video1.0<iframe src="https://www.loom.com/embed/d4c28aab8fe6431c8f7dcf20587bc332" frameborder="0" width="1910" height="1432" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>14321910Loomhttps://www.loom.com14321910https://cdn.loom.com/sessions/thumbnails/d4c28aab8fe6431c8f7dcf20587bc332-bc69a2079a79f849.gif331.205This Loom discusses troubleshooting a Pickaxe community node workflow for generating images with image 2.0 and controlling “thinking” or intelligence settings. The author says one run successfully generated an image using a simple prompt, but subsequent attempts failed or produced blank results, especially after increasing intelligence to max where it appeared to get stuck thinking without generating an image or message. In later tests, using a different prompt format improved the outcome and produced an image while presumably sending the full transcript. The author seeks insight into why higher reasoning prevented image generation and how to make image thinking more detailed, noting one failed run cost nine cents.