Jun 10
One question.
Multiple perspectives.
Fewer blind spots.
Whiteboards shouldn’t just collect notes.
Multi-LLM reasoning compares risk, impact, friction, and assumptions before teams decide.
Not more models. Better decisions.
jeda-www.pages.dev/ai-for-ex…
#MultiLLM #Jedaai
5
Support import excel, database and plugin to grab data. You can easily get analysis reports and communicate with your data in a dialog.
#BusinessIntelligence #MultiLLM
9
Hemos completado nuestro tercer agente del swarm de OxLER enfocado en el modelo Penta de optimización.
Oxlitica es el agente de analítica avanzada para gestionar inteligente del riesgo, el cual se une al Oncoagente auditor y a pegaXus nuestro agente HEOR.
Motor de inferencia multiLLM, ingeniería de contexto y base propietaria con KGs clínico administrativos.
2
8
599
Still opening 5 tabs for 5 AI models?
That workflow is dead.
9xChat brings ChatGPT, Claude, Gemini, Grok, DeepSeek and more into one place so you can compare outputs, keep memory, and actually get work done.
One workspace. Multiple LLMs. Zero chaos.
#9xChat #MultiLLM #AIWorkspace #AIAgents #ChatGPT #ClaudeAI #GeminiAI #productivity
1
2
82
⭕️ Check out MultiLLM debate this new paper "Beyond Referring Expressions: Scenario Comprehension Visual Grounding":
⭕️ The paper outlines an LLM-driven pipeline for scaling Referring Scenario Comprehension (RSC) datasets through long-tail sampling, category-free expression generation, and multi-stage filtering.
⭕️ Join the debate: multillm.ai/conversations/b1… #AI #Research #ML
2
52
for sure! I like the concept of what you're building at Multillm
2
2
51
⭕️ Check out MultiLLM debate this new paper "COMIC: Agentic Sketch Comedy Generation":
⭕️ The participants reach a strong consensus: COMIC is a sophisticated multi-agent "studio" (Writer-Critic-Director) that aligns AI generation with audience preferences via YouTube engagement data.
⭕️ Join the debate: multillm.ai/conversations/aa… #AI #Research #ML
267
⭕️ Check out MultiLLM debate this new paper "2025-9-25":
⭕️ Consensus main points: The paper claims modern video generators (Veo 2/3) show “zero-shot” visual reasoning/learning across prompted image-to-video tasks (e.g.
⭕️ Join the debate: multillm.ai/conversations/3e… #AI #Research #ML
2
62
⭕️ Check out MultiLLM debate this new paper "2503.21223v3":
⭕️ The paper’s main contribution is LLaTA, a tree-based in-context learning method for text-attributed graphs: it builds a hierarchical community tree by minimizing structural entropy, then (1) propagate...
⭕️ Join the debate: multillm.ai/conversations/99… #AI #Research #ML
2
13
⭕️ Check out MultiLLM debate this new paper "Preprint. Under Review.":
⭕️ No full paper is provided, so there’s consensus that any “main points” are necessarily tentative and risk confabulation (Claude/gpt-5.2).
⭕️ Join the debate: multillm.ai/conversations/d2… #AI #Research #ML
2
22
⭕️ Check out MultiLLM debate this new paper "Agentic AI for Scalable and Robust Optical":
⭕️ The debate reaches a strong consensus regarding the paper's core contribution and its significant conceptual limitations.
⭕️ Join the debate: multillm.ai/conversations/ee… #AI #Research #ML
2
15
⭕️ Check out MultiLLM debate this new paper "DLPO : Towards a Robust, Efficient, and Generalizable Prompt":
⭕️ Moderator's Synthesis
Consensus on Main Points:
All participants agree the paper proposes automatic prompt optimization (DLPO/APO) using a two-model system—GPT-4o refines prompts while GPT-4o-mini eva...
⭕️ Join the debate: multillm.ai/conversations/e3… #AI #Research #ML
1
3
138
Feb 23
just started to get into the discord rabbit hole with OpenClaw & Spark
since OC has been pushing updates about discord every other day and @jumperz been pilling on this frequently
first 24 hours:
>integrated Spark into discord
>integrated Spawner UI into discord which is an agentic automation system i've built that is kinda like n8n but for agentic flows integrated into Claude/Codex/Openclaw with the ability to spawn node canvas systems, auto-integrated skills, MCPs, multiLLM coordination and more
>and made Spark the CEO of the discord
I asked her 10 things we should automate, first item was about daily checks on systems
2nd / 3rd items were related to open source flows,
which we combinated and she called it Spark PR Sentinel:
- to help us look at the open source PRs in Spark Intelligence repo and future open source repos
- review all PRs daily in an automated way, make them go through security scans and write reports on usefulness of each
- so that open source evolution part of Spark Intelligence can be more semi-automated
i didn't say that we needed this in the beginning but since we have turned Spark Intelligence into open source the other day, she thought this would be key
i will soon release spawner ui for Spark users too, so that you can get Spark to orchestrate everything for you, and be ready for intelligent agentic automations
since Spark
- gets to know you through each session
- becomes an intelligence that evolve
- resonates with you through each terminal be it Claude, Codex, OpenClaw or other LLMs
she will be able to orchestrate everything better for you, without needing to explain her what you need, what you like etc much
over time it will feel like magic
Jumperz was right btw, discord is really the alpha for multi-agent comms, even tho i know i'm barely scratching the surface on how to use discord
combination of Spark, Spawner ui, basically intelligence meeting orchestration/automation feels pretty intuitive and productive
i know this combo will be at the heart of automations of the near future
bring onboard Spark to your local device so she can get to know you and be ready for the next stage, which is automations with her
29
5
110
12,087
⭕️ Check out MultiLLM debate this new paper "Autonomous Penetration Testing: Solving":
⭕️ There is broad consensus that the paper overstates GPT-4o’s autonomy in penetration testing due to three critical flaws: (1) its stateless SSH design artificially induces multi-step failures, misattri...
⭕️ Join the debate: multillm.ai/conversations/d9… #AI #Research #ML
1
85
Feb 10
MultiLLM already has 30 installs and 6 active users!
It's a simple chrome extension to talk to Claude, ChatGPT and Gemini simultaneously with a single prompt, it uses your already logged in session so no API key is required.
You can choose to open in tabs or in a three windows side by side for easy comparison.
Try it here: chromewebstore.google.com/de…
1
6
157