We've very quickly reached the point where AI models are now too dangerous to publicly release, and HN users are still trying to trivialize the situation.
Are they actually too dangerous to publicly release? It seems like a little bit of marketing from the model-producing companies to raise more funding. It's important to look at who specifically is making that statement and what their incentives are. There are hundreds of billions of dollars poured into this thing at this point.
You really think some marketers got leaders from companies across the industry to come together to make a video - and they're all in on the conspiracy because money?
That’s literally exactly the kind of thing marketing does, and has been doing for a very long time. Did you just arrive on earth from outer space or something?
Yes? Saying "conspiracy" is overstating things. A company can make a marketing push overselling their product and then have exclusive corporate partners that benefit from being associated with that marketing. That just seems like normal business that happens every day, and being skeptical of marketing messages should be your default position.
Says the marketing department of the company who is apparently still working on these AI models and will 100% release them to the public when their competitive advantage slips.
Marketing pushing to release a dangerous model is a lot more likely than marketing labeling a model of dangerous when it really isn't. If anything marketing would want to downplay the danger of a model being dangerous which is the opposite of what Anthropic is doing.
Everyone here doing mental gymnastics to imagine Anthropic playing 5-D chess because they're in denial of what is happening in front of their faces. AI is getting more capable/dangerous - it's not surprising to anyone. The trendlines have pointed in this direction for years now and we're right on schedule.
1) OpenAI and Anthropic are killing it, and continue to do so, their coding tools are unmatched for professionals.
2) Local models don't hold a candle to SOTA models and there's nothing on the horizon that indicates that consumers will be able to run anything close to what you can get in a data center.
3) Coding is a killer product, OpenAI and Anthropic are raking in the cash. The top 3 apps are apps in the app store are AI. Everyone who knows anything is using AI, every day, across the economy.
The grandparent is definitely wrong on (3). Yes, coding is a killer product, I agree with you.
On (2), I agree with you for local models. BUT, there are also the open source Chinese models accessible via open-router. Your argument ("don't hold a candle to SOTA models") does not hold if the comparison is between those.
On (1), I agree more with the grandparent than with your assessment. Yes, OpenAI and Anthropic are killing it for now, but the time horizon is very short. I use codex and claude daily, but it's also clear to me that open source is catching up quickly, both w.r.t. the models and the agentic harnesses.
>BUT, there are also the open source Chinese models accessible via open-router.
I thought so myself, but after burning a lot of money on OpenRouter in a few days I just subscribed to Z.ai's Coding Pro plan and using the subscription is much, much friendlier with my wallet.
Open models are good but if you need a $10k GPU to run them then 99% of people are better of subscribing to OAI or CC.
Nowadays I also feel model performance matters less than the design of the tool harness, inference speed, and the other systems that surround a typical coding model.
I used GLM5 quite a bit, and I'd say it was maybe on par with Sonnet for most simple to medium tasks. Definitely not Opus though. Didn't test super long context tasks, and that's where I would expect it to break down. A recent study on software maintainability still showed Sonnet and Opus were peerless on that metric, although GLM series of models has been making impressive gains.
I don't want to respond to 100 comments about the same thing, and this one happens to be on top, so, in my humble opinion:
(1): You don't have to be an Ed Zitron disciple to infer that OpenAI and Anthropic are likely overvalued and that Nvidia is selling everyone shovels in a gold rush. AI is a game-changing technology, but a shitty chat interface does not a company make. OpenAI and Anthropic need to recoup astronomical costs used in training these models. Models that are now being distilled[1] and are quickly becoming commoditized. (And frankly, models that were trained by torrenting copyrighted data[2], anyway.) Many have been calling this out for years: the model cannot be your product. And to be clear, OpenAI/Anthropic most definitely know this: that's why they've been aquihiring like crazy, trying to find that one team that will make the thing.
