Former meta prick, sorry employee:
> Meta's strategy
Meta has no strategy. The people that made llama work left two years ago. MSI are a bunch of charlatans who are trying to change the culutre and tooling of meta from the outside (I can see their point, trying to work outside of the defaults is incredibly hard.)
its basically a bun fight, with all of your productive workers being poached by frontier labs, and everyone else is either a new grad, burnt out, or a master bullshitter.
I disagree. The strategy was unclear until now. Spending billions on acquisitions and targeted hires, clearing ranks, playing catch-up on coding and multimodal, open sourcing weak models, the AI girlfriends. It was all very confusing.
It is clear now. Alex and co have refocused Meta on Personal AGI or pAGI. The AI that runs your life. It feels appropriate for a social media company that hosts your photos and group chats. Most users will not pay for pAGI so the tokens need to be cheap. Muse is targeted at intelligence per dollar. Many sites do not offer API access. The marketing is focused on computer usage. Installing software adds friction. It runs securely in the cloud. If this isn't the right AI play for Meta then I don't know what would be.
It isn't the right play.
Part of the reason Meta is scrambling for purpose is how much privacy laws and changes have hurt them while their platforms fail to attract new generations. Ads are still selling, but they can no longer tie them to actual sales. Users are still logging in, but aging (i.e. failures to capture the younger generation).
Personal AGI has massive hurdles to overcome to be anywhere near the levels of profitability their ailing ads network is. This will bring in new data for them to mine for sure. But it will basically have to outrun new legislature that will likely limit just how "personal" these systems can get as they seek the profit they need to sustain growth via said data.
From the outside looking in, I don't think Meta has a clear vision. I think it's a stretch to say that this one announcement makes it clear. There are AI announcements from Meta every month, recent ones more focused on coding (Muse Spark/Muse Code)
Their ads team wants AI to improve their targeting and ad creation. Some want to have frontier AI including coding. Others might want personal AI agent. Then there's the flirting with renting their compute
And I'm not even taking shots at them. I think Google is exactly the same. I think OpenAI is largely the same (though lately more following Anthropic footsteps and focusing on productivity). And I think the whole field is so new and allows change in so many places - that it's sort of normal or fine to throw things at the wall and see what sticks.
But nothing as of now says to me that Meta has a clear strategy around llms
What’s Meta’s moat? Content and doom scrolling…
I would say they should put ALL their focus on Next Generation content generation… I honestly think in the ~10 years we’re going to see cinema completely flip from high budget blockbusters to way cheaper, personalised “Choose Your Own Adventure” content but where the story changes/rewrites itself in realtime based on constant user feedback (thumbs up from popups mid movie, facial expressions etc).
Meta could be the next generation’s Disney, Warner Brothers, Universal etc
>The AI that runs your life
it was always "personal AI", thats why Zuck wants AR so much.
The problem is, before about 2025, there were a bunch of clear systems and procedures that stopped your personal data from being used for dodgy things. For example the browser inside oculus didn't emit any site specific usage metrics. This means that they couldn't use your VR porn habit to target you.
The same with the raybans. There was a perm block on processing user data offsite/outside of meta systems. (which is why we couldn't use FAIRs AWS cluster to do research on research data, because even though it was paid for data, it was still technically 'user' data because it had non-public domain PII [peoples faces, voices and locations in it]) So getting access to any kind of data that came in from raybans was a massive "nope never"
Then we find out that someone convinced Zuck to let people train on private rayban data? and they send it to fucking third party annotators, and new fucking annotators at that. That to me smells like MSL. To get how much of a big fucking deal that is, the original raybans had two cameras, so that FAIR could do some research or other. but because they didn't want to make a separate infra just for research, they couldn't get permission to use any of the data.
Meta had a working always on assistant that could locate, transcribe and recall any thing inside the offices (assuming you wanted to carry a fannypack with a jetson in it) it was always personal.
its now a case of how much shit they can get away with, its not like the FCC are going to audit them anymore.
