It still seems like something is missing from all these frameworks.
I feel like an average human wouldn't pass some of these metrics yet they are "generally intelligent". On the other hand they also wouldn't pass a lot of the expert questions that AI is good at.
We're measuring something, and I think optimizing it is useful, I'd even say it is "intelligent" in some ways, but it doesn't seem "intelligent" in the same way that humans are.
If a human cares about the work, they can often outperform an LLM because they will keep at it until the work meets their standard of quality. Whereas the LLM will guess and then wait to be corrected. As a recent tweet I saw said: it’s amazing how fast the software bottleneck went from writing code, to reviewing code.
I think we’ll need to split the concept of intelligence into the capacity to accomplish a task and the capacity to conceive and prompt a task. If the former is called “intelligence” then LLMs are intelligent.
But what then do we call the latter? I think the idea of an AI that can independently accomplish great things is where people talk about “general” intelligence. But I think we need a label more specific, that covers this idea that successful humans are not just good at doing things, they originate what should be done and are not easy to dissuade.
On the other hand, AI being very good at everything while select humans may only be very good at some things is likely also a quality we want to retain (or, well, achieve).
What I’d be most worried about is an “us vs them” mentality developing. It doesn’t break first but it can develop almost naturally. 10 -> 50 people tends to mean “team of teams” rather than “a team”.
Once this dynamic takes root, it’s much harder to stomp out. I’d be more worried about systemic issues like this than the short term pain from growth.
I agree with the criticisms of modern macOS. However, let’s not forget their competitor is Windows. So even if they aren’t going great, it’s not like they have stiff competiton. Thats before even getting into hardware where Microsoft is even further behind.
It’s unfortunate that it’s so hard to disrupt this space.
As a long time user of all 3, it's hard to understand why people would use Linux as a main driver, especially if they aren't just programmers, but also dabble in video or music.
Really? When every thread about linux on a laptop includes comments like “palm rejection doesn’t work but I use an external mouse so I don’t care” and “the laptop gets hot in my backpack because sleep/suspend doesn’t work so I just power it down” you’re still not sure why people are using a windows or a mac?
At this point I’m beginning to suspect that linux users have stockholm syndrome. Or is Tux standing behind you with a gun?
I know, I know, you don’t have any of those problems. You have the Blessed Configuration that has the right tradeoffs for you. But like… most people don’t want to spend waking hours reading Archwiki to figure out why their wifi drops when they move from one room in their house to another…
Yeah, like, it's not even that difficult to re-compile your media player in order for it to support a new format you want, like FLAC.
Yes yes, I can already hear the naysayers "it's not that easy". It actually is! Just make sure you have the appropriate GLIBC version and a specific version of either clang or GCC that is compatible (hey it's Linux, you can choose!). Then do the usual ./configure --with-openssl=<CUSTOM_SSL_LIB_BECAUSE_THE_STOCK_ONE_IS_TOO_OLD>, make, make install (remember to use sudo on that one because we write some system files).
Honestly, the whole process took me just two hours from start to finish. Easy peasy.
I'd much rather do that than buying hardware that is massively overpr... oh, you're saying they're cheaper than Linux laptops now? Idk man ... I would still not buy any of that, those are definitely for brainwashed cattle.
Except Apple laptops usually aren't subsidised by mobile phone operators, offered for free in exchange of a five year contract, where the owner is actually the mobile operator due to device locking.
On the contrary, it’s unfortunate that this space can’t just be an island of stability, where things keep working the way they always have and new features are added unobtrusively. Instead we get surface-level change for the sake of change as well as slop and surveillance nobody asked for, and are still missing conveniences that have been table-stakes on mobile for years. It would be a a step in the right direction if Windows would simply stop disrupting itself.
> it’s unfortunate that this space can’t just be an island of stability, where things keep working the way they always have and new features are added unobtrusively
There are plenty of Linux operating systems that prize stability over feature richness. They work fine if that's your workflow. Most folks' workflows, unfortunately, are not that.
Interesting, cutting way back in the product they renamed the whole company for.
They feel a bit directionless to me. They are still making money but even their AI attempt feels half hearted. I think they are really trying but I’m not sure they can build the engineering muscle to move in new areas with the brand damage they’ve sustained.
I was at Intel for a while and there was one glaring problem - they have one product that spins off a huge amount of cash. This means a few things: First, that one product is really where the things that matter happen. But second, they have all this money and they don't know what to do with it, they can't spend it all on their core product because that looks terrible - they're already throwing off money, investing more probably just makes your company look bad (you're spending more to get the same revenue). SO instead you have to take that money and make bets. But not just any bets. You need a bet that (a) matters if it pays off, and (b) looks favourable compared to the core business. So you buy Mcafee and Altera and MobilEye, 5G was the future once...
So to take the Meta example, they need something that is going to have revenue upside similar to Meta advertising revenue (one of the most profitable things in the universe), and that has better margins that the advertising business (basically impossible).So the only logical thing to do is to make grotesquely large bets on things that are extremely speculative. You can't bet on things that are well known - because nothing known has the properties from earlier that you're looking for, and you can't bet small because you've got to convince people you're the pay off is of a similar size to your existing business.