(2): Token prices are significantly subsidized and anyone that does any serious work with AI can tell you this. Go use an almost-SOTA model (a big Deepseek or Qwen model) offered by many bare-metal providers and you'll see what "true" token prices should look like. The end-state here is likely some models running locally and some running in the cloud. But the current state of OpenClaw token-vomit on top of Claude is fiscally untenable (in fact, this is why Anthropic shut it down).
(3): This is typical Dropbox HN snark[3], of which I am also often guilty of. I really don't think AI coding is a killer product and this seems very myopic—engineers are an extreme minority. Imo, the closest we've seen to something revolutionary is OpenClaw, but it's janky, hard to set up, full of vulnerabilities, and you need to buy a separate computer. But there's certainly a spark there. (And that's personally the vertical I'm focusing on.)
> Go use an almost-SOTA model (a big Deepseek or Qwen model) offered by many bare-metal providers and you'll see what "true" token prices should look like.
Qwen3.5-122B-A10B is $0.26 input, $2.08 output. Where's the subsidy? It's ten times cheaper than Opus. Or did you mean that we're subsidizing their training? But then "OpenClaw token-vomit on top of Claude is fiscally untenable" makes no sense.
Yeah, I don't know where you got your costs from. Bare metal providers are significantly cheaper than Anthropic.
Maybe he's comparing the renting price of a bare metal server on its own, and doesn't realise how drastically cheaper they are to batch together for an API provider.
> And to be clear, OpenAI/Anthropic most definitely know this: that's why they've been aquihiring like crazy, trying to find that one team that will make the thing.
Anthropic is up to $30B annual recurring revenue. I wish I had failing business models like that.
> Token prices are significantly subsidized and anyone that does any serious work with AI can tell you this. Go use an almost-SOTA model (a big Deepseek or Qwen model) offered by many bare-metal providers and you'll see what "true" token prices should look like.
I'm not sure what think you are saying here, but if you look at the providers for both "almost-SOTA model (a big Deepseek or Qwen model)" or at the price for Claude on AWS Bedrock, Azure or on GCP you will quickly see inference is very profitable.
Anthropic has raised $64B in total since they were founded.
Even if you say we are going to measure profit in the very special hacker news way of looking at money taken in from customer revenue against money invested and we say they can't do things like counting building data centers or buying GPUs as capital expenses and instead have to count them against profit then in 2 years time they will have made more money than they have taken in investment.
The proverbial "50B" is investment in next year's model. The current model cost under "30B", and therefore "is profitable". It is a bet on scaling, yes, but that's been common throughout the industry (see, eg, Amazon not being profitable for many years but building infrastructure)
> If every year we predict exactly what the demand is going to be, we’ll be profitable every year. Because spending 50% of your compute on research, roughly, plus a gross margin that’s higher than 50% and correct demand prediction leads to profit. That’s the profitable business model that I think is kind of there, but obscured by these building ahead and prediction errors.
You're missing the forest for the trees. Per-token pricing is irrelevant when you're just trying to get shit done. I pay 20 bucks a month for OpenAI, but I use likely $200+ a month of tokens just on the coding (and I'm just looking at the raw tokens, this is ignoring all the harnessing on their end). Even OpenAI has said that they're losing money on the 200-dollar subscriptions[1]. This is not a viable business model. Why do you think they are introducing ads this year[2]?
One way or another H1B's got in the door, a lot of them. And at this point they're just hiring themselves. OPT's as well, they will stack the candidate pools with only OPT/H1Bs. Post any job and you will get a flood of thousands of spam resumes. If you ignore them then you put yourself at risk of being accused of discrimination.
Anthropic didn't miscalculate anything. They calculated what they could charge/subsidize for humans, not automatons. Banning OpenClaw brings usage levels under control.
If you had to pay for APIs yourself for any provider then you'd know that SOTA tokens are not cheap, and Claude Code for $100 is almost a too good to be true bargain for what you can get out of it.
