AI on Aneesh Sathe

Fri, 24 Jul 2026 21:48:36 +0000

The new Opus model was released. It is supposed to be less susceptible to prompt injection.

On Thickets, Enclosure, and Leaving

Tue, 21 Jul 2026 00:00:00 +0000

Over the past few months I have discovered and written about the idea of thickets and have been wondering what the practice of thicketing looks like. At its core the idea behind thicketing is to create and dwell in systems that resist capture and defy legibility. Weirdly this is not some anti-establishment move but a simple desire to just be and return to a state where you had your own corner, your friends, and your thoughts. Over the past few hundred years, with the ascent of modernity, legibility and legibilizing have been treated as the norm and a default good thing. However the pattern which has emerged (not just on the internet) is that systems position themselves as commons, invite people to enjoy themselves and slowly initiate capture mechanisms under the guise of some or the other security. Not all of this is bad, much of the modern world would not function without security and minimal enclosures.

The Birth of Thickets

Sun, 19 Jul 2026 00:00:00 +0000

The legibility regime runs one operation. It converts Space into Place, and everything else it does follows from that.

Space is the undifferentiated field. It holds every possibility and offers no orientation, which is why standing in it feels less like freedom than like vertigo. Draw a boundary around part of it, name what is inside, and you have made a Place. A farm is a Place; beyond the fence is forest. Drawing that boundary is the first act of thought, because you cannot know anything you have not first cut out from everything else. The trouble begins when the operation runs everywhere, on everything, at once, until it stops being thought and becomes weather.

The Lightening of Intent

Sat, 25 Apr 2026 00:00:00 +0000

I. Ceolfrith on the road #

In late spring 716, an English abbot named Ceolfrith left the twin monastery of Wearmouth-Jarrow in Northumbria for the city of Rome. He was seventy-four. He carried with him a manuscript Bible, the Codex Amiatinus, which he intended to present as a gift to Pope Gregory II. The Codex contained the entire Latin Bible in two volumes, weighed about seventy-five pounds, and required something close to one thousand calfskins to produce. Ceolfrith had commissioned three of these “pandects” at his abbey; this was the survivor he meant to deliver in person. The other two pandects are lost. The Codex Amiatinus was misattributed to a sixth-century Italian scribe for more than a thousand years; only in 1888 did Giovanni Battista de Rossi establish that the dedication page had been altered and the manuscript was Northumbrian work. He died at Langres in eastern Burgundy on 25 September 716, six months into the journey, before reaching Rome. His companions carried the Codex on. It sits today in the Biblioteca Medicea Laurenziana in Florence and is the oldest complete Latin Bible in the world.

The Cosmos and the Model

Fri, 10 Apr 2026 00:00:00 +0000

I. The Averaging Machine #

In 2025, researchers at USC ran a study that produced a result nobody expected. Atari et al., “AI-Powered Homogenization of Scientific Reasoning,” Trends in Cognitive Sciences (2026). They gave people reasoning tasks, some with LLM assistance, some without. Individual performance improved. Every metric showed it: more accurate responses, faster completion, fewer errors. The LLM made each person better.

Then they looked at the groups.

The Octotypic Mind

Fri, 27 Mar 2026 00:00:00 +0000

The Scholars of Dejima #

In the 1770s, a samurai, a physician, and a Confucian scholar gathered around a Dutch anatomy textbook none of them could properly read. They were attempting a translation of the Tabulae Anatomicae. Sugita Genpaku and his collaborators, the rangaku-sha, “Dutch studies scholars” of Tokugawa Japan. The shogunate had sealed the country for over a century. European knowledge trickled in through one pinhole: the Dutch trading post on Dejima, a fan-shaped artificial island in Nagasaki harbor.

