To be completely honest, setting aside the raw model performance of Gemini, using Google’s AI ecosystem overall feels like a failure.
From JULE to AI Studio to the Gemini web chat, everything feels fragmented. The pricing is inconsistent, tools are slow, and even basic issues like the AUTO ACCEPT bug in Antigravity have been left unresolved for weeks. Every time I try to decide whether to use Vertex AI or something else, I end up confused because the pricing models are all different.
There is no unified UX, so I genuinely don’t know what I’m supposed to treat as the “standard” way to use Google’s AI.
Even when using Gemini in the web interface, the per-response token limits feel restrictive, which makes the answers incomplete or unsatisfying.
I’m actually a big fan of companies like Google and Microsoft. I’ve learned a lot from Google’s documentation, and I regularly follow their engineering blogs. So I’m not coming from a negative bias—I want Google to succeed.
But regardless of the technology itself, the overall experience is difficult to use because everything feels so fragmented.
I do think building something like a “Claude Code”-style harness is a good direction, but I’m also worried that Google will end up abandoning many of its existing projects along the way.
I subscribe to and actively use multiple AI tools across the board, but honestly, my overall experience with Google’s AI products has been quite disappointing.
To be completely honest, setting aside the raw model performance of Gemini, using Google’s AI ecosystem overall feels like a failure.
From JULE to AI Studio to the Gemini web chat, everything feels fragmented. The pricing is inconsistent, tools are slow, and even basic issues like the AUTO ACCEPT bug in Antigravity have been left unresolved for weeks. Every time I try to decide whether to use Vertex AI or something else, I end up confused because the pricing models are all different.
There is no unified UX, so I genuinely don’t know what I’m supposed to treat as the “standard” way to use Google’s AI.
Even when using Gemini in the web interface, the per-response token limits feel restrictive, which makes the answers incomplete or unsatisfying.
I’m actually a big fan of companies like Google and Microsoft. I’ve learned a lot from Google’s documentation, and I regularly follow their engineering blogs. So I’m not coming from a negative bias—I want Google to succeed.
But regardless of the technology itself, the overall experience is difficult to use because everything feels so fragmented.
I do think building something like a “Claude Code”-style harness is a good direction, but I’m also worried that Google will end up abandoning many of its existing projects along the way.
I subscribe to and actively use multiple AI tools across the board, but honestly, my overall experience with Google’s AI products has been quite disappointing.