RE: https://dobbs.town/@hobbs/116032781720531564

dear #lazyweb

hit me with your favorite RSS feeds for #homelab #selfhosting #linux #opensource #computing #programming #computerscience #cpu #microarchitecture #electronics #robotics #ai #llm #vlm #mllm #cognitivescience #consciousness #complexity #psychology #jung #philosophy #astronomy #cosmology #physics #chemistry #biology #books #literature #anthropology #jrpg #retrogaming #survival #outdoors #hunting #homesteading #gardening

i need to enrich my feed reader.

UniGen 1.5: el modelo de IA de Apple que puede ver, crear y editar imágenes

#MLLM #ML #AI

https://mecambioamac.com/unigen-1-5-el-modelo-de-ia-de-apple-que-puede-ver-crear-y-editar-imagenes/

Chào mọi người! Công cụ thay thế cục bộ **Super-Bot** vừa được ra mắt: tự viết và chạy mã (Python), tự phục hồi lỗi khi crash, kiểm tra trực quan qua ảnh chụp màn hình. Thử thành công với Ray Tracer và game Snake. Mua một lần thay vì đăng ký hàng tháng! Ý kiến bạn về AI cục bộ vs đám mây?
#AgentStudio #SuperBot #AI #MLLM #ĐạiLýTựĐộng #CôngNghệViệt

https://www.reddit.com/r/LocalLLaMA/comments/1ponuog/agent_studio/

Не все чувства одинаково полезны: как искусственный интеллект объединяет информацию из разных источников

Новое исследование показывает, что мультимодальные модели искусственного интеллекта неравномерно полагаются на визуальные и текстовые данные, что может приводить к ошибкам при обработке противоречивой информации.

https://habr.com/ru/articles/973866/

#kandinsky #gigachat #mllm #сезон_ии_в_разработке

Why AI Sucks At Telling Time... and why this should concern us for autonomous vehicles and more.

#News #TechNews #AI #MLLM #AIlimitations #SelfDriving #MedTech

https://youtu.be/t2Cn0zGRkME

Daily podcast: Why AI Sucks At Telling Time... and why this should concern us for autonomous vehicles and more.

#News #TechNews #AI #MLLM #AIlimitations #SelfDriving #MedTech #podcast

https://soundcloud.com/nickaesp/acr

Kết quả test nhanh cho thấy việc sử dụng Oculink eGPU không gây ảnh hưởng tiêu cực đến hiệu năng khi chạy các dự án LLM cục bộ. Thậm chí, việc kết hợp RTX 3090 qua Oculink với RTX A6000 còn mang lại hiệu suất đáng kể.

#LocalLLaMA #eGPU #Oculink #AI #MLLM #gpu #vietnam
#trítuệnhântạo #máytính

https://www.reddit.com/r/LocalLLaMA/comments/1oqe1kq/no_negative_impact_using_oculink_egpu_a_quick_test/

BDH (Baby Dragon Hatchling) đã được port sang MLX cho Apple Silicon! 🚀 Mã nguồn, tài liệu và script huấn luyện đã sẵn sàng. Model này tương thích với M1/M2/M3. Weights sẽ sớm được upload lên Hugging Face.

#LocalLLaMA #MachineLearning #AI #MLLM #HọcMáy #TríTuệNhânTạo

https://www.reddit.com/r/LocalLLaMA/comments/1o108q5/mlx_port_of_bdh_baby_dragon_hatchling_is_up/

Ước tính chi phí sử dụng token của OpenAI: 1T tokens cho GPT-5 có thể tốn khoảng 3 triệu đô la. Một số lượng lớn người dùng đã tiêu thụ tổng cộng 112,62 triệu đô la tokens, chiếm khoảng 3% doanh thu năm 2024 của OpenAI.
#OpenAI #GPT5 #AI #MLLM # trí_tuệ_nhân_tạo

https://www.reddit.com/r/LocalLLaMA/comments/1o0or4w/how_much_does_1t_tokens_cost_how_much_did_all/

At the @bifold.berlin conference "AI-based methods in the humanities", I have just attended a great talk by Seid Muhie Yimam of Hamburg University who confirmed my impression that there is a kind of momentum in this area at the moment. He mentioned many datasets, publications and shared tasks on African Languages. I will list them (bit by bit) in this thread.

