261–270 of 507 posts
Matt Levine suggested that this was an aquihire. And the weird setup where they only buy non voting shares is to not trigger any regulatory review
Do regulators actually fall for these sort of things in the US? One would expect companies to be judged based on following the spirit of the law, rather than nitpicking and allowing wide holes like this.
Anyone know what scale does these days beyond labeling tools that would make them this interesting to Meta? Data labeling tools seem more of a traditional software application and not much to do with AI models themselves that would be somewhat easily replicated, but guessing my impression is out of date. Also now apparently their CEO is leaving [1], so the idea that they were super impressed with him doesn't seem to…
Scale has also built massive amounts of proprietary datasets that they license to the big players in training. Meta, Google, OpenAI, Anthropic, etc. all use Scale data in training. So, the play I’m guessing is to shut that tap off for everyone else now, and double down on using Scale to generate more proprietary datasets.
Earlier quoted context omitted.
Reverse acquisition? IE, similar to how Disney "bought" Pixar, but much of Pixar's IP overshadows Disney's IP; or how Apple bought Next and the current MacOS is basically NextOS under the hood. It's a technique that companies do to avoid disruption: Buy early stage startups, and by the time they could "disrupt" the parent company, the parent company's management is ready to retire, and the former startup's management…
In what way does Pixar's IP overshadow Disney's? Listing the highest-grossing media franchises [1], Mickey Mouse, Winnie the Pooh, Star Wars, and Disney Princesses are on #2-#5 respectively, while Pixar's top spot is #16 with Cars. [1] https://en.wikipedia.org/wiki/List_of_highest-grossing_media...
Earlier quoted context omitted.
I think this is kind of a philosphical distinction to a lot of people: the assumption is that a computer that can reason like a smart person but still runs at the speed of a computer would appear superintelligent to us. Speed is already the way we distinguish supercomputers from normal ones.
I'd say superintelligence is more about producing deeper insight, making more abstract links across domains, and advancing the frontiers of knowledge than about doing stuff faster. Thinking speed correlates with intelligence to some extent, but at the higher end the distinction between speed and quality becomes clear.
Similarly, "deeper insight" may be surfaced occasionally simply by making a low-intelligence AI 'think' for longer, but this is not something you can count on under any circumstances, which is what you may well expect from something that's claimed to be "super intelligent".
This is a weird deal? Meta buys a non-controlling stake and says no customers will be affected but the CEO and others are leaving Scale for Meta. Meta also says they won’t have access to competitor data but at 49% ownership they get major investor rights? Sounds like an acqui-kill to me?
WCGW with Mark in charge of Ai? Nothing. How could a little "Like" button end up destroying world peace? Absurd.
... What world peace are you speaking of?
I haven't read the article yet (I'm about to) but I'm curious: what are Yann's views on this? He's been pretty vocal that LLMs won't lead to AGI. Edit: read the article. No mention of Yann. What kind of journalists are these people, to not get viewpoints from different angles? They might as well just reproduce press releases and be done with it.
You can believe LLMs won't lead to AGI and still believe that spending billions to have a best in class model will allow you to make products that will recoup that investment.
Reminds me of "Don't dig for the gold, sell the shovels". Could also be read: > Meta spends 10% of last year's revenue to acquire 49% of a top AI data company and poach their leadership, to ensure they are a key player in what could be a ~5-trillion dollar industry by 2033. Meta has a history of this. Acquiring Oculus (and leaning in on VR), Ray-Ban partnership (and leaning in on AR)... etc. These all just seem like…
Bringing up Oculus and VR is quite fitting, because I think it's the same problem. Meta is attempting to find their next business, but like with social media they don't really have a plan. It worked out well with Facebook and some of the purchases surrounding social media, but it was never a clear path to profit, so they slapped ads on it. Why does Meta want VR to work? Create the Meta-verse? We're back at why, what…
It wouldn't surprise me if at least some of that data is being piped back to Meta. Data that can latter be used by LLMs to train on.
Even if this isn't enabled on consumer models, on the corporate side it can make sense. Say you're a risk adjuster for a factory. Walk around with your VR headset. In real time MetaOshaHelper can identify issues, you can tag them yourself.
Then send the video back to your on prem LLM for data processing. New hires get a VR headset which can use this data for help on boarding.
Or... Robots will use the data and replace human workers entirely.
This is a weird deal? Meta buys a non-controlling stake and says no customers will be affected but the CEO and others are leaving Scale for Meta. Meta also says they won’t have access to competitor data but at 49% ownership they get major investor rights? Sounds like an acqui-kill to me?
The host on this podcast[0] had a good point about the "investment". It was really a merger, but framed as an investment to sidestep regulators. Key attributes: * CEO works for meta * almost but not quite a majority stake taken 0: https://podcasts.apple.com/us/podcast/world-bank-cuts-u-s-gr...
Is Wikipedia is correct, Scale AI does not work on AGI at all, but has clearly pivoted to surveillance and military applications like so many others: https://en.wikipedia.org/wiki/Scale_AI A drone army would come in handy for suppressing dissent in the Gulf monarchies or in LA.