The Kernel and the Ark

Sun, 11 Jan 2026 08:28:42 +0000

I. The Wall and the Infinite #

It is possible that the history of the modern West hinges on a single, melancholic misreading of Voltaire. When Candide, exhausted by the Lisbon earthquake and the brutalities of the Seven Years’ War, finally withdraws to the banks of the Propontis to utter his famous dictum—“Il faut cultiver notre jardin”—he is not proposing a program of agricultural management. He is issuing a plea for containment. To cultivate a garden, in the shadow of such overwhelming chaos, is an act of stoic resignation. It is an admission that the world is too vast, too violent, and too unintelligible to be governed by reason. One builds a wall against the infinite, and within that limited circumference, one tends to the soil. The garden is a refuge from nature.

The Deep Dark Terroir of the Soul

Fri, 26 Dec 2025 02:17:37 +0000

This is the third and final part of the Thicket Series:
Part 1: Logic of the Thicket and the Unsearchable Web
Part 2: The Architecture of Resistance

The history of the working subject might be best understood not as a ledger of wages or a sequence of industrial breakthroughs, but as a study in the migration of the Master. In the eighteenth century, the Master was a concrete presence, a figure residing in the castle or the cathedral, distinct from the worker by a physical and social chasm. One knew where the authority lived because one could see the smoke from its chimneys. By the nineteenth century, this figure had moved into the factory office, closer to the rhythm of the machine but still identifiable by the suit and the watch. The twentieth century saw a further dissolution; the Master became atmospheric, blending into the very walls of the institutions that housed us—the schools, the hospitals, the barracks.

The Architecture of Resistance

Tue, 23 Dec 2025 05:38:32 +0000

The seventeenth-century Hague, the mid-twentieth-century Levant, and the digital terraforming of 2025 have a shared preoccupation with the “Average.” Whether it is the theologian’s way or predictive stats, control begins by smoothing out the landscape. The project of power is a project of cartography and illumination—an attempt to banish the dark corners where the unmapped might grow. Thus, the history of resistance, of being “against the world”, is less a history of rebellion than a history of seeking cover.

The Shelter as Epistemic Engine

Wed, 17 Dec 2025 06:35:42 +0000

This is a continuation of my ongoing exploration of places and spaces. Previously: We need homes in the delta quadrant, Thinking with places, Problems are places questions are spaces.

Introduction: The Terror of the Open Field #

We tend to think of “Space” as a vacuum—an emptiness waiting to be filled. But geographically and philosophically, Space is actually a condition of high-entropy potential. As Yi-Fu Tuan famously articulated, space is “freedom,” but it is also “possibility without orientation.” It is the open field where everything is possible, which means nothing is yet distinct.

The Tortured Artist Is So Yesterday

Mon, 08 Dec 2025 05:08:29 +0000

41 years ago, Samuel Lipman wrote that an artist’s life is a “constant—and constantly losing—battle” against one’s own limits. That image has lasted because print culture taught us to imagine the artist as a solitary figure whose worth is measured by the perfection of a single, final work. Print fixed texts in place, elevated the individual author, and made loneliness part of the creative job description.

That world is slipping away.
And with it, the tortured artist.

Four Early-Modern Tempers for a World That Can Summon Itself

Sat, 06 Dec 2025 10:12:37 +0000

This is a partial synthesis of the books read through 2025 in the Contraptions Book Club.

We live in a moment when the whole of human culture has become strangely available, no longer just an archive but something that behaves like a responding presence. A sentence typed into a search bar or messaging window returns citations and, more strikingly, continuations: pastiche, commentary, new variations of ideas that never existed until the instant we requested them. The canon now behaves more like a voice than a library. It is easy to treat this as convenience, yet summoning culture alters our relation to meaning in ways we are only beginning to see. The question is no longer whether we can find the relevant text, but what it means to think in a world that can generate its own echoes.

The Small God of the Internet

Sat, 02 Aug 2025 05:26:46 +0000

It was a small announcement on an innocuous page about “spring cleaning”. The herald, some guy with the kind of name that promised he was all yours. Four sentences you only find because you were already looking for a shortcuts through life. A paragraph, tidy as a folded handkerchief, explained that a certain popular reader of feeds was retiring in four months’ time. Somewhere in the draughty back alleys of the web, a small god cleared his throat. Once he had roared every morning in a thousand offices. Now, when people clicked for their daily liturgy, the sound he made was… domesticated.

