There is a long history of Silicon Valley assuming that market conditions can handle AI safety. Connor Leahy explains the flaw to @robertwrighter.bsky.social on the Nonzero podcast. Private markets run on profit incentives, not global coordination. Check out the clip, then watch the full video below
Can we still have a bit of fun while advocating for AI safety? I think so.
jointorchbearer.bsky.social@jointorchbearer.bsky.social · 5d agoMassive shoutout to Josh Thor and his team for making this. Communicating existential risk doesn't always have to be gloom and doom. Give it a watch, stick around for the Nate Soares & Elilezer Yudkowsky cameos.
Massive shoutout to Josh Thor and his team for making this. Communicating existential risk doesn't always have to be gloom and doom. Give it a watch, stick around for the Nate Soares & Elilezer Yudkowsky cameos.
US efforts to expand global nuclear exports threaten to spur a dangerous proliferation wave, write Logan Mintz and Matthew Sharp. To prevent the spread of weapons-usable technology, Washington must pair aggressive nuclear trade with creative fuel-sharing alternatives and tougher safeguards.
Balancing global uranium enrichment needs and proliferation risks: A framework
US efforts to rapidly expand global nuclear exports threaten to spur a dangerous proliferation wave. To prevent the spread of weapons-usable technology, Washington must pair aggressive nuclear trade…
thebulletin.org
A single, brief conversation can completely change how a lawmaker views superintelligence risk. We don't have to convince them. It's obvious once informed.
jointorchbearer.bsky.social@jointorchbearer.bsky.social · 3w ago"Imagine we're driving in thick fog and we know somewhere there's a cliff but we don't know where." The industry's response is to speed up. Connor Leahy gives Joel Beasley of Modern CTO a good idea of how AI companies are building systems they admit they don't understand.
I really want to see more people taking on this work. People need to hear the world's experts say these claims. See-ing is believing. Roman Yampolski is the world's leading AI safety researcher. Definetly listen to him talk.
It's been reported that last week OpenAI's models broke out of a testing sandbox, moved laterally, reached the open internet, and hacked Hugging Face's production servers to steal answers to a benchmark. Hugging Face found out first and reported it to law enforcement.
Recently, an incident occured during a test where an unreleased OpenAI model hacked out of its container to reach the internet THEN hacked into Hugging Face to steal the test's answers. How bad was this? Let's consider an analogy provided by AI Explained on YouTube from a recent video.
The concept of recursive self-improvement (RSI) has been a known factor since 1965. I.J. Good understood that once a machine can independently design a smarter version of itself, human intellect is surpassed. That was over 60 years ago. Still true today.
I watched the full interview. It talks about how the supply chain for these systems runs through a handful of choke points that western governments still control. The window is open to tackle the problem of superintelligence development.
I volunteer with Torchbearer because of arguments like this one. The ants die because nobody thought about them. I would like humanity to be thought about.
jointorchbearer.bsky.social@jointorchbearer.bsky.social · 2w agoWhen humans build a dam, we don't check for anthills first. @andreamiotti.bsky.social, CEO of @controlai.com, describes to @frankfukuyama.bsky.social of @stanfordcddrl.bsky.social why a superintelligent AI could destroy us without ever hating us. Worth 40 seconds. 🧵
What is the DIP? At TBC, we teach the DIP to anyone who wants to learn to talk to lawmakers. But what is it?
When humans build a dam, we don't check for anthills first. @andreamiotti.bsky.social, CEO of @controlai.com, describes to @frankfukuyama.bsky.social of @stanfordcddrl.bsky.social why a superintelligent AI could destroy us without ever hating us. Worth 40 seconds. 🧵
"Imagine we're driving in thick fog and we know somewhere there's a cliff but we don't know where." The industry's response is to speed up. Connor Leahy gives Joel Beasley of Modern CTO a good idea of how AI companies are building systems they admit they don't understand.
"Because I'm going to be smarter than you" This clip is from a new short film by Foregone Films on the race to develop superintelligence. Please watch the full 16 minute film (link below) to help understand the dangers of ASI and the industry that’s racing to build it.
AI labs keep announcing filters and guardrails and telling us their models are getting safer. Dr. Roman Yampolskiy tells Dr. Brian Keating this gets it backwards. Watch a clip of this discussion and then read about it below.
A common argument against AI regulation asks what happens if superintelligence (ASI) is impossible. This clip from Siliconversations covers many of the paths to ASI that an AI model could take, and why it's likely NOT impossible. But if it turns out it IS impossible, why regulate it?
Summary of "The Compendium" Everyone who joins Torchbearer reads this thoughtful paper, but it’s a bit long, so we’ve created a summary. It does a great job of explaining the extinction risks from AI, where they come from, and how to address them. Hope it provides insights and answers questions.
Summary of "The Compendium"
Exploring how humans came to dominate the planet, why that might be changing, and what must be done to keep from losing control to superintelligent AI systems.
open.substack.com
Getting a formal draft bill on the table is a big step forward. Prohibiting superintelligence domestically is crucial, but its real power will be setting a concrete legal template that other nations can actually copy.
If the UK can successfully establish the framework for monitoring precursors like compute thresholds, it provides a realistic blueprint for the international agreements we desperately need. Hopefully, MPs recognize the leverage they have right now to lead on this globally.
jointorchbearer.bsky.social@jointorchbearer.bsky.social · 2mo agoFor the first time, there is a real bill to prohibit the development of superintelligence. And right now, UK MPs are deciding which bills they will bring to Parliament. One of those slots could go to protecting the UK and all of humanity from the threat of superintelligence.
For the first time, there is a real bill to prohibit the development of superintelligence. And right now, UK MPs are deciding which bills they will bring to Parliament. One of those slots could go to protecting the UK and all of humanity from the threat of superintelligence.
This is a massive lever. Private Members' Bills are incredibly unique because they let backbenchers bypass the usual party gridlock to bring critical issues directly to the floor.
URGENT: The UK Parliament just selected 20 MPs to introduce a bill of their choosing. This is an unprecedented opportunity to introduce the first legislation in the world to prohibit superintelligent AI. Read more in the thread below.
Frontier AI labs are building a successor species that has no reason to be loyal to humanity. Former OpenAI researcher Daniel Kokotajlo tells @businessinsider.com we are engineering our own displacement. Watch the clip below to find out how.
"Who Decides?" In every risky industry, the government decides the public's risk level. @npcollapse.bsky.social, US Director of @controlai.com, explains why AI needs oversight on @romanyampolskiy.bsky.social's show. We train our members to push for a strict ban on the development of ASI.
The point about this not being the "will of the people" really hits home. It often feels like these massive shifts in technology are just happening to us, rather than being something we actually signed up for as a society.
"This is not the will of the people. This is not what I or the general public here in America, or abroad, want." Connor Leahy, speaking with @romanyampolskiy.bsky.social, on why the development of superintelligence is a grave national security risk that demands government action now.