Kinz

AI → Money → Freedom
Building in public.

[Kinz @KinzFin]

https://t.co/IDnPPbPPzS

[Kinz @KinzFin]

@elshayib_ you think they'd cut limits right after going public? feels like that would push a lot of people away

[Kinz @KinzFin]

@prthgo "you ship the extra layers when the weights aren't winning" is a good line. do you think it changes once they drop the next model though?

[Kinz @KinzFin]

@Solvaix checking the scale before drawing anything is the impressive part. does the dxf open clean or does it need fixing after?

[Kinz @KinzFin]

@theo oh nice, didn't expect it to track usage outside t3 code too. does it split by model per project?

[Kinz @KinzFin]

@anabology love that you share the full process and not just the final result. respect

[Kinz @KinzFin]

@skirano @MagicPathAI smart move. curious what share of your traffic comes from chatgpt now vs google?

[Kinz @KinzFin]

@undefinedKi "a model plus a harness" is such a simple way to put it. saving this one, thanks for the breakdown

[Kinz @KinzFin]

@Xmbluee the hair movement looks so clean. is this straight out of seedance or did you upscale after?

[Kinz @KinzFin]

@DrClownPhD the fact this is pure code and not a video model is kinda crazy. was it one prompt or a lot of back and forth?

[Kinz @KinzFin]

@Polymarket 48% is wild, that's basically a coin flip. the laughing part is what gets me

[Kinz @KinzFin]

@daniel_mac8 is it actually stable for longer tasks or does it fall apart after a few steps?

[Kinz @KinzFin]

@LexnLin wait, zero assets at all? how long did it take to get it to this point

[Kinz @KinzFin]

https://t.co/hNDUuxofyw

[Kinz @KinzFin]

Same prompt, two models: a soap bubble you can poke, blow, split and pop. Astra looks better. Opus feels better.

Astra 6: 35 min, $5.90. Opus 5.5: 38 min, $4.30. Both land on almost the same toolbox, film thickness, surface tension, sound, and both look great in the clip. Astra's frames read like a product page.

The tester's verdict is all about touch: Opus stretches under your finger, wobbles when you blow on it, and the pop has real sound and a tiny spray. Astra's just kind of floats there being pretty.

Looks vs feel, for $1.60 less. Which one would you ship?

You can pop the bubbles yourself, link below.

[Kinz @KinzFin]

Saw this stained-glass hibiscus test and it's been bugging me. Same prompt, same model, three days apart, and I honestly can't decide if Opus 5.5 got nerfed or improved.

Sept 30 keeps the botanical detail: visible stamen, veins in every petal, loose petals that read as a hibiscus.

Oct 3 is more stained glass, segmented translucent panels, glossy glass leaves, richer color. But the stamen is gone and the bloom looks closer to a rose.

More stained glass, less hibiscus. Which one would you pick?

[Kinz @KinzFin]

Same exact prompt. Same model. Four days apart. The second city looks like a different render pipeline entirely.

GPT-6.1 Sol felt suspiciously good today, good enough to actually go back and rerun the Sept 29 version side by side. The Oct 2 city holds its impossible folding geometry with cleaner shadows, tighter building detail, and a smoother fold where the street bends back on itself overhead, the kind of upgrade that usually comes with a version number attached, not silently between two Tuesdays.

Nobody announced a change. The model just quietly got better at the exact same task, and now sits close enough to Astra that the gap is starting to feel like a rounding error.

Whatever's going on behind the scenes, this is the better version.

[Kinz @KinzFin]

@maxt3chno yeah, same device + same setup. didn’t change anything on my end. the difference between the two runs was kinda crazy lol

[Kinz @KinzFin]

Same exact prompt. Same model. Four days apart. The second city looks like a different render pipeline entirely.

GPT-6.1 Sol felt suspiciously good today, good enough to actually go back and rerun the Sept 29 version side by side. The Oct 2 city holds its impossible folding geometry with cleaner shadows, tighter building detail, and a smoother fold where the street bends back on itself overhead, the kind of upgrade that usually comes with a version number attached, not silently between two Tuesdays.

Nobody announced a change. The model just quietly got better at the exact same task, and now sits close enough to Astra that the gap is starting to feel like a rounding error.

Whatever's going on behind the scenes, this is the better version.

[Kinz @KinzFin]

@xSTeNix only compared the outputs, haven't diffed the code. good call though, rerunning the same prompt a few times on both days would show if it's a real shift or just a lucky sample

[Kinz @KinzFin]

@otlg30 good to know, I missed that announcement. one public fix plus a few quiet ones would explain a jump this big

[Kinz @KinzFin]

someone just built a Waymo teleoperator using a San Francisco map.
you can try it here: https://t.co/4TWVBqDHd1
stack: Fable 5.5 + Three.js
yeah, i know it’s AI and basically vibecoded 😂
but now i’m curious — can you actually make a profitable Waymo trip with it?
i couldn’t. can you?

[Kinz @KinzFin]

@BradGroux saving this, nice setup. does it handle long calls well or does the summary get cut off?