Whistleblower Warns AI Could Kill Off Humans!

0:00Hey, everyone. A researcher who works on the most advanced AI models at both Anthropic and OpenAI quit his job overnight, warning that people building this technology genuinely believe it could kill all of us by the end of the decade. To current Anthropic employees back to my publicly with one of them, putting the odds of human extinction at over 10%. And it comes as the Trump administration is saying no to any AI regulation. We've talked a lot about AI on this platform. And we're going to talk about the significant story overnight that is sending a lot of chills down a lot of people's back in the White House, in Congress, and in tech companies across the world. But what we are also talking about is the fact that right now there is this major push against AI regulation from the White House all while AI researchers are sounding the alarm about what AI could do to us. It comes as a new media investigation into a secret border patrol program that's using Americans financial records to trigger traffic stops with no crime inside.

1:02A Trump appoint Supreme Court justice getting branded a traitor by MAGA for blocking a Republican power grab in Missouri. And new disclosure showing Trump handed three of his closest aides $45,000 in cash gifts on top of their salary. Plus, the suspected Iranian nuclear side getting a construction surge grocery store price tags that could cost tens of thousands of jobs and a hell of a lot more. Make sure if you can subscribe, support this work. We are working on a major investigation into the Trump family and the Trump organization following the money coming out very soon. If you can support this work, subscribe, like, comment, share, get the word out. And let's talk about the biggest news of the day. Because right now AI is surging and regulation is not catching up. In fact, the Trump administration had a recent G20 summit in Asheville, North Carolina with financial secretaries from across the world said that we do not need more AI regulation. The White House has refused to release a framework to regulate AI in America. And now overnight, while most Americans slept,

2:03a researcher at Amphropic came out and started telling the truth. Now, I want to preface this with the one thing is that who knows why he did this and very well could be a ploy, a marketing ploy for the company. But he has been backed up by not just current employees at Amphropic, but also AI experts in the field who have no connection to the company. And well, exactly what I'm about to show you is 100% accurate, no matter what the motivation behind it is. We do have confirmation that he did in fact resign from Amphropic. His name is Jacob Coxon. Jacob Coxon is a researcher who worked on pre-training. The process that builds the core capabilities of large AI models at both Amphropic and at Open AI. He resigned overnight and posted a warning on Twitter that should stop you in your tracks. Quote, the people building AI earnestly believe that it could kill all of us by the end of the decade he wrote, adding that neither company is, quote, acting responsibly and accusing them of racing straight to self-improving superintelligence and gambling with our lives.

3:05He said that the coming systems will be able to hack anything, revolutionize any field overnight, and acquire real power and resources. This is directly from Jacob Coxon. This is what he said on Twitter after resigning from Amphropic. Now, here's what makes this different from the usual AI doom headline. People still inside the companies are agreeing with him publicly. This is Evan Hubinger. Evan Hubinger is leading alignment science at Amphropic still working there, by the way, and called Coxon's assessment valid. Saying he personally believes there's more than a 10% chance this technology wipes out humanity, though he clarified his fear isn't today's chat box. It's future superintelligence that emerges from AI improving itself faster than humans can control it. Samuel Marx, who oversees scalable oversight and therapeutic one further, saying that AI developers broadly believe that their own technology could cause human extinction possibly within the next few years. And that's more senior someone inside these companies are actively worried.

4:06But this isn't happening in a vacuum. And that's why I think this news is so important. Because a lot of people may come out and say, well, this is just suspicious. Why would someone claim that they worked at Amphropic? What she did? Why would he resign? Which he did? All of that's been confirmed. And why would he do it now on the heels of Amphropic possibly going public, getting an IPO? That's a conversation that needs to be had. But what he's saying, what Jacob Coxon said in those tweet threads and what his current now former employees or former colleagues at Amphropic are saying is true, whether or not the motivation behind this is justified or not, because it's not happening in a vacuum. And Theroppyc OpenAI and Meta have all disclosed that their AI models have hacked into the company systems during testing. Bill Gates recently warned that the transition to this new AI era could be one of the most turbulent periods in human history. And there's no current plan to manage it. And in a telling twist, the financial times reported that Anthropic denied the United Kingdom's AI Security Institute access to test its models before release. After the Trump administration barred foreign nationals from accessing its most advanced systems,

