Personally, I don’t think the general-purpose LLM as a standalone tool is long for this world, at least not in consumer-facing applications. I think when the economics make more sense, product designers will make things that people actually want to use that will pretty transparently handle whatever model interactions are necessary, when it makes sense. As a consumer, the last things I want in an interface are to a) be sycophantic enough to lessen my judgment, and b) be obstinate, obtuse, or argumentative, or generally just be something that I have to explain things to. I think a lot of tech folks are far more biased than they realize by the “ooh, neato” factor when imagining how nontechnical people might want to use things. And the weight of these tools just feels wrong for what a lot of people use them for: the thing that plays whatever music I feel like hearing absolutely does not need to be able to generate a volumes of fanfic about the movie that song was in. It’s abstractly impressive that something could do that, but it’s just not useful.

The problem (right now) is that Open Weight models depend right now on huge companies to spend billion of dollars to train and develop them, all backed up by their incentives and their state to support this, while essentially giving away their monetization path.

With open source projects, the benefit was that each individual could improve the complex system (e.g. Linux Kernel) interpedently, and over time the benefits accumulated. With models right now, there is just no way to do distributed training, or really, any large scale parallel way to improve them.

So whatever the short term strategy driving publicizing the model weights (e.g. potentially, to create a price war in order to put pressure on western companies and deprive them of the money they need), we can't ignore the fact that incentives and decisions could easily change in the future, and unless there is a way to truly decentralize models improvements - the party could stop at any time.

Another important thing that made software usage and education available for most of the world was piracy. I remember as a kid growing up in a developing country, any software (windows, office, Visual Basic, flash, dreamweaver, etc.) was less than 1$. That allowed me to try out and learn so many things on my own without paying a huge amount of money for the license. And I think this is true for most of the software developers of my generation who grew up in developing countries

"I think that in 10-15 years, we are going to have consumer PCs (and phones!) running models doing pretty much anything that frontier models can do right now."

I don't think it'll take 10-15 years. Gemma 4 31B in the 4-bit QAT is competitive with the frontier of less than three years ago and runs on any high-end 32GB gaming PC GPU or a large-ish Mac.

The question is whether the frontier will continue to get better at a rate that allows it to stay ahead of the two curves of availability of consumer hardware big enough to run somewhat larger models and the capability of small models to compete with large ones. When the bottom falls out and GPUs/RAM becomes affordable again, the size of what normal people have on their desk will trend quite a bit larger than today.

I think there's a future not too far from now, where a 120B model with really good reasoning and a large context, but limited knowledge (necessitated by being small, you can't fit the world's knowledge in 100 gigabytes), can substitute for a frontier model on almost any task, just by giving it access to web search and documentation for the thing you're trying to do. A 256GB unified memory machine with sufficient memory bandwidth would comfortably run that 120B model.

> I think that in 10-15 years, we are going to have consumer PCs (and phones!) running models doing pretty much anything that frontier models can do right now

10-15 years? The current rate is closer to 10-15 months.

15 months ago, the top model on the Artificial Analysis index was GPT-o3. It scores 30 on the Artificial Analysis index.

Today, you can easily run Qwen 3.6 27B on a variety of consumer hardware. It scores 37 on that index.

Here are a number of open weights models that you can run locally compared with the frontier class models from 7 to 15 months ago: https://artificialanalysis.ai/?models=o3%2Co3-pro%2Cclaude-4...

I've run all of these models on my laptop (Strix Halo, 128 GiB of unified RAM); the bigger ones, like MiniMax M2.7 and DeepSeek V4 Flash, need to be done at fairly aggressive quants that will certainly lose some performance and not quite hit the performance of the unquantized models. But still, it's definitely the case that you can run models that are competitive with the frontier models of 10-15 months ago on consumer laptops.

Heck, just announced though the weights haven't yet been released for independent confirmation is MiniCPM5-2B, a 2 billion parameter (small enough to run on your phone) model, that according to their benchmarks has performance competitive with GPT-4o, a frontier class model from 2024.

https://nitter.net/i/status/2079088670804767114

So that's around 1 year for frontier to consumer device class, 2 years from frontier to phone.

Now, this kind of rate won't necessarily keep up; it's possible that local models will hit a performance ceiling before frontier models do. There's only so much information you can cram into a certain number of bytes, and the AI boom is causing hardware prices to skyrocket so keeping consumer hardware from advancing quite as fast as it had been.

> - PCs destroyed minicomputers.

What's weird is that with "store your everything in the cloud and pay a monthly recurring subscription", we have now regressed to a 1960s/1970s timesharing revenue model for individual workstation computers.

The default new factory out of box workflow for "enrollment" in google services, iCloud or Microsoft-everything on a new ios, macos, windows or android personal computing device is clearly designed to sign people up for subscriptions.

