Longreads + Open Thread

Longreads

• Steven Levy in Wired on the story of agents, particularly the project now known as OpenClaw. If nothing else, this has been a useful exercise in understanding how much people care about average latency versus access at random times, and it turns out that providing enough of both has convinced at least a few C-suite types to shift a substantial amount of their time to telling agents what to do. This has led to some distortions—Garry Tan has chosen to lean in to the narrative and disclose that he's on track to spend at least a million dollars on inference this year. But it wouldn't be at all crazy for an organization as profitable as Y Combinator to add up the cost of all the time spent there and find that more than a million dollars' worth of it had specifically been spent on CEO whims. That's what CEOs are selected for—good taste in expensive whims! A surprising number of senior executives have Github profiles that used to be mostly gray and are now a festive riot of green, but this is the exact cohort that had the most access to technical talent that could implement whatever they wanted. So, for the narrow slice of people who employ lots of programmers and are still addictive to agentic vibecoding, the best way to pattern-match this is to AI therapy and AI romantic companions: there are some things that you can economically afford to ask for, but that are too awkward to ask a real person for, and LLMs have finally provided an outlet.
• Scott Alexander reviews The Dialectical Imagination, a book about the Frankfurt School. It's great fun. Intellectual history from a generation or two ago really hits above its weight in providing a lot of educational value along with the expected historical value: it's long ago enough that it's clear who was right and who was wrong, and because the moments of high drama occur when a movement is at its peak, they always look incredibly petty in retrospect. Alexander ties this to a timeless pattern: the downwardly-mobile descendants of fairly rich people develop an extremely elaborate theory as to why that downward mobility is not their fault.
• Tyler Cowen interviews Toby Wilkinson, of The Last Dynasty fame, on Egypt, Ptolemaic and otherwise. One of the pleasures of ancient history is that they sometimes stumble upon a local minimum that's actually better than whatever system they could design in advance. The Ptolemaics subsidized Alexandria as a city uniquely suited to commerce, but also sold off the right to collect taxes in less fraught places. There was some noisy implicit Laffer Curve at work, where there was a tradeoff between state-sponsored projects and things the private sector would do on its own. One of the nice things about Egyptian history is that there's so much of it, so your sample size is bigger. If there's a general pattern in how power gets distributed, Egypt probably has the biggest sample size.
• Jay Caspian Kang continues his series about universities in the age of AI. If you truly enjoyed the cringe comedy of the original UK version of The Office, be sure to read the whole thing. The main message of the academics interviewed is that, one way or another, they're resigned: a few of them have decided to teach students as if they're going to have AI agents all the time, and many more have tried to find ways for their students to demonstrate that they've somehow exerted educational labor. This is a pretty natural split: AI massively magnifies some people's output, and replaces other people's entirely. The signaling function of education is more valuable than ever, because some new grads’ labor has been replaced while some of the graduates are vastly more productive than previous generations of 22-year-olds. So a school that embraces this, and thinks of its mission as credibly demonstrating that their graduates are worth hiring can do very well. Of course, schools are supposed to do more than train people to be diligent worker bees, but many of the anecdotes in this piece indicate that, given an opportunity for pursuit of wisdom for its own sake, a large fraction of students would rather outsource that dreary task to LLMs. If a college can’t fulfil its economic purpose or find enough students who care about its more holistic role, it doesn’t have a good reason to exist.
• Irene Zhang in Chinatalk on the prehistory of China's AI and robotics revolution. This is a very interesting piece from a political economy perspective: any time there's a net increase in productivity, but it leads to a decrease in the need for some jobs, there's a political question about who is responsible. Labor-substituting technologies have minimum value when there's a surplus of labor. And they respond appropriately when demand moves in the other direction.
• In this week's Capital Gains, we look at the market for alpha: if you can pay someone to beat the market by $1m, why would they let you pay anything less than $1m? As it turns out, the market for alpha is like any other market: if you want excess returns, you need to either take on more risk or accept less liquidity than whoever sets the benchmark.
• A Read.Haus reader asks why the markets are so blasé about the Strait of Hormuz, given how bad previous disruptions have been. I'm confused on this one, too! One possibility is that the market is pricing in enough overall growth from AI to offset some of the negative impact of expensive energy (and of tariffs, and other geopolitical risks). Maybe they S&P would counterfactually be a thousand points higher or something. Another possibility is that we've learned correct lessons from the last few big shocks—the US economy is resilient enough to survive a pandemic, inflation, and tariffs, and will survive the next thing, too. The US is relatively well-positioned in this respect, but the sample size is too small to say that disruptions don't matter. We might have just had a lucky streak.

Open Thread

• Drop in any links or comments of interest to Diff readers.
• What's everyone's guess for the next big bottleneck in AI?

Diff Jobs

Companies in the Diff network are actively looking for talent. See a sampling of current open roles below:

• Lightspeed-backed team building the engineering services firm of the future is looking for founding members of technical staff excited about working alongside civil engineers to translate their domain expertise into the operating system that powers the next era of great American infrastructure. If you’re an engineer with strong product intuition, who's energized by access to users, and excited by the prospect of transforming how we design and construct our built world with frontier AI, this is for you. (NYC, SF or Remote) 
• Series-A defense tech company that’s redefining logistics superiority with AI is looking for a MLE to build and deploy models that eliminate weeks of Excel work for the Special Forces. If you want to turn complex logistics systems into parametric models, fit them using Bayesian inference, and optimize logistics decision-making with gradient descent, this is for you. Python, PyTorch/TensorFlow, MLOps (Kubernetes, MLflow), and cloud infrastructure experience preferred. (Salt Lake City or NYC)
• A well-funded, Series C startup building the platform and agent primitives to drive operational transformation at large, complex institutions (starting with higher education) is hiring platform engineers. The work spans distributed systems, applied AI, and full-stack infrastructure, focused on deploying reliable agents that meaningfully bend institutional cost curves. (Remote)
• A hyper-growth startup that’s turning the fastest growing unicorns’ sales and marketing data into revenue (driven $XXXM incremental customer revenue the last year alone) is looking for a senior/staff-level software engineer with a track record of building large, performant distributed systems and owning customer delivery at high velocity. Experience with AI agents, orchestration frameworks, and contributing to open source AI a plus. (NYC)

Even if you don't see an exact match for your skills and interests right now, we're happy to talk early so we can let you know if a good opportunity comes up.

If you’re at a company that's looking for talent, we should talk! Diff Jobs works with companies across fintech, hard tech, consumer software, enterprise software, and other areas—any company where finding unusually effective people is a top priority.

And: we're now actively deploying capital into early-stage companies through Anomaly. Our focus is on defense, logistics, robotics, and energy. If you'd like to chat, please reach out.