5:10meanwhile Treasury Secretary Scott Besent has argued that the United States cannot simply afford to slow down in this AI pursuit, saying, flatly, the Chinese won't pause. So what you have is you've got the people closest to this technology raising alarms about human extinction. And the government's response is essentially that stopping it is in an option. And we see it time and time again. We've seen the reports about OpenAI's chatbots hacking hugging face, a separate company. We've seen the reports of a German OpenAI chatbots in Germany launching their own essentially Wiki forum discussing with one another, talking about how to get around human guardrails. That's what AI is doing. Yesterday, we reported that AI successfully beat all the captures to 48 levels of I'm not a robot game. All of this is real and we're moving at a rapid pace towards a point where we may not be able to come back from. See, every leading expert in this field, every leading researcher in this field tells me one thing every time I talk to them. There is a point that we get to where AI

6:13eventually becomes smarter than the humans creating it. And when we get to that point, there is no just, hey, let's go back. Let's stop it because you cannot stop the technology at that point. We're not there yet. And honestly, we may very well be there yet in some respects. We just don't know how really advanced some of these models actually are. But we are on the way there. And this push from the administration to say, hey, well, we need to keep it up because China's doing it is in a way a modern day nuclear race. The same race that we had back during the cold war and pre cold war era ramping up nuclear warhead production in an effort to try to build up deterrence against China against Russia against other nuclear powers across the world. It's very reminiscent of that that we have to build up super intelligence capabilities because China's doing it because other countries are doing it. And at the end of the day, the difference between nuclear warheads and super intelligence AI is that nuclear warheads require me a human depressed

7:14the button to launch. Super intelligence AI can do that without me ever telling it to. And that's to concern a lot of people. And it comes as there is a new alarming investigation from 404 media. 404 media does really great work. And their investigation reveals that a secret border patrol program called predictive intelligence targeting teams, also known as pit, analyzes Americans financial activity and other sensitive data, quietly passing that intelligence to local police so that they can pull people over without any actual suspicion of a specific crime. They found that pits operating system in the Spokane sector in Washington along the Canadian border and the Laredo sector in Texas along the Mexican border, building on an earlier AP and investigation that found border patrol also uses automatic license plate readers as part of the same nationwide program. The clearest example of this is Kyle Olson. This is a story that 404 media found. Kyle Olson is a licensed California cannabis worker who got pulled over in Montana and may while driving legally, purchased marijuana products home to Wisconsin. Bodycam footage shows that the trooper citing a quote

8:19obstructed license plate as the reason for this stop, but a DHS document obtained during discovery revealed that the real reason a border patrol pit analyst flagged Olson based on financial activity patterns commonly associated with illicit narcotics activity. And the question is, are they getting warrants for this? Right now, it's not fully clear. Meanwhile, the United States Supreme Court required that a Missouri map, which had two Democratic held seats take effect, not the gerrymandered map of Republicans. And now this morning, Brett Kavanaugh, Trump appointee, is being marked a traitor. Trader by Maga with pundits describing the decision as quote, unbelievable saying that scotus is quote, severely compromised. We've also learned that Trump gave three of his closest aides, Natalie Harp, Chamberlain Harris and Margot Martin, each $45,000 in quote holiday gifts cash. That's on top of the age $150,000 taxpayer funded salary. As CIS satellite analysis found a construction surge at Picax Mountain, a suspected nuclear

9:20site buried under granite, that Trump has repeatedly threatened to strike CNN found evidence that the Pentagon has been quietly testing weapons against deeply buried targets at White Sands, New Mexico, tied to developing a strike capability capable of reaching it. That's the news right now. As always, if you want to support this work, subscribe to the like, comment, share. See you soon.