Hacker News Reader

Comment by light_triad | Hacker News Reader

They are positioning themselves as champions of AI open source mostly because they were blindsided by OpenAI, are not in the infra game, and want to commoditize their complements as much as possible.

This is not altruism although it's still great for devs and startups. All FB GPU investments is primarily for new AI products "friends", recommendations and selling ads.

https://www.joelonsoftware.com/2002/06/12/strategy-letter-v/

View on HN

Meta does a good thing

HN spends a day figuring out how it’s actually bad

It’s not actually bad, OP’s point is that it is not motivated by altruism. An action can be beneficial to the people without that effect being the incentive

Of course, it's not altruism; it's a publicly traded corporation. No one should ever believe in any such claims by these organizations. Non-altruistic organizations can still make positive-impact actions when they align with their goals.

No one said it was bad. It's just self interested (as companies generally are) and are using that to have a PR spin on the topic. But again, this is what all companies do and nothing about it is bad per se.

Nothing they're doing is bad, and sometimes we benefit when large companies interests align with our own. All the spiel about believing in open systems because of being prevented from making their best products by Apple is a bit much considering we're talking about Facebook which is hardly an 'open platform', and the main thing Apple blocked them on was profiling their users to target ads.

So you think FB did this with zero benefit to themselves? They did open source so people could improve their models and eventually have a paid tier later either from hosting services or other strategies

The linked article already spelled out the benefits

By virtue of it being Meta, it's automatically bad.

If we lived in a sensible world we'd have nuked Meta into a trillion tiny little pieces some time around the Cambridge Analytica bullshit.

They've been working on AI for a good bit now. Open source especially is something they've championed since the mid 2010s at least with things like PyTorch, GraphQL, and React. It's not something they've suddenly pivoted to since ChatGPT came in 2022.

They are giving it "for free" because:

* they need LLMs that they can control for features on their platforms (Fb/Instagram, but I can see many use cases on VR too)

* they cannot sell it. They have no cloud services to offer.

So they would spend this money anyways, but to compensate some losses they just decided to use it to fix their PR by contenting developers

They also reap the benefits of AI researchers across the world using Llama as a base. All their research is immediately applicable to their models. It's also likely a strategic decision to reduce the moat OpenAI is building around itself.

I also think LeCunn opposes OpenAI's gatekeeping at a philosophical/political level. He's using his position to strengthen open-source AI. Sure, there's strategic business considerations, but I wouldn't rule out principled motivations too.

Yes LeCun has said he thinks AI should be open like journalism should be - that openness is inherently valuable in such things.

Add to the list of benefits to Meta that it keeps LeCun happy.

I think people massively underestimate how much time/attention span (and ad revenue) will be up for grabs once a platform really nails the "AI friend" concept. And it makes sense for Meta to position themselves for it.

yes ... I remember when online dating was absolutely cringe / weird thing to do. Ten years later and it's the primary way a whole generation seeks a partner.

It will seem incredibly weird today to have an imaginary friend that you treat as a genuine relationship but I genuinely expect this will happen and become a commonplace thing within the next two decades.

> they were blindsided by OpenAI

Given the mountain of GPUs they bought at precisely the right moment I don't think that's entirely accurate

> Given the mountain of GPUs they bought at precisely the right moment I don't think that's entirely accurate

If I remember correctly, FB didnt buy those GPUs because of Open AI, they were going to buy it anyway but Mark said whatever we are buying let's double it.

Yeah, still not entirely clear what exactly they're doing with all of it...but they certainly saw the GPU supply crunch earlier than the rest

AI is not a "complement" of a social network in the way Spolsky defines the term.

> A complement is a product that you usually buy together with another product. Gas and cars are complements. Computer hardware is a classic complement of computer operating systems. And babysitters are a complement of dinner at fine restaurants. In a small town, when the local five star restaurant has a two-for-one Valentine’s day special, the local babysitters double their rates. (Actually, the nine-year-olds get roped into early service.)

> All else being equal, demand for a product increases when the prices of its complements decrease.

Smart phones ar a complement of Instagram. VR headsets are a complement of the metaverse. AI could be a component of a social network, but it's not a complement.

Intentions are overrated. Given how many people with good intentions fuck up everything, I'd rather have actual results, even if the intention is self-serving.

Reply on news.ycombinator.com