Top
Best
New

Posted by sethkim 3 days ago

The End of Moore's Law for AI? Gemini Flash Offers a Warning(sutro.sh)
113 points | 73 comments
simonw 3 days ago|
"In a move that at first went unnoticed, Google significantly increased the price of its popular Gemini 2.5 Flash model"

It's not quite that simple. Gemini 2.5 Flash previously had two prices, depending on if you enabled "thinking" mode or not. The new 2.5 Flash has just a single price, which is a lot more if you were using the non-thinking mode and may be slightly less for thinking mode.

Another way to think about this is that they retired their Gemini 2.5 Flash non-thinking model entirely, and changed the price of their Gemini 2.5 Flash thinking model from $0.15/m input, $3.50/m output to $0.30/m input (more expensive) and $2.50/m output (less expensive).

Another minor nit-pick:

> For LLM providers, API calls cost them quadratically in throughput as sequence length increases. However, API providers price their services linearly, meaning that there is a fixed cost to the end consumer for every unit of input or output token they use.

That's mostly true, but not entirely: Gemini 2.5 Pro (but oddly not Gemini 2.5 Flash) charges a higher rate for inputs over 200,000 tokens. Gemini 1.5 also had a higher rate for >128,000 tokens. As a result I treat those as separate models on my pricing table on https://www.llm-prices.com

One last one:

> o3 is a completely different class of model. It is at the frontier of intelligence, whereas Flash is meant to be a workhorse. Consequently, there is more room for optimization that isn’t available in Flash’s case, such as more room for pruning, distillation, etc.

OpenAI are on the record that the o3 optimizations were not through model changes such as pruning or distillation. This is backed up by independent benchmarks that find the performance of the new o3 matches the previous one: https://twitter.com/arcprize/status/1932836756791177316

sethkim 3 days ago||
Both great points, but more or less speak to the same root cause - customer usage patterns are becoming more of a driver for pricing than underlying technology improvements. If so, we likely have hit a "soft" floor for now on pricing. Do you not see it this way?
simonw 3 days ago|||
Even given how much prices have decreased over the past 3 years I think there's still room for them to keep going down. I expect there remain a whole lot of optimizations that have not yet been discovered, in both software and hardware.

That 80% drop in o3 was only a few weeks ago!

sethkim 3 days ago||
No doubt prices will continue to drop! We just don't think it will be anything like the orders-of-magnitude YoY improvements we're used to seeing. Consequently, developers shouldn't expect the cost of building and scaling AI applications to be anything close to "free" in the near future as many suspect.
vfvthunter 3 days ago|||
I do not see it this way. Google is a publicly traded company responsible for creating value for their shareholders. When they became dicks about ad blockers on youtube last year or so, was it because they hit a bandwidth Moore's law? No. It was a money grab.

ChatGPT is simply what Google should've been 5-7 years ago, but Google was more interested in presenting me with ads to click on instead of helping me find what I was looking for. ChatGPT is at least 50% of my searches now. And they're losing revenue because of that.

mathiaspoint 2 days ago||
I really hate the thinking. I do my best to disable it but don't always remember. So often it just gets into a loop second guessing itself until it hits the token limit. It's rare it figures anything out while it's thinking too but maybe that's because I'm better at writing prompts.
bigbuppo 2 days ago|||
It's almost like there's an incentive for them to burn as many tokens as possible accomplishing nothing useful.
thomashop 2 days ago||||
I have the impression that the thinking helps even if the actual content of the thinking output is nonsense. It awards more cycles to the model to think about the problem.
wat10000 2 days ago||
That would be strange. There's no hidden memory or data channel, the "thinking" output is all the model receives afterwards. If it's all nonsense, then nonsense is all it gets. I wouldn't be completely surprised if a context with a bunch of apparent nonsense still helps somehow, LLMs are weird, but it would be odd.
barrkel 2 days ago|||
This isn't quite right. Even when an LLM generates meaningless tokens, its internal state continues to evolve. Each new token triggers a fresh pass through the network, with attention over the KV cache, allowing the model to refine its contextual representation. The specific tokens may be gibberish, but the underlying computation can still reflect ongoing "thinking".
yorwba 2 days ago||||
Attention operates entirely on hidden memory, in the sense that it usually isn't exposed to the end user. An attention head on one thinking token can attend to one thing and the same attention head on the next thinking token can attend to something entirely different, and the next layer can combine the two values, maybe on the second thinking token, maybe much later. So even nonsense filler can create space for intermediate computation to happen.
Wowfunhappy 2 days ago||||
Wasn't there some study that just telling the LLM to write a bunch of periods first improves responses?
mathiaspoint 2 days ago|||
Eh. The embeddings themselves could act like hidden layer activations and encode some useful information.
jgalt212 2 days ago|||
I hate thinking mode because I prefer a mostly right answer right now over having to wait for a probably better, but still not exactly right answer.
sharkjacobs 3 days ago||
> If you’re building batch tasks with LLMs and are looking to navigate this new cost landscape, feel free to reach out to see how Sutro can help.

