Meta rolls out an updated AI assistant, built with the long-awaited Llama 3


Meta Meta AI, built using the long-awaited open source Llama 3 large language model (LLM) for AI assistant platform. The company says it’s “the smartest AI assistant you can use for free right now.” As for usage scenarios, the company offers the ability to help users study for tests, plan dinner, and plan nights out. You know how to dig. This is an AI chatbot.

Meta AI, however, has expanded to almost every corner of the company’s entire portfolio after one trial. . It is still available on Instagram, but now users can access it on Messenger, Facebook feeds and Whatsapp. The chatbot also has its own dedicated web portal, wait for it, . You don’t need a company login to use it this way, although it won’t generate images. Recently released also integrate with the bot soon with Quest headset integration.

On the topic of image generation, Meta says it’s now faster and will produce images as you type. It also handles custom animated GIFs, which is pretty cool. We hope it can successfully create images of people of different races. We found it a few weeks ago, because it seemed biased towards creating images of people of the same race, even when asked otherwise.

Meta is also expanding its global reach with this update, as Meta AI is coming to more than a dozen countries outside of the US. These include Australia, Canada, Ghana, Jamaica, Pakistan, Uganda, etc. However, there is one major caveat. It’s only in English, which doesn’t seem very useful for a global audience, but whatever.

As for security and reliability, the company says that Llama 3 is trained on an expanded data set compared to Llama 2. It has also used synthetic data to create long files for training, and claims that it excludes all data sources that contain large amounts of data. “high volumes of personal information about individuals.” Meta says it conducted a series of evaluations to see how the chatbot would handle areas of risk, such as conversations about weapons, cyberattacks and child abuse, and make adjustments as required. In our short test with the product, we already experienced hallucinations, as can be seen below.

The meta AI makes a mistake in the recipe. The meta AI makes a mistake in the recipe.

Engadget/Carissa Bell

Artificial intelligence has become one of the CEOs of Meta together with in a secluded Hawaiian compound, but the company is still reaching out to OpenAI and, to a lesser extent, Google. Meta’s Llama 2 never impressed users due to its limited feature set, so maybe this new version of the AI ​​assistant will be lightning in a bottle. At the very least, it should be able to capture lightning in a bottle, or rather, slightly scratch the image of someone else’s lightning in a bottle.



Source link

Leave a Reply

Your email address will not be published. Required fields are marked *