DEV Community

#llm

Posts

đź‘‹ Sign in for the ability to sort posts by relevant, latest, or top.
Why RL Training Collapses on Long-Horizon Agents

Why RL Training Collapses on Long-Horizon Agents

Comments 2
10 min read
What 78K attack samples taught me about catching prompt injection

What 78K attack samples taught me about catching prompt injection

Comments
2 min read
I built a spend cap for LLM calls. It failed by 4.2x under parallel load.

I built a spend cap for LLM calls. It failed by 4.2x under parallel load.

1
Comments 2
5 min read
Predicting the Speed of a 276B Model Streamed From an SSD

Predicting the Speed of a 276B Model Streamed From an SSD

2
Comments 2
8 min read
Your Knowledge Graph Is Wasting 70% of Its Tokens

Your Knowledge Graph Is Wasting 70% of Its Tokens

1
Comments
2 min read
Most of Your LLM Spend Is Wasted on Calls That Don't Need a Frontier Model

Most of Your LLM Spend Is Wasted on Calls That Don't Need a Frontier Model

7
Comments 12
4 min read
A 2-Token Prompt and a 39,966-Token Bill: Measuring What My Agent Actually Costs

A 2-Token Prompt and a 39,966-Token Bill: Measuring What My Agent Actually Costs

2
Comments 1
5 min read
We made SigNoz's LLM observability actually reachable…

We made SigNoz's LLM observability actually reachable…

Comments
8 min read
Is AI Limited by the Exact Superpower That Made It Powerful?

Is AI Limited by the Exact Superpower That Made It Powerful?

Comments
7 min read
Tracing Voice AI is Hard: How I Instrumented Streaming LLMs with OpenTelemetry and SigNoz

Tracing Voice AI is Hard: How I Instrumented Streaming LLMs with OpenTelemetry and SigNoz

Comments
4 min read
Stop Validating AI Agents Only at the Start: Introducing Mid-Chain Governance

Stop Validating AI Agents Only at the Start: Introducing Mid-Chain Governance

Comments
4 min read
🛡️ Arquitectura de Defensa para Agentes de IA: Cómo asegurar tus LLMs contra Prompt Injection, Tool-Poisoning y Fugitividad.

🛡️ Arquitectura de Defensa para Agentes de IA: Cómo asegurar tus LLMs contra Prompt Injection, Tool-Poisoning y Fugitividad.

3
Comments
3 min read
LangChain’s tool routing is a bloated mess. So we built a <10ms local semantic registry to replace it at US Neural.

LangChain’s tool routing is a bloated mess. So we built a <10ms local semantic registry to replace it at US Neural.

Comments
1 min read
Turn any PDF into clean Markdown with a self-hosted Docling API

Turn any PDF into clean Markdown with a self-hosted Docling API

1
Comments
2 min read
Agentes 101 - 02: El ciclo del agente como State Machine (y Hierarchical State Machines)

Agentes 101 - 02: El ciclo del agente como State Machine (y Hierarchical State Machines)

Comments
5 min read
đź‘‹ Sign in for the ability to sort posts by relevant, latest, or top.