To add to this, Fireworks (and many competitors, e.g. enumerated here[0]) offer zero-data retention inference, to anyone. DeepSeek cloud does not. It's part of their terms that anything you send them is fair game for them to train models on.

[0] https://openrouter.ai/deepseek/deepseek-v4.1-flash

(The badges next to companies indicate their privacy terms, i.e. green/yellow/red)

It’s pretty clear despite what they want you to think Anthropic and OpenAI train on everyone’s data who has a subsidized plan. Only way you get ZDR is by paying API prices. Also note that it is zero retention, not zero training.

> "Also note that it is zero retention, not zero training"

The "green" badges explicitly say the opposite:

> "Private: this provider does not train on prompts and does not retain prompt data."

I.e. for the example of Fireworks, this is the precise language they use to describe ZDR:

> "More technically: prompt and generation data exist only in volatile memory for the duration of the request. If prompt caching is active, some prompt data (and associated KV caches) can be stored in volatile memory for several minutes. In either case, prompt and generation data are not logged into any persistent storage."

https://docs.fireworks.ai/guides/security_compliance/data_ha...

How is it clear? I don't fully trust them either, but what indications are there?

Spoiler, it's not clear by any means. There's no evidence to back up the claim, it's just their feelings... which is fine, but I definitely wouldn't call it clear.

My slight suspicion is that the that somehow train on derived data from the sessions. In a way that they are technically in the clear.

If one is not American, why prefer American? Chinese fine by me, I trust the US providers just as little.

There's also European providers for these models. More than one!

Could you recommend me a few?

Scaleway, Mistral, Lyceum, Verda, Silo AI, Siili, Upcloud, CSC, Gcore, Nebius...

I haven't found that Mistral serve DS.

They do not yet. They are testing it but they go with GLM right now.

I’ll add Melious.[1]

I feel like I’m getting good value out of my subscription and they have the cheapest GLM-5.3 API rates I could find in Europe. I moved to them recently from Mistral for the better prices.

They do have a slightly different daily inference capacity allowance, but it’s actually not too bad once used to it. I was forced to up my game and figure out what GLM-5.3 Flash and Deepseek-4.1 Flash could actually do for me, rather than just throwing everything at GLM-5.3. Quite happily chugging along with a nice balance across them now.

I am hoping Melious make the new DS 4.1 Flash 1017 available soonest.

[1] https://melious.ai

Biased obviously, but sference.com. We deployed the new checkpoint immediately. Lmk if you have any thoughts, I can send some credits...

There's even a European OpenRouter competitor

https://www.edenai.co/

Also https://cortecs.ai/detailedServerlessView/deepseek-v4.1-flas... (I like their site bit more over Eden)

Thanks, i didn't know about these guys. Is their pricing competitive?

you can check for yourself

e.g. https://www.edenai.co/models/deepseek-v4-1-flash

Which ones? And how is the price?

I support people making choices with full knowledge of what they are doing and what the consequences are, as much as it is possible. What I oppose is making choices without such knowledge.

I figure (hope) buying from DeepSeek directly benefits them the most so they can release even more improvements and open-weight models.

Plus only one country has threatened to annex mine and it wasn't China.

Assuming canadian or dane, but maybe there are other places he’s threatened. There are canadian inference providers running deepseek, and definitely eu ones idk about ones based in denmark.

The economic sovereignty of any nation, but especially smaller ones like canada and denmark, is harmed by sending our money out of the country when there are options available in country.

> The economic sovereignty of any nation, but especially smaller ones like canada and denmark, is harmed by sending our money out of the country when there are options available in country.

I respectfully disagree. I subscribe to the comparative advantage economic principle, so I'm fine with international trade.

What I am saying isn’t inconsistent with comparative advantage. International trade is fine, but only when it’s not a commodity easily obtained inside your own nation. Open model inference is a commodity. There is no real lost opportunity cost from using canadian owned GPUs or chinese owned ones. In fact, as a country with abundant cheap power, and physical space, canada’s probably a better place than many to run inference. That’s what I meant by “when there are options in country”.

I hear you, but as I said, I hope that buying from Deepseek gets more Deepseek. Maybe they get the same amount (or more) from a Canadian provider, and if so I might reconsider.

The 'clean power' would factor in a well, but I haven't really looked in to it. I know China is installing a ton of renewable and has a carbon market while Canada's building a new pipeline and building Alberta data centres running on natural gas, so I'm not willing to take on faith that switching is cleaner.