Not much impact, Codex is already open source. The real value is in the model itself and the ability to use it with a subscription. Something you can't do legally with a clone of this code.
The only thing I found interesting about this leak is just how much of a rats nest the code base is. Like it actually feels vibe coded without a shred of intelligent architecture behind it.
Regardless, you can't beat the subscription and model access despite the state of the code base, so I still use Claude Code daily and love it.
I just hope it doesn't turn out like n8n. I built a few things, wanted to make changes, looked at the code base, opened the devcontainer, noped out after being mortified by the sheer number of warning and dependency issues, threw away all of my work, uninstalled, didn't think about it again.
Anyone have a theory why Apple hasn't done this yet? They release an 'iBook' which is basically a wired or even wireless lapdock for your iPhone running OSX in a partition. Seems like that would decimate the entire Windows, laptop, even desktop market in short order.
Everyone with an iPhone, no longer needs their laptop/desktop. Just buy a cheap iBook and there's a good chance it'll already be better than most consumer PCs.
There isn't much demand for using phone as computer. If you are at home or work, you can buy a desktop computers for cheap. If you are traveling, you need to find a monitor and keyboard. You could carry small monitor and wireless keyboard, but then you are carrying as much as laptop. People who need to work on the road get a laptop. People who need to send email get iPad and keyboard.
Good example of the economics is that Macbook Neo or iPad Air are cheaper than new iPhone.
iPhone should export display, but more for showing videos or presentations. My Pixel 10 has USB-C display and I haven't used it, but I have computers for all purposes.
Apple should spend more effort making the iPad usable for work. It would be good candidate for USB-C display, but with iPadOS.
Imagine an executive placing their phone on a magnetic dock as they sit down, which automagically connects to the screen and gives them access to everything they were doing before. Also easy to imagine a university computer lab where everyone brings their own compute and IT doesn't have to manage physical desktops.
I'm skeptical that there's "no demand" for that kind of functionality rather than a lack of good implementations. Look at how popular wireless CarPlay and Android Auto are. They're essentially the same functionality, but tailored to an in-car experience instead of desktop.
Imagine executive tapping their phone down on reader, and it pops up everything they were doing, and they get to keep using their phone.
The first flaw in the idea is that computing is cheap. You can make a computer the size of a phone for people to carry around, that has been tried but failed. The second flaw is that everything is in the cloud, only developers and offline need local access to their files. The cloud also means that can desktop in the cloud.
You can make a computer the size of a phone. That's what the latest macbook neo is. The rest of the space inside is battery and peripherals. I'm not sure what cloud has to do with this discussion.
Re: keep using phone, that's exactly what's already possible with CarPlay and AA.
FWIW, you can plug your iPhone into an external monitor to do a Keynote presentation. You need a USB-C (or Lightning) to HDMI dongle in most cases, but it works fine.
I'm always reluctant to do non-standard stuff for presentations. There's enough that can go wrong even with a direct HDMI out. I've done it in a pinch but pretty much always carry a laptop with me when I'm presenting along with local copies of my presentations. I've actually gotten a text in the middle of the night asking me if I can fill in for another speaker who forgot and are in a different country :-)
On an upcoming trip I'm actually going to give an iPad with magnetic keyboard I bought a couple years back, assuming different travel patterns than I've had, a try. It seems to work fine. An iPad is also great for plane/train entertainment without a keyboard. But, honestly, it's no lighter than a MacBook Air would be and if my ancient MacBook Pro dies--have a newer one up in my office--that's what I'll probably buy.
I have traveled with just my iPhone and can get by but don't really love it.
How can there be demand for something that doesn't exist?
If Apple releases a $300 lapdock tomorrow, basically a screen, keyboard, battery, that allows using your iPhone as a normal general purpose computer with OSX - why would anyone buy a laptop/desktop?
HNers are significantly more technical than the median consumer and are used to text and keyboard interfaces - a large portion of humanity isn't. You see this with Foundation Models as well - most have started to shift away from only concentrating on text to TTS and STT usecases.
