Hello all, it's been a while since I posted one of these newsletters. Mainly, I wanted to share links to a pair of articles I’ve had published over the last couple of months. Hope you find them interesting!

First, in The American Prospect, I had a piece arguing that Congress ought to revitalize the long-dead Office of Technology Assessment – its old in-house body of science and technology experts – before it rushes into building a bespoke regulatory regime for artificial intelligence technologies. 

Here’s the link: https://prospect.org/2026/09/18/artificial-intelligence-congress-office-tchnology-assessment/

This came out a few days after 27-year-old coder Jacob Coxon left the AI giant Anthropic and started warning that AI might eradicate humanity by the end of the decade. In a very strange turn of events for someone accused of abetting Armageddon, Anthropic CEO Dario Amodei said within days “yeah, could happen!” and started pressuring Congress to buttress his company’s safety regime right now, while there’s STILL TIME.

I’m extremely skeptical of all this, but my point in the Prospect piece wasn’t that Coxon or Amodei are wrong. I don’t know if AI is going to wipe out humanity by 2030, in the same way that I don’t know whether aliens will invade the Earth by 2030. But I don’t think we should rush into integrating a brand new, investment-intensive, ultra-unprofitable, and planet-burning industry into the fabric of the federal government basically because its creators are bad at cybersecurity. (To say nothing of the cult(s) for Yale nerds that came up with a lot of this “alignment” and “existential risk” stuff in the first place.) Let’s get some actual tech experts to independently assess the issue first, then think about legislating from there.

On a related note: the AI bubble really does look like it’s starting to pop after CoreWeave fell into junk bond territory last week and Oracle tried to invoke force majeure clauses earlier this month. Apparently investors are now rushing into credit default swaps to bet against AI and data center companies. 

Americans for Financial Reform (where, disclaimer, I used to work) and the Open Markets Institute put out a hell of a report on just how big and just how bad the AI bubble might be. This stuff has fascinated me since AI burst onto the scene a few years ago, given my obsession with financial crises. But seeing it all laid out in one place is pretty chilling. If things keep going sour, we’ll have to see who’s left holding the bag…but it could get ugly.

If you’re looking to get caught up on AI bubble analysis, check out Ed Zitron and Dean Baker (whose think tank I also used to work for, another disclaimer.) Plus, I truly recommend everyone read Cory Doctorow’s phenomenal book The Reverse Centaur’s Guide to Life After AI. It’s super approachable, short enough to read in a weekend, and will answer questions you didn’t know you had.

Second, in The Nation, I had a piece about centrists who argue that Congressional Democrats shouldn’t investigate the companies that have collaborated with or paid off the Trump administration. Why? Because in the words of one Congressman, Trump “had a gun on their head.” 

Here’s the link: https://www.thenation.com/article/politics/congress-trump-oversight-hannah-arendt/

That phrase really irked me. It got me thinking about Hannah Arendt, the philosopher and Holocaust survivor who wrote extensively about whether and how one is responsible for what they do under fascist and totalitarian regimes. Her book Eichmann in Jerusalem explores legal, moral, and psychological culpability in the trial of a major Nazi bureaucrat. The follow-up essay “Personal Responsibility Under a Dictatorship” offers a nuanced, but full-throated defense of morally judging people for their actions – especially powerful people living through a fundamental shift in the society that brought them to power.

The nub of my essay is reading “Personal Responsibility” for any lessons it can offer on how to think about the fascist collaborators in our time. As I say in the piece, this country badly needs truth commissions and criminal prosecutions for what has taken place over the last two years, and really long before that. Arendt spent the latter part of her life trying to make sense of what happened when senselessness took hold in the country of her birth, so she is one of many thinkers with a lot to teach us.

Finally, I should also share the biggest news: I passed the bar exam! I’ll be sworn in as a full-blown licensed attorney later this week. It’s a big accomplishment, the capstone of three years of hard work. Yay for me.

I do have an angle on this besides self-congratulation: As an Oregonian, I happened to be part of the first-ever cohort to take the new “NextGen” Uniform Bar Examination, which will be rolling out to most other U.S. jurisdictions over the next few years. So, what is the “NextGen,” and is it any good?

The bar exam is a very weird institution. It aims to test general legal knowledge, but no single practitioner relies on their “general legal knowledge,” and no one uses all of the subjects they have to learn to pass the bar – any attorney who does leveraged financial transactions and religious liberty cases and criminal defense has a bizarre practice indeed. As a result, many onlookers accuse the bar exam of being a pointless memorization exercise. Test-takers immediately forget most of the things they learned to pass the test, and the skills that help them succeed at the bar exam have little to do with the research, analysis, and persuasive abilities that make or break actual lawyers.

In response to these points, the “NextGen” exam offers reform, not revolution. It reduces some of the memorization, and tries to test analysis and application skills more. Many questions are structured as though the examinee is a first-year attorney doing typical first-year attorney work. Instead of asking you to recite old common-law rules that few or no actual jurisdictions still follow, you might be provided with a model statute. Instead of being given a wall of facts and told “Analyze this,” you might be given a wall of facts and asked to draft an office memo or a letter to a client.

