Viewing profile — eli

Qwen 3.8 Max is very strong at troubleshooting and code review.

I assume/hope this is about prices going up for the next release of Pro

And all that just to allow internet access for npm and pypi? If you've got the bandwidth and disk space, it's very easy to make an offline mirror of both.

That's got a significant selection bias. Claude and ChatGPT and Gemini and other subs do not go through openrouter.

I use https://pi.dev/ which works fine out of the box but is fairly minimal and intended to be customized. There are many extensions. OpenCode or oh-my-pi might make more sense if …

It's not exactly what you asked for, but the public transit and multimodal focused apps are generally much better at this stuff. I like CityMapper and Transit. Neither is good for …

It's not enough that it's better? Many providers will host it and will compete on price. It also can't easily be taken away because one company (or one government) decides they don…

I believe it. It's extremely good at troubleshooting. I gave Qwen and Kimi K3 the same annoying, complicated, intermittent bug to track down. Kimi did a bit better in understanding…

The amount of abuse of free VMs must be absolutely immense even with credit card verification. I’m not sure there needs to be additional secret motives.

They're usually leased. So yeah they probably owe Flock new ones.

They bill out the ass to lease them

I'm not familiar with North Dakota in particular, but at the federal level and in some states ALL works created by the government are exempt from copyright. The public paid for it,…

According to a few tasks from my little personal coding benchmark it's very good at coding and kinda bad at web design. (Also excellent at "draw me a picture" one-shot prompts, for…

To be fair, "people with no skills inflating their own value" is what LinkedIn has always been like. But I guess LLMs are uniquely well positioned for that task.

These are the API rates. They don’t train on prompts from the API. In what sense are you the product?

I mean, yeah. Alternatively I’ve seen the approach of starter with the cheaper/faster model but give it a tool to handoff or talk to a strong model if it gets stuck.

This is the improved iOS parental controls. When they shipped it (and for years afterward) it simply did not work https://www.wsj.com/tech/personal-tech/apples-parental-contr... It…

Yup, trying to be really strategic about testing. I didn't end up sticking with it, but I tried requiring test cases to cite a matching clause in the task assignment. But also: the…

Max is lot of extra reasoning. I wonder how many fewer tasks it solves on high. I bet that costs quite a lot less.

Oh sorry I meant the framework. Though I could see posting a couple of the tasks.

I have actually been planning to open source the framework. Only benchmarks I care about are the ones that look like real work I do. So it makes it easy to trawl your own repos loo…

In my tests, it averages to much cheaper than Opus 4.8 on real tasks on account of being smarter and more token efficient. I have a benchmark to build a game engine from a set of w…

It's a new and improved version of an existing model? I don't think it's intentionally befuddling.

That's a plausible explanation but I'm not seeing evidence for it. I have a personal benchmark suite of 14 real, non-public tasks. Opus 5 and Fable tied on 10, Opus won on 3, and F…

I’m not convinced it wasn’t a PR stunt from the start