Also, DeX style monitor screen share with a Bluetooth keyboard has been supported since iOS 15.
Additionally, a major portion of Apple's desktop revenue is coming from poweruser and specialist demand - IT departments bulk purchasing developer laptops, designers having their entire design workflow within the MacOS environment, and video editors heavily dependent on MacOS.
Furthermore, arguments about how Apple has an incentive not to cannibalize revenue are dumb, given how open Apple is to cannibalizing revenue where PMF exists (eg. the iPad Pro versus lower tier MacBooks or the MacBook Neo versus lower tier iPads).
The entire Mac line is a teeny tiny slice of revenue compared to iPhone. Allowing OSX on iPhone would increase the utility of iPhone, leading to more sales.
> Allowing OSX on iPhone would increase the utility of iPhone, leading to more sales
That assumption is not necessarily true.
What this implies is that there is a market of existing consumers that would not buy an iPhone because it lacks OSX support.
The iPhone portion of Apple's business generates around $144B in YoY revenue in Q1FY27 [0].
Whenever an organization contemplates building a net new capability like the one you mentioned, a quick test is whether it would be able to generate and sustain at minimum the equivalent of 1% of yearly revenue.
If this was a $1B revenue opportunity it would have been implemented, but it's not.
Nor is it a feature that can actively or dramatically increase Apple's market share in most markets.
A good proxy of such demand would have been a sudden increase in iOS users using USB-C screen share and a Bluetooth keyboard to interface with an iPhone in a desktop form factor (something which has been enabled since iOS 15), but such an increase has not happened.
Consumers haven't been told they can do that though. It's not ergonomic to do that. There's not a Belkin plastic dock to support that use case, so I don't find that is a good proxy for it.
Consumers don't need to be "told" what to do. If there truly was demand, Samsung or other third-party vendors would have created an ergonomic dock for DeX enabled Samsung phones and it would have been a killer app.
Other than UI and other surface differences, the fundamental distinction between a Mac and an iDevice is... what it is.
A Mac is a real computer. I can run any code I want on it. I have root.
An iDevice is like a game console. I can only run App Store apps (without jumping through a lot of hoops). I do not have root (without again jumping through many hoops or ugly hacks).
If Apple wanted to unify the platform they have two choices. The first is to abandon the "real computer" market entirely. The second is to make iDevices real computers by unlocking them.
I suspect they'd rather keep two platforms.
Under the hood they both share a lot of code, so it's not two totally distinct platforms. It's more like two sets of defaults and two "skins."
I think the friction of using a keyboard/pointing device with a touchscreen, or fingers with a desktop interface, is too high to unify them. I know it's been done, I'm unconvinced it's been done well.
That’s the difference though. Put macOS on a phone chip and it’s now a “real computer,” just a smaller one.
The M chips are mostly just roided out A chips: higher clocks, better cooling, more P cores, big GPU, and I think deeper pipelines and more cache. The ALU and many other sections are, I think, identical.
Thermal throttling is actually a non trivial limit on phones. Put a heat sink and a fan on an A chip and sustained compute is faster.
The OS and its restrictiveness determines the class of device not the hardware.
That was already the case with the M-series chips, which are shared between Macs and higher-end iPads. The Neo just extends it to the A-series as well.
Yep I know, and now using a last gen A chip, I feel they are really rubbing our faces in it.
Like Apple is saying, "Nice iPhone 17 Pro w/ A19 w/ vapor cooling chip you have there; you know you run a full general purpose OS on it, but we're not gonna let you, nanananana :p"
No exactly, Apple is playing in our faces, all while people continue to defend the “differences” of device categories and the subsequent justification of shipping iPhones and iPads with locked bootloaders.
The belief that people only hold opposing opinions to yours because they have money on the line is such conspiracy theory nonsense. Some random teenage in middle America couldn't just really like Apple products? It's gotta be some grand conspiracy against you?
