posted ago by AlfredicEnglishRules ago by AlfredicEnglishRules +4 / -0

AI: Business

Jeff Bezos and other big names are investing in a humanoid worker bot company. This could be in tech, but all of the claims are based on AI.

https://archive.ph/EfWMs

How AI is helping eco tourism. It's more about tourism in general.

https://archive.ph/E0Kkd

NVidia gives blueprint for data centers of the future

https://archive.ph/VJh9a

Microsoft Copilot will be in regular PCs next year. Intel confirmed it.

https://archive.ph/cpn5i

OpenAI decides to keep Brockman after review

https://archive.ph/BfBgE

Several companies are going after NVidia software to stop their own decline.

https://archive.ph/YTCoU

AI may interview you for a job soon

https://archive.ph/NUq6u

Tumblr is making deals to use art from its website for AI

https://archive.ph/TFRxe

_B: Stability Inc

MidJourney bans Stability AI workers from its servers because they were stealing stuff to improve their own AI

https://archive.ph/ULQ6X

Stability AI has major people leaving it. The article reads like it was paid for by Open AI.

https://archive.ph/vxVuh

The CEO of stability is leaving. He wants the company to go into the black and hates centralized AI

https://archive.ph/NeBxP

More stories on the company not doing well. This is a big AI imaging group, and it looks like others are trying to destroy them.

https://archive.ph/yuErC

A long article about Stability falling apart.

https://archive.ph/snbeV

AI: Papers

Diffusion World Models is an attempt to create a standard for training data on how the world looks

https://archive.ph/UKPlM

ASCII art hacks AI

https://archive.ph/vjICI

How about an AI that insults you and does what it wants? A paper proposes making one

https://archive.ph/6gZdh

Adversarial Attacks on LLMs

https://archive.ph/2BmZ0

A long article explaining how diffusion models work

https://archive.ph/lAQcK

AI teaching other AI is now possible

https://archive.ph/dQaLd

Fine-tuning LLM by pruning layers

https://archive.ph/TxZiL

Emergent Behavior by LLMs is predictable

https://archive.ph/Z0EOB

LoRa Hub is a cross task AI

https://archive.ph/3miMf

New matrix systems maths discovers easier way to analyze in AI

https://archive.ph/F1TCT

Single step creation

https://archive.ph/i7T1l

LLMs are so complex, most people don't know how they are doing what they are doing. There is new research studying it as a natural phenomenon instead of a constructed one

https://archive.ph/gYyVt

Transparent Layers!

https://arxiv.org/html/2402.17113v1

LLMs use linear functions to decide what to use.

https://archive.ph/T8JhV

_P: Fun Names

ZigZag Mamba is a more efficient diffusion model

https://archive.ph/SqlGO

StableDrag let's you move characters with a mouse in AI images to do what you want. The lion will turn it's head, or the building will swivel based on use.

https://archive.ph/0z8PS

SDXS is a Diffusion Model designed to make images in a second

https://archive.ph/GYP2h

SwitchLight has better physics and training for human lighting

https://archive.ph/1YAAg

_P: Chemical Design

Dual Diffusion systems in creating new chemical drugs

https://archive.ph/EoqER

Machine Learning for Beer design

https://archive.ph/ZR0XT

Using ChatGPT for RNA sequence analysis

https://archive.ph/TENwK

AI: 3D

CSM AI lets you create and edit 3D creations using AI. Then you can animate and even render them using the same website. This link is the best I could find on the subject. The announcement is huge, but getting to it is hard.

https://archive.ph/Cp022

Vast Gaussian for large model creation

https://archive.ph/xucJ8

3D point cloud systems and Blender. Could be in general 3D

https://archive.ph/Z6Jx1

A quick intro to NeRF studio

https://archive.ph/jjsjO

Kiri engine turns Gaussian Splats into PLY models

https://archive.ph/0xTjg

RadSplat is an updated efficiency system using Gaussian Splatting.

https://archive.ph/ZGdmW

Stability AI has a new image to 3D system that can be run on your computer. Not mine, it's a potato, but yours would be great.

https://archive.ph/iYlES

A long list of 3D AI projects

https://archive.ph/KA2vT

3D AI has gotten a lot more control for it, including changing the textures.

https://archive.ph/yZhAD

Latte3D makes a character in seconds

https://archive.ph/YRJAF

AI: Video

HUGS Human Gaussian Splatting lets you make a rendered version of yourself or others for a 3D character

https://archive.ph/GW4e8

Haiper is a new video creator.

https://archive.ph/94Wb3

Motion Latent Decomposition

https://archive.ph/27YuN

Stability AI video creation

https://archive.ph/w68AR

TRIP, pretraining images for video

https://archive.ph/0sy99

A long article on the future of AI Video

https://archive.ph/KfeIm

_V: Sora

The good and bad of Sora

https://archive.ph/vBvCW

People can't tell the difference between AI and real videos

https://archive.ph/xl6ZT

Best videos so far from Sora

https://archive.ph/nkibf

More videos from Sora

https://archive.ph/DE1L6

Other companies are working to make better videos. Ethics is questioned in what will happen next

https://archive.ph/8NEje

OpenAI admits it doesn't know how Sora makes videos.

https://archive.ph/H4qjk

Videos made by outside artists using Sora

https://archive.ph/hg1gD

A really long article explaining how Sora and most AI creators work

https://archive.ph/cWk50

How will Sora affect the rendering scene?

https://archive.ph/UsQgw

_S: Outside Testing

Artists are now trying Sora out

https://archive.ph/DESGb

Another article on those artists

https://archive.ph/ef1zs

First impressions from those artists

https://archive.ph/IE1oY

_S: Hollywood

Sora is being shown off to Hollywood types

https://archive.ph/4wDRD

OpenAI is trying to convince Hollywood that Sora will help, not destroy. Just let it in, just a little.

https://archive.ph/SF4RB

AI: Sound

New Google AI creates music by prompt

https://archive.ph/nrpxh

Adobe has a new sound creation program

https://archive.ph/qMIX4