Why Every Biotech Research Group Needs a Data Lakehouse

Tue, 29 Jul 2025 06:04:54 +0000

start tiny and scale fast without vendor lock-in

All biotech labs have data, tons of it. The problem is the same across scales. Accessing data across experiments is hard. Often data simply gets lost on somebody’s laptop with a pretty plot on a poster as the only clue it ever existed. The problem is almost insurmountable if you try to track multiple data types. Trying to run any kind of data management activity used to have large overhead. New technology like DuckDB and their new data lakehouse infrastructure, DuckLake, try to make it very easy to adopt and scale with your data. All while avoiding vendor lock-in.

Work or Play? Ludic Feedback Loops

Mon, 28 Jul 2025 05:57:58 +0000

In his substack post today, Venkatesh Rao wrote about reading and writing in the age of LLMs as playing and making toys respectively. In one part he writes about how the dopamine feedback loop from writing drove his switch from engineering to writing. For him, writing has ludic, play-like, qualities.

I have made almost all my “career” decisions as a function of play. I originally started off with a deep love of plants, how to grow them and their impact on the world. I was convinced I was going to have a lot of fun. I did have some. My wonderful undergrad professor literally hand held me through my first experiments growing tobacco plants from seeds. But that was about it. My next experiment was with woody plants and growing the seeds alone took 6 months, and by the end I had 4 measly leaves to experiment with. I quickly switched to cell biology.

Briefing: The State of Explainable AI (XAI) and its Impact on Human-AI Decision-Making

Thu, 24 Jul 2025 06:10:10 +0000

This post is a sloptraption, my silk thread in the CloisterWeb. The post was made with the help of NotebookLM. You can chat with the essay and the sources here: XAI NotebookLM Chat

I. Executive Summary #

The field of Explainable AI (XAI) aims to make AI systems more transparent and understandable, fostering trust and enabling informed human-AI collaboration, particularly in high-stakes decision-making. Despite significant research efforts, XAI faces fundamental challenges, including a lack of standardized definitions and evaluation frameworks, and a tendency to prioritize technical “faithfulness” over practical utility for end-users. A new paradigm emphasizes designing explanations as a “means to an end,” grounded in statistical decision theory, to improve concrete decision tasks. This shift necessitates a human-centered approach, integrating human factors engineering to address user cognitive abilities, potential pitfalls, and the complexities of human-AI interaction. Practical challenges persist in implementation, including compatibility, integration, performance, and, crucially, inconsistencies (disagreements) among XAI methods, which significantly undermine user trust and adoption.

AI: Explainable Enough

Wed, 23 Jul 2025 05:50:31 +0000

They look really juicy, she said. I was sitting in a small room with a faint chemical smell, doing one my first customer interviews. There is a sweet spot between going too deep and asserting a position. Good AI has to be just explainable enough to satisfy the user without overwhelming them with information. Luckily, I wasn’t new to the problem.

Coming from a microscopy and bio background with a strong inclination towards image analysis I had picked up deep learning as a way to be lazy in lab. Why bother figuring out features of interest when you can have a computer do it for you, was my angle. The issue was that in 2015 no biologist would accept any kind of deep learning analysis and definitely not if you couldn’t explain the details.

My Road to Bayesian Stats

Tue, 22 Jul 2025 06:32:55 +0000

By 2015, I had heard of Bayesian Stats but didn’t bother to go deeper into it. After all, significance stars, and p-values worked fine. I started to explore Bayesian Statistics when considering small sample sizes in biological experiments. How much can you say when you are comparing means of 6 or even 60 observations? This is the nature work at the edge of knowledge. Not knowing what to expect is normal. Multiple possible routes to a seen a result is normal. Not knowing how to pick the route to the observed result is also normal. Yet, our statistics fails to capture this reality and the associated uncertainties. There must be a way I thought.