2/x

#HisTag25 #MLLM #LLM #MultilingualDH

In der Sektion über #GlobalHistory from a global perspective geht's gerade um die Begrenzungen von LLMs für "low-resourced" languages. Da tut sich allerdings viel - nicht bei OpenAI, Google, Meta & Co. aber andernorts. Ich suche später noch weitere Links, für den Moment muss es dieser tun:

https://doi.org/10.1038/s43588-025-00865-y

#HistTag25 #MLLM #LLM

OmniEVA: Bridging the 2D–3D Gap in Embodied AI

New paper introduces OmniEVA, a versatile embodied planner that pushes the boundaries of multimodal large language models (MLLMs) for robotics and spatial reasoning.

Results: OmniEVA achieves state-of-the-art performance across 2D/3D reasoning benchmarks and outperforms existing models in object navigation tasks.

Paper: https://arxiv.org/pdf/2509.09332v1
Project: https://omnieva.github.io/

#EmbodiedAI #Robotics #LLM #MLLM #3DVision #AIResearch #AIPlanning

And consider following the authors Jiahui Geng (MBZUAI), Thy Thy Tran (UKP Lab/Technische Universität Darmstadt), Preslav Nakov (MBZUAI), and Iryna Gurevych (UKP Lab & MBZUAI)

See you in Vienna! #ACL2025 !

(4/4)

#MLLM #AISafety #Jailbreak #Multimodal #ConInstruction #ACL2025 #LLMRedTeaming #VisionLanguage #AudioLanguage#NLProc

"OpenAI Furious DeepSeek Might Have Stolen All the Data OpenAI Stole From Us"

Headline of the week. 🥰

OpenAI shocked that an AI company would train on someone else's data without permission or compensation.

404media.co/openai-furious-dee… (no-charge subscription wall for full article)

#OpenAI #DeepSeek
#AI #LLM #MLLM
#GenAI #GenerativeAI

Survey: Multimodal #LLMs like GPT-4V are redefining AI by excelling at tasks like image-based storytelling & OCR-free math reasoning, hinting at AGI potential. This paper reviews their progress, architectures, and challenges while exploring new horizons for research. 🚀 #AI #MLLM

#開源分享 一款開源的多模態影片理解大模型：PPLLaVA，它提供了一個統一的框架，可以同時有效處理長、短片任務
1k上下文，性能優於LLaVA-Next-Video 8k上下文，吞吐率提高了8倍

它通過解決影片冗餘問題，實現了統一處理的能力，可以處理從幾秒到幾小時的各種長度影片

支持影片內容理解和分析、影片場景描述、影片問答交互和多輪對話推理

專案地址： github.com/farewellthree/PPLLaVA

#MLLM #影片理解模型 #PPLLaVA

@peer Dann wäre das #LLM ein Multimodal Large Language Model (#MLLM). Genau das ist der Punkt: Wenn Du in einem Raum sitzt, in dem chinesisches Radio läuft, lernst Du nicht Chinesisch. Ein LLM schon. Es lernt ganz anders als wir. Es lernt nur die Distribution von Sprachteilen. Wir lernen mit Grounding. Das wird in der KI auch kommen, aber jetzt ist es noch nicht so weit und deshalb sind die LLMs noch nicht der Beweis, dass Chomsky falsch lag, aber 1) wussten wir das schon vor den LLMs und 2) machen die LLMs das auch für Laien und hardcore Chomskyaner (die vorher einfach die Literatur nicht gelesen hatten) plausibel.

Дослідники Apple представили Ferret-UI, нову модель штучного інтелекту, яка значно покращує розуміння та взаємодію з мобільними інтерфейсами.

#AI #Apple #MLLM

https://thetransmitted.com/ai/doslidnyky-z-apple-proponuyut-ferret-ui-novu-mllm-shho-znachno-pokrashhuye-rozuminnya-ta-vzayemodiyu-z-mobilnymy-interfejsamy/

Great blog: https://stratechery.com/about/
... interesting article:

Gemini 1.5 and Google’s Nature
https://stratechery.com/2024/gemini-1-5-and-googles-nature

#NLP #LLM
#Gemini (Google DeepMind, successor to Google Bard | see also: GPT-4 (OpenAI)
#MLLM (multi-modal large language models)
#PromptEngineering #RAG (retrieval-augmented generation)

Like, don’t get me wrong, I’d be excited if Apple came out with some LLM or MLLM stuff for iOS18, but I don’t want them to do it just because everyone else is doing it.

To be honest, the only thing I REALLY want is for Siri to not be so….dumb. 🤷🏼‍♂️

#Apple #iOS18 #Siri #LLM #MLLM