Flux
PulseAugur coverage of Flux — every cluster mentioning Flux across labs, papers, and developer communities, ranked by signal.
- developed by Black Forest Labs 100%
- instance of Diffusion Models 90%
- instance of alphaXiv 90%
- used by Martin Scorsese 90%
- developed Diffusion Transformers 90%
- uses Martin Scorsese 90%
- used by Wan2.1-14B 90%
- used by Stable Diffusion 3 70%
- used by Diffusion Models 70%
- used by Diffusion Transformers 70%
- developed Diffusion Models 70%
- used by alphaXiv 70%
- 2026-10-04 product_launch DoorDash's Flux cloud platform is now handling 130,000 engineering tasks monthly, automating code reviews and expanding into CI triage and maintenance workflows. source
- 2026-09-06 product_launch DoorDash moved engineering agents to its Flux cloud platform, which automates engineering tasks and code reviews. source
- 2026-06-13 regulatory Adafruit Industries filed a lawsuit against Flux AI alleging violations of the Computer Fraud and Abuse Act. source
- 2026-06-03 product_launch Martin Scorsese has partnered with Black Forest Labs to promote their AI storyboard-generating app, FLUX. source
24 day(s) with sentiment data
-
New DiRotQ method enhances 4-bit quantization for Diffusion Transformers
Researchers have developed DiRotQ, a new 4-bit quantization framework for Diffusion Transformers (DiTs) that significantly reduces memory and computational costs while preserving image generation quality. The method uti…
-
Adafruit sues AI board designer Flux for false advertising
Adafruit Industries has accused Flux, a company specializing in AI board design, of false advertising. Adafruit claims Flux misrepresented its user base, stating that Flux falsely advertised having 1.1 million users. Th…
-
Open-source tools integrate ComfyUI with Nuke and After Effects
A team at KENT Post Production has developed and open-sourced a suite of tools designed to integrate ComfyUI into professional VFX pipelines. These tools include a "Director's Console" for managing ComfyUI workflows and…
-
Kubernetes Engineers Need Advanced Skills Beyond Container Orchestration
The landscape for Kubernetes engineers is evolving beyond basic container orchestration, with a focus shifting towards more advanced skills. Engineers will need to master areas such as cloud-native security, observabili…
-
Home lab user deploys Cofer on Kubernetes with AI agents
An individual is preparing to deploy a system called Cofer on Kubernetes using GitOps principles with Flux. This deployment involves AI and LLM agents that are validating manifests for the Cofer system, which is being s…
-
FluxC: New Zero-Shot Codec Leverages Flux for Image Compression
Researchers have introduced FluxC, a novel zero-shot codec that leverages Flux as a compression prior. This method analyzes coding cost, bit allocation, and reconstruction techniques to optimize rate-distortion-percepti…
-
Ohm Raumzeit and Gaia-Studio release "Enzyklika 2026 - Book 1"
Ohm Raumzeit and Gaia-Studio have released "Enzyklika 2026 - Book 1 - A silhouette of Truth," a full-length instrumental album. The music is described as Symphonic Speed Metal, with orchestral and cinematic influences. …
-
Text-to-image models struggle with object counts beyond ten
A new arXiv paper investigates how text-to-image models handle number words, particularly when requesting more than ten objects. The study found that while models like SD3.5-Medium and FLUX.1-dev struggle to accurately …
-
User seeks help with Flux Klein 9b model modifications
A user on Reddit is seeking assistance with creating "refmods" for the Flux Klein 9b model. They are encountering difficulties using the Hugging Face CLI for this process and are looking for guidance from the community.
-
TailSteer framework improves multi-turn image editing robustness
Researchers have developed TailSteer, a novel training framework designed to enhance the robustness of multi-turn image editing in foundation models. This method focuses on preserving the integrity of early denoising st…
-
New EmoReward system enhances emotional expression in AI-generated videos
Researchers have developed EmoReward, a novel self-evolving training mechanism designed to improve emotional expression in generated videos. This system adaptively calibrates reward weights by distilling evidence reliab…
-
MagoStudio open-sources AI video generation tool Mago
MagoStudio has released Mago, an open-source tool designed for creating long-form AI-generated videos. The tool, which can be run on a local GPU or via cloud models, was used to produce an episode of "Yokai" from a scri…
-
User seeks advice on training consistent AI identity with FLUX LoRA
A user is seeking advice on training a LoRA (Low-Rank Adaptation) model for FLUX.1-dev to create a consistent AI version of themselves. They are encountering issues with maintaining facial structure and identity across …
-
Mago Studio releases open-source AI production suite with local-first capabilities
Mago Studio has released an open-source, local-first AI production suite that offers end-to-end pipelines for image and video generation. The tool includes features for script-to-film conversion, project management, and…
-
FLUX.1-dev AI model achieves 64% speedup on RTX 4090 via hardware tuning
A developer has benchmarked the FLUX.1-dev generative AI model on an RTX 4090 GPU under Windows, achieving a 64.2% increase in inference throughput. This optimization was accomplished by reconfiguring the GPU's hardware…
-
AI art generated with Flux/Psychoseed pipeline on Strix Halo hardware
A user named greyqueen shared artwork generated using a two-stage AI pipeline. The process involved "Flux dev1-q4/psychoseed" and took approximately 34 seconds per image. The setup utilized "2x Strix Halo AI 9" processo…
-
User seeks workflow for 90s cel-anime motion without AI distortion
A user on Reddit is seeking advice on creating a hybrid pipeline for generating 90s cel-animated style content, inspired by the anime Jin-Roh. They have successfully implemented background parallax and a cel-shader effe…
-
INTRA sparse attention boosts transformer efficiency with novel relay-communication
Researchers have introduced INTRA, a novel sparse attention mechanism designed to improve the efficiency of transformer models. INTRA employs an interleaved, non-contiguous token approach, viewing attention as a relay-c…
-
New ReSeed method restores diversity in distilled diffusion models
Researchers have developed a new method called ReSeed to improve the diversity of images generated by distilled diffusion models. They found that the first denoising step is crucial for maintaining diversity, as distill…
-
New distillation methods enhance LLM reasoning and image generation
Researchers are developing advanced on-policy distillation (OPD) techniques to improve language model performance, particularly in areas like mathematical reasoning and text-to-image generation. New methods such as Stud…