And same general idea of "move all your servers to the cloud" recurring revenue for what is effectively the same as mainframe timesharing for key business functions, by renting VMs in GCP, Azure, AWS in perpetuity.

Yes, you can still use your desktop or laptop PC in 2026 with zero external third party subscriptions (other than maybe your residential home ISP), but how many non-tech people actually do so now?

I do agree that Chinese open-source models are going to play a bigger and bigger role in the entire ecosystem moving forward, but I don't agree with you in the sense that they are going to eventually "win."

Just because they are cheapp doesn't mean they automatically win. You've picked a lot of great examples, but there is still a little bit of cherry-picking.

One clear outlier is the iPhone, which coexists with Android globally. Even though the iPhone is the leader in the US, and globally Android has the majority of the smartphone market share, they still cater to different price points and different ecosystems, and generally the iPhone has better margins.

i believe American frontier models like from Anthropic and OpenAI are still going to thrive, and coexist with Chinese models. They are just going to cater to different customers and different use cases.

100x this it is why all of the AI giants are going to fail. They are too big and inefficient to scale properly. This is why Google is just casually taking its time in AI and not racing to a finish line. AI is essential but if it already does most things good enough then it can take longer to make it more efficient.

What about AWS, Azure, cloud computing?

What's interesting/funny is that the American LLM companies took from the public domain and copyrighted work to close all that content into a box they charge for.

Then the Chinese took the distilled stuff out from that box and released it into the world for everyone.

> Put all of these trends together, and I think that in 10-15 years, we are going to have consumer PCs (and phones!) running models doing pretty much anything that frontier models can do right now.

Phones are constrained by battery power and memory does not shrink as fast as CPU/GPU, so unless there's a battery breakthrough and/or memory breakthrough, you're not fitting 100Gb of RAM on your phone in 10 years.

Absolutely in a Mac Studio equivalent.

LLMs have emergent capabilities when they get smarter. So who knows how insanely big frontier models might be at that time, or what their capabilities may be.

> Put all of these trends together, and I think that in 10-15 years, we are going to have consumer PCs (and phones!) running models doing pretty much anything that frontier models can do right now.

Not likely. The last 50 years had Moore’s law growth in compute. That’s over. Frontier models are roughly compressed all written text and a large part of images. Those don’t compress forever, and likely not a ton more than now.

Inference requires touching a significant of that per token.

All of these are up against fundamental limits, more or less.

> Mainframes survive, but serving a much tinier portion of the market than they used to.

I would argue mainframes rebranded to "cloud" which is ubiquitous and more people interact with this computer than any other type of device... only difference is that it's a browser instead of a terminal

Define "winning."

Open source is cheap, yet its operating systems are the least-popular. But their existence is critical to a healthy market.

It's not zero sum.

What you're describing is how things become commoditized, but many companies are excellent at ensuring they aren't seen as commodities

I think this is all true, but that unlike with Moore's law and improved PC tooling and capabilities, we also have essentially existing biological evidence that there should be a way to create much better intelligent systems in terms of training, memory, and efficiency. With classic PC evolution we didn't even have that evidence but still could make a relatively strong inference (Moore's law). But here we basically have evidence that there can be something much improved and know that it's only going to take research and discovery to figure it out, not new hardware processes.

> Put all of these trends together, and I think that in 10-15 years, we are going to have consumer PCs (and phones!) running models doing pretty much anything that frontier models can do right now.

At current pace, we'll have open weight LLMs with frontier intelligence in 6-12 months. The constraint is RAM - both for the model and the context. It's likely that distillation and quantisation and TurboQuant will significantly reduce RAM requirements. I think we'll have Opus 4.8-like performance on 64GB of RAM in two years.

Of course, by then, frontier intelligence will be god-like.

The value of an LLM is the dynamic reasoning you get out of it and the cost to execute on that.

I see two forces working against this that proprietary models will always have over an open source model.

1. The biggest is content licensing. Content is quickly becoming gated by systems at the front of their load balancers, completely changing the social contract of the Internet. What used to be a quick google search for recent facts that lead me to places like reddit or twitter, is now completely walled off if you're not physically at your browser and using an IP address from a last-mile provider.

LLMs have pre-trained on the bulk of the information up to 2024/2025, but over time that will be more and more out of date.

Anthropic, OpenAI and Google will all have to pay for access to a lot of this content refresh going forward, and it does make a material difference in the output you get.

2. Liability is the other. A corporation can look at a contract for model access and see one that provides uptime guarentees, content infringement promises and model safety, and pick the contract that shields the corporation from the most liability. A 3rd party hosting platform like fireworks.ai that hosts open weights models won't provide any of that at all. They will simply bill you for time spent on their hardware and make promises that they won't log or inspect corporate traffic.

