You didn't answer the question, you answered a different question: "why would someone want to work, just in general?" The question that was posed was, why would someone who has already chosen to retire, or who is already fully occupied, or who is sick, want to work?
>You can't create a tiered paid plan for the whole world that fits everyone.
I mean, you can. Electricity is already sold that way. Subscribers with uncharacteristic usage spikes don't get blackouts, they get a slightly larger bill, and perhaps get moved up a tier.
Very valid. My comment was fixated around the fact that big tech has the addiction to have subscriptions for everything. It's common that you provide generic subscription plans for the masses and supply "call us" custom plans for the specific (usually corporate) needs. If anthropic doesn't provide that or vibe coders are too cheap to do that, then those are issues, but the subscription models are itself valid. It is certainly misleading to a degree, but we've stopped complaining about this a while ago.
It's pretty stupid because as others in this thread have pointed out it's already not a flat plan. Even from their side it makes zero sense to bill things this way rather than based on usage. It's not like a VPS where your VM shares the hardware, which consumes electricity more or less regardless of what you use the machine for.
The difference being that an LLM request is not an operating system. Since they're compartmentalized and ephemeral, you can very easily distribute requests among your available hardware so that you can switch off machines during periods of low activity.
But that would be using (a special Claude code version) of the API; as it stands now, I have tried the current api for fun and I hit $200 well within an hour. So if they would charge for real use, no one would use it as there are competitors who have less harsh limits with tier plans still. If all go away then I will be running open models on vast.ai or so as those are now viable (been testing with glm 5 and it's great for coding). So tier subscriptions cannot go away as it will end those companies fast.
>If gyms faced a situation where people would go and spend 18 hours working out every day for a month, they would probably change how they billed things.
Not the best example. The upkeep cost of a gym is pretty flat regardless of how much people use the facilities. Two people can't use a single machine at the same time make it wear out twice as fast. The price of memberships is not correlated to usage, it's inversely correlated to the number of memberships sold.
Two people can't use a machine at the same time is the issue. If you have 50 machines and 200 customers all of whom want to be in the gym 18 hours per day that's quickly going to lead to cancelled subscriptions. Now you need more space and machines or some other way to balance things.
Agreed, but it's an indirect causal link, not a direct one. If the demand far outstrips the possibly supply the demand will have to go down, and it can either go down by people accepting that they can't be in the gym as much time as they would like, or as you say by memberships being cancelled (in which case the price may go up or something else might change).
>Two people can't use a single machine at the same time make it wear out twice as fast
The machine doesn't care about the number of people using it. If it's constantly being used, it will wear out faster. You are conflating "we price based on expected under-utilization" with "costs don't scale with usage." Those are different things.
The inverse correlation you talk about isn't relevant here - People buy gym memberships intending to go, feel good about the intention, and then don't follow through. The business model is built on that gap. That's pretty specific to fitness and a handful of similar industries where aspiration drives purchase.
Anthropic doesn't sell based on a "golly gee I hope people dont use this" gap - they sell compute. Different business.
> Anthropic doesn't sell based on a "golly gee I hope people dont use this" gap - they sell compute. Different business.
There is nothing anywhere hinting at that.
They don’t sell compute. They sell a subscription for LLM token budgets that they hope people don’t use because the compute is vastly more expensive than what they charge or what users are ever willing to pay.
Especially with enterprise subscription plans the idea is for customers to never utilize anywhere close to their limits.
>If it's constantly being used, it will wear out faster.
Yeah, but there's an absolute limit to that, beyond which the cost doesn't keep increasing. Beyond that point, the QoS goes down (queues).
>You are conflating "we price based on expected under-utilization" with "costs don't scale with usage."
I'm not conflating anything, I'm responding to what you said:
>If gyms faced a situation where people would go and spend 18 hours working out every day for a month, they would probably change how they billed things.
Why would a gym need to change how they bill things if all their customers were aiming for maximal utilization, when their costs would barely see any change? I doubt your typical gym operates on razor-thin margins.
Gym costs absolutely scale with usage. Equipment wears faster under heavier use. Cleaning and maintenance staff hours scale with how much the facility is used. Consumables like towels, soap, and chalk go faster. HVAC runs harder. The reason gyms can offer flat-rate pricing is that they bet on under-utilization, not that costs are flat.
Setting that aside, even if we accept your argument that gym costs barely scale with usage, then that makes gyms a bad comparison case for Anthropic, whose costs directly scale with usage. You can't use the gym model to defend Anthropic's pricing decisions if the two cost structures are nothing alike.
I'm arguing that both gyms and Anthropic have usage costs that scale with usage, but gym business model assumes a large margin of under-utilization and there's a hard cap to "power user" - I think both of those extremes don't apply to Anthropic's situation. Under-utilizers aren't paying for AI they have a free tier. There's also a natural ceiling on how much any one person can use a gym. There's no equivalent constraint on API usage.
> The reason gyms can offer flat-rate pricing is that they bet on under-utilization, not that costs are flat.
