BitRiftTech's profile picture. Software Engineer & AI enthusiast

Matthew Johnson

@BitRiftTech

Software Engineer & AI enthusiast

Matthew Johnson reposted

Laser Cutting Blade (80W) 🔥😳 #BoysWithToys


Matthew Johnson reposted

Top 7 local AI models you can run on a laptop: 1. Qwen3 Coder 30B 3A: best for coding basically anything 2. Gemma 3n E4B: small, versatile model that even runs offline on a phone 3. Magistral Small 1.2: multimodal with vision capabilities and pretty solid for coding 4. Hermes…


Matthew Johnson reposted

If you're learning math for AI/ML, try out Deep ML. It's a leetcode style platform for ML, focusing on code implementations for a variety of problems > Lin Alg / Calc / stats / prob > ML / DL / RL > Optimization / CV / NLP Solve a qs as you learn it! deep-ml.com

TheGlobalMinima's tweet image. If you're learning math for AI/ML, try out Deep ML. It's a leetcode style platform for ML, focusing on code implementations for a variety of problems
> Lin Alg / Calc / stats / prob
> ML / DL / RL 
> Optimization / CV / NLP

Solve a qs as you learn it!

deep-ml.com

Matthew Johnson reposted

SAVE THESE SITES FOREVER!!! All the links inside:

swapnakpanda's tweet image. SAVE THESE SITES FOREVER!!!

All the links inside:

Matthew Johnson reposted

this is beyond mindblowing for me. somebody built a 5 million param language model inside minecraft, trained it, equipped it with basic conversational ability. probably the best thing i have seen entire month.

tokenbender's tweet image. this is beyond mindblowing for me.

somebody built a 5 million param language model inside minecraft, trained it, equipped it with basic conversational ability.

probably the best thing i have seen entire month.

Matthew Johnson reposted

Hey ambitious devs, this weekend you should expand your skills by building your own LLM from scratch with Python. 🐍 freeCodeCamp just dropped a 6-hour intermediate course on this. Here are this week's five freeCodeCamp resources that are worth your time: 1. freeCodeCamp just…

ossia's tweet image. Hey ambitious devs, this weekend you should expand your skills by building your own LLM from scratch with Python. 🐍

freeCodeCamp just dropped a 6-hour intermediate course on this. Here are this week's five freeCodeCamp resources that are worth your time:

1. freeCodeCamp just…

Matthew Johnson reposted

I spent the past month reimplementing DeepMind’s Genie 3 world model from scratch Ended up making TinyWorlds, a 3M parameter world model capable of generating playable game environments demo below + everything I learned in thread (full repo at the end)👇🏼


Matthew Johnson reposted

🏆📚This 200-Page LLM Paper Is a 𝗚𝗼𝗹𝗱𝗺𝗶𝗻𝗲 — and it’ll save you months 𝗣𝗿𝗼𝗺𝗽𝘁𝗶𝗻𝗴, 𝘁𝗿𝗮𝗶𝗻𝗶𝗻𝗴, 𝗮𝗹𝗶𝗴𝗻𝗺𝗲𝗻𝘁 — finally crystal clear. If you don’t have time to read all 200+ pages, here are the most valuable 𝘁𝗮𝗸𝗲𝗮𝘄𝗮𝘆𝘀 ↓ 》…

MaryamMiradi's tweet image. 🏆📚This 200-Page LLM Paper Is a 𝗚𝗼𝗹𝗱𝗺𝗶𝗻𝗲 — and it’ll save you months
𝗣𝗿𝗼𝗺𝗽𝘁𝗶𝗻𝗴, 𝘁𝗿𝗮𝗶𝗻𝗶𝗻𝗴, 𝗮𝗹𝗶𝗴𝗻𝗺𝗲𝗻𝘁 — finally crystal clear.
If you don’t have time to read all 200+ pages, here are the most valuable 𝘁𝗮𝗸𝗲𝗮𝘄𝗮𝘆𝘀 ↓

》…

Matthew Johnson reposted

Introduction to Reinforcement Learning. The only book you need to fundamentally understand how agents learn optimal behavior through trial and error. 100% free to read.

