Sign in

Daniel Van Zant

@danielvanzant.bsky.social
16 followers 19 following 121 posts

Building AI that amplifies human expertise instead of diminishing it. Founder at www.sylvum.com PhD Student at www.mpcrlab.com

PostsRepliesMedia
Daniel Van Zant @danielvanzant.bsky.social · 05/02/2026
4/ This is the first in a six-part series I'm writing for the Wolfram Institute on new kinds of science organizations. wolframinstitute.org/output/how-d...
wolframinstitute.org
How Democratic Coarse-Graining Filters Against Some of the Research That Matters Most — Wolfram Institute
Daniel Van Zant This article is part of a series of articles I am releasing as part of a research project with the Wolfram Institute. This project is intending to answer the question: How can DeSc...
000
Daniel Van Zant @danielvanzant.bsky.social · 05/02/2026
3/ I call this the coarse-graining problem, borrowing from Stephen Wolfram. Funding agencies are computationally bounded observers. They cannot track every investigation in detail, so they rely on proxy indicators. Transformative research often resists this. The details matter.
100
Daniel Van Zant @danielvanzant.bsky.social · 05/02/2026
2/ There is an explanation. Public research funding in a democracy requires public justification. Agencies defend budgets to legislators who defend votes to constituents. This creates pressure toward research that can demonstrate understandable progress within politically relevant timeframes.
100
Daniel Van Zant @danielvanzant.bsky.social · 05/02/2026
1/ I did a historical analysis of five transformative technologies from 1950 to now. Research periods with long time horizons and high uncertainty were about three times more likely to experience sustained underfunding.
112
Daniel Van Zant @danielvanzant.bsky.social · 18/08/2025
When I find an interesting article/paper/etc I don't read it right away, I save it in a folder on my computer. Once I have enough, I run a small script to get it printed as a glossy magazine with an AI-designed cover (here's the latest) and read through everything at my leisure.
000
Daniel Van Zant @danielvanzant.bsky.social · 13/08/2025
I am presenting online for the Wolfram Institute at 2pm ET today. Will be talking about an AI system that I am building and how it might be used to validate abstract theories with purely computational "Wolframian-style" experiments. m.youtube.com/watch?v=S9qz...
m.youtube.com
Presentation | Daniel Van Zant | How AI-assisted Organization can Accelerate Science Theory-Building
YouTube video by Wolfram Institute
000
Daniel Van Zant @danielvanzant.bsky.social · 21/07/2025
Got to have a great conversation with my advisor Dr. Elan Barenholtz on his theory of language! youtu.be/P-yfQLDM5pA?...
youtu.be
Is Memory Real? UWaterloo Discussion w/ Prof. Elan Barenholtz & Dan Van Zant | Ekkolapto Salon
YouTube video by Elan Barenholtz, PhD
020
Daniel Van Zant @danielvanzant.bsky.social · 09/07/2025
Setting up a weapons shop inside Rome's walls makes you a respected merchant. Outside the walls, you're seen as a barbarian threat, but you're free from Rome's taxes and politics. Research has the same geography. It's worth knowing which side of the wall serves your purposes.
010
Daniel Van Zant @danielvanzant.bsky.social · 11/06/2025
One minute clip from a short talk with Dr. Elan Barenholtz about how he uses AI in his work. Check out the full video here: www.danielvanzant.com/p/an-extra-c... . Also worth checking out an interview he did with Curt Jaimungal of Theories of Everything: curtjaimungal.substack.com/p/the-theory...
110
Reposted by Daniel Van Zant
Daniel Van Zant @danielvanzant.bsky.social · 08/06/2025
For anyone interested in this question this guy: substack.com/@cyberneticist sent me this excellent resource: people.eecs.berkeley.edu/~jrs/highd/ over on Substack.
substack.com
Cyberneticist | Substack
Writing about the higher level theory of computation and cybernetics from a critical perspective
032
Daniel Van Zant @danielvanzant.bsky.social · 08/06/2025
For anyone interested in this question this guy: substack.com/@cyberneticist sent me this excellent resource: people.eecs.berkeley.edu/~jrs/highd/ over on Substack.
substack.com
Cyberneticist | Substack
Writing about the higher level theory of computation and cybernetics from a critical perspective
032
Daniel Van Zant @danielvanzant.bsky.social · 07/06/2025
Being unbiased in science communication seems almost impossible. I admire scientists who, instead of pretending to be unbiased, work to be transparent about their biases. I.e. "I came up with this theory so I prefer it," "my field looks at this thing this particular way," etc.
