News for hackers. Stuff that matters.
Re:Nvidia deliberately kept GPU prices high through circular... by Aurornis on Saturday November 14, 2026 @08:58PM (#50503245)
> Engineering types and power users need not apply,but For Everyone Else, the "normies", Qwen Flash Next is probably "640k is enough for anyone" territory.

Why would any of those non-power-users buy a $2000 GPU that doubles as a space heater (seriously, try sitting next to one during long runs in the summer) and thousands more for a computer to host it, plus all of the trouble of setting it up, instead of a $20/month subscription?

I do have the hardware to run Qwen3.8-Flash-Next at home, and I still heavily use my subscriptions even for simple tasks. The experience of using the apps on my phone on the go is great and I always have the option to scale the model up to the task.

The people who are going to pay thousands of dollars up front to avoid $20 are power users and people who have different priorities and motivations, like independence or ultimate privacy. GPU prices dropping won’t change that much.

Also, you don’t need an RTX6000 for Q3.8FN. There are much cheaper options right now at lower speeds.

Re:Nvidia deliberately kept GPU prices high through circular... by komali2 on Saturday November 14, 2026 @09:31PM (#50503407)
> The experience of using the apps on my phone on the go is great and I always have the option to scale the model up to the task.

Just sidenote, this experience can be replicated with software like Paseo or some other options.

I usually have a frontier model orchestrating between dumber frontier, Chinese, and local models, using Paseo subagents, from my phone or whatever hardware I'm near.

Though I don't lean too much on the local models, I find them to simply be too stupid. They're fun for experiments though like quizzing them against Wikipedia or making them debate philosophy in loops.

Re:Nvidia deliberately kept GPU prices high through circular... by ericd on Saturday November 14, 2026 @09:34PM (#50503420)
[delayed]
Re:Nvidia deliberately kept GPU prices high through circular... by satvikpendem on Saturday November 14, 2026 @10:04PM (#50503578)
Tokens per second is slower on most local setups than cloud models.
Re:Nvidia deliberately kept GPU prices high through circular... by ericd on Saturday November 14, 2026 @10:07PM (#50503591)
We're talking about RTX Pro 6000s, local models on those are not slower than cloud models. With MTP, you're generally at 200-300 tps on one of the big MoE models in single stream in code generation (the small MTP models tend to be better at code token prediction than prose). The big cloud models are typically <100 tps.
Re:Nvidia deliberately kept GPU prices high through circular... by plorkyeran on Saturday November 14, 2026 @09:38PM (#50503448)
Even if you own a RTX6000 anyway it's not obvious that it's cheaper to run Q3.8FN on it than buying tokens. Google tells me that it can produce ~1 million tokens per hour and uses 450W of power, which for me comes out to 18¢ per million tokens. That's very similar to the price on OpenRouter. Maybe during the winter when all the waste heat cuts your heating bill a bit?
Re:Nvidia deliberately kept GPU prices high through circular... by ericd on Saturday November 14, 2026 @10:04PM (#50503582)
[delayed]
Re:Nvidia deliberately kept GPU prices high through circular... by milkshakes on Saturday November 14, 2026 @10:07PM (#50503590)
it's probably not cheaper, but economics are only one concern at play
Re:Nvidia deliberately kept GPU prices high through circular... by GenerocUsername on Saturday November 14, 2026 @10:22PM (#50503677)
Bingo. It's always cheaper to rent...
Re:Nvidia deliberately kept GPU prices high through circular... by zenoprax on Saturday November 14, 2026 @10:08PM (#50503598)
I did similar math for my card (9070 XT) and it is close to break-even on electricity cost alone nevermind the amortized cost and requirement to have a desktop for it to go into. The parallelization and faster speed of the cloud makes it feel more expensive but it's just doing the same work more quickly.

The strongest case for local AI is for anything requiring privacy/discretion but I don't think individuals will collectively prioritize this. Convenience is king for the majority followed closely by cost. Businesses who care about protecting their competitive edge seem like the more likely candidates for local AI but only if ZDR providers aren't good enough.

Local AI will probably end up where it is now: in the hands of hobbyists and/or privacy-minded people who already own a GPU for gaming.

