@shengtang135754 I use agy as a worker for codex.
Sol as the main lead,
Luna for delegations and tasks when agy quotas are out,
Sonnet 5.5 as a senior worker and owner,
Opus 5.5 as a consultant/signoff with Veto over everything,
Gemini 3.8 (high) for everything.
@shakhzod1001 Sol for leading,
Astra as a consultant/signoff,
Luna for almost everything else.
Let Sol be the owner and accountable for: 1. Quality. 2. Quota saving.
@younesbag1 Micro-Task Offloading.
و Multi-Model Red Teaming. تجربة الـ Pipeline على نماذج بمعماريات مختلفة تكشف لك الـ Blind Spots، والـ Hallucination وثغرات الفهم اللي ممكن تكون مخفية في انحيازات Opus المعتادة.
وطبعا السرعة وعدم تعقيد الأمور over-engineering
@bridgemindai Cancelled my subscription. Moving to a better usage limits.
If Anthropic followed, Chinese AI companies will be my sole providers.
Currently I have ChatGPT, CLAUDE, and GEMINI subscriptions. But no more US monopoly for me just to keep some balance.
@synthwavedd I have 7 days in my Claude subscription plan.
Every month I switch between Claude and ChatGPT. With Gemini as an reliable worker.
But GLM has my attention.
@baran__robin@claudeai Tested the new Claude Fable model. Might still lean toward Opus 4.8 , at least it follows prompts accurately. Anyone else seeing instruction drift with Fable?