In Intel's case they lost focus on the core business and so that died and their other bets didn't matter because the core business was dead. With Meta the core business in't dead, but it's only a matter of time before it's seriously threatened and so they're going to attack that threat with everything they've got - and they have a tonne of resources.
In 2026 we need to update our mental model of Google. Google has been wildly successful at adding diversification. Around 40% of Google’s profit (depending on the quarter) comes from non-search income.
They build a wildly successful cloud platform, they’re expanding their subscription services, they’ve got enterprise offerings, etc
The trick is that Google accepted that none of their other business would likely have the margins and volume that search has, but they did it anyways.
But Google actually knows how to do research and how to apply it to products. Meta's AI research hasn't produced anywhere near as many state of the art products /revolutionary achievements.
Google knows how to do research - and at the very least lets other people figure out the products, and then becomes the #3 or #4 player.
Both GCP and Gemini are products of this. Modern cloud was arguably built by Google (think Chubby, GFS, Bigtable as building blocks) - they just spent 10 years ceding it to Amazon before competing.
Since we're comparing to Meta, you just have to look at the state of their publicly facing products that feature AI. Google has better AI models (Gemini, Nanobanana) and they've integrated them successfully into way more products than Meta has.
Meta spends a lot of money on AI research with little to show for it. As imperfect as Google may be, they're still doing much better.
I was thinking more of their primary revenue source / money printer being their ads business like Meta then they also spend billions from it on all kinds of other bets.
They already attacked it with everything they've got lmao
As in, in 2012. They outright replaced people's email addresses in their profile (makes it harder to reach people outside the walled garden, makes it harder to transfer your credentials to a competing service) and I've heard Google+ links got blocked
Zuckerberg is many things, not everything he's accused of (Trump/Cambridge Analytica) might be entirely accurate but he is at least partly a bit of a scumbag
The Meta/Oculus bet was never about VR or gaming. It is to solve Zuck's greatest fear of being beholden to platforms controlled by others. What it was supposed to be was basically what Microsoft missed out on with mobile, losing to Google and Apple. VR and AR could be that next platform to own and control. They would love to have all the data that Google and Apple enjoy. That's what Quest -> Horizon was about, change a gaming device into a mainstream entertainment and work-friendly one. This is all driven out of fear and lack of control. It would suck for them if OpenAI owns the most successful personal AI device.
They could release a subsidized Facebook phone which is just some android pos with Facebook rammed into it for way less than the 70 billion or whatever they burned on VR.
I still find it funny that Apple did exactly the usual Apple thing of coming in way later and yet inventing a UX for this kind of stuff good enough that Meta immediately started scrambling to copy it.
Having a baked in hardware userbase and dev team helps a lot. The hardest part of inventing new things is getting feedback from real people who really think deeply about this stuff, who also have lots of power within the company to say no.
> Interesting, cutting way back in the product they renamed the whole company for.
It was clearly the wrong bet. He pumped something like $100B into the endeavour (Meta Quest / VR / Horizons) and it is just slowly dying as we speak. He has to give up on it, although I am sure it will be called a "pivot" into AR glasses.
> He pumped something like $100B into the endeavour (Meta Quest / VR / Horizons) and it is just slowly dying as we speak.
Literally never met anyone who used or liked the Horizon thing, VRChat in comparison is more popular and doesn't feel like a soulless corporate husk: they also have quite the variety of worlds, from party games, to someone building a whole jet/chopper flight combat simcade world; ofc all of them are a bit jank, but lots of cool stuff and very expressive avatars.
Meta Quest, on the other hand, seems like a really good piece of tech - I still have my Quest 2 (because I'm broke as hell), but I enjoyed even that one, albeit maybe with a slightly more comfy head strap than the default one and the Virtual Desktop app cause their Link app doesn't support Intel Arc GPUs. The tracking is good, the experience of all sorts of stuff in VR is nice, games like H3VR or VTOL VR are great, as is Into The Radius VR! At the same time, I can see why it never saw super widespread adoption - tricky to develop for and also a somewhat limited audience.
Also the productivity situation just isn't there, closest I got to a good productivity setup (out of curiosity) was the Immersed app before they messed it all up by removing support for physical monitors - I could have my 4 physical monitors in VR surrounded by whatever I want and some virtual monitors and just lock in, it was kind of zen despite the technical limitations. It seems like people got promising tech in place... and then never really wrote good software to take advantage of it. Even Virtual Desktop has artificially enforced monitor limits in VR.
I hope VR tech continues to progress (especially lightweight headsets) no matter what happens to Meta.
Yeah, it was a bizarre decision. There isn't a clear ROI on games and that's what Horizon Worlds has been the whole time. There's no equation that says a 100M game automatically makes 100x more than a 1M game on average. If anything the equation is sub-linear. 100B just doesn't seem like the right size for a game investment.
It's supposed to be a Roblox competitor, which does print money, though probably not to the extent of how much they invested.
The problems are 2 fold:
People/kids don't want to put on a VR headset to play Roblox. I guess they're conceding this point by pivoting to mobile.
