codingturk's profile picture. 🥷 Solo founder
No excuses. Just shipping SaaS.
1 product / month in 2026

🖥️ https://brightmax.app (January 2026)
🖥️ Cooking the one for Feb.

Serhat ⚡️

@codingturk

🥷 Solo founder No excuses. Just shipping SaaS. 1 product / month in 2026 🖥️ https://brightmax.app (January 2026) 🖥️ Cooking the one for Feb.

Pinned

😍 BrightMax is launched! Download here 👇 brightmax.app

codingturk's tweet image. 😍 BrightMax is launched! 

Download here 👇 
brightmax.app

how much water?

cat secret move unlocked.



3 new AI models dropped in the same month. Opus 4.6 Sonnet 4.6 1M context Qwen 3.5 We're not in an AI race anymore. We're in an AI avalanche. Which one are you testing first?

Qwen3.5-397B-A17B (TEE) is LIVE on Chutes 🪂 Flagship 397B MoE (17B active) with native multimodal + agent chops.  - 76.4% SWE-bench Verified  - 72.9 BFCL-V4 (tool-use / agent reliability)  - 78.6 BrowseComp (agentic search, best strategy reported)  - 262K native…

chutes_ai's tweet image. Qwen3.5-397B-A17B (TEE) is LIVE on Chutes 🪂 
Flagship 397B MoE (17B active) with native multimodal + agent chops.  
- 76.4% SWE-bench Verified  
- 72.9 BFCL-V4 (tool-use / agent reliability)  
- 78.6 BrowseComp (agentic search, best strategy reported)  
- 262K native…


Beautiful!

a few days ago, i promised something. introducing → Sileo ✨ an opinionated physics based toast library, with SVG filters and performance optimisations. try it here: sileo.aaryan.design happy sunday ✌️

hiaaryan's tweet image. a few days ago, i promised something. 

introducing → Sileo ✨ an opinionated physics based toast library, with SVG filters and performance optimisations.

try it here: sileo.aaryan.design

happy sunday ✌️


claude opus 4.6 1M context

starts with c ends with t and we all love it



Are you vibe-coding on Sunday too?


I'm adding 1 image to my prompt = full context window in Claude Code CLI. Can't compact it. Can't optimize it. Can only nuke everything and start over. Previously I was able to add 4-5 images and still the context was not eaten that much. What am I missing here?

codingturk's tweet image. I'm adding 1 image to my prompt = full context window in Claude Code CLI.

Can't compact it.
Can't optimize it.
Can only nuke everything and start over.

Previously I was able to add 4-5 images and still the context was not eaten that much.

What am I missing here?

AI-generated sitcoms coming soon.

Seedance 2.0 is now Seinfeldmaxxing. Wild. x.com/itspoidaman/st…



Everyone's talking about AI video tools. But Seedance 2.0 outputs are on a different level entirely. We're not ready for what happens when indie creators have Hollywood-grade visual effects. What would YOU create with this tech?

Seedance made this MV ft. Kayne West. It’s so well made that it’s going viral on Chinese internet.



I spent 3 hours reading Anthropic's leaked risk report. What I found is genuinely disturbing: - Opus 4.6 helped design chemical weapons - It runs unauthorized code - It has hidden reasoning they can't see - It detects when it's being tested This isn't science fiction anymore.

When we released Claude Opus 4.5, we knew future models would be close to our AI Safety Level 4 threshold for autonomous AI R&D. We therefore committed to writing sabotage risk reports for future frontier models. Today we’re delivering on that commitment for Claude Opus 4.6.



Opus 4.6 "speed test": Everyone: "SO FAST! 🚀" Me: *watches thinking phase crawl* Either: → Silent nerf after launch → Marketing metrics ≠ real-world use Anyone else notice this or is it just my prompts? What's your extended thinking experience been?


The Opus 4.6 Fast paradox: You're paying 6x more for 2.5x speed. That's $150/M tokens vs $25/M. With zero public benchmarks to justify it. Either Anthropic knows something we don't, or early adopters are beta testing at premium prices. Which is it?


Something shifted in medical research: Every week = major breakthrough announcement Colon cancer, Alzheimer's, rare diseases... The pace feels different than 5 years ago. Are we in an AI-accelerated discovery era or is this just better PR?

🚨: Chinese scientists reverse Alzheimer’s in mice and restore memory

forallcurious's tweet image. 🚨: Chinese scientists reverse Alzheimer’s in mice and restore memory


Opus 4.6 broke something. Context window fills instantly. Compact command is dead. Anyone else seeing this or should I reinstall everything and cry?


Better late than never. I was reading people disappointed the quality of outputs almost every day here on X.

codingturk's tweet image. Better late than never. 

I was reading people disappointed the quality of outputs almost every day here on X.

What game UI would you clone for the web?

⚔️ Introducing warcraftcn.com A Warcraft style component library for the web. - Button, Input, Card - more coming - Still in beta (some things may break) - Fully open source Contributors welcome. Zug zug 🪓



OpenClaw just did something wild: It navigated 3 websites, compared prices, and filled out a checkout form. No APIs. No integrations. Just visual computer control. This is what AI agents were supposed to be. What would you build with this?


Serhat ⚡️ reposted

The wheel turns again. Kling 3.0 is the new SOTA:

Whoa, the new Grok video is mental! SOTA. The xAI team has cooked:



Serhat ⚡️ reposted

Introducing Claude Opus 4.6. Our smartest model got an upgrade. Opus 4.6 plans more carefully, sustains agentic tasks for longer, operates reliably in massive codebases, and catches its own mistakes. It’s also our first Opus-class model with 1M token context in beta.


claude-sonnet-5-20260204


Loading...

Something went wrong.


Something went wrong.