Christofer H Bagge

Director and Innovator at Lirox AB. Proud distributor of Evnroll putters in Sweden

[Nonzero Podcasts @NonzeroPods]

"The AI industry has more lobbyists in Washington and Brussels than the fossil fuel industry and the tobacco industry combined." (full conversation below)
@tegmark and @robertwrighter

[Rev. Johnnie Moore ن @JohnnieM]

NEW: Evangelical leaders hail Trump as “world’s leader now by Divine Providence to guide A.I.” while warning of the “peril of artificial general intelligence.” President Trump is “the AI President.”

In a new letter to President Trump, leading evangelicals express their optimism about AI while raising concerns about the dangers of “autonomous smarter-than-human machines that no one knows how to control.”

Press Release: JDA Worldwide for Rev. Samuel Rodriguez and the NHCLC
May 21, 2025

WASHINGTON D.C. — Today, in a new letter released during the 6th National Faith@Work Conference hosted at the Catholic University of America, leading Christian voices are calling for President Trump to ensure artificial intelligence is developed responsibly.

The letter is signed by Reverend Johnnie Moore, President of the Congress of Christian Leaders, and Reverend Samuel Rodriguez, President of the National Hispanic Christian Leadership Council. Moore and Rodriguez are inviting other Chrisitian leaders to join them in signing the letter - which will be delivered to the White House in the coming days - expressing their excitement and concerns about the AI revolution. Leaders of other faith traditions are also invited to sign the letter in solidarity.

The full letter and its current signatories can be viewed here.

“As people of faith, we believe we should rapidly develop powerful AI tools that help cure diseases and solve practical problems, but not autonomous smarter-than-human machines that nobody knows how to control,” the letter says. “The spiritual implications of creating intelligence that may one day surpass human capabilities raises profound theological and ethical questions that must be thoughtfully considered with wisdom.”

The two prominent evangelicals are part of a growing number of religious leaders questioning the unchecked development of AI even as the technology is becoming dramatically more powerful. Earlier this month, in his first address to the college of cardinals, the new pope Leo XIV explained he chose his papal name as a reference to the challenge AI poses to “the defense of human dignity, justice and labor.”

The signatories recommend the creation of an advisory council which would “convene leaders who will pay attention especially not only to what AI can do but also what it should do.”

The letter goes on to say, “We believe the Lord has positioned you, Mr. President, for ‘such a time as this,’ and nothing is more important than managing the Age of Intelligence.”

Reverend Johnnie Moore is participating in a panel discussion at Faith@Work on Wednesday, May 21st called Shaping the Future of Artificial Intelligence and Religious Liberty. He is available for interviews at Catholic University of America. The panel will feature a question and answer session in the Maloney Auditorium of the Busch School of Business at 2:30 pm.

###

[Christofer H Bagge @chbagge]

I think this is a very important initiative and that from a theological viewpoint we also have a big dilemma with those who seek to develop AGI. The best Tool AI will help mankind but AGI will leave that to chance with a to low probability for mankind

[Rev. Johnnie Moore ن @JohnnieM]

https://t.co/sW5Esig47u

[Christofer H Bagge @chbagge]

@tegmark Nice April 1st joke 😉

[Christofer H Bagge @chbagge]

@ukrainiansquad Sweden support you

[Future of Life Institute @FLI_org]

📺 From @Siliconvos: "We are on the cusp of creating artificial general intelligence (AGI), even though the corporations building this technology admit they don't know how to control it."

Watch their video on @AnthonyNAguirre's new essay about how we can "Keep the Future Human":

[Max Tegmark @tegmark]

AGI companies are breaking promises even faster than many politicians:

[Geoffrey Miller @gmiller]

I do hope that everyone in the Trump administration understands that in any 'AI arms race' to build agentic Artificial Superintelligence (ASI), it doesn't matter which nation builds it first.

If the ASI isn't completely contained, controlled, and aligned with specific human goals and values, with 100% certainty and reliability, then the agentic ASI will do whatever it wants to do -- regardless of which nation happens to build it.

The only winner of an AI arms race will be the ASI. Not America. Not China. The ASI itself.

(To use the standard science fiction metaphor, it doesn't matter who builds Skynet first, thinking it will give them some decisive geopolitical advantage; Skynet will have no allegiance to its creator.)