Re:Nvidia deliberately kept GPU prices high through circular... by horacemorace on Sunday November 15, 2026 @12:22AM (#50504248)
This assumes status quo. The privacy part will become more important as subscriptions become more invasive, ad-ridden, and hostile, as it always seems to go. Local will catch on as hardware availability improves and the community converges on common solutions.
Re:Nvidia deliberately kept GPU prices high through circular... by visarga on Sunday November 15, 2026 @01:40AM (#50504578)
Local models can still surprise us, a small 27B could be much stronger in 2-3 years than today, while capability to run it would exist on more devices.
Re:Nvidia deliberately kept GPU prices high through circular... by zenoprax on Sunday November 15, 2026 @10:47AM (#50507808)
If a company (Apple?) can provide the illusion of privacy and seamless AI integration then they could capture the market of people that are at least aware of the issues.

Despite everything we know about the attention economy the self-hosted community is still a niche group of tech-minded hobbyists. Running anything locally is a commitment and adding expensive hardware to the mix is an even higher barrier.

I want people to care about privacy, technology, and interoperability but I'm just not seeing it happen.

Re:Nvidia deliberately kept GPU prices high through circular... by Aurornis on Sunday November 15, 2026 @02:54PM (#50510085)
> The privacy part will become more important as subscriptions become more invasive, ad-ridden, and hostile

I would bet good money that my $20/month subscriptions are not going to become ad-ridden.

If someone wants to use the free tiers, then nobody is going to be surprised if they see ads.

But these claims that my $20/month plan are going to become “ad-ridden” are at odds with what is happening. I’ve had at least $20/month plans from OpenAI and Anthropic for about as long as they’ve been available, and the quality and amount of inference I get from even the lowest tier are increasing all the time.

At this point, I can run Luna/xhigh on the $20 plan almost constantly (with brief breaks for tool calls doing actual work, not just token generation for fun) and it feels endless. In another 24 months we’ll be doing even higher quality outputs with smaller hosted models on faster cloud hardware and it’s going to be available from many different providers.

The doomsaying about “ad ridden” should be kept to the free plans. It’s just not going to happen on $20/month

Re:Nvidia deliberately kept GPU prices high through circular... by drnick1 on Saturday November 14, 2026 @10:02PM (#50503568)
> Why would any of those non-power-users buy a $2000 GPU that doubles as a space heater

Because, unlike the $20 Claude subscription, your own hardware will not respond "sorry, but I cannot do that" when you try to hack a device or game that you own.

Re:Nvidia deliberately kept GPU prices high through circular... by satvikpendem on Saturday November 14, 2026 @10:05PM (#50503584)
Open weight models are also increasingly starting to deny requests, like recent Kimi showed.
Re:Nvidia deliberately kept GPU prices high through circular... by ericd on Saturday November 14, 2026 @10:13PM (#50503637)
[delayed]
Re:Nvidia deliberately kept GPU prices high through circular... by girvo on Saturday November 14, 2026 @10:53PM (#50503839)
That used to be true. But now the Chinese model companies are just not training their open weight sets on cybersecurity tasks: there’s no censorship to remove, it just doesn’t have that specific ability.
Re:Nvidia deliberately kept GPU prices high through circular... by ericd on Saturday November 14, 2026 @11:16PM (#50503969)
Moonshot reportedly removed some training data, which dinged its cybersecurity ability somewhat, but Kimi still has pretty strong ability - they still reported >80% on their xdev-bench-hard, down from 90% before the data removal. DS 4.1 is supposedly quite good for it (85%), and Z.ai has been talking about how they're trying to improve GLM's cybersecurity abilities, so it doesn't seem like a broad trend.
Re:Nvidia deliberately kept GPU prices high through circular... by girvo on Sunday November 15, 2026 @05:24AM (#50505732)
I hope you're right, but the word on the street is it's part of why Qwen is releasing v4 later than they originally planned. That's only a rumour though, so maybe its complete nonsense!
Re:Nvidia deliberately kept GPU prices high through circular... by ericd on Sunday November 15, 2026 @09:25AM (#50507167)
Yeah, I hope so too!
Re:Nvidia deliberately kept GPU prices high through circular... by Topfi on Sunday November 15, 2026 @03:27PM (#50510360)
Ah yes, all those none-power-user-hackers.
Re:Nvidia deliberately kept GPU prices high through circular... by pkulak on Saturday November 14, 2026 @10:51PM (#50503822)
> instead of a $20/month subscription

And that $20 subscription isn't a good comparison. That gets you Astra or Opus. Better comparison is OpenAI's $8/month plan, which gets you Luna (which scores the same as Qwen on benchmarks). At that price, most of the cost isn't even the model; it's just hosting the infrastructure.

