Advancements in Waste Management With the integration of vision, sensors, and AI, waste management has become more efficient and precise. ZenRobotics’ systems can sort various types of waste, from bulky construction debris…
Veo 3
Veo 3 is available in Poland. You can access it in the Cloud console and start to type your prompts to get awesome results (probably better than mine so far 😂 )….
News for 2025-07-04
AI Model Comparison Researchers have compared various AI models, including GPT-4o, to evaluate their understanding of images. The study found that while GPT-4o can perform well in multiple tasks, it still lags…
Gemini CLI
I’ve tested briefly Gemini CLI in the free version – I asked it to copy and generate a dashboard from a link, I followed some questions from it, and also gave some…
n8n
Starting with Docker local installation: https://docs.n8n.io/hosting/installation/docker/#prerequisites It works quite fine locally, and it has a lot of integrations Slack, MySQL, Google Calendar or Dropbox to just mention a few. It can also…
AI News
https://generatech.ai – aggregation platform for different AI tools: Text, Image, Video, Audio, Chat and and more! https://academy.openai.com – free registration learning events from OpenAI, different topics AI and ML https://huggingface.co/learn – 9…
Vibe coding
Here I can gather some useful links about vibe coding. In general, I know what it is, but techniques and tools are different for different programmers, and still, it’s not a simple…
Docker Desktop MCP toolkit
It took me some time to localise the menu with developer settings – it’s available from the system menu, not the settings near the profile icon.
Hunyuan 3D 2.1, nVidia PartPacker, Sparc3D
Tencent has released Hunyuan 3D 2.1, an open-source model for generating 3D assets, including PBR materials. This model can synthesize objects with cinematic textures and realistic light interactions. https://huggingface.co/spaces/tencent/Hunyuan3D-2.1https://huggingface.co/tencent/Hunyuan3D-2.1/tree/main nVidia PartPacker: A…
HiDream
I’m testing new model for AI pictures called HiDream for same prompt HiDream (fast, 5-bit) vs Flux: HiDream E1 full takes much longer to render, and consumes above 16GB of VRAM, giving…