Fair point. Personally i don’t see value in closed models, be they chinese or american. So, it’s pretty much against my religion to pay the model creator to provide inference. Even a situation where model makers license models to providers seems better than deepseek hosted deepseek provided models.

Also, re: power. I said canada has lots of cheap power, clean was your word.

RE: clean power - China is by far the biggest leader in getting renewables off the ground, but they are also still comfortably the biggest polluter, so it's difficult to compare.

The unfortunate truth behind at least the solar PV boom in China is that it's driven in large part by the exploitation of Uyghurs in the Xinjiang region. One of the reasons they can sustain the pace they're going at is because while nowadays PV modules exports increasingly gravitate towards offshore fabs (Vietnam, Malaysia), domestically all the biggest PV players make full use of modern slavery within Xinjiang's re-education camps.

more on the topic: https://www.shu.ac.uk/-/media/home/research/helena-kennedy-c...

>International trade is fine, but only when it’s not a commodity easily obtained inside your own nation

It might be a minor nitpick, but that part is just not how the macroeconomic principle works. Abundance of a good/service within a country is irrelevant to the concept of a comparative advantage - this is counter-intuitive but but that was David Ricardo's entire point. It's not about whether production is more efficient domestically, it's whether comparative allocation of resources between multiple countries is more efficient by specialization within the different goods.

Even if you can produce "good 1" cheaper in Country A vs Country B, if Country A can instead produce "good 2" that Country B cannot, it's better for Country A to produce "good 2" domestically and import "good 1".

Now, I still agree with the original notion. It's unnecessarily risky to use services from "democratically-challenged" nations (especially imperialist ones), when safer options are available. But there's no need to twist macro 101 concepts to justify that view. Modern trade policy takes more into account than basic market efficiency.

> There are canadian inference providers running deepseek

As someone who's in the market for Canadian inference (data sovereignty reasons), the pickings are slim for DSF 4.1 man.

China has annexed (or it claims so) several countries already, so it sounds like you're only concerned with yourself - probably not what you wanted to achieve.

(But I'd concede that in the field of genocide China doesn't come up as bad)

It surely does. Official DeepSeek provider logs and trains off your data.

There are multiple others in EU, and in US. Cheaper, more expensive, slower, faster. So much choice.

Completely unrelated, but it's so odd to see upside down '8's in the first column of the base tables on their pricing page.

I'd rather my data go to China and directly to labs that release open weight models (DeepSeek is MIT licensed still); instead of closed labs and the NSA/CIA; who has far better ability to hurt me than China does.

Without getting too much into politics; over the past several decades, one country is going around starting wars, regime changes and other interventionism, or initiating economic warfare (Huawei sanctions because they were getting too competitive to US companies; GPU restrictions; tariffs; etc), and that's not one administration: it's US policy over the past few decades.

The biggest problem with this is that I can't run my company's own internal benchmarks against this API, because they will be trained into the next version of the model, making them worthless.

Have you ever wondered why Manus founders gave up billions of dollars from Meta and “voluntarily” moved back to Beijing? One country does things out in the open, while the other operates in the shadows. The latter worries me more.

It worries me far more what people are ok with us doing in broad daylight.

> Have you ever wondered why Manus founders gave up billions of dollars from Meta and “voluntarily” moved back to Beijing?

This country wouldn't be such a shithole if we were able to bring the wealthy to heel. I suspect most people on earth see China as the more competently governed country. Understandably, this forum tends to select for people who view private investment positively, but this is clearly insufficient by itself to produce functional societies.

You mean the project that literally had an anus as its mascot? Wasn’t that a joke project? Maybe the government was admonishing them for naming their company “man-anus”

You are not really narrowing it down

Which one? https://velvetshark.com/ai-company-logos-that-look-like-butt... https://news.ycombinator.com/item?id=48956924

I mean, someone has to say it, but Anthropic is the one who has an anus as a mascot. More specifically, Claude. It even winks at you, and clinches, too. I'm sorry that I have to be the one to report this to anyone reading this, I truly am. As the saying goes, don't kill the messenger.

There are two kinds of AI company logos: animals and assholes.

> One country does things out in the open,

boy oh boy, if you think the US doesn't do anything in the shadows. The last 100 years of world geopolictical history would beg to disagree

Things in the OPEN?!

Assange, Snowden, Manning, just three of many and the massive things that came from that? And god knows what Musk & friends are up to.

Talk about being completely brainwashed by the US narrative and oblivious of recent history.

Have a look here:

https://ourworldindata.org/grapher/human-rights-index-vdem

lol the US was a .60 in 1789 during slavery. this chart is useless.