020
Daniel Van Zant @danielvanzant.bsky.social · 06/06/2025
Hyperdimensional spaces (like the ones that LLMs and other AI systems work in) have emergent properties that are more than just "scaling up 3 dimensions a lot". Is there a single resource that lists these unique properties. Or even better one that gives a good intuition of them? #MLSky
120
Daniel Van Zant @danielvanzant.bsky.social · 21/05/2025
I've definitely found myself being unfairly negative about LLM's. I suspect it's because I'm completely epistemically exhausted from distinguishing the 0.1% of truth from the hype, and my priors subconsciously become "If someone is saying LLM's can do a thing they are wrong.".
020
Reposted by Daniel Van Zant
Ian Phillips @ianbphillips.bsky.social · 20/05/2025
5 years since our first pilot, and 25,000 participants later, I'm super happy this work with Makaela Nartker, @chazfirestone.bsky.social and Howard Egeth on inattentional blindness is now out in eLife! A little 🧵 of what we found... 1/12 elifesciences.org/articles/100...
elifesciences.org
Sensitivity to visual features in inattentional blindness
As a group, inattentionally blind participants can successfully report the location, color, and shape of stimuli they deny noticing, and exhibit a systematic bias to report not noticing.
48338
Reposted by Daniel Van Zant
Daniel Van Zant @danielvanzant.bsky.social · 21/05/2025
Wikipedia is also a good evaluator of how much I can trust the LLM. If the topic has multiple fleshed-out Wikipedia articles with lots of sources then the LLM is usually mostly right. If there's just one two-paragraph article that cites one paper then usually the LLM output has many hallucinations.
032
Reposted by Daniel Van Zant
Kevin Zollman @kevinzollman.com · 20/05/2025
There's so much polarization around LLMs. They are way overhyped, I agree. But I also use them semi-regularly now. Here's a thread of genuine use cases where I find them helpful. Please add your own!
79022
Daniel Van Zant @danielvanzant.bsky.social · 21/05/2025
Wikipedia is also a good evaluator of how much I can trust the LLM. If the topic has multiple fleshed-out Wikipedia articles with lots of sources then the LLM is usually mostly right. If there's just one two-paragraph article that cites one paper then usually the LLM output has many hallucinations.
032
Daniel Van Zant @danielvanzant.bsky.social · 14/05/2025
Demo of an AI tool that displays relevant information live as you are having a technical conversation or presentation. The display is based on either a query you gave it beforehand where it can grab and process relevant documents from the internet. www.danielvanzant.com/p/augmenting...
danielvanzant.com
Augmenting technical conversations with AI (demo)
A quick demo of a project I am working on for an AI tool that displays relevant information live as you are having a technical conversation or presentation.
000
Daniel Van Zant @danielvanzant.bsky.social · 08/04/2025
asteriskmag.com/issues/09/gr... Very cool article going into detail on what the best strategies might be for humans to take their first steps into becoming a multi-planet species. Always excited about anything on how we could get one step closer to Star Trek.
asteriskmag.com
Greening the Solar System—Asterisk
A future where life flourishes beyond Earth is closer than you think. How, precisely, will we get there?
000
Daniel Van Zant @danielvanzant.bsky.social · 02/04/2025
I will be presenting my work building an artificial intelligence cognitive co-pilot to help neuroscientists create better theories next week, April 9th, 10am EST. Anyone is welcome to join the Zoom: fau-edu.zoom.us/j/3806783624... . The talk will also be recorded and posted on Youtube.
010
Daniel Van Zant @danielvanzant.bsky.social · 31/03/2025
1. Figure out what the current self-maintenance mechanisms are. 2. Figure out how to circumvent those self-maintenance mechanisms. 3. Setup new self-maintenance mechanisms for the new desired state. 4. Wait for the system to transition on its' own to the new desired state.
