re-thc

> They should've just lead with real, up to date data, because it's good, not the silly old tactics like comparing to Opus 4.8 when 5.0 is out in many of their charts

It's what people know. Opus is just the common target.

> Smarter and cheaper than Luna xhigh, not as smart but less expensive than Luna max. Smashes deepseek v4 flash

The problem with this and DeepSWE is it goes for a very specific profile. I'm not convinced DeepSWE is any accurate in actual work. It's surely a different signal (compared to some that allow cheating) but it has its own issues, e.g. weak harness.

Luna is great at following instructions but bad instructions or anything not covered = death.

Deepseek is more analytical. Good for bug tracking.

GLM is a better all rounder in some ways. Better at creativity.

> The export controls were revoked before

Zai is on another "export control" list outside the broader 1. Doesn't help.

Which 9/10 times hasn't been great anyway (stock reaction).

> From a biased source, but would be big if true. I've had great results with GLM 5.2.

It's at least close (even if not better) from the Ox Alpha runs. For the price it's definitely great.

2. There was a new checkpoint. Official.

With vision on top

the outperform Fable was a mid (not completed) benchmark run. Real results were lower.

> _China doesn't need to invade Taiwan._

All your points consider what China can do but not how others would respond. Blocking Taiwan when the whole AI economy is running on it will invoke lots of countries to go up in arms. There was a quote from US military that it'd cost less to fight than watch Taiwan fall (invasion or not).

> 4. China is close to producing industry-grade EUV

No where near close. At least not production grade at massive scale.

> * Establish a total naval and air blockade of Taiwan using their fleets stationed in Ningbo and Zhanjiang.

Doesn't work. If they do that the US, JP and others can also blockade around Shanghai, Hong Kong and the other ports.

> They won't need TSMC soon

Taiwan is not just TSMC. There are thousands to millions of things involving Taiwan you're likely not aware of and they might not even be AI or semi related.

> this model is around 64B and can run on laptops.

We're lucky if it'd fit in 1 DGX Spark. Laptops - nah, unless you mean like an M5 Max with 128GB of RAM then maybe.

> Thats the reason behind the hype.

The hype is imagine DeepSeek Flash before the price increase with even better performance. It'd be like unlimited Sonnet.

> why some G employees were vague posting about it

They were just trolling.

> So I can only assume codebases became more bloated, websites became more heavy on resources.

If we assume this is true, it at least meant if you didn't upgrade you'd get even worse performance.

Option 4: the claimed capacity is not true. Real world usage hasn’t reached anywhere close to it.

Possibly next Longcat by Meituan too.

> got what seemed to me like a neutral, well-balanced reply

Ask if Taiwan is a country instead :)

> Model is suspiciously fast

> aren't representative of models from the big Chinese labs

There were reports that China has let Nvidia's chips through, so this might be it. Testing both the chip and infrastructure.

GLM 5.3 is out and does even better in this area, so…

> The bun rewrite is looked to as proof that this sort of full-throttle vibecoding is the future of software development

To who? You?

> To prove that, a proper release is needed

For who? Who is trying to prove what?

> It's Jarred that keeps saying it'll be released tomorrow

There's no official blog... he's just rambling on x?

> Github insights for the last week

That's the problem with the rest of your post. Bun is no longer VC-fueled open source project. It got sold to Anthropic.

> There's a huge partisan eagerness to declare the Claude Code release

Hence there isn't. The Claude Code release is for its own use.

It's 1/2 baked in everything it tries to do.

You need a "separate" tool called npm to install packages. It's dead slow so they added corepack. Another extra tool. Then you use it to install yarn or pnpm. Another extra tool. It might still be slower than bun...

In the past you need to transpile typescript with other tools. It can only run js out of the box. The later fix was a wasm build, which means slower than it could be. Another 1/2 job.

Many more examples in many areas but anything Node tries to do itself feels 1/2 baked.

Bun and Deno are faster than Node because it is a 1/2 job at its core. It has v8. It has the tools. It could be done but no. It did nothing for years until Deno and Bun came on the scene to nudge it.

> The article answers this. hint: they're not shipping.

They're shipping though.

Claude Code and many others use it. There's just not been a "public" release.

It feels worse but is it actually worse? Opus has always made mistakes.

> There’s an assumption that you can spin up the infra and acquire customers within that margin

Only Nvidia and approved friends can at the moment. Nvidia can even backstop your loan required.

> Anthropic's Project Glasswing is supposed to find them quite a while ago?

Someone still has to run it. The analysis and fix could be someone's machine but not committed / published.

> Apple will release M7 MacBook Pros / Mac Minis next year

The latest on Apple is TSMC is stuck on the next iPhone due to lack of RAM. Good luck getting any Macs. Memory shortage is getting worse.

> implies no significant Gemini Flash developments until the next year.

Gemini 4 is apparently just around the corner so unless there's a 3 month delay... there's at least a new Flash update.

> For Google, this is still gemini-3.1-pro-preview, right?

Flash is better than Pro for now.

> over DS V4 Flash/Pro? 13-26x cheaper with comparable intelligence, and available across many different inference providers.

That's why DS4 already had a huge price hike announcement.

I'd argue ds4-flash is closer to terra than luna. Hence different price class.

> haven’t actually used Fable or even Opus 5 yet

I've used plenty of Opus and Fable. Still do.

> Sol is laughable by comparison

Not really, it depends. Sol is better and useful in some areas. Definitely not all.

Fable is gimped just by those "guardrails" that silently downgrades you to Opus 4.8. Not only do you pay extra for Fable but your caching can be easily messed up. It also doesn't just find all the bugs or is bug-free. Sol has spotted lots of Fable issues and vice versa. Fable also costs 2-100x as much.

> I don’t constantly swear at and call stupid

That's not a judge of anything. There are models that may be stupid and you can swear at it, but if they still get the job done for 1/10th the price... maybe that's all you're paying for.

> If I was Chinese, I'd probably trust Grok more than a local AI company.

Nah. There are more established companies (e.g. Tencent, Alibaba, etc) and academia (e.g. Moonshot, Zai, etc) involved than in the US (comparatively). Also there are more Chinese AI researchers involved than non-Chinese (whether they physically sit in China or not).

> It's just an obvious fact.

What's the fact? Facts require proof, right? Where is in it?

> China is clearly the US' main adversary.

This?

It's clearly documented Trump and friends randomly made that policy up in the 1st term. Can you tell from the current term? There's been more effort spent on non-China matters, e.g. Middle East related than China.

> it just comes down to which government do your self-interests align with best

Why do you have to pick 1? Most normal people, US citizens or not wouldn't. Tesla has a gigafactory in China. Apple is trying to buy Chinese memory. Meta tried to buy Manus AI. What adversary?