Has anyone switched to Gemini CLI? It's so important but also exhausting keeping up with which model is the leading edge. Especially since every model has different idiosyncrasies you have to learn to work with it effectively.
Currently my ranking is
* Cursor composer: impressively fast and able but not tuned to be that agentic, so it's better for one-shot code changes than long-running tasks. Fantastic UI.
* Claude Code: Works great if you can set up a verifiable environment, a clear plan and set it loose to build something for an hour
* Grok: Similar to cursor composer but slower and more agentic. Not currently using.
In the end, the model only ran `git bisect` (if we're to believe the video at least) for various pointless reasons, it isn't being used for what it's usually used for. Why did it run bisect at all? Well, the user asked the LLM to use `git bisect` to find a specific commit, but that doesn't make sense, `git bisect` is not for that, so what the user is asking for, isn't possible.
Instead of the model stopping and saying "Hey, that's not the right idea, did you mean ... ?" so to ensure it's actually possible and what the user wants, the model runs its own race and start invoking a bunch of other git commands, because that's how you'd find that commit the user is looking for, and then finally does some git bisecting stuff just for fun, it had already found the right commit.
I think I see the same thing when letting LLMs code as well. If you give them some work to do that is actually impossible, but the words kind of make sense, and it'll produce something but not what you wanted, I think they're doing exactly the same thing, bypassing what you clearly instructed so they at least do something.
I'm not sure if I'm just hallucinating that they're acting like that, but LLMs doing "the wrong thing" has been hitting me more than once, and imagining something more dangerous than `do a git bisect`, it seems to me like that video is telling us Gemini 3 Pro will act exactly the same way, no improvements on that front.
Also, do these blog posts not go through review from engineering before they're published? Besides the video not really showcasing anything of interest, the prompt itself doesn't make any sense and would have been caught if a engineer who uses git at least weekly reviewed it before.
It seems like you don't have access to Gemini 3.
Learn more at https://goo.gle/enable-preview-features
To disable Gemini 3, disable "Preview features" in /settings.
• 1. Switch to gemini-2.5-pro
• 2. Stop
Note: You can always use /model to select a different option.
Google never disappoints with their half-ass-launches.
> For Google AI Ultra subscribers and paid Gemini and Vertex API key holders, Gemini 3 Pro is already available and ready to enable. For everyone else, we're gradually expanding access through a waitlist.
So, not available yet. Tried with a free API key and I did not have access. I do have access on a paid API key, but I'd rather not use that with an agent. The rate limits page the docs link to currently has no info on gemini-3-pro rate limits. Seems to me this is really only for users on the $200/mo subscription plan. Somewhat odd, given the model is already GA in every other coding agent as I understand
15 comments
[ 2.6 ms ] story [ 42.7 ms ] threadCurrently my ranking is
* Cursor composer: impressively fast and able but not tuned to be that agentic, so it's better for one-shot code changes than long-running tasks. Fantastic UI.
* Claude Code: Works great if you can set up a verifiable environment, a clear plan and set it loose to build something for an hour
* Grok: Similar to cursor composer but slower and more agentic. Not currently using.
* ChatGPT Codex, Gemini: Haven't tried yet.
Going to download Gemini CLI right now™ and see how it performs™ against Cursor, Claude Code, Aider, OpenCode, Droid, Warp, Devin, and ForgeCode.
In the end, the model only ran `git bisect` (if we're to believe the video at least) for various pointless reasons, it isn't being used for what it's usually used for. Why did it run bisect at all? Well, the user asked the LLM to use `git bisect` to find a specific commit, but that doesn't make sense, `git bisect` is not for that, so what the user is asking for, isn't possible.
Instead of the model stopping and saying "Hey, that's not the right idea, did you mean ... ?" so to ensure it's actually possible and what the user wants, the model runs its own race and start invoking a bunch of other git commands, because that's how you'd find that commit the user is looking for, and then finally does some git bisecting stuff just for fun, it had already found the right commit.
I think I see the same thing when letting LLMs code as well. If you give them some work to do that is actually impossible, but the words kind of make sense, and it'll produce something but not what you wanted, I think they're doing exactly the same thing, bypassing what you clearly instructed so they at least do something.
I'm not sure if I'm just hallucinating that they're acting like that, but LLMs doing "the wrong thing" has been hitting me more than once, and imagining something more dangerous than `do a git bisect`, it seems to me like that video is telling us Gemini 3 Pro will act exactly the same way, no improvements on that front.
Also, do these blog posts not go through review from engineering before they're published? Besides the video not really showcasing anything of interest, the prompt itself doesn't make any sense and would have been caught if a engineer who uses git at least weekly reviewed it before.
> gemini
It seems like you don't have access to Gemini 3. Learn more at https://goo.gle/enable-preview-features To disable Gemini 3, disable "Preview features" in /settings. • 1. Switch to gemini-2.5-pro • 2. Stop Note: You can always use /model to select a different option.
Google never disappoints with their half-ass-launches.
Gemini 3 Pro Preview Live in AI Studio - https://news.ycombinator.com/item?id=45967211 - Nov 2025 (385 comments)
Gemini 3 Pro Model Card - https://news.ycombinator.com/item?id=45963670 - Nov 2025 (263 comments)
> For Google AI Ultra subscribers and paid Gemini and Vertex API key holders, Gemini 3 Pro is already available and ready to enable. For everyone else, we're gradually expanding access through a waitlist.
So, not available yet. Tried with a free API key and I did not have access. I do have access on a paid API key, but I'd rather not use that with an agent. The rate limits page the docs link to currently has no info on gemini-3-pro rate limits. Seems to me this is really only for users on the $200/mo subscription plan. Somewhat odd, given the model is already GA in every other coding agent as I understand