Slice Posted 23 hours ago Share Posted 23 hours ago Tell us your story what you asked from AI. Tell us non-standard approach to the dialog. Ask a question how to use AI. Tell us what AI do you prefer and why. Show images created with AI. Tell us funny story related to AI. 3 Link to comment https://www.insanelymac.com/forum/topic/362922-welcome-to-discuss-artificial-intelligence/ Share on other sites More sharing options...
Cyberdevs Posted 16 hours ago Share Posted 16 hours ago For me Claude is the go to source, it really does the job for me. I've tested ChatGPT, DeepSeek, Gemini and Grok and Claude by far is the best one for my needs. I've tested some local AIs but never get the results I was looking for. Besides my GPU (RX6800 XT) seems to be not that suitable for large projects. On macOS it's nearly impossible to do anything heavy but in Windows I used Zluda and Stable Diffusion Forge for AMD and it generates images with 720x1080 between 11 to 21 seconds based on the Loras and the Models used. For Text generation I've tested Ollama and LM Studio which can utilize AMD GPUs if you have ROCm installed and the result is much better than macOS. Stability Matrix has an application for Windows and macOS (Apple Silicon only) which can download and install various Generative AI tools and it works pretty nice for Image and Text Generation. I've recently discovered another app called LynxHub which works on Intel Macs as well and has plenty of repos to install but haven't used it much. 2 Link to comment https://www.insanelymac.com/forum/topic/362922-welcome-to-discuss-artificial-intelligence/#findComment-2851342 Share on other sites More sharing options...
Irish_Man Posted 15 hours ago Share Posted 15 hours ago I have to agree with @Cyberdevs. I've tried several AI's but Claude Code is the best for my needs. It made my life much easier when I was setting up my server. It even fixed few issues that I had with my Hackintosh. 2 Link to comment https://www.insanelymac.com/forum/topic/362922-welcome-to-discuss-artificial-intelligence/#findComment-2851343 Share on other sites More sharing options...
MacNB Posted 13 hours ago Share Posted 13 hours ago This is funny... I have been using an App for processing analog video (https://github.com/StuartCameronCode/VapourBox) The chap who created it is called Stuart Cameron. On his GitHub repo, the other developer is someone called Claude I thought...hmmm Claude does not say much on this repo... It's only when I clicked on the Claude's name I realised it's an AI tool by Anthropic 🤦♂️ I have just been initiated into AI and who the hell is Claude 🤣 1 Link to comment https://www.insanelymac.com/forum/topic/362922-welcome-to-discuss-artificial-intelligence/#findComment-2851345 Share on other sites More sharing options...
Max.1974 Posted 13 hours ago Share Posted 13 hours ago I use VS Code with Chat Copilot and several AI tools, including Claude 4.6. But sometimes GPT-5.4 surprises me by doing many things better than Claude. 1 Link to comment https://www.insanelymac.com/forum/topic/362922-welcome-to-discuss-artificial-intelligence/#findComment-2851346 Share on other sites More sharing options...
Cyberdevs Posted 12 hours ago Share Posted 12 hours ago I would like to hear any success stories on using local LLMs for coding, if anyone has any experience please share your stories. I've used Gemma, Qwen, Omnicoder but none of them could do what Claude, Gemini, Grok and GPT can do. 1 Link to comment https://www.insanelymac.com/forum/topic/362922-welcome-to-discuss-artificial-intelligence/#findComment-2851347 Share on other sites More sharing options...
Oyecomova Posted 8 hours ago Share Posted 8 hours ago 8 hours ago, Cyberdevs said: Para mí, Claude es la mejor opción; realmente cumple su función. He probado ChatGPT, DeepSeek, Gemini y Grok, y Claude es, sin duda, la mejor para mis necesidades. He probado algunas IA locales, pero nunca he obtenido los resultados que buscaba. Además, mi GPU (RX6800 XT) no parece ser muy adecuada para proyectos grandes. En macOS es casi imposible hacer algo pesado, pero en Windows usé Zluda y Stable Diffusion Forge para AMD y genera imágenes de 720x1080 entre 11 y 21 segundos, dependiendo de los procesadores Lora y los modelos utilizados. Para la generación de texto, he probado Ollama y LM Studio, que pueden utilizar las GPU de AMD si tienes ROCm instalado, y el resultado es mucho mejor que en macOS. Stability Matrix cuenta con una aplicación para Windows y macOS (solo para Apple Silicon) que permite descargar e instalar diversas herramientas de IA generativa y funciona bastante bien para la generación de imágenes y texto. Recientemente descubrí otra aplicación llamada LynxHub que también funciona en Macs con procesador Intel y tiene muchos repositorios para instalar, pero no la he usado mucho. My modest RX 580 performs very well for processing batches of up to 10 GB of images and preparing them in CMYK TIFF format for offset printing... in minutes... with AI "Aiarty Image Enhancer"... on macOS Tahoe. Link to comment https://www.insanelymac.com/forum/topic/362922-welcome-to-discuss-artificial-intelligence/#findComment-2851348 Share on other sites More sharing options...
Cyberdevs Posted 1 hour ago Share Posted 1 hour ago 6 hours ago, Oyecomova said: My modest RX 580 performs very well for processing batches of up to 10 GB of images and preparing them in CMYK TIFF format for offset printing... in minutes... with AI "Aiarty Image Enhancer"... on macOS Tahoe. That's nice to know that the RX 580 is still up to the tasks for certain jobs, I haven't used my RX 580 in any serious AI related processing. Link to comment https://www.insanelymac.com/forum/topic/362922-welcome-to-discuss-artificial-intelligence/#findComment-2851356 Share on other sites More sharing options...
Recommended Posts