faitswulff

Ah, so that's where the profit comes from. Very payment-processor like move, makes sense that Stripe picked them up.

Interestingly, OpenRouter will serve inference at cost for a lot of models.

For readers wondering, OpenRouter isn’t capable of caching as effectively as DeepSeek is because they will, for instance, switch inference providers in the middle of a session.

Deepseek is cheap and fast, but it couldn't follow the basic directions I gave every other model (GLM, Qwen, GPT, Claude) to use red/green TDD. Put me off of using DS.

The UI design gap is something I’ve noticed as well, in things as simple as ASCII diagrams. Claude has a more human touch. All the diagrams GPT 5.6 generated for me were dressed up lists with too many pipe symbols.

To add to your point:

> "Quite frankly, the biggest issue we are now having is not a compute glut, but it's power and it's sort of the ability to get the builds done fast enough close to power," he told the show's hosts. "So if you can't do that, you may actually have a bunch of chips sitting in inventory that I can't plug in. In fact, that is my problem today. It's not a supply issue of chips. It's actually the fact that I don't have warm shells to plug into."

> - Satya Nadella

https://redmondmag.com/blogs/generationai/2025/12/microsoft-...

And the will

Astra's looks the worst to me on mobile, though

Everyone's still bottlenecked on foundries, not designs.

Knowledge is for the machines, you see; we are simply here to tend to their electrons. At least for now.

But is the grammar load bearing?

And the majority shareholder, at that!

Politics means putting job creating infrastructure projects in red states to curry favor with Republican demographics.

The same reason scam artists describe themselves as businessmen

I’m not very good at mathematics, but it seems like Kevin should take his girlfriend on trips more often for the good of all mathematicians.

Heard on the grapevine that the OpenAI blip was a cloudflare issue

This video compares the Mediatek Dimensity to the Xring 03 (at the given timestamp 329s in) and finds significant differences in physical layout and performance: https://youtu.be/BCCJuOopLK8?t=329

China actually penalizes, imprisons, and even executes its billionaires, so that’s different.

> For Codex, working with Jira and Atlassian was a hassle in my environment...

I hated this so much. Both acli, which my agents have to rediscover how to use from --help every time, and the MCP, which I have to reauthenticate against frequently. I have replaced both with a "skill" that just describes where to find an Atlassian API key and which version of the API to use. Works perfectly every time.

The currency code is CNY or Chinese Yuan. Yuan itself is like “dollar,” with many different countries calling their currency “dollars.” Renminbi refers to the Chinese currency as a whole rather than any particular unit.

Data centers would be less contentious in the U.S. if energy was plentiful and clean. As it is, it externalizes costs onto the public and strains profit-optimized electrical grids while contributing to local pollution and global warming.

Another reason this may be true, selling hardware may be the only durable moat after a while

It’s also a flywheel for China’s homegrown chips industry.

Nationalism is a cancer

I wish this had some code or pseudo-code tests!

That simplistic characterization is still essentially "China-bad." The CCP is the same government responsible for lifting historically unprecedented numbers of people out of abject poverty. Does it make up for other human rights violations? No. But "CCP bad" flattens a complex and powerful political organization into a fairy tale boogeyman.

> Didn't they bring hundreds of millions out of poverty, and built amazing cities and facilities in the past 30 years?

Yes, but China-bad ideology demands that we ask ”at what cost?”

> I can’t conceive of the leadership of the Python or Rust or any other community I’ve ever worked with doing something like that.

At least for Rust, part of this is because there is no BDFL, by design. Python has moved away from the BDFL model as well, IIRC. Individual contributors can get mad and write angry personal attacks, but there’s no face of the language like there is with Zig.

Also throw in GLM 5.2 for good measure

> That depends on if car was originally made for left-hand or right-hand traffic

I was referring to i.e. Arabic vs English