gemma-4-12B-agentic-fable5-composer2.5 V2 is out.the agentic upgrade to the model trained on Fable 5's reasoning. Running it now with TurboQuant llama.cpp on a single RTX 4060( 8 GB VRAM) at 30 tokens/second with full 25000 context and reasoning:# The benchmarksv2 is built for coding + agentic work. writing code, running commands, using tools, debugging, multi step technical tasks. The clearest signal is tau2 bench telecom, an agentic tool use benchmark whose diagnose → fix → verify loop mirrors
0:16
gemma-4-12B-agentic-fable5-composer2.5 V2 is out.the agentic upgrade to the model trained on Fable 5's reasoning. Running it now with TurboQuant llama.cpp on a single RTX 4060( 8 GB VRAM) at 30 tokens/second with full 25000 context and reasoning:# The benchmarksv2 is built for coding + agentic work. writing code, running commands, using tools, debugging, multi step technical tasks. The clearest signal is tau2 bench telecom, an agentic tool use benchmark whose diagnose → fix → verify loop mirrors
136.6K views1 week ago
x.comAlok
Getting closer and closer to vibe coding video games customized by users. Soon enough we’ll probably just pay for the hardware and have thousands of custom servers or/and games. A sort of elevated Roblox.
1:00
Getting closer and closer to vibe coding video games customized by users. Soon enough we’ll probably just pay for the hardware and have thousands of custom servers or/and games. A sort of elevated Roblox.
2.8M views2 weeks ago
x.comTHE SHORT BEAR
imo using ai tools like this is fine as long as it helps the devs (+ it's optional), and it'll get normalized over time anyway, same way AI coding tools went from "bad quality" to "cheating" to standard in just 3 years once people realized how much faster they get things done with itbecause unfortunately whether people wanna accept it or not, we (including game dev studios) either adapt to these tools or get left behind EVENTUALLYand there's a VERY high chance your favorite app, website or game
5:36
imo using ai tools like this is fine as long as it helps the devs (+ it's optional), and it'll get normalized over time anyway, same way AI coding tools went from "bad quality" to "cheating" to standard in just 3 years once people realized how much faster they get things done with itbecause unfortunately whether people wanna accept it or not, we (including game dev studios) either adapt to these tools or get left behind EVENTUALLYand there's a VERY high chance your favorite app, website or game
1.5M views2 weeks ago
x.comHYPEX
$8k in 8 days.I live-streamed myself building CNVS.Some days I go for 8 hours. It’s the best vibe coding platform on the planet and no one is even close. I am creating a Mac OS app that is equal parts beautiful ( Liquid Glass, live backgrounds, dynamic agent organization ). And raw power ( voice control with gpt realtime 2 or nvidia parakeet fully local, cross agent memory system, mcp and cli )The platform inspires you to build amazing things.We have been building in the dumpster for too long (
0:50
$8k in 8 days.I live-streamed myself building CNVS.Some days I go for 8 hours. It’s the best vibe coding platform on the planet and no one is even close. I am creating a Mac OS app that is equal parts beautiful ( Liquid Glass, live backgrounds, dynamic agent organization ). And raw power ( voice control with gpt realtime 2 or nvidia parakeet fully local, cross agent memory system, mcp and cli )The platform inspires you to build amazing things.We have been building in the dumpster for too long (
108.2K views1 week ago
x.comMax Blade
You build stuff and host on GitHub pubically? Paste this into a coding-agent session and point it at your own GitHub account.This is happening way too much. ROTATE YOUR KEYS. Review my public GitHub repositories for accidentally exposed environment secrets.Scope:- Only inspect repositories I own or explicitly authorize.- Focus on public repos first.- Check current files and git history.- Look for API keys, tokens, private keys, database URLs, OAuth secrets, webhooks, cloud credentials, .env file
0:10
You build stuff and host on GitHub pubically? Paste this into a coding-agent session and point it at your own GitHub account.This is happening way too much. ROTATE YOUR KEYS. Review my public GitHub repositories for accidentally exposed environment secrets.Scope:- Only inspect repositories I own or explicitly authorize.- Focus on public repos first.- Check current files and git history.- Look for API keys, tokens, private keys, database URLs, OAuth secrets, webhooks, cloud credentials, .env file
3.4M views2 weeks ago
x.comJarrad Grigg
While ENPIRE w/ @nvidia @_wenlixiao @DrJimFan enables coding agents to explore algorithms and improve policies for a given real-world task, RATs asks: what can agents learn before a human specifies the task?Through curiosity-driven play, agents propose tasks, hill-climb toward solutions, and accumulate reusable, transferable skills. When a human later requests a new task, the agents retrieve and compose these skills to solve it.RATs explores an analogue of pre-training for embodied coding agents
1:37
