AI - AI Dynamics

→ View original post on X — @askalphaxiv

Mamba 3: State Space Models with RoPE for Improved Sequence Modeling

By

@askalphaxiv

–

18 March 2026 4h39

mamba 3: Mamba with RoPE! "Improved Sequence Modeling using State Space Principles" They show that state space models can have both speed and performance! In this new iteration, Mamba now has better recurrence design, complex-valued state tracking, and a MIMO update. It

GPU Availability Planning for Weekend Deployment

By

–

18 March 2026 4h29

maybe we'll do it this weekend, it depends on how many spare gpus we have

Diffusion Model Release: Early Testing Phase for Image Video Generation

By

–

18 March 2026 4h29

pretty much every image and video model is using diffusion, this model release is a early test phase so we can gather feedback. if you have any examples of a moodboard you'd like us to look at the performance of pls send us the code, we're going to play with those more

Freedom Over Profit: Independence in Tech Development

By

–

18 March 2026 4h26

we liked the freedom to set our own direction and not be money focused

→ View original post on X — @garymarcus

Poetry for Our AGI Times: Creative Expression and Advanced AI

By

@garymarcus

–

18 March 2026 4h12

https://
garymarcus.substack.com/p/poetry-for-o
ur-agi-times
…

→ Voir le post original sur X — @thegautamkamath

AI will change everything, few realize its future influence

By

@thegautamkamath

–

18 March 2026 4h06

See https://
journals.plos.org/plosone/articl
e?id=10.1371/journal.pone.0331871
… and https://
x.com/icmlconf/statu
s/2022760600477073876
…. More details coming from @icmlconf tomorrow.

5.2 and 6.1 Model Usage Rates Among Users

By

–

18 March 2026 3h50

tell us about your 5.2 and 6.1 usage! people say this but overall the usage on these models is quite low right now

→ View original post on X — @futurepedia_io

Higgsfield Launches Soul Cinema Cinematic Image Model

By

@futurepedia_io

–

18 March 2026 3h47

Higgsfield just rolled out Soul Cinema, its most cinematic image model yet ✋🧩🤚

After strong feedback from the preview, it’s now fully available; bringing more realistic, film-like visuals.

New Pro+ users can get up to 10,000 free generations to try it out.

Works best when… pic.twitter.com/WSZ2P4aclZ
— Futurepedia – Learn to Leverage AI (@futurepedia_io) 18 mars 2026

Higgsfield just rolled out Soul Cinema, its most cinematic image model yet After strong feedback from the preview, it’s now fully available; bringing more realistic, film-like visuals. New Pro+ users can get up to 10,000 free generations to try it out. Works best when