TheArcane

Oligarchs don't face consequences in an oligarchy silly

> Although if you have killer robots,

we already do..

Sonnet high is slower in my experience, and much more expensive

Doesn't hold a handle to Deepseek v4.1 Flash

I have both platform.deepseek.com (hooray off-peak pricing) and the $100/m Claude subscription and boy do I find myself mostly using DS-flash pretty much every time.

There's something to be said for the speed with which it gets word done - I take fewer "it's compiling" breaks, and lose my train of thought less frequently. What's crazy is that the speedy work it does is pretty much always correct - which is a reason why I don't feel the need to spin up claude-code and "ask the expert".

This speed & good-work combo is the DS signature IMO

there we go again, the king of "China BAD" boogeyman

Have you run into their concurrent agent limits?

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

> 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

> 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

This wasn't my experience. GPT 6.1 Sol is ~6-7x more expensive for similar intelligence. My experience matches [this](https://api-docs.deepseek.com/img/v4.1_261017_perf_cost_en.p...)

opencode and deepseek harness GUI (dsh-tui when I need to)

> What would you prefer, 30t/s 50k output tokens needed for solution, or 300t/s, 500k output tokens needed for solution?

The best way to answer this question: [Pareto frontier or cost/per task curve](https://api-docs.deepseek.com/img/v4.1_261017_perf_cost_en.p...)

img from https://api-docs.deepseek.com/news/news261017/

I have both platform.deepseek.com (hooray off-peak pricing) and the $100/m Claude subscription and boy do I find myself mostly using DS-flash pretty much every time.

There's something to be said for the speed with which it gets word done - I take fewer "it's compiling" breaks, and lose my train of thought less frequently. What's crazy is that the speedy work it does is pretty much always correct - which is a reason why I don't feel the need to spin up claude-code and "ask the expert".

This speed & good-work combo is the DS signature IMO

$250K deposits / car

If you don't buy into "America good, China bad" narrative, this new entrant & release by Inclusion Ai is a lot more exciting by every measurable metric.

https://github.com/inclusionAI/Ling

I cheer every time there's a new Deepseek, Qwen, Kimi, Z.ai release. Intelligence shouldn't be owned and charged for by a few

If it were up-to these silicon valley tech bros, they'd find a way to meter and charge for the air we breathe.

certainly better than this one. Look at Africa under western neocolonialism vs China's infrastructure loans

masking the active genocide & war crimes in Palestine

For instance, try asking American models about the Palestinian genocide

anecdotal but we heavily use Qwen to build our own models on top of

> Apple somehow reigns supreme still.

That's because America ban anything that starts to compete, like Huawei or Chinese car companies

boo hoo china bad, buy my more expensive and shitty american product

> This is literally the coolest thing we as a species achieved that doesn't serve self-preservation purposes.

couldn't be more wrong

> He's one of the most empathetic individuals I know.

You should try knowing more people

aka the paradox of tolerance https://en.wikipedia.org/wiki/Paradox_of_tolerance

chat.qwenlm.ai has quickly risen to the preferred choice for all my LLM needs. As accurate as Deepseek v3, but without the server issues.

This makes it even better!

> Don't have to expose my home IP address and open a port for the world to start banging on

Is that really an issue if all you're exposing is the VPN port? Wireguard for instance has industrial-grade encryption. Even open port 51820 should be fine

Bit of a waste to host this locally and still need an openai api key. This would be great with Ollama

And they did all this under a GPU embargo? We're witnessing a Cuba moment.

I'm confused as to how you haven't found R1 to be much better. My experience has been exactly like that of the OP's