I don't have any reason to doubt the reasoning this article is doing or the conclusions it reaches, but it's important to recognize that this article is part of a sales pitch.

sethkim 3 days ago||
Yes, we're a startup! And LLM inference is a major component of what we do - more importantly, we're working on making these models accessible as analytical processing tools, so we have a strong focus on making them cost-effective at scale.
sharkjacobs 2 days ago||
I see your prices page lists the average cost per million tokens. Is that because you are using the formula you describe, which depends on hardware time and throughput?

> API Price ≈ (Hourly Hardware Cost / Throughput in Tokens per Hour) + Margin

samtheprogram 2 days ago||
There’s absolutely nothing wrong with putting a small plug at the end of an article.
sharkjacobs 2 days ago||
Of course not.

But the thrust of the article is that contrary to conventional wisdom, we shouldn't expect llm models to continue getting more efficient, and so its worthwhile to explore other options for cost savings in inference, such as batch processing.

The conclusion they reach is one which directly serves what they're selling.

I'll repeat; I'm not disputing anything in this article. I'm really not, I'm not even trying to be coy and make allusions without directly saying anything. If I thought this was bullshit I'm not afraid to semi-anonymously post a comment saying so.

But this is advertising, just like Backblaze's hard drive reliability blog posts are advertising.

apstroll 2 days ago||
Extremely doubtful that it boils down to quadratic scaling of attention. That whole issue is a leftover from the days of small bert models with very few parameters.

For large models, compute is very rarely dominated by attention. Take, for example, this FLOPs calculation from https://www.adamcasson.com/posts/transformer-flops

Compute per token = 2(P + L × W × D)

P: total parameters L: Number of Layers W: context size D: Embedding dimension

For Llama 8b, the window size starts dominating compute cost per token only at 61k tokens.

lemming 2 days ago||
Can anyone explain the economics of Anthropic's Max plan pricing to me? I have friends on the $100/month plan using well over $800 of tokens per month with Claude Code (according to ccusage). I certainly don't use Claude Code as much if I'm not on a flat rate plan, the cost spirals out of control very quickly. I understand that a subscription makes for more predictable revenue and that there will be people on the Max plan not using Claude Code 24/7, but the delta between what the API costs and what using the Max plan with Claude Code costs just seems too great for that to be an explanation. I don't think that user/mindshare capture can fully explain it either, Code is free and the cost of switching to something else if pricing later changes is just too low. I don't get it.
ido 2 days ago|
We’re in an LLM bubble and their money is cheap as they’re drowning in investor money and have to spend it + show growth. If it doesn’t make economic sense you probably can’t count on it to last once the bubble bursts.
fusionadvocate 2 days ago||
What is holding back AI is this business necessity that models must perform everything. Nobody can push for a smaller model that learns a few simple tasks and then build upon that, similar to the best known intelligent machine: the human.

If these corporations had to build a car they would make the largest possible engine, because "MORE ENGINE MORE SPEED", just like they think that bigger models means bigger intelligence, but forget to add steering, or even a chassi.

dehugger 2 days ago||
I agree. I want to be able to get smaller models which are complete, contained, products which we can run on-prem for our organization.

I'll take a model specialized in web scraping. Give me one trained on generating report and documentation templates (I'd commit felonies for one which could spit out a near-conplete report for SSRS).

Models trained for specific helpdesk tasks ("install a printer", "grant this user access to these services with this permission level").

A model for analyzing network traffic and identifying specific patterns.

None of these things should require titanic models nearing trillions of parameters.

furyofantares 2 days ago|||
This is extremely theorycrafted but I see this as an excellent thing driving AI forward, not holding it back.

I suspect a large part of the reason we've had many decades of exponential improvements in compute is the general purpose nature of computers. It's a narrow set of technologies that are universally applicable and each time they get better/cheaper they find more demand, so we've put an exponentially increasing amount of economical force behind it to match. There needed to be "plenty of room at the bottom" in terms of physics and plenty of room at the top in terms of software eating the world, but if we'd built special purpose hardware for each application I don't think we'd have seen such incredible sustained growth.

