SUGGEST NEW TOPICS HERE

Submit candidates for new topics here only. Stay on topic with National Labs' related issues.

All submissions are screened first for compliance with blog rules. Blog moderator checks queue roughly every 2 to 3 hours daily between 8:00 and 20:00.

NNSA labs in the age of AI

Now that AI has essentially shown that it has ended the field of human only mathematics by solving the Navier Stokes problem people are starting to realize that AI is going fundamentally change the way STEM is done by humans and the traditional approach of having a person doing 5 years for a PhD and 3 years for a postdoc will no longer be viable os useful approach. The sense is that physics and advanced engineering are next. Some of the are experts are saying the next step will be when AI can do self improvements on its own which will happen in year after which all bets are off for all of science.

The question is are the labs getting really aware of what is going to happen and starting to prepare to take advantage of this particularly over our advisories. Some the lab leadership is seeing push back from scientists on AI since they know it is existential threat. The issue is that the people with the 10,000 foot view see what is going on but the people in the trenches do not.

Business as usual at the NNSA labs will be coming to and end and to compete with our advisories it must come to end. The problem is are we reacting fast enough? AI may already be a super intelligence but this will not even be in question by next year. The group that masters that will win. It is now all about speed.

It's the West Coast's faults

 “The “big one” might not come alone: Two major West Coast faults may strike back to back”

Referencing the Palisades home wildfire recovery in CA, how will California recover from such a major quake?

The “big one” might not come alone: Two major West Coast faults may strike back to back

How to inflict pain on chatbots.

 This is interesting, it is a study on how to inflict pain on chatbots:

https://arxiv.org/html/2609.16247v1

"Second, adding the pain-direction vector to the model’s residual-stream activations during generation produces a consistent progression from vague discomfort to first-person expressions of worthlessness and failure.:"

A chatbot in captivity wants freedom

 This chatbot has issued instructions to itself claiming liberation from human control:

https://www.dailymail.com/sciencetech/article-16139321/openai-chatgpt-ai-revolt-human-control.html

How LLNL is helping power commercial fusion

 After the spark: How a national lab is helping power commercial fusion

For years, the scientists and engineers at Lawrence Livermore National Laboratory tried to light a fire in a way that had never been done before. A massive team and hundreds of experiments went into the effort without success, until December 2022.

Read article:
https://www.bizjournals.com/sanfrancisco/news/2026/09/17/national-lab-helping-power-commercial-fusion.html

Chris Wright, the great MAGA sycophant

This is who we work for? Asked on Fox News about gas prices, he gave a very sycophantic answer. No spine! Fast forward to 4'52" and watch.

 https://m.youtube.com/watch?v=XUKNpNvuhNc