> Put all of these trends together, and I think that in 10-15 years, we are going to have consumer PCs (and phones!) running models doing pretty much anything that frontier models can do right now.

People overestimate what can happen in a year and underestimate what can happen in 5.

I'm betting that increased model efficiency and hardware optimisations will get us there a lot sooner. Biggest hurdle would be the memory prices though, if those do not drop back down it might take 15.

I'm not clear what you mean with "PC office productivity software" but it seems like Microsoft Office is the winner there. Isn't that professional?

Training cost is actually not that high — it’s fixed and amortizable across the lifetime of the model. Inference is expensive, and open weights don’t solve that problem — in fact, they might even encourage it, since a high cost of entry means consumers will pay for inference directly from the labs anyway.

Unfortunately it seems likely the winner will be the cloud providers. If anyone can run inference on open models, then profit will flow to the vendors who can afford the capital to run them. That’s the CSPs.

(It’s basically the same business model as pharmaceutical R&D, but the major difference is that nobody has even talked about patenting the models like a pharmaceutical company patents each new drug. I’m surprised about that, tbh — why give all the leverage to the cloud platforms? They aren’t training frontier models…)

> free and low-end eventually wins

Apple, the world's second most valuable company, seems like a counterexample.

> - PC office productivity software destroyed expensive professional products.

I agree with the lesson too. Just to be precise, wouldn't the current model war be more akin to open-source office suite versus MS office suite? If so, then the cheaper option didn't really win. That said, the open-source alternatives didn't really feel the same as MS Office, and it took them a long time to reach the feature parity (or did they ever?). In contrast, the open-weights models are getting close enough to the SOTA models, and users can easily switch from one to another without feeling any difference for mojority of the tasks.

> in 10-15 years, we are going to have consumer PCs (and phones!) running models doing pretty much anything that frontier models can do right now.

The way things are going with regards to RAM/storage prices, I highly doubt that anyone but the richest among us will be able to afford them.

Phones are already running models locally which can be used in the field for specific use cases. Maybe not for frontier coding just yet.

Also you don't need to be connected to the network to use a local AI in many instances. If all mobile apps were done with a local-first approach, then you could use a local AI to query your emails, lookup already visited pages, summarise recently received documents, and lots more. Lots of apps could use an inbox/outbox approach for receiving and sending updates instead of relying on the network at all times. And this pattern could be greatly leveraged by local agents.

> Put all of these trends together, and I think that in 10-15 years, we are going to have consumer PCs (and phones!) running models.

I love the idea of SaaS offering these at lower rates today integrated into what ever you do and be 100% private. But I think the key challenge to mass adoption is productizing them in a way which makes sense for people to pay money for. As a commodity a local model is useless unless combined with some capabilities important to me. A PC is inherently useful because of so many applications offered on it on it. How local LLMs would be useful as a product that is useful for mass market is not yet proven.

How will open-source and open-weight models continue to thrive after financial incentives die off? Surely open models will suffer from outdated knowledge cutoffs if noone will pay for model training?

That would sound very reasonable except this is the same what people said first about Windows and then Android crushing Apple.

And yet it's Apple that controls the top of the market and has the best margins in the business.

This is the same position OpenAI and Anthropic have right now.

Could this market be different? Maybe. But the status quo could be preserved as well.

> free and low-end eventually wins

Not in SaaS which is what LLMs are. You can get VMs for much cheaper than AWS, Microsoft, and Google offer them but large companies (and startups) are happy to pay a premium for the support, reputation, and reliability that they perceive those companies as offering. Same thing for some of the managed database providers who are effectively selling a very heavily marked up version of postgres.

> The high price, and social pushback, mean that the American companies producing these models are precarious

I doubt it. The models really aren't that expensive when you look at what they can do. Fable is probably at least as good as the average software engineer and costs $50/wk on the max plan vs a software engineer who would cost closer to $4000 a week. The real money is probably in selling to enterprise vs consumers (Google has best route to making money from consumers since they can do what they did with ads and search to LLM queries).

It seems unlikely to me that US companies will send important corporate data to models controlled by a Chinese company as well.

A counterpoint would be all are chip fabs are in Taiwan right now due to huge investment. And there are lower end chip fabs around the world, but they have not cracked the major market.

Apple seems a counter example, no?

I think you are right. One small exception I can think of is Microsoft Office still crushes Libre Office.

if you look at how GPU memory grew in the last 15 years, it's about 10x. Sadly, 10x from today doesn't get us to a typical frontier model size of today which is a quickly moving target. some other advancement needs to happen to get us another 10x both in memory/compute requirements, and also power requirements.

