Would love your thoughts on these takes and if I’m missing anything.
My general-use go-to’s of late are Opus4.6 for logic, code, copy, and useful work. ChatGPT is noticeably worse at most things if unmonitored, so I use it for deep research, it’s good at that. But that result gets scrubbed by Opus 4.6 before use. Grok is fantastic for quick images, image edits and current info queries. Again, all critical info is scrubbed by a second model for competent pushback against hallucinations. Gemini is the worst at hallucinations. I barely use it any more unless I’m needing something that I know Google has heavy access to, like web results specifically cited, or their Nano pro 2 image generation via AI Studio API by Leon. Everything changes so much, and so fast, that I know better than to get stuck in a rut.