I see neural networks and even LLMs as being potentially similar. They're general purpose, a small set of technologies that are broadly applicable and, as long as we can keep making them better/faster/cheaper, they will find more demand, and so benefit from concentrated economic investment.

fnord123 2 days ago||
They aren't arguing against LLMs They are arguing against their toaster's LLM to make the perfect toast from being trained on the tax policies of the Chang Dynasty.
furyofantares 2 days ago||
I'm aware! And I'm personally excited about small models but my intuition is that maybe pouring more and more money into giant general purpose models will have payoff as long as it keeps working at producing better general purpose results (which maybe it won't).
cruffle_duffle 2 days ago||
That’s just machine learning though!
sharkjacobs 3 days ago||
> This is the first time a major provider has backtracked on the price of an established model

Arguably that was Haiku 3.5 in October 2024.

I think the same hypothesis could apply though, that you price your model expecting a certain average input size, and then adjust price up to accommodate the reality that people use that cheapest model when they want to throw as much as they can into the context.

ryao 2 days ago||
I had the same thought about haiku 3.5. They claimed it was due to the model being more capable, which basically means that they raised the price because they could.

Then there is Poe with its pricing games. Prices at Poe have been going up over time since they were extremely aggressive to gain market share presumably under the assumption that there would be reduced pricing in the future and the reduced pricing for LLMs did not materialize.

simonw 3 days ago||
Haiku 3.5 was a completely different model from Haiku 3, and part of a new model generation.

Gemini Flash 2.5 and Gemini 2.5 Flash Preview were presumably a whole lot more similar to each other.

mossTechnician 2 days ago||
Is there an expectation Haiku 3.5 is completely different? Even leaving semantic versioning aside, even if the .5 symbolizes a "halfway point" between major releases, it still suggests a non-major release to me.
simonw 2 days ago||
Consumers have no idea what Haiku is.

Engineers who work with LLM APIs are hopefully paying enough attention that they understand the difference between Claude 3, Claude 3.5 and Claude 4.

cmogni1 3 days ago||
The article does a great job of highlighting the core disconnect in the LLM API economy: linear pricing for a service with non-linear, quadratic compute costs. The traffic analogy is an excellent framing.

One addition: the O(n^2) compute cost is most acute during the one-time prefill of the input prompt. I think the real bottleneck, however, is the KV cache during the decode phase.

For each new token generated, the model must access the intermediate state of all previous tokens. This state is held in the KV Cache, which grows linearly with sequence length and consumes an enormous amount of expensive GPU VRAM. The speed of generating a response is therefore more limited by memory bandwidth.

Viewed this way, Google's 2x price hike on input tokens is probably related to the KV Cache, which supports the article’s “workload shape” hypothesis. A long input prompt creates a huge memory footprint that must be held for the entire generation, even if the output is short.

trhway 2 days ago|
That obviously should and will be fixed architecturally.

>For each new token generated, the model must access the intermediate state of all previous tokens.

Not all the previous tokens are equal, not all deserve the same attention so to speak. The farther the tokens, the more opportunity for many of them to be pruned and/or collapsed with other similarly distant and lesser meaningful tokens in a given context. So instead of O(n^2) it would be more like O(nlog(n))

I mean, you'd expect that for example "knowlegde worker" models (vs. say "poetry" models) would posses some perturbative stability wrt. changes to/pruning of the remote previous tokens, at least to those tokens which are less meaningful in the current context.

Personally, i feel the situation is good - performance engineering work again becomes somewhat valuable as we're reaching N where O(n^2) forces management to throw some money at engineers instead of at the hardware :)

llm_nerd 2 days ago||
Basing anything on Google's pricing is folly. Quite recently Google offered several of their preview models at a price of $0.00.

Because they were the underdog. Everyone was talking about ChatGPT, or maybe Anthropic. Then Deepseek. Google were the afterthought that was renowned for that ridiculous image generator that envisioned 17th century European scientists as full-headdress North American natives.

There has been absolute 180 since then, and Google now has the ability to set their pricing similar to the others. Indeed, Google's pricing still has a pretty large discount over similarly capable model levels, even after they raised prices.

The warning is that there is no free lunch, and when someone is basically subsidizing usage to get noticed, they don't have to do that once their offering is good.

jasonthorsness 3 days ago||
Unfounded extrapolation from a minor pricing update. I am sure every generation of chips also came with “end of Moore’s law” articles for the actual Moore’s law.

FWIW Gemini 2.5 Flash Lite is still very good; I used it in my latest side project to generate entire web sites and it outputs great content and markup every single time.

antirez 2 days ago|
I think providers are making a mistake in simplifying prices at all costs, hiding the quadratic nature of attention. People can understand the pricing anyway, even if more complex, by having a tool that let them select a prompt and a reply length and see the cost, or fancy 3D graphs that capture the cost surface of different cases. People would start sending smaller prompts and less context when less is enough, and what they pay would be more related to the amount of GPU/TPU/... power they use.
More comments...