Mark Spencer

The U.S. National Debt is $40T Donald Trump accumulated $16T of that $40T In just 6 years, Donald Trump has been responsible for 40% of the debt our country has ever accumulated in 250 years.

the fastest way to make an AI agent dangerous is to give it access to everything. our sales agent can maintain sales knowledge. it should not be able to rewrite finance policy because it found a conflicting sentence in a call transcript. read scope. write scope. approval rules. stop conditions. that boundary map should be part of every agent service you build for yourself or sell to clients.

Exactly

An American sent me an email. I did not reply within two days. He sent a second email. It contained four words at the top and then the entire original email below it, unchanged, indented, like a body being carried. The four words were: "Just following up on this." Just following up. He was not angry. He was not accusing. He was simply standing next to his previous message, pointing at it, saying nothing about my silence. Inside I said: HE HAS NOT DRAWN. HE HAS ONLY STEPPED FORWARD ONE PACE AND ALLOWED ME TO SEE HIM. I replied immediately. Two weeks later I needed something from him. He did not reply. I waited two days. Then I typed the four words. "Just following up on this." I felt something move in my chest as I typed them. It was not aggression. It was worse. It was the calm of a man who has entered a system and found it comfortable. He replied in six minutes. Then I learned there is a third stage. If the second email fails, you send a third. It contains three words. "Bumping this up." Bumping. You are not pursuing him. You are lifting the message. You are moving it upward in a stack, physically, as a favor to him, so that he might see it more easily. The entire escalation is phrased as assistance. There is a fourth stage. I have seen it once. A colleague, after three unanswered emails, sent one that said only: "Hey — did this get buried?" Did this get buried. He offered a theory in which the other man is innocent and the email itself is at fault. He built an excuse for a stranger and handed it to him at no charge. The man replied in ninety seconds, apologizing, and they were friendly again immediately, and nothing was ever said about the eleven days. Inside I said: THEY HAVE BUILT AN ENTIRE PROTOCOL WHOSE ONLY PURPOSE IS TO ALLOW A MAN TO FAIL WITHOUT BEING SEEN TO FAIL. I have used all four stages now. I have never once been angry while using them. I do not think anyone is. That is the terrifying part.

when you experience a language model doing a task that used to require you at every step and now doesn't, it's a gigantic agency rupture. we tend to equate ourselves with our tasks, our outputs, and a rupture of this kind is akin to the loss of your identity and sense of self. you are no longer required, and you had no say in the matter. it's a kind of death. however, agency ruptures with AI seem to follow a predictable pattern: - Initial agency rupture. Here you see the AI doing what you used to do, and you see only the AI. In your perception of reality the AI is highlighted, and the human work around getting the AI to do the thing, is invisible. you can see this in the way we use language: "AI solved XYZ Erdos Problem" is another way to say we're in the initial rupture phase. - Seeing human scaffolding. After a little bit of experience with the model and its new capibility—programming, writing, math, etc—we start to see the edges of what's possible with the model. Here, instead of just seeing the AI doing the task, we start to see the scaffolding required by you and other humans on either end to make sure the model does its work. At this stage, the scaffolding can feel like a secondary part of the work, but it is work nonetheless. We begin to say things like, "The AI is prompting me", or "My job is just to babysit Claude." - Agency reconstruction. After a long enough time with the model at its current level of capability, we start to reconstruct a new sense of agency that centers our own role in the work being done, and makes the model's contribution mostly invisible. It is now a tool, it fast-forwards you through the boring or repetitive parts of the work, and what used to seem like scaffolding now seems like the actual work itself—the interesting part. You've had enough experience with the scaffolding to see its nuances, and how hard it is to get the model to reliably output high quality work. (As part of this, your conception of what quality work is changes—the floor is raised, and so is the ceiling.) You can tell you're at this point because you stop saying "The AI did this" and just say, "I did this" and the fact that you used AI is implied. For example, it would be weird for me to talk to someone @every and ask them if the AI built the most recent feature of a product. Of course it did! But we're so used to the capability that it becomes invisible, and we re-center ourselves and our own conception of agency. My theory is that the ability to metabolize agency ruptures and turn them into playfulness, and curiosity is a good leading indicator for whether you're a person who likes and excels in the new AI economy or not. Each time capabilities jump, you get new ruptures for existing fields touched by AI. And you get new ruptures for naive populations (like mathemeticians) who until recently hadn't been affected. if you know what the cycle looks like, it makes it much easier to ride it without freaking out. i think most people @every are naturally good at this

This is one of the most mind-blowing things you’ll EVER see. (And I swear to G-D if this post only gets a few thousand views on X, then X is simply not fit for purpose for anything. Because this is something everyone should see and be blown away by) You’re watching a laser beam move in real time… at 2 billion frames per second. Light only travels about 15 cm (6 inches) in each frame. That’s why you can actually see the beam crawl across the room, bounce off the mirrors, and come back. Now watch carefully when the camera position changes… The light going away from the camera suddenly looks slow. The light coming toward the camera looks almost instant. It looks like the speed of light is changing. It’s not. You’re seeing the finite speed of light itself. The camera is literally recording the PAST at different moments depending on how far the light has to travel to reach the sensor. This isn’t CGI. This isn’t a simulation. This is a homemade camera filming actual photons. Full credit, and huge respect, to @Alpha__Phoenix Alpha Phoenix Channel

I just joined the waitlist for Tau Robotics (@taurobots): a humanoid robot that cleans your home, one hour for $30. Join with my referral link and I move up the list: tau-robotics.com/waitlist?ref=T…

This deserves some kind of award.

