← sources

Is it me who is thinking Claude actually the closest thing we have to AGI, or are we just coping?

source: reddit (reddit:ClaudeAI)published 2026-10-24 06:05ingested 2026-10-24 10:56

retained text

Claude feels like the closest thing we have to a genuine intellectual collaborator. Its nuance, contextual reasoning, and coding capability make GPT models look robotic and mechanical at times.

That said, calling it **AGI** is a massive stretch:

1. **The Good:** Unmatched writing tone, phenomenal long-context coherence, and top-tier code architecture reasoning.
2. **The Bad:** Zero autonomous agency, persistent hallucinations under complex edge cases, no persistent memory across sessions, and aggressive refusal guardrails. It’s an elite reasoning tool, but it's still fundamentally a reactive next-token predictor, not an autonomous mind.

Where do you think it actually lands on the AGI scale? Has anyone built a fully autonomous workflow with it that didn't eventually derail?

document fetched 2026-10-25 00:22 · canonical https://reddit.com/r/ClaudeAI/comments/1xe0xb6/is_it_me_who_is_thinking_claude_actually_the