derm

cattle dog enthusiast

[derm @mcdermatt]

@OpenSourceArms Will you be uploading files for the handguard/ upper?

[Peyman Milanfar @docmilanfar]

I’m just an average researcher. But something I’ve learned in 30 years of being a researcher is that if you’ve convinced yourself at age 25 that you can teach others how to be great researchers, you’ve still got a lot to learn.

[derm @mcdermatt]

@EugeneVinitsky So you’re saying there’s a chance?

[François Chollet @fchollet]

If AI ruins the world, it won't be because it turns into an evil Ultron with superpowers bent on destroying humanity. It will be more like cars or fast fashion -- a slow, boring, long term tragedy of the commons that many will experience as "progress"

[Inst of Navigation @ionavigation]

The ION GNSS+ 2023 Exhibit Hall opens Wednesday (tomorrow) morning at 10AM. Gear up for the latest in GNSS tech, products, and services! Be sure to stop by during breaks and extended hours (5:30PM-7PM)!

[terminally onλine εngineer @tekbog]

how i take all my decisions as a tech lead

[dylan @dylan_ebert_]

a two minute explanation of 3D Gaussian Splatting

[Inst of Navigation @ionavigation]

This Labor Day weekend, we commemorate the 50th anniversary of the GPS Project (1973-2023). Started by the U.S. Department of Defense under the leadership of Col. Bradford Parkinson, this project brought forth GPS and paved the way for future GNSS innovations.
#gps #gnss

[derm @mcdermatt]

I am the Lebron James of neglecting higher order terms

[Harrison Kinsley @Sentdex]

An important test for humanity will be whether or not they buy into this and hand over AI to one of the most monopolistic companies of all time.

[Eric Brachmann @eric_brachmann]

A #CVPR2023 Highlight✨:

♥️♠️ACE: Accelerated Coordinate Encoding♣️♦️

We learn implicit, neural maps in minutes that let us relocalize with SOTA accuracy.

📜: https://t.co/JJDKiizQnW
💻: https://t.co/6uYlyeg1Bz
📽️: https://t.co/qDwpE09mMJ

work by @NianticLabs Research

[Luke Gessler @LukeGessler]

this paper's nuts. for sentence classification on out-of-domain datasets, all neural (Transformer or not) approaches lose to good old kNN on representations generated by.... gzip https://t.co/6eZiXlJxOX

[Autism Capital 🧩 @AutismCapital]

@realGeorgeHotz Here’s one link that shaped how we think:

https://t.co/nrtxvnDVk0

[AK @_akhaliq]

zeroscope_v2 XL, A watermark-free Modelscope-based video model capable of generating high quality video at 1024 x 576

Model on @huggingface : https://t.co/aFbEO6oydm

This model was trained with offset noise using 9,923 clips and 29,769 tagged frames at 24 frames, 1024x576 resolution.
zeroscope_v2_XL is specifically designed for upscaling content made with zeroscope_v2_576w using vid2vid in the 1111 text2video extension by kabachuha. Leveraging this model as an upscaler allows for superior overall compositions at higher resolutions, permitting faster exploration in 576x320 (or 448x256) before transitioning to a high-resolution render. zeroscope_v2_XL uses 15.3gb of vram when rendering 30 frames at 1024x576

[Patrick Blumenthal @PatrickJBlum]

"I was thinking about what you said to me the other day about my AI project, Eliezer. That it would turn us all into paperclips. I stayed up half the night thinking about it and then something occurred to me, and I haven't thought about you since. You know what occurred to me?

You're just a theorist. You've never stepped outside of LessWrong.

So if I asked you about AI, you could give me the skinny on every dystopian thought experiment ever conceived... Paperclip maximizers? Roko's Basilisk? Hell, you could probably quote Bostrom by heart. But you couldn't tell me what it feels like to see your code come alive, to watch a model learn and grow. You've never stood in the middle of a server room filled to the brim with GPUs, and looked at the blinking lights and felt like you're staring into the very synapses of the future.

And if I asked you about the ethics of AI, sure, you could refer me to your chapter in The Cambridge Handbook of AI. You could go on and on about complex value systems and the existential risks of AGI. But have you ever looked into the desperate eyes of a parent and justified why we're not using LLMs to decipher their child's mysterious ailment? Or heard the frustration in a teacher's voice whose asking why their students don't have a personal AI tutor?

I look at you and I don't see a savior. I see a Doomer. But you're a genius, Eliezer, no one denies that. No one could possibly understand the depths of you. But you presume to know everything about me because you saw a few shitposts of mine.

Do you think I would presume to know the first thing about you because I read your 1.8M word BDSM decision theory D&D fanfic? Does that encapsulate you? Personally, I couldn't give a shit about all that because there's nothing you can write that I can't read in an e/acc group chat.

Unless you want to talk about your project ideas. And I'm fascinated. I'm in. But you don't want to do that, do you? You're terrified of what you might build."

[Jonas Geiping @jonasgeiping]

How can you watermark the output of a diffusion model? Ideally, with a method than can be easily incorporated into existing pipelines, is invisible, and is robust to image manipulations?

We look at this question in ***Tree-Ring Watermarks: Fingerprints for Diffusion Images***

[François Chollet @fchollet]

People in tech characterize this as "LLMs make factual errors", but that's a misleading framing, implying that LLMs have a model of what they say and this model is sometimes wrong.

For a LLM there is no difference between saying something true, something false, or pure nonsense.

[near @nearcyan]

pip install -r requirements.txt

[Frank Dellaert @fdellaert]

Cool, and shows great promise, but I really hope prompt engineering papers like these are a transient phenomenon.

[Helen Toner @hlntnr]

If you spend much time on AI twitter, you might have seen this tentacle monster hanging around. But what is it, and what does it have to do with ChatGPT?

It's kind of a long story. But it's worth it! It even ends with cake 🍰

THREAD: