Iman

"this rarely works" is very optimistic. This is a much needed justification to shut down one's consciousness.

YouTube video by The Royal Institution Information, Evolution, and intelligent Design - With Daniel Dennett

youtu.be/AZX6awZq5Z0?...

Firefighters Race to Save a Treasured Sequoia Grove in California

What a stat: Worsening wildfires have killed nearly a fifth of the world's mature sequoias just since 2020 www.nytimes.com/2025/09/08/u... / www.fs.usda.gov/about-agency...

I don't know about this one, but some of these faulty ai search responses cannot be reproduced.

Table of records in France by Meteociel

HISTORIC HEAT WAVE IN EUROPE

FRANCE >42C with several records in secundary stations.

SWITZERLAND 35.5 Visp tied its August record.

Hundreds if not thousands of records will be pulverized in Western and Eastern Europe next week:
France might attack its national August record.

My condolences.

Fascinating.

The true value of AI lies more in thorough testing and evaluation than in clever coding or prompting. — Dr. Shea Brown on AI auditing through the scientific method Podcasts 4 Brainport, featured by Radio 4 Brainport, 747AM and DAB+ · Episode

Everyone can prompt.

But can you convince yourself and others that the system is doing what it's supposed to do?

My conversation with Dr. Shea Brown, astrophysicist turned AI auditor.

open.spotify.com/episode/6Plj...

Such a brilliant connection

Top F.D.A. Official Overrode Scientists on Covid Shots

"...F.D.A. staff members concluded that the vaccines were safe and effective based on clinical trials of the shots tested in thousands of people. But Dr. Prasad wrote that there could still be vaccine-related injuries that have yet to be discovered."

Gift Link.

deepdivepod.eu

If any of these apply, check out the latest episode of my podcast:

You're into algorithms (esp. game theory)

You're in the Netherlands

Your child is soon starting high school/middelbare school

You're curious how algorithms shape society

Link in the comments!

New research from MIT found that those who used ChatGPT can’t remember any of the content of their essays.

Key takeaway: the product doesn’t suffer, but the process does. And when it comes to essays, the process *is* how they learn.

arxiv.org/pdf/2506.088...

The bigger issue is that the entire AGI dream may not be the best way to approach intelligence. This is a historical take and there are better ways to think of intelligence as complex cooperative systems, why autonomy is so fascinating?

Not feasible unfortunately. Many ML problems are NP complete, but their average cases are not hard. That means AI models often don't work in the worst-case. They are generally ok-ish on average. They are also vulnerable to adversarial attacks and they develop sub-goals and deception.

Same feeling here.

Mathematics in Biology Biology has turned into a quantitative science. The core problems in the life sciences today involve complex systems that require mathematical expression, ye...

If you're planning a course with mathematical methods content, or use such methods in your own work, please take a look at "Mathematics in Biology", by Meister, Lee, and Portugues, published at MIT Press. 1/2 @portugueslab.bsky.social mitpress.mit.edu/978026204940...

What can be the incentive?

I am surprised at how positive he sounded on LLMs.

Many recent posts on free energy. Here is a summary from my class “Statistical mechanics of learning and computation” on the many relations between free energy, KL divergence, large deviation theory, entropy, Boltzmann distribution, cumulants, Legendre duality, saddle points, fluctuation-response…

In my case it is accurate enough to pinpoint restaurants that are nearby. Tried it multiple times in different countries. But if I ask "do you know where I am" it says no.

I am not saying it is using your IP to detect the image. I am saying these two abilities together open doors for extensive user manipulation. It can use IP to understand what the user sees and use that to convince/influence the user.

That might actually explain why everybody eventually dies.

Already did. Not sure why you say that.

Now you can also ask for a restaurant near you, and it knows where you are (based on your IP)

Seeing Atoms, Shaping Futures: Dr. Remco Schoenmakers on Asimov, AI, Science, Purpose & Strategy Podcasts 4 Brainport, featured by Radio 4 Brainport, 747AM and DAB+ · Episode

Interviewt Dr. Remco Schoenmakers, senior director en AI strategy lead at Thermo Fisher Scientific.

open.spotify.com/episode/2CrW...

AI, Art and the Human Touch: Mei-Li Nieuwland on Creativity in the Age of Machines Podcasts 4 Brainport, featured by Radio 4 Brainport, 747AM and DAB+ · Episode

Link to the episode: open.spotify.com/episode/3r2n...

A few days ago, I shared a conversation (🔗🎧⬇️) with Mei-Li Nieuwland about how Gen-AI is affecting art and artists. Less than a week later, we are in the middle of a situation that has left one of the greatest filmmakers in animation history feeling “disgusted.”

#ai #genai #Miyazaki #StudioGhibli

AI, Science and Society Conference - AI ACTION SUMMIT YouTube video by IP Paris

Great that the Paris summit is surfacing some of our more nuanced thinkers.

www.youtube.com/live/W0QLq4q...

Worth listening to Mike Jordan in this panel.