Elon Musk’s X.AI is raising a cool $1 billion.
X.AI has filed with the US Securities and Exchange Commission (SEC), signaling its intent to raise as much as $1 billion in equity. According to its SEC paperwork, X.AI has already secured $135 million from four backers - though the filing is doesn’t say who they are.
X.AI launched in July with the not-so-humble mission of uncovering “the nature of the universe.” Since that time, it has rolled out Grok, a snarky ChatGPT competitor. (Hey, you’ve got to start somewhere.)
Perhaps most interestingly, Musk has spoken at some length about how X.AI is being trained on content shared to X. And to that point, Musk has also stated publicly that X investors will own 25% of X.AI - which is probably a good thing as a lot of advertisers have quit X because of a certain F-bomb assault.
Why it matters: It’s Musk - the man who made electric cars cool, put reusable rockets in space, and (for many) trashed the Artist Formerly Known as Twitter. With this in mind, X.AI could end up being the greatest thing ever - or another folly that ultimately hurts Musk’s backers.
Google has announced a new generative AI model that it claims can outperform OpenAI’s ChatGPT across multiple benchmarks.
Dubbed Gemini, it comes in three iterations: Nano - built for on-device tasks, Pro - which is more powerful but not too resource intensive, and Ultra - the most capable, but least efficient.
Google says Gemini Nano will make its way to certain Android applications, while Pro is being folded into Bard. The Don’t be Evil mega-corp says Ultra will arrive sometime in 2024, serving as the backbone of Bard Advanced - a multimodal service, capable of interpreting and producing images, text, audio, and video.
Why it matters: If Google’s Gemini models live up to their promise, Bard Advanced could dislodge ChatGPT’s crown.
Meta, IBM, AMD, and roughly 50 other companies and organizations have teamed up to establish a working group focused on “open innovation and open science in AI.”
Dubbed AI Alliance, the collective says it wants to help developers and researchers “accelerate responsible innovation” in the AI arena. What that means in practice is a little fuzzy, but the AI Alliance promises it will look like “scientific rigor, trust, safety, security, diversity and economic competitiveness.”
Why it matters: Maybe we should be skeptical about a bunch of competitive entities joining hands and promising to do good with a disruptive technology. But then, what if they stick to their word? What might the future of AI development look like?
Calm, the wind-down app, is garnering a batch of headlines for rolling out a new sleep story narrated by Jimmy Stewart of _It’s a Wonderful Life _fame. And if you’re confused by that statement because Stewart died in 1997, then the words “AI Jimmy Stewart” should clear everything up.
Resurrecting dead celebrities for fresh appearances is nothing new of course; we’ve had hologram Tupac Shakur and uncanny valley Peter Cushing in Star Wars: Rogue One. But what’s most interesting about the revival of Stewarts vocal cords is the company who came up with the technology.
Respeecher is a Ukranian startup that specializes in recreating people’s voices. Beyond its Hollywood work, which includes de-aging voices for Star Wars shows, Respeecher has used its AI-powered speech cloning technology to translate songs into other languages and even replicate the voices of laryngectomy patients.
Why it matters: Respeecher’s approach to AI voice emulation demonstrates both the entertaining and practical uses of such technology.