Thunderful Games’ LEGO Bricktales is coming to Meta Quest on December 7th, priced at $29.99 USD. Not a simple side-scroller, this title offers the most intuitive brick-by-brick building in a LEGO video game yet, as players can see their creations come to life in a three-dimensional world.
Starting at $299 USD, the new Ray-Ban Meta smartglasses build upon its predecessor and adds an ultra-wide 12 MP camera for improved photos as well as 1080p videos up to 60 seconds. Plus, you can directly share photos from your glasses with a simple “send a photo” voice command.
You’ve seen AudioCraft, now check out Meta’s new SeamlessM4T. This cutting edge multilingual and multitask model is capable of seamlessly translating as well as transcribing speech and text for nearly 100 languages. A text encoder, based on the NLLB model, has been trained to understand text in nearly 100 languages and produce representations that are useful for translation.
MusicGen is fun to play around with, but the all-new Meta AudioCraft takes generative AI to the next level. This tool lets you easily generate high-quality audio as well as music from text using three models: MusicGen, AudioGen and EnCodec. AudioGen was basically trained on public sound effects and can generate audio from text prompts.
Meta Avatar calls are set to launch and they can be used for real-time calling when using Messenger as well as Instagram when you’re not camera-ready. A new feature on Facebook and WhatsApp will soon let you take a live selfie and generate an avatar in seconds. You’ll then be able to customize the avatar to best represent yourself.
Threads by Instagram images first leaked back in May, but now, it’s officially launching in the US on July 6th to take on rival Twitter. The app is described as a way for communities come together to discuss everything from the topics you care about today to what will be trending tomorrow, and most importantly for some, your Instagram username can be kept the same without having to register again.
Cornell’s EchoSpeech uses AI and sonar to read silent speech while Meta Voicebox is a generative AI model capable of replicating voices from scratch in six languages. This can be used to give natural-sounding voices to virtual assistants or non-player-characters in the metaverse and allow visually impaired people to hear written messages from friends read by AI in their voices.
Photo credit: Mashable / Meta
Elon Musk has agreed to fight Mark Zuckerberg in a cage match. This rivalry all started after Musk posted a Tweet that he would be up for a cage match if Zuckerberg is. The Meta CEO replied with a screenshot of the Tweet and the caption ‘Send Me Location’.
Think of Meta’s MusicGen AI as ChatGPT for tunes rather than imagery. This single stage, auto-regressive Transformer model was trained over a 32kHz EnCodec tokenizer with 4 codebooks sampled at 50 Hz. Unlike Google’s MusicLM, MusicGen doesn’t require a self-supervised semantic representation, thus it’s capable of generating all four codebooks in one pass.