AI Assistants: Copilot vs. Claude
Two tools, different strengths, one workflow
One of the unexpected joys of this hobby has been discovering how AI can support the process — not by replacing judgment, but by sharpening it.
I've been using two tools regularly: Microsoft Copilot in Edge and Claude AI. Over time, I've noticed they approach problems differently. Copilot is structured, procedural, and deeply integrated with my browser. Claude is more interpretive, better at narrative, and surprisingly good at iterating on documents and code. Neither is complete on its own, but together they cover a lot of ground.
Here's a comparison based on how I've used them for Bortle 9 imaging.
| Category | Microsoft Copilot in Edge | Claude AI |
|---|---|---|
| Overall Style | Structured, procedural, workflow‑driven. | Interpretive, narrative, insight‑driven. |
| Strength in Bortle 9 Noise Interpretation | Good at diagnosing causes (skyglow, gradients, sensor heat) and prescribing steps. | Excellent at identifying faint structures or gradients; may occasionally overinterpret. |
| Image Annotation | Conservative; accurate when context is provided. | More intuitive; often identifies subtle nebulosity or dust patterns. |
| Workflow Support | Exceptional at checklists, step‑by‑step sequences, and repeatable processes. | Strong at conceptual reasoning and explaining why a workflow matters. |
| Planning (Targets, Filters, Integration) | Great at structured planning tables, seasonal lists, and filter‑based decision trees. | Great at high‑level strategy, multi‑night planning, and filter philosophy. |
| Real‑Time Troubleshooting | Excellent at diagnosing issues (focus, tilt, dew, tracking). | Good conversational partner for interpreting what you're seeing. |
| Post‑Processing Guidance (GraXpert & Siril) | Strong at procedural workflows (background extraction, stacking, color calibration). | Strong at interpreting results and suggesting refinements. |
| Creative Writing (Captions, Blog Posts) | Consistent, structured, SEO‑friendly, metadata‑aware. | Best‑in‑class for poetic, reflective, narrative writing. |
| Visitor Engagement (Website) | Great for alt text, metadata, consistent tone across pages. | Great for storytelling, emotional resonance, and gallery narratives. |
| Browser Integration | Deep integration with open tabs (targets, tutorials, your drafts). | Can fetch URLs and search the web; no live tab context. |
| Filter Strategy Under Bortle 9 | Provides structured recommendations (dual‑band, narrowband, avoid LRGB). | Excellent at explaining nuanced filter behavior and tradeoffs. |
| Interpretation of Faint Detail | Accurate but cautious. | Bold and intuitive; benefits from user verification. |
| Strength with Seestar / Smart Scopes | Good at explaining limitations and comparing results. | Great at interpreting Seestar images and writing educational notes. |
| Strength with ASIAir | Good at explaining settings, autorun sequences, and troubleshooting connections. | Strong at conceptual guidance and interpreting results from imaging sessions. |
| Code and File Generation | Can generate code snippets in conversation. | Can generate, iterate, and deliver downloadable files (HTML, documents). |
| Best Use Cases | Planning, troubleshooting, structured workflows, metadata, consistency. | Interpretation, storytelling, creative writing, iterative editing, file creation. |
| Limitations | Less intuitive with faint detail; less creative without guidance. | No live browser context; can be overconfident in image interpretation. |
This comparison reflects where I am today. As I learn more — and as these tools evolve — the balance may shift. For now, I'm grateful to have two capable assistants helping me make the most of every hour under bright city skies.
Clear skies,
Pete