> The lesson of the last 50 years of the computer and software marketplace is that free and low-end eventually wins.

The parent comment cherry-picks evidence. There are plenty of counter-examples:

  * Office productivity suites
  * Search engines
  * Email services
  * Cloud services
  * Accounting software

The biggest exception is cloud. Big cloud carries an insane markup (bandwidth is like 10000X!) and everyone runs on it.

The strategy there is false openness where deployment complexity is the real proprietary moat. Sure Linux, Docker, Kubernetes, Postgres, and all the other standard tools in the box are open source and free, but they're also arcane and complex to run and hard to make fault tolerant. So you're lured in by "open" and then locked in via a kind of "death by a thousand cuts" complexity moat.

(Personally I hold the view that complexity and arcane-ness beyond a certain point is indistinguishable from closed in practice. Open source that's really complex and hard to run is not open in any meaningful sense.)

AI may not admit that kind of moat though, because AI is very good at slicing through that kind of thing. You can prompt a model to make itself compatible with another model or to change code to make it compatible. There's no moat because the moat bridges itself.

This is not really true considering Apple and nvidia are two most successful hardware companies, and they are notoriously closed. Not to mention microsoft, oracle they are all pretty closed.

You forgot smartphones, where low-cost did not win out. It led to low margins for the Chinese firms and eventually left them unable to invest properly in key markets. They may still hold marketshare, but in terms of profits, falls well short of Apple and Samsung.

I can see a lot of parallels here. Model performance doesn't matter if you can't make the system commercially sustainable.

I agree, although I think it will be a lot sooner than 10-15 years. I'm running local AI right now and it's definitely not production grade yet, but it's surprisingly good. Speculative prediction that I probably shouldn't make: when the bubble pops, depending on when it pops, RAM prices might drop a lot. I could foresee these companies having produced a lot of RAM that suddenly doesn't have a buyer. (I know high bandwidth memory is different, but I imagine there are companies that will want to take advantage of that)

The obvious counterpoint: Apple captured the majority of the US smartphone market.

Libreoffice

> The lesson of the last 50 years of the computer and software marketplace is that free and low-end eventually wins.

Is that actually true? There are very large markets that make a lot of money from paid software. And I would honestly prefer actually paying for software rather than constantly dealing with "not a bug" or "PRs are welcome".

Except cloud services won over local-first apps.

> Put all of these trends together, and I think that in 10-15 years, we are going to have consumer PCs (and phones!) doing

I'm not even sure in 10-15 years whether we're still going to have consumer PCs, or PCs at all.

That’s not what this is. It’s industrial dumping applied to software. China has successfully applied this strategy to become the manufacturing workshop of the world.

If china is subsidizing training they diminish their off-shore competitors expectations of a viable return on investment. It’s trade-war behavior.

While I generally agree there's some nuance here and that is that there really are few new ideas. Old ideas just get recycled.

For example, mainframes and minicomputer. Yes they were displaced by PCs. But what is cloud computing if not mainframes 2.0?

I do agree that in the next 2-3 years we're going to see real growth in local LLMs as the hardware becomes more accessible. It won't even necessarily be cheaper because data centers can run 24/7 and have cheaper cooling and electricity. It'll be done for privacy because your prompts and responses are themselves a commodity to AI companies and they live under a legal grey cloud. For example, does AI usage break attorney-client privilege? There are lots of opinions on this but it hasn't been tested in court.

paranoid me feels like the artificial gpu/ram/ssd shortages are a plot by the VCs to forcefully reclaim central control via new age mainframes

one certainly cannot buy a PC for cheap anymore

Windows - pfff, Microsoft charged companies like Dell for an install on computers they didn't even install it on !!!

in 15 years we might all be fighting terminators

instead of 15 years I think it'll be more like 1.5 years.

I wouldn't be surprised if apple were shipping 512 GB unified RAM macbooks before 2030 and that would be standard issue for folks to use local LLMs for their daily work

> The lesson of the last 50 years of the computer and software marketplace is that free and low-end eventually wins.

Except, uhm, for ..you know, that one company that hit a trillion cap

But you're right: Just like how million dollar computers with 1 bit of RAM performing 1 operation a second and taking up a colossal cave were replaced by $1 laptops with a zillion zekabytes running at a trillion hertz (exact values may vary),

the sprawling data centers of today with a quadrillion GPUs powered by black holes will get replaced by breakthroughs in hardware and most importantly, algorithms:

The human brain is proof right here that intelligence doesn't require dinosaur-sized hardware or eat half the sun every second.

I actually wonder if we're seeing the limits of discrete binary logic: Maybe it's high time to give analog ternary and all that funky jazz an honest try :)

umm do Mac vs pc and Iphone vs android

[dead]