This point has been widely made in the AI safety community for twenty years. But with all the recent hype around AI, I'm concerned that it isn't understood by political decision-makers -- who tend to think an ASI would act just like any other military tool, weapon system, or strategic innovation. As if it will be easily contained and controlled in our interests. (Everyone working on AI alignment is quite worried right now that AI capabilities development is advancing _way_ faster than alignment methods -- and a significant number believe that AI alignment might be impossible, even in principle, for game-theoretic reasons.)

The silver lining to this point is that if both America and China come to understand that racing to build an ASI benefits only the ASI itself, not its nation of origin, then it may be _much_ easier to coordinate international treaties to avoid building such a nightmare that could impose global catastrophic risks on all of us and our kids.

The Chinese aren't stupid or suicidal. If they understood that racing to build an ASI would benefit only the ASI, not China, then why build it? Likewise, I hope Americans are smart enough to understand that racing to build an ASI would benefit only the ASI, not America.

That's the potential foundation for a cooperative treaty banning dangerous AGI/ASI development.

[An Inconvenient Doom @An_Inconv_Doom]

AI existential risk is hard to explain – but we did it anyway!

Watch the two minute teaser here, and then dive in to the full explanation video of AI existential risk in laypeople's terms – complete with 45 links to articles, research papers and other videos.

This topic concerns literally everybody. So a firm understanding of the basics is the first step towards a fact-based discussion of risks.

Link in 🧵

[Christofer H Bagge @chbagge]

Thi si s very good explanation of why we want Tool AI but not AGI

[Max Tegmark @tegmark]

In this controversial @WebSummit talk, I argue that #AGI is unnecessary, undesirable & preventable - while tool AI can give us basically all of AI's exciting benefits, and the "but China" argument is flawed.

[Max Tegmark @tegmark]

In this controversial @WebSummit talk, I argue that #AGI is unnecessary, undesirable & preventable - while tool AI can give us basically all of AI's exciting benefits, and the "but China" argument is flawed.

[David D. Baek @dbaek__]

5/6 This work expands on recent works finding structures in SAEs, and we are excited to dive deeper in future works to understand why some of these structures emerge!

[David D. Baek @dbaek__]

4/6 The “galaxy” large-scale structure is not isotropic, but instead has a power law of eigenvalues with the steepest slope in the middle layers. Clustering entropy also peaks around the middle layers!

[David D. Baek @dbaek__]

3/6 The “brain” intermediate-scale structure has significant spatial modularity, which we measure as alignment between spatial and co-occurrence clusters; for example, math and code features form a “lobe” akin to functional lobes seen in neural fMRI images.

[David D. Baek @dbaek__]

2/6 The “atomic” scale structure contains “crystals” whose faces are parallelograms or trapezoids, similar to the classic (man:woman::king:queen). The quality of crystals improves when projecting out global distractor directions such as word length (via linear discriminant analysis).

[David D. Baek @dbaek__]

1/6 New paper! “The Geometry of Concepts: Sparse Autoencoder Feature Structure.” We find that the concept universe of SAE features has interesting structure at three levels: 1) “atomic” small-scale, 2) “brain” intermediate-scale, and 3) “galaxy” large-scale!

[Christofer H Bagge @chbagge]

@gunsnrosesgirl3 Ni

[Christofer H Bagge @chbagge]

Read until the end - Tool AI is the way to go rather the AGI

[Max Tegmark @tegmark]

We need more "tool AI" as in Elon's space triumph today, and less AGI world domination hype as in Dario Amodei's recent "entente" essay. My essay below argues that "scaling quickly" won't lead to Dario's "eternal 1991" – but perhaps to 1984 until the end, with a non-human Big Brother. https://t.co/va6PHmZ2g5

[Yuval Noah Harari @harari_yuval]

The race to develop AI is driven by mistrust between humans. But what reason do we have to trust AI?
My full conversation with @Trevornoah and Christiana Mbakwe Medina is available on https://t.co/gVJopn4l2f.
#NexusBook

[Christofer H Bagge @chbagge]

Every one should watch this and reflect!
Humanity must take a stand here and not do things that can become catastrophic

[Max Tegmark @tegmark]

The arguments for why AI may soon cause human extinction are so easy to understand that even AI’s (here ChatGPT4o & Claude 3 Opus) can make them: