Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Getting a Gemini API key is an exercise in frustration

ankursethi.com

Post reply on HN
Print view View original

221–230 of 361 posts

Re: Getting a Gemini API key is an exercise in frustration

#221

Post by MrOrelliOReilly »

rozap wrote: post #177

I wasted several hours this week going around in the exact same circles. We have a billing account, but kept hitting a gemini quota. Fine. But then on the quota page, every quota said 0% usage. And our bill was like $5. Some docs said check AI studio, but then the "import project from google cloud to AI studio" button kept silently failing. This was a requests per minute quota, which was set at 15 (not a whole lot...…

I have been fighting the same bizarre quota demon. Scripts kept timing out due to quota limitations, but I haven't been able to find any indication of a limit in the console. Finally gave up and switched to Claude, since they at least have a sane interface for API keys and billing!

Re: Getting a Gemini API key is an exercise in frustration

#223

Post by mijoharas »

odie5533 wrote: post #210

The difficult process is on purpose. You're too small. You're just going to waste their customer support resources and only give them maybe a couple hundred dollars. They're hoping you give up and go away.

We're a reasonably sized company. Recently we needed to change our google cloud payment details.

Finance users had changed in the meantime, so I navigate and create an iam user, ok, billing administrator is a thing, great.

Oh, they said it didn't work? alright, there seems to be a project billing administrator as well as an organisation billing administrator? weird, ok let's try that.

Hmm... it still didn't work? let's look around a little more. Ok, within the billing account (that they're a billing administrator to) and within the organisation (that they're a billing administrator to) there is a tab called "payment users". This seems to be _separate_ from their IAM users, and the person needs to be added there (as well as? instead of? who knows) and _then_ they can change the card details.

UX is especially crap here (for google cloud billing).

Let's not even get started on the whole vertex vs. aistudio stuff. Also when one of the gemini's came out their python library worked while their curl docs, and their ruby client didn't so we had to read the source of the python library to figure out what it actually did under the hood to test it out. (this was a while ago, I think they might've gotten better since but the documentation/devex was really bad at at the time)

Re: Getting a Gemini API key is an exercise in frustration

#224

Post by altbdoor »

yawnxyz wrote: post #65
obmelvin wrote: post #43

I don't understand the multiple posts / comments I've seen about this. I google `gemini API key` and the first result* is this docs page: https://ai.google.dev/gemini-api/docs/api-key That docs page has a link in the first primary section on the page. Sure, it could be a huge CTA, but this is a docs page, so it's kinda nice that it's not gone through a marketing make over. * besides sponsored result for AI Studio (Ma…

I've to this day never been able to pay for Gemini through the API, even though I've tried maybe 6-7 times If you bring it up to Logan he'll just brush it off — I honestly don't know if they test these UX flows with their own personal accounts, or if something is buggy with my account.

To Logan's credit though, his team made and drove a lot of good improvements in AI studio and Gemini in general since the early days.

I feel his team is really hitting a wall now in terms of improvements, because it involves Google teams/products outside of their control, or require deep collaboration.

Re: Getting a Gemini API key is an exercise in frustration

#225

Post by downsplat »

hobofan wrote: post #213
CSMastermind wrote: post #147

Earlier quoted context omitted.

Some other fun things you'll find: - The models perform differently when called via the API vs in the Gemini UI. - The Gemini API will randomly fail about 1% of the time, retry logic is basically mandatory. - API performance is heavily influenced by the whims of the Google we've observed spreads between 30 seconds and 4 minutes for the same query depending on how Google is feeling that day.

> The Gemini API will randomly fail about 1% of the time, retry logic is basically mandatory. That is sadly true across the board for AI inference API providers. OpenAI and Anthropic API stability usually suffers around launch events. Azure OpenAI/Foundry serving regularly has 500 errors for certain time periods. For any production feature with high uptime guarantees I would right now strongly advise for picking a mo…

Yeah at $WORK we use various LLM APIs to analyze text; it's not heavy usage in terms of tokens but maybe 10K calls per day. We've found that response times vary a lot, sometimes going over a minute for simple tasks, and random fails happen. Retry logic is definitely mandatory, and it's good to have multiple providers ready. We're abstracting calls across three different APIs (openai, gemini and mistral, btw we're getting pretty good results with mistral!) so we can switch workloads quickly if needed.