000
Daniel Van Zant @danielvanzant.bsky.social · 31/03/2025
Most systems that exists do so because they are good at maintaining their current state in the face of entropy. If you want to change the state of a system on the long-term you have to do three things:
100
Daniel Van Zant @danielvanzant.bsky.social · 27/03/2025
www.newscientist.com/article/2472... UK tech secretary's official chatgpt records were requested and given under freedom of information(FOI) laws. Under most FOI laws, text conversations between two people on are exempt. Should conversations with a chatbot be treated the same?
newscientist.com
Revealed: How the UK tech secretary uses ChatGPT for policy advice
New Scientist has used freedom of information laws to obtain the ChatGPT records of Peter Kyle, the UK's technology secretary, in what is believed to be a world-first use of such legislation
000
Daniel Van Zant @danielvanzant.bsky.social · 25/03/2025
7/7 Read the full proposal on my blog open.substack.com/pub/danielva.... I'd love some constructive criticism on this idea. #EffectiveAltruism #ScienceFunding #metascience
open.substack.com
Breakthrough Incentive Markets
Aligning financial incentives with scientific progress to solve humanity's most urgent challenges
000
Daniel Van Zant @danielvanzant.bsky.social · 25/03/2025
6/7 Everyone does what they do best: The public identifies important problems through donations, markets efficiently allocate resources, and researchers focus on innovation instead of grant-writing.
100
Daniel Van Zant @danielvanzant.bsky.social · 25/03/2025
5/7 The key innovation: Investors' returns are higher when solutions come sooner. This creates powerful incentives for speed, collaboration, and funding approaches too radical for grants but too public-good-oriented for industry.
100
Daniel Van Zant @danielvanzant.bsky.social · 25/03/2025
4/7 Breakthrough Incentive Markets create outcome pools for specific scientific problems. Investors buy positions, fund promising research, and earn returns when breakthroughs happen.
100
Daniel Van Zant @danielvanzant.bsky.social · 25/03/2025
3/7 What if we harnessed the same efficient mechanisms that drive innovation in the private sector to fund science that solves our most important problems?
100
Daniel Van Zant @danielvanzant.bsky.social · 25/03/2025
2/7 Science today is trapped: academic funding favors incremental progress over radical ideas, while industry only invests where profits exist. Critical research falls through this gap.
100
Daniel Van Zant @danielvanzant.bsky.social · 25/03/2025
1/7 New on the my blog: Breakthrough Incentive Markets - a third path for science funding that combines democratic problem prioritization with market-driven resource allocation. open.substack.com/pub/danielva...
open.substack.com
Breakthrough Incentive Markets
Aligning financial incentives with scientific progress to solve humanity's most urgent challenges
100
Daniel Van Zant @danielvanzant.bsky.social · 19/03/2025
Have been enjoying participating in these "salons" which are long-form semi-structured conversations with a neat group of scientists and thinkers. I have appeared in these two: youtu.be/lzLg4lEnQuY?... youtu.be/TN4aMXWpnoI?... and there are others on that Youtube channel.
youtu.be
000
Daniel Van Zant @danielvanzant.bsky.social · 17/03/2025
arxiv.org/abs/2502.203... Maybe LLM's are using symbolic mechanisms built through transformers? Raw transformers are only the "machine code" explanation for how LLM's incredible performance. I really enjoy papers like this that form theories about what the "software" might be.
arxiv.org
Emergent Symbolic Mechanisms Support Abstract Reasoning in Large Language Models
Many recent studies have found evidence for emergent reasoning capabilities in large language models, but debate persists concerning the robustness of these capabilities, and the extent to which they ...
000
Daniel Van Zant @danielvanzant.bsky.social · 12/03/2025
open.substack.com/pub/elanbare... Start of a series my advisor is writing on his theory of cognition. Even though I disagree with some parts of it, this theory is well-argued, well-evidenced, compelling and accessibly written. It deserves serious thought and consideration.
open.substack.com
Memory Isn't Real (Part 1)
The autoregressive theory of cognition changes everything we thought we knew about knowing
000
Daniel Van Zant @danielvanzant.bsky.social · 09/03/2025
Paper comprehensively benchmarking different models and methods for RAG use-cases. If you're wondering "what is the best reranker for X" or "what is the best architecture for Y type of questions" there is probably a table that answers that question. arxiv.org/pdf/2502.202...
arxiv.org
000
Daniel Van Zant @danielvanzant.bsky.social · 09/03/2025
Excellent interview from the leader of a research org that is managing to “think big” without becoming based purely around ideas and do applied work without becoming completely profit-focused. I love seeing stuff like this out there in the world. substack.com/home/post/p-...
