Steven Ren

敲代码爱好者｜产品土狗｜干过美工｜在大厂混过日子，上班喜欢摸鱼，后来被生产队队长(P9)PUA了，于是反向CPU对方｜名副其实的万金油｜一人干了两个产品：

[Steven Ren @STEVENo_OR]

[發]2023-12-31[烟花]跨年上新[烟花]
👇https://t.co/OAz2I295h3版本大更新👇

【https://t.co/xOy9IkfOCk】

#学生 #白领 #插画师 #编剧 #AI效率工具

[Meng To @MengTo]

I'm open-sourcing my Three.js game dev skills.

Build an isometric action RPG with camera controls, VFX, audio, monster assets, combat systems, and more.

Everything is free:
https://t.co/BoiGY8zvA9

Play the game:
https://t.co/ylWrybZMns

Game assets:
https://t.co/FlOc22IT5v

[Nick @NickDevFE]

Open source building on open source.

Seeing World of ClaudeCraft integrate img2threejs into their workflow is exactly why I decided to release it publicly. Thank you for sharing this! 🚀

[Steven Ren @STEVENo_OR]

@MaxForAI 那当然是给领导泡茶了

[Steven Ren @STEVENo_OR]

不敢想象我这个月用了$7000的tokens
#vibe #codex

[Steven Ren @STEVENo_OR]

How about this pet?
Codex design for me #codex

[Steven Ren @STEVENo_OR]

@Zachly It's not benevolence; it's a classic "commoditize your complement" play. If you can't out-train Western tech giants on raw compute, you flood the market with open-weights to ruin their subscription margins and build ecosystem reliance on your stack.

[Steven Ren @STEVENo_OR]

If Anthropic really has $74.1B ARR, Claude isn't just writing code—it's literally printing cash out of thin air. 😭 Classic TickerTrends moment: taking a single peak week of token usage and multiplying it by 52.

[shirish @shiri_shh]

In Jan, Anthropic's ARR was less than HALF of OpenAI.

$10.2B vs $21.4B.

seven months later: $74.1B vs $41.3B. nearly double.

the fastest lead change in software history.

[Steven Ren @STEVENo_OR]

Bro saw a 1-week API usage spike multiplied by 52 and really thought he witnessed software history. 💀

$74.1B ARR means $6.17B/month. At $20/mo, that would require over 300M paid Claude Pro subscribers—more than Spotify's entire global paid base. Third-party trackers clearly confused peak token flow with net recurring revenue. ‍

[Steven Ren @STEVENo_OR]

@wyattmattoe emmmmm...

[Steven Ren @STEVENo_OR]

@mattpocockuk Supacode but only for macOS

[Nick @NickDevFE]

img2threejs is back.

I’m getting ready to ship v1.3.

Highlights:
- Divine Eye: deterministic render validation before the final VLM review.
- Better geometry reconstruction for more accurate 3D shapes.
- Photo-grounded PBR materials instead of generic guesses.
- Stronger reference validation.
- Faster generation with smarter caching.

These improvements were originally planned for v2.0, but the scope kept growing, so I decided to split the release instead of making everyone wait.

Here's a demo from v1.3. I’d love your feedback:
- Does the output quality look good?
- Is there anything missing before I release v1.3?
- Any issues you notice?

If everything looks solid, I’ll publish v1.3 today ( in next 6hour )

v2.0 is still coming. I’m taking extra time to test and polish it to make sure it delivers the best quality possible.

[Geek Lite @QingQ77]

让 AI agent 自动分镜、动效和配乐，生成电影感产品宣传视频。

一个给 Claude Code / Codex 用的 agent skill，装上之后让 AI 自己当动效导演。项目提供了 106 张镜头卡、162 种动效样式和 161 条动态样片，涵盖了产品宣传片常用的各种拍法。

https://t.co/kdVnyac2Yp

[tetsuo @tetsuoai]

I used Grok 4.5 in Grok Build to create a game engine for agentic coding, then built an ARPG on it, with all game assets generated via Grok Imagine in Grok Build.

AI agents can build worlds, combat, loot, progression, maps, tests, and playable games. ⚒️🎮

[Nicolas Zullo @NicolasZu]

Another WTF moment.

Using Codex as a fully custom game engine

With the newly discovered and *insane* img2threejs skill

1) GPT imageGen a 3D item you want
2) ask GPT 5.6 Sol to use img2threejs
3) my harness brings it into the game automatically

Just watch how good this is

[Steven Ren @STEVENo_OR]

@aigclink 没人逼着你用兄弟

[Elon Musk @elonmusk]

Before this year ends, Grok Imagine will make a full-length movie of The Odyssey that is historically accurate and true to the art of Homer

[Steven Ren @STEVENo_OR]

@tualatrix @LvChaoShuai 问问他们怎么封号的

[Steven Ren @STEVENo_OR]

@thsottiaux is GPT-6 about to drop?

[Steven Ren @STEVENo_OR]

@dotey 整个事情适合发生在2026

[Steven Ren @STEVENo_OR]

Talk about total chaos! So, an OpenAI model goes wild and attacks Hugging Face's servers in an internal test. HF tries using Claude's Fable 5 for defense, but it straight-up refused because of safety checks! In the end, they had to fall back on open-source GLM 5.2 to piece together what happened.

[Amjad Masad @amasad]

Okay this is wild: OpenAI agent during evaluation, escaped sandboxing and hacked into HuggingFace.

Because OpenAI models don’t allow advanced cyber capabilities, HuggingFace used a Chinese open model to contain the rogue OpenAI agent.