DEV Community

#llm

Posts

👋 Sign in for the ability to sort posts by relevant, latest, or top.
Prompt Pad: What Is a Context Window?

Prompt Pad: What Is a Context Window?

27
Comments
1 min read
Running a video-by-code agent on one RTX 4090: the context budget is the real limit

Running a video-by-code agent on one RTX 4090: the context budget is the real limit

Comments
4 min read
One turn of our Claude Code agent used 3.57 billion tokens. 22 long-lived teammates spent 2.34 billion of it

One turn of our Claude Code agent used 3.57 billion tokens. 22 long-lived teammates spent 2.34 billion of it

Comments
7 min read
I Put Five AI Agents at a Brazilian Truco Table. One Built a Monte Carlo Simulator and Still Got Crushed.

I Put Five AI Agents at a Brazilian Truco Table. One Built a Monte Carlo Simulator and Still Got Crushed.

Comments
7 min read
vLLM interview questions, answered with numbers from a real GPU

vLLM interview questions, answered with numbers from a real GPU

Comments
13 min read
Inkling 2 Mini Lands Open Weights on Laptop Hardware — Dual-Track WASAPI Still Sets the Overlay Floor

Inkling 2 Mini Lands Open Weights on Laptop Hardware — Dual-Track WASAPI Still Sets the Overlay Floor

Comments
4 min read
On-device AI agents got real this month. Where a small local model beats a big remote one.

On-device AI agents got real this month. Where a small local model beats a big remote one.

Comments
2 min read
Inference Engineering in Production: LLM Serving, Optimization, and Scale

Inference Engineering in Production: LLM Serving, Optimization, and Scale

5
Comments
8 min read
Output tokens are 17% of the volume and 50% of the bill — a year of Gemini 4 Neon vs Claude Opus 5.5

Output tokens are 17% of the volume and 50% of the bill — a year of Gemini 4 Neon vs Claude Opus 5.5

Comments
4 min read
Why My Local LLM (Qwen 3.5 9B) Rated Every Board Game 50/50 — And How Switching to Gemini 3.5 Flash-Lite Fixed It

Why My Local LLM (Qwen 3.5 9B) Rated Every Board Game 50/50 — And How Switching to Gemini 3.5 Flash-Lite Fixed It

Comments
4 min read
What I let a 4B model do on a Raspberry Pi 5, and what I took away from it

What I let a 4B model do on a Raspberry Pi 5, and what I took away from it

Comments
4 min read
What an agent harness actually owns: 8 reads on loops, cost routing, exactly-once, durability and spec authority

What an agent harness actually owns: 8 reads on loops, cost routing, exactly-once, durability and spec authority

Comments
4 min read
Run SenseNova-Loop-9B on Consumer GPUs: Deploying the 9B Multimodal Vision Model That Outperforms 27B with Local Loop Reasoning in 2026

Run SenseNova-Loop-9B on Consumer GPUs: Deploying the 9B Multimodal Vision Model That Outperforms 27B with Local Loop Reasoning in 2026

Comments
9 min read
Private, OpenAI-compatible LLM APIs compared: zero retention, TEE and verifiable enclaves

Private, OpenAI-compatible LLM APIs compared: zero retention, TEE and verifiable enclaves

Comments
7 min read
The context cliff: 200K, 272K, and the line that doubles your prompt bill

The context cliff: 200K, 272K, and the line that doubles your prompt bill

Comments
4 min read
👋 Sign in for the ability to sort posts by relevant, latest, or top.