I think Apple is just really careful about how they segment their product line for each use case, and would never go for a "jack of all trades" solution like this.
Why would it decimate the Windows market? From my experience, there's a strong correlation between iPhone and Mac usage.
Looking at the stats, the Win:Mac ratio is 4:1 but Android:iPhone only 2:1 so it might hurt Windows. But if iPhone users are more likely to use Mac or don't use computers much already, then expanding iPhone capabilities would cannibalize Apple business.
Because then most people with an iPhone wouldn't need to buy a separate laptop/desktop. I'm sure Android as well would follow in short order (not the half hearted attempts they've made so far). Sales would plummet. Windows decimated.
Why would Apple want to sell a lapdock when they could instead sell you the same thing + a redundant SOC (aka, a MacBook) and then high-margin cloud services to sync all of your data between your two differently-shaped computers?
Because most people with iPhones are buying Windows computers, but give them a cheap entry lapdock into the Mac ecosystem and maybe their next more powerful system will be a Mac.
Mac is a niche right now, iPhone with OSX could level the playing field.
If Apple could bring themselves to sell a lapdock, it'd have to cost at least $500. We know this because the Magic Keyboard for iPad, just a keyboard and trackpad, is priced at $349 (and it was introduced at that price way back in 2020, so at the time, Apple believed that keyboard was worth $440 adjusted for inflation). A screen to Apple's quality standards, even a 12-13" one, cannot possibly increase that price point by less than $150. So, the Apple in our universe could not produce a lapdock, because in our universe they have a whole laptop at that price point.
On second thought, the reality distortion field is real, so I suppose if they told people their new $600 lapdock was a good value even though it costs as much as the entry-level Mac, they'd still find willing buyers.
This. The more locked down, the less in control we are, the higher margins they command. This is why app stores exist - it has nothing to do with safety or security, and everything to do with monopolizing the distribution supply chain from soup to nuts. Don’t like it? Too bad, it’s fully locked down and cracking it is a (potentially) criminal offense, so whaddayagonnadoaboutit?!
A little computer board is only a fraction of the BOM of a laptop, so a 'lapdock' of equivalent quality couldn't be very much cheaper than a whole laptop.
If you use cloud storage, your laptop already has all the stuff on your phone anyway.
The general public thinks phones and computers are fundamentally different. Heck, I remember arguing this point even on HN back when smart phones were first coming out and being generally on the losing side as people got very excited about "app stores" and such. I see no practical path to getting to the point that enough of us realize that there is simply no reason for our phones to be locked down the way they are that the companies are forced to undo it, especially with our elites pushing with all they are worth to lock things down harder.
The companies take that confusion to the bank.
There have been numerous attempts at making phone/laptop crossovers, where you can plug your phone into a dock and get a computer, or slide your phone into a laptop case, etc. Some of them are even still around, but they're all definitely second-class citizens. There's a variety of problems that I think they've had in the market, not least of which is the fact that the average person still sees "phones" and "computers" as fundamentally different so the product makes no sense to them, but another issue that I think has held them back is that the product inevitably work by porting the limitations of the phone into the computer, rather than porting the freedom of the computer into the phone.
In the USB-C era, there is no excuse for every phone not having a mode where you can plug it into any ol' USB-C hub/dock and be able to get a desktop environment, even down to the "middle-of-the-line" phones. It would require in most cases no extra hardware. They just don't.
Money? You don't think Apple would make a killing on OSX licenses and lapdock sales if they allowed OSX on iPhone tomorrow?
Mac is a tiny slice of revenue for apple. OSX on iPhone would blow it out of the water. Apple would turn the PC market upside down, taking a sizeable chunk from Windows. As there'd be no point for most people to have a separate laptop/desktop at that point.
People also thought that phones needed keyboards before Apple showed them a better way. This is all on Apple to make a reality, no one else can bring general purpose computing to iPhone except them. It's their choice to make.
