Eigenhector
@eigenhector.bsky.social
Machine learning, meditation, magic, muscles, mathematics, metta, meta, memes, manifolds.
Researcher in Machine Learning and Contemplative arts.
https://scholar.google.com/citations?user=KjOV76MAAAAJ&hl=en
Recsys book author https://amzn.to/3ANPx6e
Researcher in Machine Learning and Contemplative arts.
https://scholar.google.com/citations?user=KjOV76MAAAAJ&hl=en
Recsys book author https://amzn.to/3ANPx6e
Pinned
Eigenhector
@eigenhector.bsky.social
· Nov 19
Hello new friends! My day job is a research lead in machine learning on population dynamics, public health and climate. However I am also doing a grad degree in contemplative arts so I might sometimes write about machine learning and then some random mystic practice from an ancient text FYI.
Nice to be in robotics while everyone else is going crazy over language models
July 2, 2025 at 8:17 PM
Nice to be in robotics while everyone else is going crazy over language models
Are there use cases of GenAI people like that isn't generic AI slop? I've found short stories narrated on YouTube good if the content I written by humans and if the voiceover is good but bad if the stories are AI generated.
January 31, 2025 at 6:54 PM
Are there use cases of GenAI people like that isn't generic AI slop? I've found short stories narrated on YouTube good if the content I written by humans and if the voiceover is good but bad if the stories are AI generated.
Yeah category error. The traditions I study don't consider intelligence to be the ineffable mystery but consciousness or sunyata or nirguna Brahman. Intelligence in traditions like yoga are hardly mentioned maybe under name and form and considered the outer layers of the mind I believe.
i mean, it doesn't _have_ to be an ineffable mystery to the human mind. but it's just as vibes-driven to assert that it simply _can't_ be.
consider also the possibility that it's a category error to expect a nebulous term like intelligence to map in any sensible way onto "fundamental principles"
consider also the possibility that it's a category error to expect a nebulous term like intelligence to map in any sensible way onto "fundamental principles"
January 5, 2025 at 4:16 PM
Yeah category error. The traditions I study don't consider intelligence to be the ineffable mystery but consciousness or sunyata or nirguna Brahman. Intelligence in traditions like yoga are hardly mentioned maybe under name and form and considered the outer layers of the mind I believe.
Huh I didn't realize paprika is chili that got adapted by Hungarians. It's amazing how the peoples of the world adapt plants to their local environment.
January 2, 2025 at 10:16 PM
Huh I didn't realize paprika is chili that got adapted by Hungarians. It's amazing how the peoples of the world adapt plants to their local environment.
Netflix's Pantheon seems to borrow a lot of themes from non dual traditions interesting story
January 2, 2025 at 5:26 AM
Netflix's Pantheon seems to borrow a lot of themes from non dual traditions interesting story
How does feed ranking work here? Without algorithmic ranking does that mean one has to have a separate account for each topic?
December 24, 2024 at 7:44 PM
How does feed ranking work here? Without algorithmic ranking does that mean one has to have a separate account for each topic?
Does this mean 95% are now reviewer number 2s 🤣
Reviewing is dead.
I just read some ICLR reviews. Absolutely junk. No insight. Written by humans, but could have been written by a chatbot or teenager who had read a few dozen reviews.
Why do we think this is useful? Only 5% (if that!) of the community has worthwhile opinions.
I just read some ICLR reviews. Absolutely junk. No insight. Written by humans, but could have been written by a chatbot or teenager who had read a few dozen reviews.
Why do we think this is useful? Only 5% (if that!) of the community has worthwhile opinions.
November 23, 2024 at 7:26 PM
Does this mean 95% are now reviewer number 2s 🤣
Reposted by Eigenhector
Why are some LLMs better at chess than others
Part 1: dynomight.net/chess/
Part 2: dynomight.net/more-chess/
Part 1: dynomight.net/chess/
Part 2: dynomight.net/more-chess/
Something weird is happening with LLMs and chess
are they good or bad?