The reason the US scores a 0.6 in 1789 comes down to how composite data is aggregated and the global baseline of the 18th century.

Yes, people have put thought into this. This is why I trust institutions and not random posters on the internet.

Is WHO a reputable institution? And was it reputable between 2019-2022?

Yes

The reason that US and its allies/friendlies always score high on these kinds of ratings while their enemies/adversaries always score low is because these institutions and 'experts' are directly funded by and beholden to these countries/interests.

Just look at Russia - it's score is highest in the 90s - the most corrupt and destitute time in its modern history, also happens to be when it was deemed somewhat friendly to US and friends.

If an institution told you to jump off a bridge...

Why link to a number, which means roughly nothing, instead of the specific details on the countries you're trying to compare? I have no idea how to break down or make my own assessment of what 0.83 or 0.18 means. Does that 0.83 for the US include all of the people being shipped off without due process to El Salvador, or do they count that as part of El Salvador's Human Right Index?

People like to bring up the Uyghurs immediately, but never really have the same smoke for the US. Also, do we include the prison population in human rights? Do we include access to healthcare? What about systems perpetuating poverty like lack of access to public transportation?

To pre-emptively be clear on my stance or potential bias, I stand by no country in the world. I was born in Iran, and my parents left when I was young to the US because they believed in the American Dream. The irony being that the dominos of why they left were initiated by the US and UK. And long before Trump was elected, but certainly immediately afterwards, I decided to leave the US once I had the means. (Also because of the lack of public transit, the $100k medical bills after an accident, being punched by a homeless person, and having my possessions stolen repeatedly and the cops giving me a :shrug:). I moved to Japan, where I experienced blatant racism for the first time. I moved to Hong Kong (and tbh no real complaints so far), but I am technically subject to the freedom of speech constraints on China.

I have no allegiance to any country or specific people. My allegiance is to all living beings.

Sorry but you're arguing against a reputable institution here.

I'm not arguing against a reputable institution, I'm arguing against using that reputation + their final number alone without going through the process to verify the methodology myself, which you didn't link to. You only showed the conclusion, which to all of us, should mean very little until we look into how that number was arrived at.

And I don't care if they're reputable, I will look at the methodology and make my own conclusion either way. Appeals to authority are not a substitute for critical thinking.

We're not talking about a few % difference here. If you want to look into it then fine, but you have to come up with some really good evidence if you think you can debunk anything here.

There's absolutely nothing in those numbers to argue against. They are simply providing fluffy quantitative declarations in order for people who don't even claim any particular knowledge or even understanding of the details of the metric (such as yourself) to argue from authority.

If they actually said something, it would be much less effective.

> We're not talking about a few % difference here.

How may freedom percent is it again? Do you lose 1% or 20% for failing to allow foreign ownership of utilities? How many percent do you gain for stopping Russian misinformation?

The metric was carefully crafted by a group of experts as means of communicating a complicated topic.

I'm not going to discuss it further if you don't have anything substantial to bring to the table.

'experts' lmao. Your entire argument boils down to an appeal to authority fallacy.

That fallacy does not apply here as the authority is legitimate and relevant.

That's just your opinion, not fact.

Eric Cartman is that you?

I.e., an appeal to authority

That line of reasoning ... everyone in this thread is now dumber for having listened to it.

Seriously - appeal to authority is a fallacy, unless it's an appeal to an authority that I happen to agree with? It's hard to take you seriously.

Instead of repeating that they are legitimate, you would've gotten further by saying Human Rights Watch has been around for X years, has reported competently on Y matters even when suppression was attempted by Z country. Their membership is made of A, B, C and this is why you should trust in their data collection as a source that we can use in casual discussion.

Right now, you're basically asking people to trust Human Rights Watch because you trust them. But if people don't trust you, why would they trust your word on what sources are legitimized?

We're talking about the country that has tried multiple times to destroy their country, look up Mossadegh for the 1950s and check out the glorious bombing of a school for 2026.

No lessons to take from the US.

PS. https://en.wikipedia.org/wiki/Foreign_interventions_by_the_U...

I know the US thinks only its citizens are human, so "The United States has been involved in hundreds of interventions in foreign countries throughout its history, engaging in nearly 500 military interventions between 1776 and 2026, with half of these operations occurring since 1950 and over 25% occurring in the post-Cold War period” is not probably part of the reputable storytelling

They're not arguing against the institution. They're arguing against the metric .