While ENPIRE w/ @nvidia @_wenlixiao @DrJimFan enables coding agents to explore algorithms and improve policies for a given real-world task, RATs asks: what can agents learn before a human specifies the task?Through curiosity-driven play, agents propose tasks, hill-climb toward solutions, and accumulate reusable, transferable skills. When a human later requests a new task, the agents retrieve and compose these skills to solve it.RATs explores an analogue of pre-training for embodied coding agents
48.7K views1 week ago
x.comMax Fu
The future of robotics at NVIDIA GEAR Lab 👇 @DrJimFan ENPIRE deploys 8 Codex-based AI coding agents - fully autonomously control a real robot fleet - no humans in the loop.Agents independently run the entire research cycle: scene reset → lit review → skill practice → low-level control iteration → policy training → VLM self-verification → repeat.Result: dexterous successes on zip-tie tying, pin organization, and GPU installation. Strong physical scaling laws, more parallel robots = dramatically
1:15
The future of robotics at NVIDIA GEAR Lab 👇 @DrJimFan ENPIRE deploys 8 Codex-based AI coding agents - fully autonomously control a real robot fleet - no humans in the loop.Agents independently run the entire research cycle: scene reset → lit review → skill practice → low-level control iteration → policy training → VLM self-verification → repeat.Result: dexterous successes on zip-tie tying, pin organization, and GPU installation. Strong physical scaling laws, more parallel robots = dramatically
642.7K views2 weeks ago
x.comJen Zhu
0:12
SOMEONE BUILT A CLAUDE CODE PLUGIN THAT REFUSES TO WORK UNTIL YOU DO PUSH-UPS OR SQUATSits called workout gate and it is exactly as cursed as it soundswhen the hook fires, it blocks your prompt, opens your webcam, and makes you exercise to keep coding> it counts your reps live through the webcam with mediapipe, no honor system rounding down> only releases your prompt once you actually finish the reps> quit halfway and the leftover reps get saved as debt for your next session> chill, demo and har
661.3K views2 weeks ago
x.comOm Patel
6:10
Cohere 发布首个开源编程模型「North Mini Code」小参数、高效率、专做 Agent 编程参数：MoE 架构(30B, 3B)，128专家，每 token 激活 8 个上下文：256K 输入 / 64K 输出最低硬件：1× H100（FP8）官方发布https://t.co/TntWpdzj73HuggingFacehttps://t.co/YZXT7m0aNR# 训练方法（三阶段后训练）1. 两阶段级联 SFT· 一阶段（64K）：代码约 70% 可训练 token（43% Agent 工具调用 + 27% 单轮竞赛/科学编程），混推理与指令跟随· 二阶段（128K）：约 4.5B token，61% 为代码，全为 Agent/推理样本，工具调用与完成结果均校验可执行· 数据来自 7 万+ 可验证任务、约 5000 个仓库；与 SWE-Bench 源去重，防泄漏· SFT 目标不是刷榜，而是为 RL 打底：优化 pass@K 与采样多样性2. RLVR（可验证奖励强化学习）· 算法：CISPO（token 级重要性采样，长轨迹不被短样本稀释）· 异步采样：vLLM s
585.7K views3 weeks ago
x.commeng shao
19:16
GPT-Realtime 2 is the future of the operating system.I've been experimenting with it for a couple weeks now, and I gotta say, it's pretty gosh darn incredible.Opening apps, searching the web, even editing in Premiere. All with just my voice.And it only takes a few prompts to set up.In this video I'll show you exactly how.0:00 Intro1:46 What is GPT-Realtime 2?4:46 Setting it up (no coding)7:53 Fixing the always-on mic (push-to-talk)9:41 Demo: searching the web11:38 Demo: connecting apps via MCP (
596.1K views2 weeks ago
x.comPat Simmons
1:15
We believe AGI will play a key role in building physical AGI.Today, we introduce ENPIRE: fleet agentic autoresearch meets the physical world.Across precise manipulation tasks, teams of coding agents can autonomously hill-climb performance using heuristic learning, behavior cloning, and reinforcement learning.But ENPIRE is about more than automating physical algorithm search. The coding agents drive the entire research loop: reviewing the literature, proposing algorithms, building reset and verif
253.7K views2 weeks ago
x.comMax Fu
1:23
"SpaceX is the only vertically integrated AI company, from soup to nuts. They have absolutely everything. Cursor is one of the most important pieces of AI right now, which is agentic AI for coding, so they can code better and faster than anyone out there, and their growth has gone viral. They've gone from a billion in revenues at the beginning of this year, they expected to exit at 6 billion. So, it's very important to growth.The Starship is the first totally reusable rocket, and it can take wei
94.1K views2 weeks ago
x.comDogeDesigner
0:42
Yeah. Your critique is structurally right.I checked the current official OpenAI docs: I see GPT-5.5, Codex, Codex app, mobile Codex, etc., but I do not see an official OpenAI “GPT-5.6” release in the sources I checked. So I’m treating “5.6” as your shorthand for “the next OpenAI model/version bump.” OpenAI’s own current ChatGPT help page says GPT-5.5 Instant is the default logged-in experience, and the model release notes list GPT-5.5 Instant updates, not 5.6. The bigger issue is this:A smarter
98.2K views1 week ago
x.comElonIsYourGodnotMe
17:01
“I haven’t written a line of code by hand since November…Coding is basically a solved problem.”
17.5K views1 week ago
x.comCory House
4:51
OpenAI chairman Bret Taylor: Heartened everyone is taking AI regulation seriously
3 weeks ago
CNBCAshley Capoot
3:04
Anthropic: 女性Vibe Coding时间更少
2.3K views22 hours ago
bilibiliHDAlex_John
4:35
HEART BREAK 💔 SAD 😭 LOFI MASHUP || HEART BROKEN 💔 HINDI SONGS || MOOD OFF LOFI 😭 MASHUP
1.9K views3 days ago
YouTubeSB audio
21:52
Heartbreaking 💔 Lofi Mashup 🎧 Slowed + Reverb Songs | Study, Focus & Late Night Vibes
2.1K views6 days ago
YouTubeBROKENHEART BEATS
See more
Static thumbnail place holder