Re: Getting a Gemini API key is an exercise in frustration

#226

Post by halflings »

CSMastermind wrote: post #147
Ozzie_osman wrote: post #145

I was recently (vibe)-coding some games with my kid, and we wanted some basic text-to-speech functionality. We tested Google's Gemini models in-browser, and they worked great, so we figured we'd add them to the app. Some fun learnings: 1. You can access those models via three APIs: the Gemini API (which it turns out is only for prototyping and returned errors 30% of the time), the Vertex API (much more stable but lac…

Some other fun things you'll find: - The models perform differently when called via the API vs in the Gemini UI. - The Gemini API will randomly fail about 1% of the time, retry logic is basically mandatory. - API performance is heavily influenced by the whims of the Google we've observed spreads between 30 seconds and 4 minutes for the same query depending on how Google is feeling that day.

"The models perform differently when called via the API vs in the Gemini UI."

This shouldn't be surprised, e.g. the model != the product. The same way GPT4o behaves differently than the ChatGPT product when using GPT4o.

Re: Getting a Gemini API key is an exercise in frustration

#227

Post by edwinjm »

Mashimo wrote: post #218
the_snooze wrote: post #95

Earlier quoted context omitted.

Parking apps don’t seem to care much for that. They know you’ll jump through their shoddy UIs and data collection because they have a local monopoly. Often with physical payment kiosks removed and replaced with “download our shitty app!” notices.

Huh, where I live you often can use many different parking apps, and the one i tried is very simple and user friendly. Start app, wait for gps, turn time wheel, press start.

Turn time wheel? How do you know in advance how long you stay? Where I live, you start and when you leave, you click stop. You also get reminders in case you forgot to stop.

Re: Getting a Gemini API key is an exercise in frustration

#229

Post by kunley »

antonvs wrote: post #143
kunley wrote: post #126

> I was fifteen minutes into writing some code by hand like a Neanderthal Tell me this isn't classism. Tell me this kind of narrative isn't a new norm

I can’t tell if you might be joking. If you are, the rest of this comment is irrelevant. That general idiom is old and fairly widely used. There was a Seinfeld episode in 1997 in which Elaine talked about “…not lurching around like a caveman.” If you’re objecting to criticism of writing code by hand, the phrase is almost invariably used in a self-deprecating way, acknowledging some inefficiency or old-fashioned behav…

Thank you. Seinfeld reference does it.

About the "hard way", everything in the article is screaming that OP's new way of coding is the hard way, not the other way around

Re: Getting a Gemini API key is an exercise in frustration

#230

Post by sylware »

You should be able to create an API account with a classic noscript/basic (x)html browser (optionally with an email, and that could be an IPv[46] literal email address, you know, self-hosted & DNS free, which is stronger than SPF...).

Then to pay for, I should be able to redeem a code bought at my local and physical monetary terminals (no credit card info input on an internet able computer, even if elf/linux and lean classic noscript/basic (x)html web browsers) that to add credits on my account. Like steam. In my country, we even have codes for age verification only (you have physical age verification like when you buy alcohol from a [bottle] shop), much easier to crack down on abuse.

Another thing could be a public "anonymous", severely rate limited, API key for 'testing purpose' or very rare usage, or a noscript/basic (x)html web site (namely a real and honest web site) with ads (text/image/videos[])... and with solid handling of HTTP refresh?

My main usage for AI would be coding. I am craving at mass porting C++ to a plain and simple subset of C code (it seems some people are getting reasonably good results, and it seems rust has a brain damaged syntax on the scale of c++), and assembly coding with very specialized code snippet.

Post reply on HN