Meta is the opposite of cool. Real name requirements, only humanoid avatars, super corpo branding, etc really seriously hold them back from competing with VRChat or Roblox. This one is terminal it'll never be fixable as long as Meta is at the helm.
There are some really good ar glasses for a couple of hundred dollars, I think they are going to end up really cheap and not the 100 billion investment that facebook needs.
Tbf I don't think they ever intended to make back their investments via the goggles. As near as I can tell the thought process was basically: "Real estate + fashion + live entertainment + art + etc is X quadrillion dollars. We could make The Virtual World and capture all that value. It would be irrational not to invest $100B!" Basically Pascal's Investment.
I'd like to think that the top minds working on AI have a higher purpose than to get the next generation hooked to a digital morphine drip. Serving soap cutting videos and giving teen girls body dysmorphia isn't a very compelling mission.
Though I'm sure many are mercenaries and will work for whoever pays the most.
>I'd like to think that the top minds working on AI have a higher purpose than to get the next generation hooked to a digital morphine drip.
That's the irony. The genius scientists are against AI used for defense, but somehow they're all-in for AI being used to getting people addicted to ads, dopamine, gambling, debt, porn, political manipulation, etc. basically everything that's guarantee to wreck society, but thank god they aren't making weapons I guess.
Can we all say a big thank you to Neal Stephenson for inspiring Zuckerberg to light tens of billions of dollars on fire in this stupid quest? Imagine what kinds of anticompetitive acquisitions or further privacy-invading tech they might’ve spent that on instead.
It makes sense to rebrand anyway, because I'm sure they don't want people to only think of "that social media site" for all of their other ventures. Just like Google rebranding as alphabet
I think for Google the more meaningful other brands are the actual product ones, like Waymo, Nest, YouTube, Calico, Verily. Those are the ones that benefit from being able to distance themselves a bit from being "a Google company" and all the baggage that comes with that, eg an assumption that it'll be shuttered at some point, will pivot massively into ads, whatever.
I don't think Meta has nearly that need. It's "other companies" are Instagram and Whatsapp, which are basically in the exact same space as Facebook.
I don't think Google really "rebranded" in the same kind of way, since Google is still their brand across the vast majority of their product offerings and the signs on Google offices still say Google. Seems like the Alphabet thing is more about letting the "other bets" be under a higher umbrella, and possibly other reasons related to financial engineering etc.
The rebrand came at a time when "Facebook" was mainly associated with either tremendous scandal (Facebook Files, ad fraud, Cambridge Analytica, Rohingya and Tigray genocides, etc.) or a social media platform increasingly dominated by the elderly.
I think it was a desperate lunge away from that toxic brand toward ANYTHING else. Zuckerberg put his money on VR, given the pandemic and the mild success of Oculus.
Betting big on the metaverse in particular was a mistake, but it might have helped keep the Facebook stink off of products like WhatsApp and Instagram, which remain pretty popular among mainstream audiences.
What is the development loop like with this? There’s a lot of folks successfully building games with agents already on the AI gamedev Discord server. So I’m wondering if there were some shorter paths to your goal. You might want to exchange notes with folks there.
TBH "The Oracle" personality doesn't always bother me. There's a bunch of stuff that gets asked in corporate chat that makes me think, "does this person know how to use Google?" I think the same can be said for the chatbots at this point.
I feel like we just need the equivalent of "Let me Google That" for LLMs.
Caveat that I'm talking about when it's something that's in depth or requires some nuance beyond just a surface level information. They don't know themselves, so rather than saying "oh, sorry I don't know, this document or person might help."
As an example I might post in our teams chat, that I've seen an issue on our physical hardware, there is a step change in the telemetry (increased vibration or some other erratic behavior on a thermocouple). Has anyone had experience with what can cause this on this type of installation.
Then you get a Copilot paste of the prompt "what causes high vibration on rotating machinery".
If somebody clearly asks a question that could just be answered by Google or a LLM prompt quickly then fair enough, but I'm after specific product knowledge from our technical team. In self reflection I might be very specific in my question so other members of the team can't go down that route as easily.
"Let me google that" already mostly gets you LLM output, because google is fairly ruined at this point. Also, if someone asks me a question, I think it would be rude of me to respond with "Maybe I know the answer, maybe I don't, but why don't you send your question to a machine that will give you an answer that just might be right by coincidence?"
Agreed. I don't even necessarily have anything against AI edited text but there's a way to sharpen your own writing and there's a way to let its voice dominate. There's a lot of idioms it tends to fall back on (em dashes being the most well known). I'm surprised that folks don't notice these and aggressively reassert their voice.
I use LLMs in my own writing because they have benefits for conciseness but it tends to be a fairly laborious process of putting my text in the LLM for shortening and grammar, getting something more generic out, putting my soul back in, putting it back in the LLM for shortening, etc. I tend to do this at the paragraph level rather than the page level.
I feel like an average human wouldn't pass some of these metrics yet they are "generally intelligent". On the other hand they also wouldn't pass a lot of the expert questions that AI is good at.
We're measuring something, and I think optimizing it is useful, I'd even say it is "intelligent" in some ways, but it doesn't seem "intelligent" in the same way that humans are.
reply