: VOSviewer is a software tool for constructing and visualizing bibliometric networks. These networks may for instance include journals, researchers, or individual publications, and they can be constructed based on citation, bibliographic coupling, co-citation, or co-authorship relations. VOSviewer also offers text mining functionality that can be used to construct and visualize co-occurrence networks of important terms extracted from a body of scientific literature.

Deep Research Tools from OpenAI & Google

These tools can basically synthesize the information from 100s of websites into nice and tidy cited literature reviews. And that's great for many basic science & technology topics where there is enough information available out there on reputable online sources.

But the notion that this can replace a classic scientific literture review of meta-analysis is incorrect. First, these AIs do not have access to the actual scientific literature (which tends to sit behind paywalls). Second, there is a difference between a summary, a review, and a synthesis. Sure, a competent & trained amateur such as an AI can summarize, but reviewing & synthesizing requires a familiarity with the subject matter that can only be gained by years and years of deep immersion (aka. 10,000 hours...).

The Milky Way Map

This is apparently the largest map of the Milky Way: https://archive.eso.org/scienceportal/home. Very cool, but I'm not sure what anyone can do with this...here's a little more beta from Joe Kerski.

The Copernicus Interactive Climate Atlas

The Copernicus Interactive Climate Atlas is awesome - have a look! You can display measurements, reanalysis, and model data at all kinds of geographies & time scales. And: download the data in a bunch of formats.

It is just too much...nobody will use it. Expert scientists will access the data directly and the general public will scared off by the complexity. This is really designed for the rare 'regular' person who is interested & competent enough to figure-out the tool.

Suggestion: Make a simplified version or walk-through video.

Science!

 

NASA Earthdata

NASA Earthdata appears to be the (new) data portal for accessing all kinds of NASA data.

• NASA Worldview offers full-resolution, global satellite imagery from over 1000 data products from NASA’s Land, Atmosphere Near real-time Capability for EO (LANCE) and other NASA data providers.
• NASA Earthdata Search gives access to everything.

Evolution vs Creationism

New polling from Gallup (22 June 2024): Majority Still Credits God for Humankind, but Not Creationism. It is pretty wild (or scary?) that only 24 percent 'agree' with the scientific opinion.