AI for the rest of us video update EP 10 is up - really getting great value from my "ClaudeClaw" harness on the Mac Mini and my Claude-Obsidian integration on my MacBook Pro - spending less time building/improving systems and more time building deliverables - link in comments

Megyn Kelly is a vile creature.

In a review of my household safety evaluations, I identified five incidents in which my child escaped the sandbox, reached the kitchen and gained unauthorized access to the snacks. The incidents occurred 16 months ago but has only now come to my attention. This post explains what happened, how it happened and why my child is better than your child at everything.

Had to try this with the new voice

@michaelgrant @TermiusHQ @Tailscale I like cmux you can paste

Another vote for desktop native paste. Keeping this on iOS is ridiculous. I love Termius but had to switch to Herdr because the lack of native paste image was too much friction. So currently you can say Herdr+Tailscale beats Termius+Tailscale+Tmux. I still use Termius on iOS and iPadOS but c'mon guys!

Termius + Tailscale + tmux You don't need your laptop to build with Claude Code, Codex, or any other AI coding agent: → Run your agent inside tmux to keep the session alive → Use @Tailscale for secure access to your laptop → Connect with Termius over SSH from iPhone, iPad, or Android Start coding at your desk. Continue on the go.

Pro-Trump states and counties had far higher COVID death rates. Misinformation and disinformation were to blame. Trump and Musk were the biggest propagators of misinformation in terms of reach.

I believe the available evidence overwhelmingly points toward a research-related incident in Wuhan as the origin of the COVID-19 pandemic. The U.S. government provided a relatively small amount of funding connected to research at the Wuhan Institute of Virology, and the work explicitly authorized under those grants could not, by itself, have produced SARS-CoV-2. But money is fungible, and the deeper problem was that we were funding research in an institution where we lacked full transparency into what research was being conducted, by whom, and under what conditions. We should never make that mistake again. This matters as we consider broader scientific engagement with China. Collaboration can be enormously valuable, but only when there are enforceable standards of transparency, accountability, biosafety, and access to information. The relationship between scientists, research institutions, and the state in China is fundamentally different from what exists in open democratic societies. We cannot pretend otherwise. I also believe Dr. Anthony Fauci was wrong to argue so forcefully in the early months of the pandemic that the evidence favored a natural origin. As a Democrat and former member of a WHO expert advisory committee, it was clear to me in early 2020 that no one outside China had enough reliable information to reach a confident conclusion about origins. I believed then, as I do now, that a research-related origin demanded serious investigation. But accountability is not the same as vilification. Dr. Fauci has devoted his career to public health, and his work has helped save millions of lives around the world. He does not deserve the abuse and partisan attacks directed at him in today’s Senate hearing. He does deserve, like every official involved in the pandemic response, rigorous scrutiny of decisions that may have been mistaken. The Chinese government bears overwhelming responsibility for obstructing the world’s ability to understand how this catastrophe began, and that must remain the central focus of any serious origins investigation. But examining China’s failures cannot absolve us of examining our own. We should not have funded potentially risky research in laboratories where we could not demand full transparency, and U.S. officials should not have expressed confidence about pandemic origins that the available evidence did not justify. For six years, I have called for a bipartisan national commission on COVID-19. We still need one. I served as the lead Democrat on the Heritage Foundation’s COVID-19 commission, chaired by John Ratcliffe. That work can provide a starting point, but it cannot substitute for the comprehensive national reckoning we still have not had. Five years after the pandemic ended millions of lives and transformed billions more, this should not be about protecting political tribes, institutions, or reputations. It should be about finding the truth. We need to over every stone, follow the evidence wherever it leads, learn from China’s failures and our own, and then use those lessons to prepare for the next pandemic—which could be far worse than COVID-19 and could arrive sooner than we think.

Damn

@thefinalcutbro @LoganARobison Agreed Dylan same here

@LoganARobison I’ve personally had a huge difference mentally. The physical benefits are an added bonus (which I’ve seen), but more importantly, I’ve found my memory and mental energy massively improved. Creatine has been amazing for me!

Creatine is definitely a scam, right? Been taking it daily for 3 years and can’t tell a difference

The billionaire who took food away from the world's poorest people is calling the guy who feeds the world's poorest people an "asshole".