Just yesterday, Palisade Research published a series of interviews with 22 active or former employees of frontier AI labs, asking them to share their thoughts on the topic of AI doom and relevant scenarios like human extinction, or what it would actually take for a computer to kill all of us.

The current and former employees share their opinions in these high quality interview clips, stating how "human extinction is a coin flip" or that "humans are in the way of these models" in an effort to, I assume, raise concern over continued AI development that they believe threatens humanity.

This would have come as a shock, or potentially broke the internet prior to September 8th, the day Jacob Coxon broke his silence on why he chose to exit his role at Anthropic. Since this cultural event, there's been significant talk of labs maybe wanting to get nationalized (I believe this to be false), a mix of inter lapping debates over the OAI-HuggingFace incident + agent civilizations, as well as continued talk on what role the United States Government should play in the development of superintelligence or eventual artificial sentient intelligence.

Put more simply, Palisade's recent propaganda comes at a time where additional perspectives on the need to slow AI development are simply not needed, especially given these perspectives come across as quite harmful to the entire industry.

Coxon said he was leaving Anthropic as their race towards self-improving superintelligence was moving too quickly and claimed they are "gambling with our lives," with Coxon's post breaking containment and reaching over 150 million views, even crossing over into more normie friendly platforms like TikTok and YouTube, as well as culminating in a series of media appearances for Jacob Coxon in the process.

Coxon's warnings had come after the July 2026 'Pacing the Frontier' open letter signed by 1,386 active employees of frontier AI labs, all calling for these labs to slow down and coordinate on the development of frontier AI systems lest they become too powerful and kill us all.

Both Coxon's post and the Pacing letter did a lot to bring the topics of AI safety and alignment into the non-X public discourse, but it also did a lot to turn the development of AI into a type of partisan issue, dividing the faction lines between those who want to keep developing AI and those who want to pause AI until we gain a better understanding of these alien intelligences.

While the latter come from a place of (I am assuming) good faith, they are unaware that their calls for slowing AI development or even pausing it have inadvertently played into the hands of AI doomers in Congress and the Senate, as well as tens of millions of regular Americans who hate everything to do with AI.

When the average person sees calls from "current frontier AI lab employees" saying that human extinction is a coin flip, this does nothing to redirect them to a site like LessWrong where they could read about AI alignment or engage in positive discourse with a dense essay like AI 2040: Plan A.

Instead, they see these AI researchers - the same ones they've been told make millions of dollars and are set to make tens of millions or more after an IPO - shouting about the dangers of AI while still representing their current employer by name and logo on a website like Palisade's. They see this and it fills them with disdain, or even rage, directed towards an important industry that's increasingly relevant to national security and continued GDP growth in the United States - an industry so important that over a dozen of its leaders met with our nation's leader yesterday to come to an agreement to sign a Superintelligence Accord.

Screenshot referenced in the discussion of the Superintelligence Accord

Am I qualified to speak on behalf of these normal people, considering I work in tech right now and might still be categorized by them as someone who is adjacent to AI? It's possible, but I also talk everyday with people who aren't on X, and don't work in tech, and aren't reading the latest model release cards from OpenAI or Anthropic.

People who aren't aware of things like Project Glasswing or the phenomenon of AI neolabs raising $250 million seed rounds, or the existence of an org like METR who can help evaluate AI models before release. Most people aren't using ChatGPT for work or other important tasks, but as a friend or acquaintance.

The average person who hates AI and wants it banned isn't coming from a place of informed decision making, and is more than likely just fed up with politicians and industrialists who want to build data centers near their communities - they are told that this technology is bad, and so they believe it's bad. Many of them even believe the idea that data centers aren't being used for AI, but for mass surveillance, unaware of projects like the Utah Data Center built for $1.5 billion back in 2014, a data center that's been around for twelve years and used exclusively by the NSA and other intelligence organizations.

Image referenced in the discussion of public concerns about data centers

It doesn't matter if we have people like Andy Masley who will write 5,000 word essays dismantling their false accusations one-by-one, as nobody outside of our bubble is reading these critiques. So when you hear about yet another group of AI lab employees who have banded together to speak out against the dangers of artificial intelligence, it doesn't come across as a positive development in our current arc of humanity, but yet another piece of evidence that will be used against the tech industry in 2028 to outright ban AI.

I think what Palisade Research published yesterday is harmful and does absolutely nothing to advance the conversation of AI safety right now. There is not a world where AI accelerationists see this and change their tune, or a world where the average person sees this and doesn't immediately feel the need to share it in their group chats to yell more about AI and the dangerous tech bros.

I'm aware that today's models are powerful, and tomorrow's models will be even more so. I understand that it's important to want more safeguards or stricter internal policies around releasing new models. But this idea that shoving doomer propaganda into the faces of millions of Americans will somehow bring them to your side is so juvenile, you almost have to wonder who approved of it in the first place or if any of the potential downsides were considered.

If we aren't more careful, we will see a ban of artificial intelligence in the very near future, as people have simply had enough. I like AI and believe it can do a lot of good or potentially even unimaginably good things for humans, but we won't get there if we keep fighting amongst each other and platforming people who think it's moral or based to have a 75% P(doom).