Saboo_Shubham_'s tweet image. Introduction to Reinforcement Learning.

The only book you need to fundamentally understand how agents learn optimal behavior through trial and error.

100% free to read.

Matthew Johnson reposted

I tasted success as a young man but was never able to sustain it, because I lived by "feel" instead of by "have to". Once I stopped making excuses and started doing things because I had to, instead of because I wanted to, everything changed. This goes HARD.


Matthew Johnson reposted

I built a free tool that calculates any X account's TweepCred score so you can find out if you're actually shadowbanned or not


Matthew Johnson reposted

Matthew Johnson reposted

I mean, you could just self-host Supabase you know?

dshukertjr's tweet image. I mean, you could just self-host Supabase you know?

Matthew Johnson reposted

Wait so Alibaba Qwen has just released ANOTHER model?? Qwen3-Coder is simply one of the best coding model we've ever seen. → Still 100% open source → Up to 1M context window 🔥 → 35B active parameters → Same performance as Sonnet 4 They're releasing a CLI tool as well ↓

itsPaulAi's tweet image. Wait so Alibaba Qwen has just released ANOTHER model??

Qwen3-Coder is simply one of the best coding model we've ever seen.

→ Still 100% open source
→ Up to 1M context window 🔥
→ 35B active parameters
→ Same performance as Sonnet 4

They're releasing a CLI tool as well ↓

Matthew Johnson reposted

This is one of the coolest interactive sites I’ve seen for learning how LLMs work! 🔥 It kicks off with a clear intro and walks you through every core component: from Embeddings, Layer Norm, & Self-Attention to MLPs, Transformer blocks, Softmax, and Output layers. link in 🧵↓


Matthew Johnson reposted

Finally Vsauce remembered his password after 2 yrs. He's my all time favourite, the way he goes to the depth in curiosity and that unique style is amazing. youtu.be/xHd4zsIbXJ0?si…

munen5647's tweet card. All The Ghosts You Will Be

youtube.com

YouTube

All The Ghosts You Will Be


Matthew Johnson reposted

Starting today you can run any of the 100K+ GGUFs on Hugging Face directly with Docker Run! All of it one single line: docker model run hf.co/bartowski/Llam… Excited to see how y'all will use it

XciD_'s tweet image. Starting today you can run any of the 100K+ GGUFs on Hugging Face directly with Docker Run! 

All of it one single line: docker model run hf.co/bartowski/Llam…

Excited to see how y'all will use it

Matthew Johnson reposted

🚀Our rStar-Coder dataset is now released! A verified dataset of 418K competition-level code problems, each with test cases of varying difficulty. On LiveCodeBench, it boosts Qwen2.5-14B from 23.3% → 62.5%, beating o3-mini (low) by +3.1%. Try it here: huggingface.co/datasets/micro…


Matthew Johnson reposted

Claude Code pro tip: Use Kimi K2 as your agent by changing the base URL and setting the API key before running claude. all prompts and code requests will now go through Kimi, in the CC terminal you're used to!

iannuttall's tweet image. Claude Code pro tip:

Use Kimi K2 as your agent by changing the base URL and setting the API key before running claude.

all prompts and code requests will now go through Kimi, in the CC terminal you're used to!

Matthew Johnson reposted

LOVE ITT! You can run Kimi K2 (1T token MoE) on a single M4 Max 128GB VRAM (w/ offloading) or a single M3 Ultra (512GB) 🔥 The model was released less than 72 hours ago - love how fast the community optimises open weights - kudos to @UnslothAI 🤗 huggingface.co/unsloth/Kimi-K…


United States Trends

Loading...

Something went wrong.


Something went wrong.