Yeah but the metric was invented by that reputable organization and given the enormous difference it's not worth any discussion unless they come up with something good in which case they should probably publish about it.

If you don't want to have a discussion don't post pointless metrics that you don't even understand.

They are arguing for you using your brain instead of tethering it to a number.

Yes! By all means, let us appeal to authority and use institutions' reputations to shut down any discussion, rather than thinking about what these figures actually mean, what criteria they're based on, or how well they reflect reality.

edit: wording

An institution mostly funded by the governments it happens to rank highly, and the open society foundation that's owned by George Soros, a man who's been obsessed with overthrowing China's government since the 80s. https://www.v-dem.net/about/funders/

No he's arguing against your arguments and winning.

This number reports, basically, how good it is to live inside a country, while the comment that you are replying to is entirely a list of foreign policy complaints (regime changes, etc).

It's kinda like meeting someone on the street who's protesting about animal cruelty in slaughterhouses, slyly handing them a chart showing them the nutrition facts of beef, and then walking off without another word as though you've settled your case.

Like, I get the point you're trying to make, but I hope you're aware it's a non-sequitur in context?

> Like, I get the point you're trying to make, but I hope you're aware it's a non-sequitur in context?

Your hope should have been that amelius was not aware that it's a non-sequitur, because that is correctable. But then they kept posting and removed all doubt.

oh please, the US economy literally runs on slavery

https://www.prisonstudies.org/world-prison-data/highest-lowe...

https://www.theguardian.com/us-news/2022/sep/27/slavery-loop...

How much of the US economy is powered by prisoners?

Depends on the sector. For example, in California's large-scale wildfire responses, forced prison labor in firefighting acts as an important auxiliary force. https://www.commonwealmagazine.org/los-angeles-firefighters-...

In Arkansas, agricultural work accounted for 17% of all prison labor assignments in 2021. https://www.prisonlegalnews.org/media/publications/Captive_L...

That doesn't answer at all.

17% of prison assignments might be 3 assignments for all I know.

I found a number in your document: "...1.2 million people incarcerated in state and federal prisons, over 65 percent work."

That's 780,000.

And there's 163 million working folks in the USA.

Prisoners power .4% of the economy.

The original context here was human rights in China, and what we see is that the US both has far higher prison population per capita, and that it is being as slave labor which is a significant factor in certain sectors of the economy.

Your whole argument about the percentage overall economy is nonsensical because most of the US economy is not actual productive economy. It's ephemeral stuff like service industry, entertainment, and software development. Obviously, you're not going to force prisoners to go on Joe Rogan podcasts. But when it comes actually producing things, 17% of the industry being done by slave labor is obviously very significant.

The original context was this:

> oh please, the US economy literally runs on slavery

Also

> 17% of the industry being done by slave labor is obviously very significant.

That's not what the stat says. It says 17% of the prison labor did that job, not that 17% of the labor for that job was prisoners.

Here's the quote for ya:

"...agricultural work constitutes a greater percentage of work assignments, such as in Arkansas, where 17 percent of job assignments for incarcerated people were in agricultural work in 2021."

But tell me again how I'm nonsensical.

The amount of mental gymnastics you're doing to justify the amount of SLAVE LABOR in the US is truly incredible. And couple that with one of the highest rates of incarceration in the world, that really tells us all we need to know about the state of human rights.

have a look at everywhere at the human cost of what the orange man is doing depending on his morning coffee. much better to send data to china than to US right now.

This is really a roundabout way of saying you have no plans to live, or perhaps even travel, to China.

I live in the US right now, my surface of attack from the US govt & ad companies is infinitely larger even if I decide to move to China some time in the future

I'm currently looking into https://greenpt.com/ for my company,it has european data residency,no training on data and green footprint for energy.

Only thing is they aren't clear on the SOC2 certification. Am I being naive?

> I'd rather my data go to China and directly to labs that release open weight models (DeepSeek is MIT licensed still); instead of closed labs and the NSA/CIA; who has far better ability to hurt me than China does.

Thank you. I've been shouting this very obvious thing from rooftops

I work in silicon valley and I feel everyone around me is suffering from China-BAD psychosis and dismiss Chinese releases

The link points to OpenRouter with multiple providers.

No Fireworks for me. They invited me to test Firepass, screwed up the onboarding, charged me money (when the account was supposed to be gratis) and when I asked them to fix that after I used up my $$, they said they closed the program. After that, no emails were responded to, and we were on a person-to-person basis. Not sure what their deal is, but I'm off Fireworks.