Phil Steitz

I invented a new word today: “downslopping”. When an idiot leader with inflated perception of their and their bots’ abilities spews slop downward in an organization.

I'm out here sweating in the theory mines, discussing gender and race and tech, and these chucklefucks just tweet it out.

A friend just pointed out Vermillion. The idea is to use Verus annotations on a Rust program but replace Z3 with Lean. This increases the sophistication of possible proofs allows simplification. Nice work. proofsandintuitions.net/2026/10/18/v... @ilyasergey.bsky.social #proof #verus #rust #lean

Yesterday after several failed attempts to get a simple spec doc from Claude, I said “make the doc concise - no cutesy emdash throwaway crap like this.” Worked!

Agent-generated mashups with wildly off data seem to be becoming the norm. I wonder how long it will be before people realize that they can reduce a lot of cost and latency by just prompting their mashup agents to make up reasonable-looking numbers.

It occurred to me today that the massive AI capex burn by big tech companies is like them all being hit with LBOs that load debt onto their balance sheets. We know how that ends.

psteitz.blogspot.com/2026/10/how-...

I like the optimism here. pierremarion23.github.io/thoughts/202...

“You are infringing on the time and brainpower of people by pushing "your" content that little/no of your brainpower went into” is such a banger

Interested in teaching your organisation to use open source collaboration methods internally and thus prepare to seamlessly work with the upstream open source projects your business is built on? Join me in attending this year's InnerSource summit. innersourcecommons.org/events/isc-2...

Cascade of stupidity worthy of AI bro today. Decided to run in the rain. Left glasses at home (worthless in rain). Went off trail (trail == wash in AZ). Picked up pace to stay warm. Boom! Probably broke my foot :(

www.youtube.com/watch?v=5YpK... There are some nice examples in this talk showing how a coherent concept of "presence" is difficult today. The concept that she attributes to Noe of "presence as availability" made me chuckle about how badly we abuse both of those terms in collaboration technology.

Jev is demonstrating the obvious. Solving simple problems in slop-space with fake humans is laughably wasteful. Years from now, we will look back at the grotesque over-dimensionality of LLM agent representations and crack up.

www.interconnects.ai/p/the-curren... One more time with feeling. Open source always wins.

If you want a Sunday morning longread I wrote up a rough version of the talk I gave at LDX3 this week on the Manager's Path in the Age of AI skamille.medium.com/the-managers...

share.google/lLm3xjtB3rSa... One of my fondest memories of grad school was when a luminary gave an amazing talk and I asked him to clarify some definitions afterward. He said, “I didn’t come here to show you the work. I came to show you the magic.”

I just found typesafe.ai I still do not have enough details to decide what to think of it. But the classifier interface feels so much more suited to a lot of tasks that people somehow stuff into the LLM pipeline. It feels a lot more like classic machine learning.

New year, new semester, new course, new textbook: davidpicard.github.io/mldl/mldl-bo... 👀 Still in draft form, but not in a bad shape.

I made an comment the other day about anomaly detection in sporadic events. See the O'Reilly book for a more complete writeup. But here is the code that demonstrates the method: github.com/tdunning/int... www.oreilly.com/library/view... @de6u99er.bsky.social #anomalydetection

A group of 25 Fields Medalists, including myself, have made a joint declaration on Math and AI: mathandai.org . We welcome additional signatories. See also this article in the Economist announcing the declaration: www.economist.com/science-and-...

“Mathematics only makes sense as a gift economy. It has no trade secrets. Substituting a monetizable ersatz culture for the one evolved through historical experience is a particularly insidious form of knowledge collapse.” Mathematician Michael Harris on the AI companies coming for math:

This post is basically my attempt to get everyone I mostly agree with to maybe stop arguing so much. There are actual villains out there for us to fight. I am sure this one blog post will solve everything.

who wants to go in on this together?

