g/c @tensor_ninja

Building things @sylowAI Joined September 2023

Tweets

147
Followers

36
Following

313
Likes

309

Ryan Dahl @rough__sea

5 days ago

People will eventually become nostalgic for early AI slop

24 7 195 16K 5

g/c @tensor_ninja

a week ago

The network is the computer

0 0 0 5 0

Andreas Kling @awesomekling

3 weeks ago

Hiring people to post under your name online is ridiculous, and we should make fun of anyone who does this.

10 3 234 14K 2

g/c @tensor_ninja

3 weeks ago

What do vibe coders do when this happens? API Error: 500 {"type":"error","error":{"type":"api_error","message":"Overloaded"}}

1 0 0 48 0

g/c @tensor_ninja

3 weeks ago

Is anyone else struggling with token rate limits/token cost when scaling an AI app?

1 0 4 126 0

g/c @tensor_ninja

2 months ago

You still use RAG, I train symmetric LoRA adapters. We are not the same.

0 0 0 33 0

g/c @tensor_ninja

2 months ago

git push --force-with-aggression

0 0 1 27 0

trash @trashh_dev

2 months ago

when claude one shots my 5 point jira ticket

59 199 7K 360K 376

Download Image

cloud @cloud11665

3 months ago

Feeling good, might cook some hand-tuned SIMD (AVX512-BF16) subroutines later

0 1 17 954 0

Umar Jamil @hkproj

3 months ago

Hardcore devs create strong LLMs. Strong LLMs create weak devs. Weak devs create dumb LLMs. Dumb LLMs create hardcore devs.

70 397 4K 132K 388

g/c @tensor_ninja

4 months ago

Day 1 using Nix on MaxOS with Nix-Darwin. It’s great and easy to set up…what’s the catch?

0 0 1 32 0

Charlie Marsh @charliermarsh

4 months ago

Slightly concerned by how much garbage I would be churning out if I was trying to use these tools without significant software engineering experience

42 45 1K 59K 61

kitze @thekitze

4 months ago

i'm genuinely sorry for vibe coders who don't understand the generated code. it must be a nightmare

249 113 3K 461K 267

@francoisfleuret I’d say, in no particular order: 1 - FlashAttention, mostly because you can avoid materialising the weights matrix. 2 - MoE - scaling laws indicate it’s always more convenient to have it. 3 - Ring Attention + RoPE - long context 4 - GQA - reduce KV cache size

8 35 683 19K 482

Observer @Observer_ofyou

5 months ago

God is always with us .

60 1K 10K 1.5M 2K

Download Image

g/c @tensor_ninja

5 months ago

Boss: "HEY! GET BACK TO WORK!" 2025 Programmers: "CURSOR AGENT!" Boss: "OH. CARRY ON."

0 0 0 39 0

Download Image

DeepStation @DeepStationAI

6 months ago

🌴 𝗠𝗜𝗔𝗠𝗜 𝗔𝗜 𝗦𝗨𝗠𝗠𝗜𝗧 𝟮𝟬𝟮𝟱 🌴 🎙️ 𝗔𝗻𝗻𝗼𝘂𝗻𝗰𝗶𝗻𝗴 *𝗦𝗻𝗲𝗮𝗸 𝗣𝗲𝗮𝗸* 𝗦𝗽𝗲𝗮𝗸𝗲𝗿 𝗟𝗶𝗻𝗲𝘂𝗽 GPTuesdays and our partners @LightningAI & @PyTorch are thrilled to unveil the first wave of exceptional speakers who will be sharing their insights during…

0 2 10 495 0

Download Image

Josh Fischer @joshfischer1108

6 months ago

Let's be real: building a company from scratch is no easy feat. Who is going to survive?

0 1 7 111 0

Download Video

DeepStation @DeepStationAI

6 months ago

🎙GPTuesdays is excited to announce our very first “South Florida AI Summit” in partnership with @LightningAI & @PyTorch Foundation! 💥 The audience will have the unique opportunity to meet speakers with experience from Microsoft, Lightning AI, @RedHat, General Motors,…