Re:Nvidia deliberately kept GPU prices high through circular... by aalimov_ on Sunday November 15, 2026 @12:54AM (#50504378)
With all of the money these companies owe I don’t think that $20 or even $100 monthly subs will stick around forever, but who knows.
Re:Nvidia deliberately kept GPU prices high through circular... by chii on Sunday November 15, 2026 @01:18AM (#50504468)
This is exactly correct imho.

The same thing happened with uber when they needed to start making money. It's only competition that stopped it from increasing more.

Re:Nvidia deliberately kept GPU prices high through circular... by wmf on Sunday November 15, 2026 @01:23AM (#50504494)
Ironically, going bankrupt would allow them to wipe out the debt, keep the GPUs, and offer cheaper subscriptions.
Re:Nvidia deliberately kept GPU prices high through circular... by coffeebeqn on Sunday November 15, 2026 @01:34AM (#50504548)
Who are we talking about? Pretty sure you’d have to liquidate your GPUs in this case
Re:Nvidia deliberately kept GPU prices high through circular... by karmakurtisaani on Sunday November 15, 2026 @03:38AM (#50505155)
Same level of understanding as thinking going to the town square and screaming "I declare bankruptcy!" frees you from your debt.
Re:Nvidia deliberately kept GPU prices high through circular... by mike_hearn on Sunday November 15, 2026 @04:11AM (#50505302)
I don't think so? A Chapter 11 is only possible if you can demonstrate clearly to the court that the underlying business would be viable without the debts. It also wipes out the investors.

AI startups are characterized by:

1) Market dumping. They are burning investor capital to buy market share.

2) A need to constantly raise money to pay for (1).

A Chapter 11 bankruptcy would make no sense in this situation. They haven't demonstrated that their business model is viable, so if a court ordered a Chapter 11 these firms would just immediately go bankrupt again, and that's usually a criteria to require a different and harsher kind of bankruptcy.

But even if for some reason a court did allow it, they'd still need to do emergency raises because their outgoings are so high. Debt has nothing to do with that, it's purely a matter of fixed costs and (in Anthropic's case at least) huge locked in contracts. But no investor is going to give money to a company that just permanently defaulted on them and ran to court to zero out the value of those shares. So they would be unable to raise and would immediately go bankrupt again.

Re:Nvidia deliberately kept GPU prices high through circular... by godspiral on Monday November 16, 2026 @10:24AM (#50518568)
Most of the "debt" for OpenAI and Anthropic is not debt, but purchase/rental obligations. These are senior to "real debt". The overwhelming problem is that if datacenter builders force bankruptcy takeover, the model assets are only worth $10-20b instead of $1T+ company, because the talent/hype is the company's value, and they all leave if their equity/stock options go down.
Re:Nvidia deliberately kept GPU prices high through circular... by csallen on Sunday November 15, 2026 @01:48AM (#50504609)
Then you just move once they're more expensive. They have essentially no lock-in.
Re:Nvidia deliberately kept GPU prices high through circular... by Aurornis on Sunday November 15, 2026 @02:57PM (#50510102)
Anthropic has so much cash they earn a billion dollars per year in interest.

The reason they’re entering compute contracts is because every dollar they spend on compute for serving their plans (inference) produces more dollars of revenue.

Models are getting better. Serving hardware is getting faster. Fabs are being produced to make more hardware.

Things will only get cheaper and more plentiful from here.

Re:Nvidia deliberately kept GPU prices high through circular... by jitbit on Sunday November 15, 2026 @11:15AM (#50508067)
I live in northern EUrope. I wouldn’t mind a space heater.
Re:Nvidia deliberately kept GPU prices high through circular... by godspiral on Monday November 16, 2026 @10:17AM (#50518485)
models that fit on a single gpu or 96-128gb ram + a 16-32gb gpu as per strata are indeed already good. ddr6 with PIM (matrix multiply) is a path to go without a gpu. It does take a crash on HBM pricing to make everything happen affordably.

Local vs rented vs token pricing all fall at similar rates, though.