Google Rolls Out Updated AI Model Capable of Handling Longer Text, Video

An anonymous reader shares a report: Alphabet's Google is rolling out a new version of its powerful artificial intelligence model that it says can handle larger amounts of text and video than products made by competitors. The updated AI model, called Gemini 1.5 Pro, will be available on Thursday to cloud customers and developers so they can test its new features and eventually create new commercial applications. Google and its rivals have spent billions to ramp up their capabilities in generative AI and are keen to attract corporate clients to show their investments are paying off. [...]

Gemini 1.5 can be trained faster and more efficiently, and has the ability to process a huge amount of information each time it's prompted, according to Vinyals. For example, developers can use Gemini 1.5 Pro to query up to an hour's worth of video, 11 hours of audio or more than 700,000 words in a document, an amount of data that Google says is the "longest context window" of any large-scale AI model yet. Gemini 1.5 can process far more data compared with what the latest AI models from OpenAI and Anthropic can handle, according to Google. In a pre-recorded video demonstration for reporters, Google showed off how engineers asked Gemini 1.5 Pro to ingest a 402-page PDF transcript of the Apollo 11 moon landing, and then prompted it to find quotes that showed "three funny moments."

a better link

[iplayfast]

https://blog.google/technology... [blog.google]

writing on the wall

[groobly]

Writing is on the wall for Google. Which is unfortunate, because don't expect Google maps or YouTube to be free any more, not many of the other freebies like gmail or docs. Especially maps is going to be a shock when you start getting peppered with ads. All that stuff only worked while Google was a monopoly. Kiss it goodbye.

Re:

[Iamthecheese]

> Kiss it goodbye.

Gladly. Google started doing more harm than good back when they changed their motto from "don't be evil" to "MWAAAHAHAHAHAAAAA" I would rather pay for a service than get it free for the cost of my soul.

Re: writing on the wall

[LindleyF]

Google products aren't free because they're feeling generous. They're free because that's the business model. Despite the "you are the product" alarmism, they do rely on user data to improve the quality of their enterprise products. For instance, the more people who use maps, the more accurate maps data can be.

Re:

[Tony Isaac]

I'm pretty sure Google isn't about to fade into the night. Yes, they're playing catch-up, but that's a long game, not a short one. ChatGPT burst upon the scene just over a year ago. That's a blink of an eye in terms of company lifespans, especially companies as big as Google. Google has all kinds of cash to throw at this problem, and all kinds of motivation.

Meanwhile, Google Search, Maps, and YouTube pretty much own their respective markets.

The death of Google has been greatly exaggerated.

1 hour of video? That's a lot of porn!

[iAmWaySmarterThanYou]

Eom.

Sure, but can it...

[PaddirN]

Generate an image of a room with absolutely no elephants in it? This seemingly easy task seems to have stumped most of the chatbots– sorry, Large Language Model AI's.

500 times

[Plumpaquatsch]

You now have to ask it something 500 times before it spills the training data.

Social context of humor

[dsgrntlxmply]

A true AI program would identify the Apollo 11 lander 1202 alarm incidents as hilarious examples of human inadequacy at design and operational verification. Stochastic parrot parlor tricks cannot do that.