dynomight.net
November 22, 2024 at 4:16 PM
Why are some LLMs better at chess than others
Part 1: dynomight.net/chess/
Part 2: dynomight.net/more-chess/
Part 1: dynomight.net/chess/
Part 2: dynomight.net/more-chess/
Reposted by Eigenhector
We've released our paper "Generating 3D-Consistent Videos from Unposed Internet Photos"! Video models like Luma generate pretty videos, but sometimes struggle with 3D consistency. We can do better by scaling them with 3D-aware objectives. 1/N
page: genechou.com/kfcw
page: genechou.com/kfcw
November 21, 2024 at 4:19 PM
We've released our paper "Generating 3D-Consistent Videos from Unposed Internet Photos"! Video models like Luma generate pretty videos, but sometimes struggle with 3D consistency. We can do better by scaling them with 3D-aware objectives. 1/N
page: genechou.com/kfcw
page: genechou.com/kfcw
Chaos Monkey (Netflix) but for government instead of compute
November 22, 2024 at 5:43 AM
Chaos Monkey (Netflix) but for government instead of compute
Reposted by Eigenhector
Our upcoming SIGGRAPH Asia paper explores refining 3D shapes using text descriptions. For example, we transform the spheres and cylinders on the left into a detailed teddy bear! 🧸
Read on for more details in the thread! 👇
Read on for more details in the thread! 👇
November 22, 2024 at 5:15 AM
Our upcoming SIGGRAPH Asia paper explores refining 3D shapes using text descriptions. For example, we transform the spheres and cylinders on the left into a detailed teddy bear! 🧸
Read on for more details in the thread! 👇
Read on for more details in the thread! 👇
I have sort of a related goal of exploring inner worlds using these models. For example the fire kasina phenomenology of the Theravada described in the Visudhimagga can be explored with a model of convolutional neural networks instead of biology. It's all a kind of psycho cosmology.
1/ I work in #NeuroAI, a growing field of research, which many people have only the haziest conception of...
As way of introduction to this research approach, I'll provide here a very short thread outlining the definition of the field I gave recently at our BRAIN NeuroAI workshop at the NIH.
🧠📈
As way of introduction to this research approach, I'll provide here a very short thread outlining the definition of the field I gave recently at our BRAIN NeuroAI workshop at the NIH.
🧠📈
November 21, 2024 at 6:38 PM
I have sort of a related goal of exploring inner worlds using these models. For example the fire kasina phenomenology of the Theravada described in the Visudhimagga can be explored with a model of convolutional neural networks instead of biology. It's all a kind of psycho cosmology.
I like not skating where the puck is heading, be in strategic spaces is another alternative so whoever has the puck can redirect it to you when the time comes.
With regard to AI
• stay open minded
• skate where the puck is heading
• don’t neglect your task-specific evals
• stay open minded
• skate where the puck is heading
• don’t neglect your task-specific evals
November 21, 2024 at 6:22 PM
I like not skating where the puck is heading, be in strategic spaces is another alternative so whoever has the puck can redirect it to you when the time comes.
Fascinating, Egyptian mythology combined with ideas of void. Reminds me of Josephine McCarthy 's concept of the abyss with all kinds of ancient beings in it.
[2/2] I’d let the comic speak for itself, but I should clarify: the giant snake is Apep, the glowing deity is Atum -- both primordial beings of Ancient Egypt w/ opposing views regarding their world.
[ #Art #Artwork #OCs #OriginalCharacters #Comics // #DEITIESProject #Atum #Apep #EgyptianMythology ]
[ #Art #Artwork #OCs #OriginalCharacters #Comics // #DEITIESProject #Atum #Apep #EgyptianMythology ]
![[Pages continued from earlier comment thread.]
PAGE 5 of 8
---
Atum continues with his admonishing remarks to Apep, while also unfurling his golden griffin wings and increasing the golden light from his magic ward.
Atum: "[...] Especially coming from an abhorrent serpent... who could not care less about the well-being of my kin."
"I will not 'humor' you with answer for the sake of your ego trip. My thoughts would be of no use to you..."
"And you cannot threaten me here. Now, begone."
Apep hisses in protest at Atum's retort, before simmering down.
Apep: "...Very well. But heed these words..."](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:x4ryjfhjxpi24yhvupfb3its/bafkreidyi2gxvtri44btykxg4mdtd3y2x6hmsveeuslr4q65hgyrntmsa4@jpeg)
November 21, 2024 at 4:20 PM
Fascinating, Egyptian mythology combined with ideas of void. Reminds me of Josephine McCarthy 's concept of the abyss with all kinds of ancient beings in it.
Hello new data science friends please take a look at the GitHub for example notebooks on how to use the PDFM embeddings github.com/google-resea...
GitHub - google-research/population-dynamics: PDFM Embeddings: location-based vectors for geo-spatial analysis.
PDFM Embeddings: location-based vectors for geo-spatial analysis. - google-research/population-dynamics
github.com
November 20, 2024 at 11:18 PM
Hello new data science friends please take a look at the GitHub for example notebooks on how to use the PDFM embeddings github.com/google-resea...
Reposted by Eigenhector
I'm up 1100+ followers on BlueSky in less than 3 weeks and down 150 followers on X posting the exact same content.
If you are working in data:
- BlueSky is the place to connect with other developers
- LinkedIn is where to reach industry contacts
- X is just not where your audience is any longer
If you are working in data:
- BlueSky is the place to connect with other developers
- LinkedIn is where to reach industry contacts
- X is just not where your audience is any longer
November 20, 2024 at 11:10 PM
I'm up 1100+ followers on BlueSky in less than 3 weeks and down 150 followers on X posting the exact same content.