substack.com
A Category Theory-Inspired BBN
How the Topos Institute tackles problems in complex systems
000
Daniel Van Zant @danielvanzant.bsky.social · 15/02/2025
youtu.be/miLqkSU_qG4?... An awesome conversation I got to have with Dr. Elan Barenholtz on one of my favorite YouTube channel. We talked about what the structure, successes, and failings of large language models could mean for cognition more widely among other things.
youtu.be
000
Daniel Van Zant @danielvanzant.bsky.social · 29/01/2025
A short piece I just wrote on a really important trade-off in state of the art RAG: substack.com/home/post/p-...
substack.com
Home | Substack
Discover and discuss great writing with the world’s smartest readers on Substack.
000
Daniel Van Zant @danielvanzant.bsky.social · 08/01/2025
7/7 This PDF is a good summary of the concept as a whole that dives a bit deeper into the math and applications once you've wrapped your mind around the previous entries in the reading list. It is also the source of the image: www.researchgate.net/profile/Alfr...
researchgate.net
000
Daniel Van Zant @danielvanzant.bsky.social · 08/01/2025
6/7 This Medium article has a nice summary of the idea as well as implementation instructions: medium.com/@akhtarhaniy...
medium.com
Self-Organizing Maps (SOMs) — Introduced, Explained, and Implemented
What SOMs are and how we can make them for a given dataset in Python
100
Daniel Van Zant @danielvanzant.bsky.social · 08/01/2025
5/7 This short article on them gives a good idea of what a U-matrix looks like and how to interpret it: users.ics.aalto.fi/jhollmen/dip...
users.ics.aalto.fi
U-matrix
U-matrix
100
Daniel Van Zant @danielvanzant.bsky.social · 08/01/2025
4/7 Wikipedia article on self-organizing maps. It is well written and has many visualizations that are helpful for getting an intuition about self-organizing maps (the algorithm that U-matrices are built on): en.wikipedia.org/wiki/Self-or...
en.wikipedia.org
Self-organizing map - Wikipedia
100
Daniel Van Zant @danielvanzant.bsky.social · 08/01/2025
3/7 Finding a good list of resources on these has been challenging so I am posting this here in hopes that someone else can save themselves some frustration in the future. If you know of any additional high quality resources on U-matrices I would love to hear about them. Recommended reading order:
100
Daniel Van Zant @danielvanzant.bsky.social · 08/01/2025
2/7 What makes U-matrices valuable is their ability to partially maintain two key properties from the original data: they show both adjacency information (which points are neighbors) as well as distance information for your set of hyperdimensional points in space.
100
Daniel Van Zant @danielvanzant.bsky.social · 08/01/2025
1/7 Have been trying to understand self-organizing map U-matrices recently. (Example image below of a U-matrix based on hyperdimensional pain data). I am using them to visualize text embedding space.
100
Daniel Van Zant @danielvanzant.bsky.social · 09/10/2024
Diving deep into something really specific and finding a whole world of complexity and useful applications waiting for you at the bottom is one of my favorite things that can happen in science. Dr. Ramakrishnan has done this to great effect.
000
Daniel Van Zant @danielvanzant.bsky.social · 10/09/2024
Have been trying to understand what it means for a system to be Turing-complete (which in theoretical computer science acts as a sort of "check" for programming languages or physical computing systems). For a system to be turing-complete it must be able to compute an arbitrary
000
Daniel Van Zant @danielvanzant.bsky.social · 14/02/2024
Anytime a researcher who is an expert in their field writes a book with the intention of being accessible, I immediately add it my reading list. Chaitin, Ericsson, Penrose, and Schwarzlose come to mind as folks who are examples of this. Any that I should add to my list?
000
Daniel Van Zant @danielvanzant.bsky.social · 13/02/2024
From a fun conversation with a friend who is working on a sci-fi novel on what some neuroscientifically realistic options for something like "mind-control" could look like assuming technology in 20 years. These are some of the cooler papers I found.
010
Daniel Van Zant @danielvanzant.bsky.social · 13/02/2024
Attack vectors of the brain: EEG to get private info from humans: doi.org/10.1007/978-3-319-70972-7_12 Remote control of genetically modified mice with X-rays: doi.org/10.1038/s41467-021-24717-1 Genetically expressible actuator that can affect neural function: doi.org/10.1038/nn.4265 #neuroscience
120