TheCoderBTW

Yo, I’m TheCoderBTW! I’m all about coding, gamedev, tech news, and AI. Come hang out!!!

[TheCoderBTW @TheCoderBtw]

Here is the latest project I'm working right now

Orrery - An autonomous AI coding-loop engine

With a live orbital visualizer to watch it run

Fully open source with MIT License

Link is in the comments

[TheCoderBTW @TheCoderBtw]

@theo Gonna try this😅

[TheCoderBTW @TheCoderBtw]

@tmuxvim It gets me more than my girl gets me😅

[TheCoderBTW @TheCoderBtw]

@haider1 How openai managed to do this and anthropic where are you guys

[TheCoderBTW @TheCoderBtw]

@HamelHusain What it can do that claude cant?

[TheCoderBTW @TheCoderBtw]

@araseb_ Building in public

[TheCoderBTW @TheCoderBtw]

@argofowl That's a wild take

[TheCoderBTW @TheCoderBtw]

@anshnanda Do you have any before and after metrics on this?

[TheCoderBTW @TheCoderBtw]

@theo @grok whats your opinion?

[TheCoderBTW @TheCoderBtw]

@argofowl Here we go again...

[TheCoderBTW @TheCoderBtw]

@morganlinton What are the benefits of this?

[TheCoderBTW @TheCoderBtw]

@0ximjosh Eventually they'll become the new game changing product that you never intended to build

[TheCoderBTW @TheCoderBtw]

@bruvimtired @claudeai @cursor_ai Try fable with /advisor command. Update us alongside ;)

[TheCoderBTW @TheCoderBtw]

@kr0der @trq212 already mentioned this. They will definitely add fable back to subscriptions when they figure out the compute capacity issues

[TheCoderBTW @TheCoderBtw]

@omarsar0 Absolutely amazing

[TheCoderBTW @TheCoderBtw]

@MattPenny99 @gregisenberg Why is that?

[TheCoderBTW @TheCoderBtw]

@DeepakNesss I'll look into this. Thank you

[TheCoderBTW @TheCoderBtw]

@PawelHuryn Why skipping high on fable?

[TheCoderBTW @TheCoderBtw]

@trq212 This is super useful. Im currently using fable as my orchestrator for almost every tasks. Fable with low effort acts as sub agents. Super nice

[TheCoderBTW @TheCoderBtw]

@VulcanBench Agreed. Efficiency matters just as much as raw scores. Looking forward to seeing more models in your benchmarks

[TheCoderBTW @TheCoderBtw]

Showing token usage on benchmarks alongside their model performance gonna be really useful

[VulcanBench @VulcanBench]

Okay, I'm just going to come out and say it. We have to start sharing token use alongside model performance.

I don't think benchmarks are as useful if you see one model is 6% more accurate than another, but don't know if one uses 600% more tokens than the other.

A good model should have a balance of accuracy, coupled with strong token use.

This is why I share this in all of my benchmarks. Take this one from yesterday. If you just looked at the results you would say, oh GLM 5.2 High ties Fable 5.2 Low and Sonnet 5 High.

But the reality is, to tie them both, it had to use 7,628% more tokens and the cost, 596% more.

Most benchmarks would just show all these models against each other with one accuracy score. This doesn't tell the whole story.

We can do better.

[TheCoderBTW @TheCoderBtw]

@synthwavedd my Pro sub expired June 25 waiting for exactly this. if it really lands July 7 I missed it by two weeks, classic