If you are working in data:
- BlueSky is the place to connect with other developers
- LinkedIn is where to reach industry contacts
- X is just not where your audience is any longer
If you are working in data:
- BlueSky is the place to connect with other developers
- LinkedIn is where to reach industry contacts
- X is just not where your audience is any longer
Reposted by Eigenhector
The grand plan to regress the brain, which according to @tyrellturing.bsky.social, is working just fine.
So is this what understanding the brain means? The better the ANN fits the brain data the more we understand the brain?
So is this what understanding the brain means? The better the ANN fits the brain data the more we understand the brain?
2) TBH, I'm sceptical of the claim that we need some kind of grand paradigm shift in the mathematics. I actually think we're making excellent progress, and have a "progressive research program" à la Lakotos working well, as articulated by Adrien and Tim in this paper:
www.nature.com/articles/s41...
www.nature.com/articles/s41...
The neuroconnectionist research programme - Nature Reviews Neuroscience
Artificial neural networks are being widely used to model behavioural and neural data. In this Perspective article, Doerig et al. present neuroconnectionism as a Lakatosian research programme using ar...
www.nature.com
November 19, 2024 at 4:15 PM
The grand plan to regress the brain, which according to @tyrellturing.bsky.social, is working just fine.
So is this what understanding the brain means? The better the ANN fits the brain data the more we understand the brain?
So is this what understanding the brain means? The better the ANN fits the brain data the more we understand the brain?
Lol
The plot thickens!!!! This is like the real housewives of the United States congress.
November 20, 2024 at 10:35 PM
Lol
Reposted by Eigenhector
I'm super excited to share that our work on AlphaQubit, a high accuracy neural network quantum error correction decoder, is now published in Nature! www.nature.com/articles/d41...
Quantum computing: physics–AI collaboration quashes quantum errors
Machine-learning strategy for quantum error correction.
www.nature.com
November 20, 2024 at 4:44 PM
I'm super excited to share that our work on AlphaQubit, a high accuracy neural network quantum error correction decoder, is now published in Nature! www.nature.com/articles/d41...
Homotopy type theory and law crafting are weirdly similar to exploring non dual mystical states via self and other duality. It's like the inside is fractally similar to the outside as claimed by some traditions due to one's consciousness experiencing both.
Interestingly enough, mathematicians are still trying¹ to figure out exactly what “equal” means. But one thing constructive mathematicians agree on: “equal” means “not apart” and also “not separate”.
1. See, e.g., homotopy type theory and the “univalence axiom”.
1. See, e.g., homotopy type theory and the “univalence axiom”.
When asked what equal means in this context, the bill sponsor could not even define the word equal anymore.
November 20, 2024 at 4:43 PM
Homotopy type theory and law crafting are weirdly similar to exploring non dual mystical states via self and other duality. It's like the inside is fractally similar to the outside as claimed by some traditions due to one's consciousness experiencing both.
Huh logged on and Windows was pushing for me to use an NPU which I thought was funny and followed up and TIL that Windows has ARM chips! Snapdragon X elite. Amazing! Last time I was running Windows on a non x86 arch was probably Dec Alpha or Intel Itanium.
November 20, 2024 at 4:13 PM
Huh logged on and Windows was pushing for me to use an NPU which I thought was funny and followed up and TIL that Windows has ARM chips! Snapdragon X elite. Amazing! Last time I was running Windows on a non x86 arch was probably Dec Alpha or Intel Itanium.
Reposted by Eigenhector
Curious to learn what research is looking like at Meta these days. Seems that most other places are shutting down research. I’d like to think that this means we will be stagnate once LLMs hit any nontrivial wall. Thoughts welcome on this too.
November 19, 2024 at 11:25 PM
Curious to learn what research is looking like at Meta these days. Seems that most other places are shutting down research. I’d like to think that this means we will be stagnate once LLMs hit any nontrivial wall. Thoughts welcome on this too.
Reposted by Eigenhector
The thingie that brings over your twitter followers worked jolly well for me. Very cool! I am following another 500 people now thanks to that…
chromewebstore.google.com/detail/sky-f...
chromewebstore.google.com/detail/sky-f...
Sky Follower Bridge - Chrome Web Store
Instantly find and follow the same users from your Twitter follows on Bluesky.
chromewebstore.google.com
November 19, 2024 at 10:28 AM
The thingie that brings over your twitter followers worked jolly well for me. Very cool! I am following another 500 people now thanks to that…
chromewebstore.google.com/detail/sky-f...
chromewebstore.google.com/detail/sky-f...
Money is an amazing invention but it seems to distort incentives in very strange and unexpected ways.
November 19, 2024 at 7:23 PM
Money is an amazing invention but it seems to distort incentives in very strange and unexpected ways.