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

Google Gemini Pro API Available Through AI Studio

ai.google.dev

Post reply on HN
Print view View original

91–100 of 118 posts

Re: Google Gemini Pro API Available Through AI Studio

#91

Post by matsemann »

chamoda wrote: post #3

Free quota looks reasonable with 60 queries per minute. On the other hand data from free quota requests will be used to improve the product. https://ai.google.dev/pricing

Typical Google behavior, where it's cheap/free in the beginning, and when you've built on top of it prices increases drastically? Like gmaps pricing.

Re: Google Gemini Pro API Available Through AI Studio

#93

Post by miguelramos »

brrrrrm wrote: post #44

why on earth did they design the Node.js and Web APIs to be slightly different and incompatible? (edit: this might just be a bug/oversight on the landing page?) Node.js: const model = genAI.getGenerativeModel({ model: "gemini-pro-vision"}); const result = model.generateContent({ contents: [{parts: [ {text: "What’s in this photo?"}, {inlineData: {data: imgBase64, mimeType: 'image/png'}} ] }] }) Web: const model = genA…

Hey! This is Miguel from Google working of these SDKS. I'm confused about this comment, Both Web and Node.js are the same. Can you clarify where you see the difference?

Re: Google Gemini Pro API Available Through AI Studio

#94

Post by tobyjsullivan »

summerlight wrote: post #85
chamoda wrote: post #3

Free quota looks reasonable with 60 queries per minute. On the other hand data from free quota requests will be used to improve the product. https://ai.google.dev/pricing

This must be a significant investment to pick up the hype; 1 qps cannot be sustainable with a free pricing tier unless their resource efficiency is 10x better than competitors.

No doubt it’s a loss leader to some degree but, in practice, very, very few customers will have sustained request rates of 60 requests per minute. Their actual usage averaged across all users will be a tiny fraction of that.

For users that get close to that sustained rate, they’re just as likely to exceed it and actually pay.

Re: Google Gemini Pro API Available Through AI Studio

#95

Post by brrrrrm »

miguelramos wrote: post #93
brrrrrm wrote: post #44

why on earth did they design the Node.js and Web APIs to be slightly different and incompatible? (edit: this might just be a bug/oversight on the landing page?) Node.js: const model = genAI.getGenerativeModel({ model: "gemini-pro-vision"}); const result = model.generateContent({ contents: [{parts: [ {text: "What’s in this photo?"}, {inlineData: {data: imgBase64, mimeType: 'image/png'}} ] }] }) Web: const model = genA…

Hey! This is Miguel from Google working of these SDKS. I'm confused about this comment, Both Web and Node.js are the same. Can you clarify where you see the difference?

as documented it looks different (async vs sync + the necessity of `text:` annotations in the Node.js version?)

updated my comment to paste in what's written in the docs

Re: Google Gemini Pro API Available Through AI Studio

#96

Post by sorokod »

MrCzar wrote: post #20
georgehill wrote: post #9

Earlier quoted context omitted.

Okay, from my end, I am still seeing Palm (EU).

Same here, Google's offering is very weird though, a simple question to Bard about what it's using also outputs the wrong result. https://imgur.com/Vb1STQe

Ah, but can you tell which part of the result is wrong? Maybe a follow up question to the LLM?

Re: Google Gemini Pro API Available Through AI Studio

#97

Post by miguelramos »

brrrrrm wrote: post #95
miguelramos wrote: post #93

Earlier quoted context omitted.

Hey! This is Miguel from Google working of these SDKS. I'm confused about this comment, Both Web and Node.js are the same. Can you clarify where you see the difference?

as documented it looks different (async vs sync + the necessity of `text:` annotations in the Node.js version?) updated my comment to paste in what's written in the docs

Thanks, that helps to understand the confusion.

The web and Node.js are the same, each of them has many function overloads so users can call functions with simplified or complex arguments, as they prefer. We are going to fix the doc and code snippets so both Web and Node.js consistently show the same code to avoid misunderstandings. Thanks a lot!

Re: Google Gemini Pro API Available Through AI Studio

#98

Post by bobvanluijt »

payzero wrote: post #26
bobvanluijt wrote: post #14

[flagged]

Is there a reason for this? Any advantages, disadvantages, etc.?

Modules are optional, so if you want to create an end-to-end RAG pipeline, you can do this from the database directly. It's basically less abstraction based on how the user wants to feed data into the prompt of the model.

Re: Google Gemini Pro API Available Through AI Studio

#99

Post by dopb »

gregsadetsky wrote: post #88

I used this just-released API (of Gemini Pro) with multimodal input to test some of the things from the infamous Gemini Demo. You can see here [ https://www.youtube.com/watch?v=__nL7Vc0OCg ] my GPT-4 recreation of that ad which went viral. Gemini Pro is... not great. In one test, I asked what gesture I was making (while showing a thumbs up) -- it said thumbs down and "The image is a commentary on the changing nature…

I think the fair comparison would be GPT3.5 (if image inputs were supported) vs Gemini Pro. It would be great to compare this with Gemini Ultra next year.
Post reply on HN