I think this is marginally better. But the NextGen is still a standardized test, which means the old standardized test strategies – memorization, essay formats, etc. – are still absolutely required to perform well. I can’t know for sure at this point, but it certainly doesn’t feel like studying for this test did anything to prepare me for appearing in court. My process-of-elimination and time-budgeting-per-question skills are not going to mean anything to a judge.

The underlying issue is “What is the point of the bar exam in the first place?” Everyone knows that you only really learn how to become a lawyer by doing it, and generally, you only start doing it once someone else hires and mentors you. 

Bar admission systems used to acknowledge that, in a sense: in the early days of the American Republic, lawyering was learned by apprenticeship, same as most other trades. Most states granted bar admission only after someone apprenticed under an existing attorney, and maybe passed a token oral test. 

But that produced a profession that was even clubbier than it is now – if anything, the clubbiness was the point. The token oral tests were more like fraternity hazing. A Missouri bar exam entailed two hours of heavy drinking and macho posturing with senior lawyers. Abraham Lincoln once reportedly administered someone’s Illinois bar exam by asking him a few questions about contract doctrine while he watched the future President take a bath. Ain’t history grand?

It was only post-Civil War that the legal profession started to, well, professionalize. The railroad, the factory, and the general incorporation statute meant that newly-massive interstate corporations suddenly needed lots of lawyers, especially multistate practitioners. This meant society had to produce a lot more lawyers, and states had to process a lot more requests to join their bar. The answer, befitting of the Industrial Revolution, was to create a standardized legal education (the JD) and standardized methods of bar admission (the bar exam). The old initiation rituals were gradually displaced by the industrial, scientific, efficient ideals of the modern age and the new corporate clientele. You didn’t have to prove you were one of the boys; you had to prove you were an up-to-standard replaceable part.

(Of course, the old boys’ club did not disappear; it just became a background institution, the thing that gets you the fancier job in this now-massive legal industry. Maybe the boys’ club dressed itself up a bit, named itself after the Founding Fathers, that kind of thing.)

Where does that leave us today? To become a lawyer, you first need to get an academic degree certifying that you have completed a standardized education in a few, standardized areas of the law. (Particularly the areas relevant to a 20th-century corporation or wealthy individual.) You then complete a standardized assessment of your standardized knowledge, and if you pass…then you spend another five years learning an actual practice area, usually from an elder practitioner.

In other words, it’s still functionally an apprenticeship model. There’s just a lot of student loan debt and bureaucratic hoops to jump through before you can even access the apprenticeship. 

Now, I’m not at all trying to say that we have it worse than pre-industrial apprentices. I’m not a historian, but I’m pretty confident saying that apprentices in the pre-industrial world weren’t exactly living comfortably. Moreover, the underlying purposes of both law school and the bar are genuinely important. Educating all lawyers in core concepts through a common curriculum enables a shared professional vocabulary, and academic study of the law as a unified system is absolutely crucial to improving it. Most importantly, the state really does need to ensure a minimum standard of competence in this field. Lawyers take another person’s life into their hands – sometimes literally for defense attorneys, but metaphorically and emotionally for just about all practitioners.

Total reliance on standardized tests, though, is a deeply flawed strategy for maintaining those minimum standards of quality. Plenty of brilliant lawyers initially fail the bar, and plenty of awful lawyers pass it easily. Some states, including Oregon, are experimenting with state-supervised formal apprenticeships as an alternative to taking the exam. I hope these programs continue to grow and refine themselves. There should be plenty of ways to demonstrate one’s competence, without sacrificing rigor.

Some argue that the bar exam is now basically pointless, because law schools today go through a difficult accreditation process with the American Bar Association that ensures high quality education and a standardized curriculum. I’m cautious about leaving everything up to the ABA, though; it’s basically a lobbying group, a professional association unaccountable to the public. I think states themselves do need some role in deciding who gets to wield the powers of their legal systems – having the professional association and the government keep each other in check is sensible to me.

To me, the biggest issue is the debt load of law school. Even when someone wants to pursue social or economic justice work, sheer need for money often pushes new grads into BigLaw, which can influence their perceptions and approach for the rest of their careers. For all of the ideological education baked into law school, debt is always one of capitalism’s most effective tools of discipline. 

The bar exam isn’t directly responsible for that, but I can say from recent experience that spending three years working toward a degree, finally graduating, and then needing to spend another five months test-prepping and waiting for results strains budgets that much more. But solving the economics of higher education, and the threat of unemployment in a modern economy, are both subjects for another day. For now, I’m okay with leaving things in a bit of an ambiguous place – the bar exam serves important ends through extremely flawed means, it’s good to develop alternatives, and more broadly, lawyers ought to put our peculiar professional traditions in a bit of historical context.