Share this page!

Enter URL or ID to Unroll

You can paste full URL like: https://x.com/threadreaderapp/status/1644127596119195649
or just the ID like: 1644127596119195649

How to get URL link on X (Twitter) App

  1. On the Twitter thread, click on or icon on the bottom
  2. Click again on or Share Via icon
  3. Click on Copy Link to Tweet
  4. Paste it above and click "Unroll Thread"!
  5. More info at Twitter Help
Min Choi Profile picture
Min Choi
@minchoi
AI Educator. 𝕏 about AI, solutions and interesting things. Showing how to leverage AI in practical ways for you and your business. Opinions are my own.
18 subscribers
Jan 30 • 13 tweets • 4 min read
Less than 61 hours ago, Alibaba shocked the world with Qwen 2.5 Max, its most powerful multimodal AI yet.

Not just AI chat, now you can generate videos, images, and preview apps with Artifacts.

10 examples: 👇 Image 1. Preview your code with Artifacts
Jan 29 • 14 tweets • 4 min read
DeepSeek R1 is wild.

But you don’t have to send your data to China.

12 ways to use it securely: 👇 1. Run FULL DeepSeek R1 model locally on $6,000 hardware

Motherboard: Gigabyte MZ73-LM0 or MZ73-LM1
CPU: 2x AMD EPYC 9004 or 9005 CPU
RAM: 24 x 32GB DDR5 RDIMM
Case: Enthoo Pro 2 Server
PSU: Corsair HX1000i
Heatsink: CoolServer AMD SP5-4U M97
Jan 28 • 13 tweets • 4 min read
DeepSeek R1 has really changed the AI LLM game.

People are creating wild use cases beyond ChatGPT. There's a major shift.

10 examples: Image 1. DeepSeek R1 is the best model you can use in VS Code
Jan 26 • 10 tweets • 3 min read
It's only been about 48 hours since OpenAI dropped Canvas update. Now you can use o1 and make React apps.

And people are going crazy creative with it.

8 wild examples: 1. Build a Windows 95 simulator
Jan 25 • 12 tweets • 4 min read
DeepSeek R1 broke the Internet just 5 days ago.

People can't believe how powerful and "intelligent" it is. Unlocking new possibilities.

10 wild examples: 1. DeepSeek R1 1.5B runs on iPhone 16, AI reasoning model in your pocket 🤯

Jan 22 • 13 tweets • 4 min read
It's only been just over 48 hours since DeepSeek dropped DeepSeek-R1, fully open source reasoning model.

And people are getting crazy creative with it.

10 wild examples: 1. AGI at home🤯

Running DeepSeek R1 across 7 M4 Pro Mac Minis + 1 M4 Max MacBook Pro
Jan 21 • 13 tweets • 4 min read
This is wild.

Hailuo AI just completely changed the character consistency game with "Subject Reference"

Just from a single image🤯

10 wild examples: 1. Character reference image using Minimax model on Replicate
Jan 19 • 11 tweets • 3 min read
Wow, Veo 2 is unreal

100% AI YouTube and TikTok influencer videos🤯

1. Makeup vlog 2. Live game streaming
Jan 19 • 13 tweets • 4 min read
Whoa, Hailuo just changed the voice synthesis game with the T2A-01-HD model.

This isn’t just voice synthesis, it’s the industry’s first emotional intelligent system AI that replicates even subtle emotions in speech.

10 wild examples (and how to try): 👇 1. Clone your own voice + lip sync
Jan 18 • 8 tweets • 2 min read
Hunyuan AI + LoRA is insane.

It can replace any actors in any film. 🤯

6 wild examples: 1. Severance to John Wick 🤯
Jan 17 • 9 tweets • 3 min read
It’s happening.

Humanoid robots are walking, running, falling and getting back up.

They’re becoming smarter, faster, and autonomous.

8 wild examples:

1. Unitree G1 jogging like a human
2. Booster T1 will probably remember this "bullying" 😬
Jan 16 • 11 tweets • 4 min read
Less than 48 hours ago, OpenAI dropped ChatGPT with Scheduled Tasks.

And people can't stop building creative automation task to have AI agents work while they sleep.

10 wild examples:

1. Daily stock market analysis 2. Improve your life by having AI agent do the work for your 24/7
Jan 16 • 12 tweets • 3 min read
Veo 2 is insane.

People can't stop creating insane cinematic and realistic AI videos just from text.

10 wild examples. 1. Makeup influencers, not real
Jan 14 • 12 tweets • 4 min read
This is wild!

Hailuo AI just dropped MiniMax-01, an industry-first foundation models with Lightning Attention.

And it supports mind blowing 4 MILLION token context size, that’s the entire Harry Potter series three times over!

Why it’s a game-changer and how to try: Image HUGE 4M Context size

MiniMax-Text-01 can handle up to 4 million tokens, that’s 32x GPT-4o, 20x Claude-3.5-Sonnet and 2x Gemini 2.0.

And achieved 100% accuracy in the 4 million token vanilla Needle-In-A-Haystack retrieval task

Performance Benchmarks

MiniMax-01 models are on par with GPT-4o and Claude-3.5 in most tasks and outperform Google Gemini on long-context tasks.

Minimal performance degradation as input length grows.

Lightning Attention Architecture

An industry-first linear attention mechanism alternative to traditional transformers.

Boasting 456B parameters with 45.9B activated per inference.

API Pricing

- $0.2 per million input tokens
- $1.1 per million output tokensImage
Image
Image
Jan 11 • 11 tweets • 3 min read
That’s a wrap!

The world’s biggest tech event, CES 2025, is over and it was incredible.

10 wild reveals you NEED to see:

1. EyeBliss by UltraReality tracks your gaze and blinking
2. Futuristic wearable robotic suit by Sony
Jan 10 • 12 tweets • 3 min read
This is wild.

Tencent just dropped SVFR (Stable Video Face Restoration). AI tool that turns low-quality videos into stunning, high-quality face videos.

10 wild examples

1. Albert Einstein 2.
Jan 9 • 18 tweets • 5 min read
Humanoid robots are here

And they already walking and working like humans

At CES 2025, NVIDIA CEO Jensen Huang said: "The ChatGPT moment for general robotics is just around the corner"

17 wild examples you NEED to see:

1. EngineAI SE01 walking around in Shenzhen, China 2. Unitree G1 humanoid robot at CES 2025
Jan 8 • 11 tweets • 3 min read
Gaming will never be the same.

NVIDIA just dropped ACE at CES 2025, a game changing AI that powers Autonomous Game Characters that think, adapt, and interact in real time.

9 wild examples:

1. AI PUBG teammate like a real player 2. AI Boss in MIR5 will learn from past battles and adapt to your tactics for unique fights
Jan 7 • 11 tweets • 3 min read
It was only Day 1 of CES 2025.

And the announcements are already mind blowing.

10 wild reveals:

1. BYD's supercar Yangwang U9 jumping 6 meters forward using "jumping suspension"
2. OMNIA, a 360° AI health mirror that scans your body, syncs data, connects to doctors, and offers AI powered insights
Jan 7 • 13 tweets • 4 min read
NVIDIA just broke the Internet at CES2025.

Jensen Huang dropped some mind blowing announcements.

Here are 10 key reveals you can't miss: Image 1. Project DIGITS, $3,000 personal AI supercomputer 1,000x the power of average laptop
Jan 6 • 7 tweets • 2 min read
We are cooked

100% AI 🤯

1. 2.