I fear that placing too much emphasis on a specific interpretation of dimensionality, or treating dimensionality as an end-all quantification of some aspect of neural computation, may lead us down the wrong path, writes @mattperich.bsky.social. #neuroskyence www.thetransmitter.org/neural-dynam...

I went outside with my new glasses today. Saw a spot on the periphery. Shook my head and it did not go away. Absolutely a feature, not a bug. Now let’s see if I can live with it.

Every. Single. Day. I think about this post every single day.

😭

Long, but worth reading www.irregular.com/research/end... But I think he is missing two key concepts: 1. Defense has an inherent information advantage that can be aggressively exploited. 2. Innovation in compensating controls is a big missing example from the differential defensive arsenal.

Here again, guardrails are too late. You need governance *in context* engineering.moniepoint.com/what-i-learn...

Who blooms at 112 degrees? Compass cactus!

Le livre "Introduction aux graphes aléatoires" (Calvage-et-Mounet, 2020) est épuisé depuis plus d'un an. Comme je reçois régulièrement des courriers sur ce texte, j'ai rédigé une seconde édition que je place en accès libre sur: www.rogermansuy.fr/pdf/GraphesA... Bonne lecture!

Unsettling indeed. One thing that jumped out at me with great irony. The example of “LLM-speak” popularizing Plato’s “dividing reality at the joints” quote. Following that quote in the Phaedrus is a rant on the difference between sounding good (rhetoric) and seeking real meaning (dialectic).

Two things I love about this: 0) the data product innovation framework is excellent. 1) it shows that the morons in the White House have had zero impact on the commitment to availability, accuracy and confidentiality in the US Census Bureau. Many thanks for that. share.google/0wHY35cZjGss...

What role does Emergence play in Neural Networks? We find that learning emergent low-dimensional representations is key for out-of-distribution generalisation. New Preprint out with @neural-reckoning.org arxiv.org/abs/2607.10430

A cluster hire here at UNSW Science, looking for cross-disciplinary Level B/C academics. Come and work in Sydney! www.unsw.edu.au/science/abou... Crossposted with @openvibe.social

This is wonderful. I especially like “Don’t be prissy.” There is no problem more sure to go away on its own than an ink smudge on your finger. open.substack.com/pub/nealstep...

This kind of analysis applies to other fields as well. The move from proof scarcity” to “proof abundance” has analogues everywhere that AI is being used to do hard technical work. The backlog of reformulation for real insight is growing rapidly. teorth.github.io/tao-web/slid...

Troubling, but not at all surprising. I am reminded of Socrates’ warning to Hipppcrates in Protagoras 314b that you can’t “return” cognitive experience to “the vendor” like bad food because you have already been changed by it.

New #AI #addon helps developers #automate everyday #programming #tasks Researchers developed Program-as-Weights (PAW), a new method that lets developers turn AI instructions into tiny, local tools, enabling powerful, private & offline AI processing scitechupdates.com/n...

Happy to be a small part of this colossal effort led by @zouharvi.bsky.social to reinvent the future of machine translation evaluation! 🔡 Join us in building the LTB and become a coauthor, help wanted!

Enterprise AI has a data ingestion crisis: unscalable pipelines, and critical document security (ACL) leaks. Launching the Open Ingestion Standard (OIS) and OpenCrawling! www.open4dev.com/journal/2026... #Java25 #RAG #AISecurity #ZeroTrust #VectorDatabase #SpringBoot #MCP #ModelContextProtocol

Turns out if you have complete and total disregard for human creative effort you eventually steal it from someone with the resources to hit back.

I’m a little late with this message having been offgrid for the last week. Happy Birthday, USA! We will never give up on you.

eshumarneedi.com/2026/07/03/z... Nice rant.

Communities around the U.S. are putting on epic fireworks shows for America's 250th birthday — but that spells trouble for dogs and cats. Here's a guide to help the animals in your life.

A coworker gave me a compliment today. She said I was a “great storyteller.” I cringed. If a child told me that I would be happy, but in the workplace…not so much. The whole “story-telling” push among the b-school gang today reeks to me of epistemic decay.