Why wouldn't you specify the tech stack for new applications?

Do you really just ask the model to pick any backend stack and pray it works?

I agree. type checking is game changing for AI development.

And you can't duct tape your way into Ruby type checking and hope to beat the likes of C#, Java, Rust, Go. TypeScript.

Rails (and Ruby) made tradeoffs to please humans. Some of those tradeoffs became downsides now with AI.

On the other side I as the one checking the messy code my AI produces still prefer a highly readable code over something more "complicated" I would argue it's easier to catch bullshit even.

For those who are curious, this is what I could gather:

    - Three.js for rendering
    - react 18+ UI built with Vite
    - charts are custom, no libs
    - main bundle size is a js file of 1MB (290KB gziped)
    - No router, no redux/zustand
    - saves using indexedDB + service-worker

How is he able to help more other than telling the truth that phones and tablets can't do what she needs?

Especially in this department, where limitations are very much fixed. Even if he wants it to work on a phone, he cannot make it do certain things if the OS doesn’t intend it to be possible. If she were trying to do things on Linux, at least you could accuse him of being too lazy to write custom software to do it. Not true on closed platforms though.

> too lazy to write custom software to do it

Hell, I’m not a coder; I don’t even work in IT. I can install Linux and configure it reasonably well for an end user, but ask me to code something up and… well, hope you like Claude’s output.

Her desktop is a 27” iMac that’s too old for OCLP to help keep it going, and I keep thinking about putting Linux on it, but I can’t imagine trying to get her onto something that isn’t either Windows (used at work) or macOS (we have full Apple ecosystem at home and her laptop is an M2 MBA) until it just won’t run Citrix and a browser capable of hitting Gmail. Or until Apple finally makes another 27” iMac, which given the RAM and storage situation isn’t happening any time soon.

> because doing new things is hard and hardware has deadlines that software doesn't.

It was nothing new tho. They broke existing software. This is caught by basic software regression test. That's engineering 101.

Seems to me like it was an explicit design decision that fixed some use cases and broke others, rather than a regression in the sense of "A bug we introduced but didn't catch." As such a regression test wouldn't have caught this.

Someone made a design decision that broke some use cases.

Just a note: their leaderboard has Gemini Flash 3.6 on #1.

And that model was not top 5 even when it was released, 3 months ago.

wow, Opus 5 (not 5.5) solar system is broken.

I wouldn't expect that.

But it's probably n=1. Still.

Looks good to me?

What browser are u viewing it on?

https://aibenchy.com/compare/minimax-minimax-m3-1-flash-high...

Their orbit is broken:

https://imgur.com/a/Hrtk3IX

Both in Chrome and Firefox.

That's so odd, for me it looks ok on Firefox/Brave both desktop and mobile: https://ibb.co/Xk4sCVDB

Oh wow, you're right. On mobile it looks correct to me.

Look around you.

You'll have a hard time finding electronic devices not made in China. And all of them have software which is harder to monitor and debug than LLMs.

So I don't think distrust for Chinese LLMs will hold.

I think Chinese models deserve the same level of trust as US models: none.

API price rates drop should slow down if open weight models are banned.

Less competition.

No. Venders will want to capture the displaced users. So they will offer lower prices.

Open weight models don’t really compete with frontier models at scale, so they don’t really compete with each other.

I don't agree. Because DeepSeek is a few months behind Sota for 1/5 of the price even if you consider open router API prices.

I highly suspect models like Sol would be much more expensive without Chinese competition.

So to a point Sota pricing is already affected depending who you ask. And the tendency is for the gap to tighten.

Some reddit comments say it's not as good as DS 4.1 Flash. But I didn't test myself and I don't trust benchmarks.