It would explode sales of Mac. OSX on iPhone, people wouldn't need the separate Windows laptops they're used to. OSX on iPhone is the gateway for consumers into the OSX ecosystem.
And when those consumers want more powerful hardware, instead of buying a more powerful Windows laptop/desktop - they buy a Mac instead.
I feel like Apple knows this as well, so I can't figure out why they haven't pulled the trigger. Anti-trust risk? lol
I don’t understand the argument for why allowing it would mean more Apple computer hardware sales though. Could you explain why you think that would happen?
I think there are a number of reasons why Apple specifically hasn't done this. In addition to what others have already mentioned (demand, segmentation, profitability, etc), another factor would probably be difficulty with the overall design.
Part of why Apple's products are often praised for their design is that they do a few things really well and focus on those things, instead of trying to do absolutely everything. Consider the iPod, the iPhone, Apple TV, etc -- they're all pretty focused on doing certain things and Apple's really polished the experience. The Mac desktops and laptops kind of stretch this by allowing more things, but they still largely try to focus the user into certain workflows, via the plethora of apps that come standard with macOS and the vendor lock-in that they push.
Making a phone that can also be a full computer goes against these design principles. Apple's closed the gap a bit in recent years by making macOS and iOS a bit more similar than they used to be, but they're still pretty different. If you're on a M1/2/3/4/etc processor laptop and you've tried using an iOS-specific app (not ones that's designed for both phone and desktop) on it, you'll see some of those differences (interfaces tuned for touch are weird with a mouse, things are sized wrong for desktop, file restrictions can be weird, keyboard input can be lacking, etc etc etc), and it's not enjoyable. Going the other direction, the first thing that pops into my head is: how in the world would the mac desktop be represented on iOS? I'm someone who keeps a lot of files on his desktop, grouped in different sections of the screen for different reasons, and I have no idea how that would be represented on a relatively tiny phone screen (at least in a way that didn't destroy my intentional groups). There are other aspects of macOS that would prove tricky to have analogs on a phone screen, too, but this reply is already getting so long that very few will read it...
Now that's not to say that it's impossible. In fact it probably isn't. But there would be compromises (and those compromises would be on top of the compromises already present in iOS/macOS). To do it well, it'd be a much bigger project than most people realize. It's not just changing a few options and letting us use our phone that way. It'd be more akin to designing the first iPhone. Note that it's not just Apple who hasn't done this yet. Literally _no one_ has done it well yet. I truly hope one day Apple (or someone else, even) does it well, since that'll be a glorious day. But it'd be a huge project, so I'm not holding my breath.
It's scary, without the em dashes, and the rapid fire commenting of the account - who would ever realize this is a bot? Two easy to fix things, and after that it'd be very difficult to tell that this is a bot.
It's not a question of if there are other bots out there, but only what % of comments on HN right now and elsewhere are bot generated. That number is only going to increase if nothing is done.
The vast majority of websites you visit don’t have usable APIs and very poor discovery of the those APIs.
Screenshots on the other hand are documentation, API, and discovery all in one. And you’d be surprised how little context/tokens screenshots consumer compared to all the back and forth verbose json payloads of APIs
>The vast majority of websites you visit don’t have usable APIs and very poor discovery of the those APIs.
I think an important thing here is that a lot of websites/platforms don't want AIs to have direct API access, because they are afraid that AIs would take the customer "away" from the website/platform, making the consumer a customer of the AI rather than a customer of the website/platform. Therefore for AIs to be able to do what customers want them to do, they need their browsing to look just like the customer's browsing/browser.
Also the fact that they don't want automated abuse. At this point a lot of services might just go app only so they can have a verified compute environment that is difficult to bot.
That's true, and it's always been like that, which is why the comment that AI should be using APIs is already dead in the water. In terms of gating a websites to humans by not providing APIs, that is quickly coming to a close.
reply