Yes. In fact i remember hearing about a gym which offered a flat-rate pricing model but explicitly excluded certain professions from partaking in it. I remember the deal was excluding police, bouncers, models, actors and air stewardesses. They had a separate more costly tier for these people. (And I think i heard about it from the indignation the deal has caused online.)
> Under-utilizers aren't paying for AI they have a free tier.
Sure they do. Free tiers suck. I may not always need to use AI, but when I need it, I don't want to immediately get hit by stupidly low quotas and rate limits, or get anything but SOTA models.
Nitpick: What you're describing is the disk cache. If a process requests more memory than is free, the OS will not page out pages used for the cache, it will simply either release them (if they're on the read cache) or flush them (if they're on the write cache).
Still the best CRT simulation I've seen is in an X screensaver called XAnalogTV. It simulates both CRT artifacts as well as NTSC channel cross-talk and analog interference. It amazes me that still no one has produced a portable version.
You forgot one thing: flickering. At 60 Hz, a CRT is murder on the eyes. A few years ago I used a CRT for the first time in like ten years and my eyes hurt almost immediately.
I was never incredibly disturbed by 60Hz though I did notice it.
You reminded me of something I had forgotten though — remember when 100Hz / 120Hz TVs first became a thing? That I noticed!
I think most of my PC CRTs ran at 72Hz / 75Hz IIRC. At least with the monitor I had I remember pushing it to 90Hz but that would add bluriness / lose clarity so I never used it at that rate.
>But a CRT isn’t a camera filming the world. Its a physical device that generates an image as an output of physical process. [...] That’s not a post-process overlay or filter effect, its an entirely different mental model of what it means to draw or render an image. I think this is why I struggled when trying to bolt this onto a modern engine. The foundations between the two models is just so fundamentally different. At this point, I was already beginning to consider my options. I was half inclined to give up.
An LCD or an OLED are also not cameras. I honestly don't understand what insight this person believes they've stumbled upon.
This is also very mystifying:
>The frame is never a single instant, its a culmination of integrations over time.
Strictly speaking, a CRT doesn't understand frames. It just fires whatever intensity of electrons is indicated by an analog signal at any given time as the magnets steer the beam across the screen in whatever pattern has been designed into them. If the tube is controlled by a digital source, there will likely be some kind of framebuffer of some size somewhere on the pipeline that stores at least a full scanline, and nowadays invariably a complete frame, so a DAC can convert the values in it to the analog signal expected by the gun.
The entire article supposedly addresses the "why", but after getting to the end, I still don't understand the why. What's wrong with Unity or Unreal architecturally that this guy's engine addresses?
From my understand after reading, he's suggesting that Unreal and Unity's post processing are just applying effects to a camera/rendered frame, when what he wanted to do is simulate the CRT itself across the renderer to the frame that hits the swapchain.
But that's nonsensical. The CRT doesn't see the graphics pipeline of, say, an SNES, it just sees an analog signal. The graphics processing is done in the digital realm, not in the analog realm. If you want to simulate a CRT, all you need is a physical model and a digital image to display, which can come from Unreal, Unity, or whatever any other engine or program or whatever. It makes literally zero sense to write an entire engine to implement a CRT simulation.
I didn't suggest it makes any sense, just explaining, based on my understanding of the article, why he analogized the game engines to renderers to cameras.
Yeah I’m with you. Hate to assume such things but with how much AI spam is out there on programmer blogs these days I kinda just give up reading the blog post once something becomes confusing. Most of the time there’s not any insight to be learned by investigating deeper.
This one also has a lot of Its not X, its Y type phrasing
As you said, the CRT just receives the frame data and turns that into a visible image. This means you can simply build a filter that transforms the final frame buffer by simulating the physics of a CRT.
A filter is not a game engine and a game engine is not a filter.
Building a custom game engine for CRTs represents a fundamental misunderstanding of how a CRT works and what the responsibility of a game engine is.
Mmm, while this person's articles are clearly AI written, they do make some sense. Their renderer keeps samples the previous frame to achieve the effect, which of course is totally possible to do in Unreal or Unity but they also seem to have their own lighting and PBR models, which might be a bit harder to achieve.
>Lighting systems are designed to remain readable under CRT-style color quantization. Sprite and mesh pipelines emphasize bold shapes and strong contrast. Even debugging tools in the engine, like the grid overlays and scene visualization systems, exist partly to help developers maintain spatial clarity and composition.
This is AI nonsense but it could be a summarisation of something real.
It's not whataboutism, it's a legitimate question. How does it increase safety on the road to reject local SSH connections by a dumb user, when that same user can mess with the car physically?
Simplest example: a driver could probably disable attentive driving checks by pasting a script in from a web search in a few minutes. Nothing like an inattentive 3750 lbs weapon.
Yeah and they could hire a professional driver or a engineer and IPO for billions a life sized driving AI powered crypto robot too. Look, like clearly google + ctrl-v scripting or running an one click deployment exe on your computer on a whim is different than physically ordering/picking up something and then installing it into a vehicle?
Of course they're different, but you're trying to argue that the former takes objectively less effort than the latter, and it doesn't. One or the other may take less effort depending on who you are and what you know.
reply