People dont even trust meta with their personal data anymore but somehow they'll use it to run every aspect of their lives? This feels like another big failure in the making. Give me an AI that doesn't use my data at all and maybe I'll use it. I already have Google AI Mode for that. Does that steal my data too, maybe, but at least it doesn't have Metas history of abusing repeatedly. They burned that bridge so many times even non techies know not to trust Meta.
Meta certainly do give the impression that they are burning through ridiculous amounts of cash with not much to show for it, and thats just from the outside.
its the same (or was) from the inside.
Someone had a poster up of Pam "this is the same thing" on the one side had the apollo space missions, and the other a picture of Zuck avatar next to the Eiffel tower. At the top it had the current RL spend (something like 45billion) they like me working in RL.
there were three teams for everything, sometimes in the same org.
Take hand interactions for example.
There was production hand detection, they were in the OS/product team. They spent a lot of time trying to make sure it was reliable and useable.
Then there was a near research team, who took old state of the art and salmi sliced it until it worked well enough on either current or next gen hardware.
Then there was the production research people, who were derived from control labs, who for some reason were not in RL-research but buried deep in "production"
Then in RL-Research, the biggest org was agios, who also did hand interaction, but hardly any of their work actually made it near to prod. They liked wrist mounted cameras with no privacy systems, because why would you need it on your wrist? its not like those head mounted twats who spent two years trying to master privacy.
You also have to demonstrate "impact" every 6 months. which is really great when your delivery pipeline is 1-7 years.
If you scroll Instagram for a few minutes you'll see what they have been working on and that are also making a ridiculous amount of cash by selling and serving ads while burning some on side projects like that.
I had to laugh to the point of tears the other day when I saw my friend scrolling through Instagram and they encountered an "Ad Break". The app simply doesn't let you scroll through another post, story or reel until you sit there like a good boy and let the ad sit for 10 seconds or whatever.
How people put up with that kinda crap and continue to use these apps daily for hours on end is mind blowing to me.
I wonder if the prevalence of adblock has helped make a lot of things like this possible. Like, it's hard to imagine people putting up with this without literally protesting in the streets. But the people who can't tolerate it have mostly gone through the effort of opting out, and those who can just put up with it. And thus, the frog boils.
Yeah and if you look at there numbers, they can just afford it. When you realize how big zuckerberg still is.
Yeah, I was surprised to see they just gave DHH a $1.5m credit. Can't imagine they'll have good ROI on that (although what do I know)
How is it a bad thing if a company goes all in on burning cash but at the time have never reported a net loss quarter since its IPO? Its their money to burn at this point who cares
But what you say is not entirely true. There is much to love about companies burning cash on R&D. That's how a lot of greatness happened.
But what is wrong in the meta situation is that while Meta officially projects its 2026 AI capital expenditures to be between $130 billion and $145 billion, investigative reports indicate its true future liabilities exceed $690 billion once unlisted back-end deals and lease commitments are factored in. They have used shadow accounting in many places to hide these things. As it's a publicly traded company so if that goes south it hits a lot of 401ks. So that in a nutshell is why its kinda a bad thing. But you are still kinda right in that its not an existential threat and it's not going to wipe Meta out even though those numbers could totally destroy the vast majority of public companies.
Can you link to those reports? How reputable are the investigators? Financial market has great incentive to stay informed. Those 401ks are their bread and butter customers, and not every bank/fund is Meta's co-conspirator in its shady deals, if any.
Its in every 10-Q and 10-K. Original poster read 1 article in the WSJ and thinks he/they uncovered a massive conspiracy. Its disclosed exactly as it should be according to GAAP accounting. They went above & beyond to disclose them in the text narrative of the filings because the GAAP requirements require them not to include them in the numbers.
