- A personal essay on the 'Token for Token' Substack by ML researcher Jack Morris argues, according to the post, that treating research like a Zen discipline is the key to surviving the field's punishing randomness.
- Independent academic papers confirm that a systemic publish-or-perish culture drives burnout in machine learning, lending real-world context to the pain points Morris's essay addresses.
- Morris's Zen framing drew Hacker News debate, with some commenters arguing his approach reflects a Western self-strengthening reading of Zen rather than traditional practice, per observed community discussion.
What Folks Are Chattering About
Well, slap my knee and call me a catfish — a little self-published blog post about meditating your way through machine learning research done gone and bubbled up to the top of Hacker News like a biscuit in hot grease. The post, titled something along the lines of a Zen take on ML research, appears on a Substack called 'Token for Token' and is attributed to a writer going by Jack Morris. At time of observation, the Hacker News submission — posted by a user identified as jxmorris12 — had pulled in 47 points and 146 comments, which is the kind of engagement that says the fishing hole ain't empty.
The core chatter is this: Morris's essay, according to the post itself, draws on Zen philosophy to argue that doing ML research is a whole lot like sitting meditation — you sit whether insight shows up or not, and most days it won't. The blog post argues that the ability to keep grinding through those dry spells, rather than any particular flash of genius, is what separates researchers who make it from those who don't. That's Morris's personal view, and it ain't been independently verified by anybody wearing a lab coat outside his own garden.
What Is Actually Known
Here's what we can hang our hat on without squinting too hard. The blog post exists, it landed on Hacker News, and it generated real community engagement — 146 comments is more than most folks get hollering into the void. Two independent academic sources, an arXiv position paper and a paper from the ML4H 2024 symposium, separately document that a systemic publish-or-perish culture in machine learning genuinely drives researcher burnout and nudges folks away from risky, long-horizon work. That structural problem is real and corroborated, even if Morris's Zen cure for it is strictly his own prescription.
There's also an older piece of context worth knowing about: researcher Josh Schulman published what he called an opinionated guide to ML research back in 2020, and that essay independently noted that jumping between problems too fast — rather than sticking with a promising idea — is a more common failure mode than most researchers realize. That establishes a recognized tradition of practitioner advice essays in the ML community, which at least tells us Morris ain't the first to try this particular brand of campfire wisdom.
What Morris's Essay Actually Claims
According to the blog post, Morris draws an explicit comparison between ML research and a Zen sitting practice, writing — and we're paraphrasing to avoid copying his words whole-hog — that on days you find insight you keep working, and on days you don't find insight you keep working anyway, because that's just what the practice is. The post argues, in Morris's telling, that scientific insights arrive seemingly at random and that sustained effort and discipline, not inspiration, are the trait that counts.
The blog post also cites, according to Morris, Noam Shazeer's SwiGLU paper as a field-internal acknowledgment that successful research ideas can emerge from unexpectedly random tinkering. That's Morris's reading of that paper, not an independent editorial conclusion. Separately, the post prescribes what Morris describes as a dual path into ML research competence — a combination of reading and learning alongside hands-on building — arguing that leaning hard on one without the other leaves you limping like a three-legged mule at a rodeo.
What Remains Unverified
A whole barn's worth of things here ain't been nailed down. Morris's institutional affiliation and credentials have not been independently verified by this publication. The Hacker News username jxmorris12 is identified as the submitter, but community traction on a link-aggregator site does not confirm the author's expertise, the accuracy of his claims, or any real-world impact the essay might have. We are taking the essay at face value as a personal opinion piece — not as a research finding.
No independent source specifically corroborates Morris's Zen framing or the particular prescriptions he offers. The connection between Zen philosophy and ML research methodology is Morris's own analogy, and it ain't been peer-reviewed, road-tested, or blessed by anyone outside his own thinking. The claim that his approach actually helps researchers survive burnout is asserted in the post; there's no data behind it.
The Disagreements Bubbling Up
Now, the comment section apparently turned into a full-on philosophical pig-wrassle, which is about what you'd expect. Hacker News commenters debated whether Morris's framing reflects a genuine Zen tradition at all, with some arguing his emphasis on training the self to grow stronger is a Western interpretation of Zen — the kind that prioritizes self-fortification — rather than the self-dissolution at the heart of traditional East Asian Seon practice. That's a real distinction, and it means the essay's philosophical scaffolding is itself contested.
The bigger structural argument is thornier, and it cuts both ways like a double-sided axe. Independent academic literature documents a genuine systemic problem in how ML conferences and incentives are structured, with the publish-or-perish culture confirmed by separate researchers as a driver of burnout. Some critics of essays like Morris's argue that telling individual researchers to just sit with the discomfort places an unfair burden on the person when the problem is really the barn roof that needs fixing, not the farmer's attitude. That tension — between individual mindset shifts and structural reform — remains unresolved in both the academic literature and the Hacker News thread.
Analysis: Why This Caught Fire, and What to Make of It
This is analysis, not reporting. The fact that a personal Zen-inflected blog post cracked 47 points and 146 comments on Hacker News probably says more about the state of ML research culture than it does about the post's prescriptions. When independently corroborated academic work confirms that publish-or-perish pressure is real and burning people out, a blog post telling researchers to just keep sitting — to find peace in the randomness — is going to land like cold lemonade in July, no matter how philosophically wobbly the Zen framing might be.
The Schulman precedent is worth noting here as analysis: practitioner advice essays in ML have an established resonance in the community, and Morris appears to be writing in that tradition. Whether his specific Zen angle holds up philosophically, or whether mindset advice is an adequate response to a structural crisis, is genuinely debatable. But the engagement numbers suggest the underlying anxiety he's writing about is real, even if his remedy ain't been tested past the front porch.
Who is doing the hollering
These links show where the chatter came from. A link is attribution, not our endorsement or independent confirmation.
- Zen and the Art of Machine Learning Researchblog.jxmo.io (Token for Token Substack) · primary
- Position: The Current AI Conference Model is Unsustainable! Diagnosing the Crisis of Centralized AI ConferencesarXiv · specialist
- Recent Advances, Applications and Open Challenges in Machine Learning for Health: Reflections from Research Roundtables at ML4H 2024 SymposiumarXiv · specialist
- An Opinionated Guide to ML Researchjoschu.net · specialist
Last checked Jun 20, 2026, 1:06 AM EDT. Talk Around Town: This story rests almost entirely on one self-published personal essay that gained Hacker News traction. The author's identity, institutional affiliation, and credentials have not been independently verified. Community interest on Hacker News does not confirm the essay's claims or its cultural impact. Treat all characterizations of the essay's arguments as attributed to the author alone.