The secret flag of content

Sun, 20 Jul 2025 06:36:28 +0000

I don’t have any fun when I use LLMs to write. It may have perceived utility: popping out a LinkedIn article or two everyday. But I bet no one is actually reading. It’s a strip mall for a thumb stroll.

LLMs suck at writing. The summaries that LLMs give with the “Deep Research” are so poor in quality that I start to skim it. Yes, I skim the thing that is already a summary.

Divine Documentation

Wed, 16 Jul 2025 05:20:46 +0000

Dad was about my age when he said that reading the manual was better than hypothesis driven button pressing. For teenage me, that took too long. Sure, I may have crashed a computer or two but following my gut got me there. Of course my gut isn’t that smart. In the decades preceding, devices had converged on a common pattern language of buttons. Once learned, the standard grammar of action would reliably deliver me to my destination.

Beyond the Dataset

Fri, 11 Jul 2025 05:41:24 +0000

On the recent season of the show Clarkson’s farm, J.C. goes through great lengths to buy the right pub. As with any sensible buyer, the team does a thorough tear down followed by a big build up before the place is open for business. They survey how the place is built, located, and accessed. In their refresh they ensure that each part of the pub is built with purpose. Even the tractor on the ceiling. The art is in answering the question: How was this place put together?

Chatbots, Bats & Broken Oracles

Sat, 05 Jul 2025 05:54:05 +0000

I had the strangest conversation with my son today. There used to be a time when computers never made a mistake. It was always the user that was in error. The computer did exactly what you asked it to do. If something went wrong it was you, the user, that didn’t know what you wanted. After decades of that being etched in today I found myself telling him that computers make mistakes, you have to check if the computer has done the right thing and that is actually ok. A computer that hallucinates also provides a surface for exploration and seeking answers to questions.

We Need Homes in the Delta Quadrant

Tue, 29 Apr 2025 07:44:02 +0000

Place is security, space is freedom. — Yi-Fu Tuan

Starfleet Log, Delta Quadrant—Classified Briefing

At the edge of the known, maps fail and instincts take over. We don’t just explore new worlds—we build places to survive them. Because in deep space, meaning isn’t found. It’s made.

I. Interruption of Infinity #

The Delta Quadrant is a distant region of the galaxy in the Star Trek universe—vast, largely uncharted, and filled with anomalies, dangers, and promise. It is where the map ends and the unknown begins. No stations, no alliances, no history—just possibility.

Dwarf Fortress, Emacs, & AI: The allure of generative complexity

Wed, 22 Jan 2025 06:21:46 +0000

There is a shared soul shard between Dwarf Fortress, Emacs, and AI that lured me to them and has kept me engaged for over a decade. For a long time, I struggled to articulate the connection, managing only to describe Dwarf Fortress as the Emacs of games. But this analogy, while compelling, doesn’t fully capture the deeper resonance these systems share. They are not merely complicated; they are complex—tools for creativity that reward immersion and exploration.

Untitled

Mon, 20 Jan 2025 04:50:19 +0000

I’m taking part in the Contraptions Book Club where we are reading City of Fortune which is about Venice. I was struck by the character of Doge Dandolo. Dude was 80+ when we saw a trade opportunity in the 4th Crusades. In the book, the author, Roger Crowley describes a brief moment when Dandolo makes a heroic rush on the banks of Constantinople’s Golden Horn during the Sack of Constantinople.

I found both the Doge and the imagery interesting so went looking for art depicting the art, there’s supposed to be lots. Unfortunately I couldn’t find any and nothing in the public domain. So I asked AI to generate something.