WSJ just needs to sound alarm bells so people keep paying for their crappy journalism, and thus everything they write becomes alarmist slop.
https://d18rn0p25nwr6d.cloudfront.net/CIK-0001326801/abad205...
pg47 of the 2Q26 10-Q, linked above.
Ultimately, the equity is really cheap even today on trailing numbers. ~12x EV/trailing EBITDA excluding-RL losses and still growing 25%-34% in each of the last 4 quarters organically at $200B+ of scale.
One of my pet peeves is that people who dont understand business forbid these companies from making investments or starting new businesses now or in the future. Meta has been sitting on $50-100b of cash on the balance sheet for years, and people lose their minds when they start to meaningfully invest it. I think theyve earned the right to diversify
> WSJ just needs to sound alarm bells so people keep paying for their crappy journalism, and thus everything they write becomes alarmist slop.
The WSJ is a ghost of what it used to be.
I would have to say that its tragic more than bad. Meta consumed whole industries that used to spend the money on a bunch of different things. Now all the revenue goes to Meta shareholders and whatever it is Zuck is wasting billions of dollars on pursuing.
I genuinely think big tech co's top leadership have very sophisticated strategies/positioning that they simply cannot communicate or explicitly canonize due to their position as spokespeople for the company (and society writ large, news media, investors, customers, employees, vendors). Of course there are a lot of bozos and incompetent people flailing around and a lot of work ultimately gets wasted (which understandably bothers line employees a lot), but is inevitable and even necessary to eg hedge product strategy/comp and take risks on ideas and people.
It would not really be useful to have that conversation with employees because it's incredibly distracting (now product strategy is up for debate with way too many cooks in the kitchen), and very few employees have the exposure or skills to meaningfully contribute even if they think they do. I saw it firsthand at Google TGIFs.
Meta's strategy seems to be "personal agents" quite consistently. Remember they tried to buy Manus? And note that Muse Spark and the meta AI platform products clearly seem to prioritize web search, computer/browser use, and vision/language tasks over coding, which is something that had to bake for a long time. Also, this product launch itself is pretty interesting:
1. It's clearly a fast follow to the current FOTM hype startup Instinct with a much more comprehensive implementation and integration with their other agent products.
2. It's kind of like openclaw, which got a lot of non-developers very excited but was basically consistently unusable. Except this agent's compute runs remotely and presumably has slightly more sane development practices. IMO it's the first main openclaw-like product that has made it to the "just works" level of usability.
3. The focus on ecommerce is actually really really important, because Meta is trying to capture the intent/demand-driven purchasing flow that Google currently owns through search. Controlling the top of funnel is what enables google to make hundreds of billions of dollars per year on search ads. Meta is an advertising company and Google's search ads business is the most lucrative and centralized/well-defended advertising market in human history.
So it is actually a really big deal that Meta is trying to go after it (at least, the CUJ, it's possible that they'd monetize the agent-driven UX differently than search ads) because it's probably their best shot at disrupting that market and one of the few growth opportunities that would actually make a dent on their balance sheet. Obviously Meta is not going to lay out all that strategy stuff explicitly because for all intents and purposes it's a distraction and shifts the conversation in an unproductive direction (the strat behind the product, rather than the product itself). But it's pretty clear if you look for it.
Edit: Actually I thought about the advertising business more and I think in the short term this is partially about attribution/conversion. In 2022 the Apple tracking changes cost Meta $10B in lost conversion metrics; an agent-driven and proxied purchasing flow has built-in attribution and funnel measurement which is very valuable in its own right!
> very sophisticated strategies/positioning that they simply cannot communicate
It's Matryoshka nested parallel construction for corporate strategy.
There's a real, coherent, aggressive strategy that only a few in the inner concentric circle know, that's ring-zero.
Then there's the strategy that ring-zero tells ring-one, which isn't the real strategy, but it's sufficient to get EVPs and VPs to execute in rough alignment with the true, ring-zero strategy.
Then ring-one does the same dance with ring-two, etc.