Posts Tagged ‘ChatGPT’

Mark Zuckerberg Has Been A Very Bad Robot Boy

Thursday, July 27th, 2023

Meta, AKA “The Artist Formerly Known As Facebook,” announced that they just lost $21 billion on their Reality Labs division, AKA the Metaverse, AKA the worst virtual reality environment since January 2022.

Meta’s second-quarter earnings showed that Reality Labs, its virtual and augmented reality development business, has lost a staggering $21.3 billion since January 2022 — and executives warned the bleeding will only get worse.

The unit recorded $276 million in Q2 sales this year — down from the $339 million it drew in during Q1, underscoring how VR and AR technology has yet to infiltrate the mainstream.

The losses were wider than analysts expected, though CFO Susan Li suggested in the report that Meta will continue to invest in the tech, which is used to power the metaverse.

“For Reality Labs, we expect operating losses to increase meaningfully year-over-year due to our ongoing product development efforts in augmented reality/virtual reality and investments to further scale our ecosystem,” Li wrote.

Just last month, Meta unveiled its Quest 3 headset for $499, which Mark Zuckerberg touted as “the first mainstream headset with high-res color mixed reality,” though it’s unclear how successful the tech has been so far.

Hint: Not at all.

Just how do you lose $21 billion? That’s a burn rate of over a billion a month. You could hire a mountain of developers and engineers for that money, maybe 100,000 or so of them even at California salary rates. Wikipedia (usual caveats apply) says Occulus only had 17,00 employees in 2022. Meta only paid $2 billion to acquire Occulas (which became Reality Labs) in the first place. Hell, you could fund over 200 startups at $100 million a pop, and it would still be more likely for any one of them to be profitable than Reality Labs.

Usually you have to be a politician to lose that much money. I wonder if Reality Labs losses might be covering up losses in other divisions. Or if the money is getting siphoned off to somewhere else entirely…

Earlier this month, Meta found itself on the defense in a copyright infringement lawsuit filed by stand-up comic Sarah Silverman and authors Christopher Golden and Richard Kadrey, who alleged that Meta’s artificial intelligence-backed language models were trained on illegally-acquired datasets containing the authors’ work.

The suit against Meta points to the allegedly illicit sites used to train LLaMA, the ChatGPT competitor the company launched in February.

Naturally, anything involving large corporations ripping off science fiction writers attracts my attention, and I used to bump into Kadrey back when I was on the SF con circuit. The same firm is also suing on behalf of Paul Trem­blay and Mona Awad.

There probably needs to be some sort of regulation on how much AI generated content can come from any particular living creator. If I feed an AI all of Paul McCarthy’s songs, and ask it to produce a new one based on those, is it copyright infringement?

I suspect a number of lawyers are going to be getting a lot of money off AI in the near future…