The Universal Library in the River of Noise

Sun, 12 Jan 2025 08:06:00 +0000

Few ideas capture the collective human imagination more powerfully than the notion of a “universal library”—a singular repository of all recorded knowledge. From the grandeur of the Library of Alexandria to modern digital initiatives, this concept has persisted as both a philosophical ideal and a practical challenge. Miroslav Kruk’s 1999 paper, “The Internet and the Revival of the Myth of the Universal Library,” revitalizes this conversation by highlighting the historical roots of the universal library myth and cautioning against uncritical technological utopianism. Today, as Wikipedia and Large Language Models (LLMs) like ChatGPT emerge as potential heirs to this legacy, Kruk’s insights—and broader reflections on language, noise, and the very nature of truth—resonate more than ever.

Jan 10, 2025 - AI Agents, Machiavelli's Study

Fri, 10 Jan 2025 08:06:00 +0000

Agents Are Not Enough #

Last year I was heavily experimenting with Knowledge Graphs because it’s been clear that LLMs by themselves fall short because of the lack of knowledge. This paper by Chirag Shah and Ryen White (you can click the heading above) from Dec 2024 expands on those shortcomings by exploring not just knowledge but also value generation, personalization, and trust.

They open the paper by casting a very wide definition of an “agent” everything from thermostats to LLM tools. While this seems facetious at first, their next point is interesting. Agents by definition “remove agency from a user in order to do things on the user’s behalf and save them time and effort.”. I think this is an interesting way to injext an LLM flavored principal agent problem into the Agentic AI conversation.

Jan. 5, 2025

Sun, 05 Jan 2025 14:00:00 +0000

Improving Research Through Safer Learning from Data #

Another one of Frank Harrell’s posts. Given my day job, and the R&D background this one is quite close to home. As a team leader on the industry side one hopes to build a culture with the team that aligns scientific rigor with company goals. Any method, statistical or cultural (in this case both), that solves for this tension will get you the most bang for your buck.

The Arrival of Composable Knowledge

Wed, 12 Jun 2024 09:13:34 +0000

Traversing through human history, even in the last two decades, we see a rapid increase in the accessibility of knowledge. The purpose of language, and of course all communication is to transfer a concept from one system to another. For humans this ability to transfer concepts has been driven by advancements in technology, communication, and social structures and norms.

This evolution has made knowledge increasingly composable, where individual pieces of information can be combined and recombined to create new understanding and innovation. Ten years ago I would have said being able to read a research paper and having the knowledge to repeat that experiment in my lab was strong evidence of this composability (reproducibility issues not withstanding).

Dancing on the Shoulders of Giants

Fri, 10 May 2024 18:31:09 +0000

In Newton’s era it was rare to say things like “if I have seen further, it is by standing on the shoulders of giants” and actually mean it. Now it’s trivial. With education, training, and experience, professionals always stand “on shoulders of giants” (OSOG). Experts readily solve complex problems but the truly difficult ones aren’t solved through training. Instead, a combination of muddling through and the dancer style of curiosity is deployed, more on this later. We have industries like semiconductors, solar, and gene sequencing with such high learning rates that the whole field seems to ascend OSOG levels daily.

Reimagining AI in Healthcare: Beyond Basic RAG with FHIR, Knowledge Graphs, and AI Agents

Fri, 03 May 2024 21:15:05 +0000

Introduction #

While exploring the application of AI agents in healthcare we see that standard Retrieval-Augmented Generation (RAG) and fine-tuning methods often fall short in the interconnected realms of healthcare and research. These traditional methods struggle to leverage the structured knowledge available, such as knowledge graphs. Data approaches like Fast Healthcare Interoperability Resources (FHIR) used alongside advanced knowledge graphs can significantly enhance AI agents, providing more effective and context-aware solutions.

On Being Good vs. Knowing Good: Perspectives on AI

Mon, 22 Jan 2024 19:41:07 +0000

In this video, Stephen Fry narrates Nick Cave’s letter, which argues that using ChatGPT as a shortcut to creativity is detrimental. Surprisingly, I found myself in agreement. Having been involved in the AI industry for over a decade, I’ve always viewed AI positively. As an entrepreneur who pitches AI to investors and customers, I liken AI to technologies like spreadsheets: they eliminate tedious tasks, but you still need to understand what you’re doing.