I have internal knowledge, I am closely affiliated with Google.
Infrastructure and scalability has been and is key, as well as technical expertise still absolutely super top notch.
Let’s put it this way: Google is the only company that knows how to find, store and utilize information beyond a specific narrowing. And I mean it really in the sense of curating, compression, long time storage, load balancing as well as compliance and world wide redundancy with a focus on speed and efficiency.
Under the hood of AI is pure engineering genius. Google might be trashed as the Search Engine giant that only displays ads now, but reconsider.
Why does all AI provider except for Google have massive problems with load time, reach, etc? Apple chose Google mainly because of the infrastructure. They eat everyone for lunch here. And they earned it.
Engineering at Google etc. are still the finest you can read about software engineering at the highest level. It is highly impressive how Google managed to not fall behind OpenAI. Who else was able to join the race? Microsoft? No. Apple? Oh well… Meta etc. won’t get there ever.
I think that Gemini is 3rd behind OpenAI and Claude but mainly because Google being Google, they kind of have no versioning for their AI and therefore the results are pretty much random in quality, less predictable than the others.
But the creativity and tooling like Nano Banana - fantastic.
There you have it. People don’t get that it is the infrastructure the moment they complain about Claude outtakes here.
Gemini is not reliable whatsoever: https://openrouter.ai/google/gemini-3.1-pro-preview/uptime (the orange chart is either AI studio or Vertex, I suspect AI studio, but it's not good either way).
The reason you don't hear people complaining (esp on HN) is because noone is using Gemini with coding agents. Claude Code, Codex (and IMO OpenCode et al with open weights models) are miles ahead of Gemini CLI/Jules/Antigravity/whatever other coding products Google have.
I tried gemini-cli.
While the model was "ok" everything else was trash.
Constant 429s or 502s for "reasons".
10 different ways to try and pay for the stupid thing and none of them clear.
My favourite was as a paying customer I could not get it to use the latest model. Sometimes it would but most times it would dump me to 2.5.
All of my experience is exactly the opposite of the gp comment is saying.
The gemini-cli repo is gong show too https://github.com/google-gemini/gemini-cli
If you don't pay for it, you don't get much in the way of quota.
Earlier on (okay, until recently), Gemini CLI's quota management didn't work very well.
Antigravity tends to have better quota management behavior.
That is what was infuriating.
It was paid for through code assist enterprise and had all the flags enabled for the "preview" models. Still the only way to get gemini 3+ was to open and close the application 5 to 10 times and sometimes you would get 3 for a bit and then get dumped back to 2.5 and no matter what you do it would not use 3.
I tossed it after spending like 3 hours total messing around the google cloud console and trying a bunch of shit from the github issues. The other offerings don't waste my time (or waste less of it anyway). If they want me to beta test their shit they shouldn't charge for it.
I noticed early on that Gemini responded multiple times faster than claude and chatgpt do, which is why I use it as my main daily LLM (claude code for coding, gemini for all general queries).
Don't you worry, Leadership will get us off this reliable Tech Island and onto the shit-infested Tech Continent that the rest of the world lives on.
This is such an interesting take. Thank you.
Seconding that, not sure if it's the full take I'd stand behind but the perspective is definitely food for thought and way more thought out than my own.