larkhere @larkhere

Joined November 2014

Tweets

14
Followers

15
Following

509
Likes

1K

Releasing the model weights and technical report of Kimi K3. Kimi K3 is our most capable model: a 2.8T MoE model with native visual understanding and a 1M-token context window. New model architecture: 2.5x the intelligence per unit of compute, not just more params. Alongside

1K 7K 46K 14.2M 14K

0 0 0 6 0

View Details

HeyGen @HeyGen

a week ago

HyperFrames -> Video Agent Our HTML motion graphics framework, now inside HeyGen Prompt what you want, our agent builds the entire video Avatar, graphics, captions, music Infinitely customizable, just ask Comment + RT "prompt" for a free month of HeyGen (must follow)

3K 2K 2K 1.7M 1K

View Details

larkhere @larkhere

a week ago

@HeyGen prompt

1 0 0 106 0

View Details

larkhere @larkhere

2 weeks ago

🚀How I’ve put ChatGPT Work to work: As a React Native + SwiftUI indie dev, I used it to rapidly prototype my personal task management app. I fed it rough notes, user flows, and screenshots — it turned them into a detailed product brief, component architecture, clean task breakdown, and even generated solid starter code for Expo + iOS push notifications. What used to take me days of back-and-forth planning now happens in one focused session. It spots edge cases I missed and keeps perfect context across files. Game-changer for shipping faster. Huge thanks to the OpenAI team — loving GPT-5.6 Sol in ChatGPT Work!

0 0 1 42 0

View Details

larkhere @larkhere

3 weeks ago

GPT-5.6 Sol is truly breathtaking. What I’ve been doing with it: • Prototyping full-stack features for my React Native + Mac apps (health tracking app) • Building complex agentic workflows and self-hosted tools (NAS automation) • Long-horizon coding sessions: planning architecture, generating production-ready code, debugging, and iterating faster than ever I love how Sol handles deep context, multi-step reasoning, and computer-use tasks with high efficiency and polished output. It truly feels like a capable technical co-founder. This model is a massive upgrade for real builders.

0 0 1 8 0

View Details

larkhere @larkhere

3 weeks ago

Switched to Codex with GPT-5.6 Sol and it’s been a game changer. What I love most: exceptional reasoning on complex tasks, strong agentic capabilities, computer use, and coding workflows. It’s significantly more token-efficient than previous models, delivers polished results faster, and feels like a real collaborative partner rather than just a chatbot. Perfect for long-horizon projects and actual shipping.