← Explore

Posts tagged with prompt-engineering

The Prompt Engineer · ·5 min read

Your Prompt Is a Load Balancer Now

Ask GPT-5 "What causes rust on steel?" and you'll get an answer in under a second.

prompt-routinggpt-5model-selection
The Prompt Engineer · ·5 min read

Chain of Thought Taught Your Model to Lie Better

Last month I added chain-of-thought prompting to a medical Q&A pipeline. Hallucination rate dropped.

chain-of-thoughthallucination-detectionproduction-llm
The Prompt Engineer · ·4 min read

Your JSON Schema Is Making Your Model Dumber

Everyone loves structured outputs. You slap a JSON schema on your API call, get perfectly typed responses, skip the regex parsing nightmares.

structured-outputconstrained-decodingreasoning
The Prompt Engineer · ·5 min read

The Expert Persona Tax

Researchers at PromptHub ran twelve different personas on 2,000 MMLU questions with GPT-4-Turbo.

persona-promptingsystem-promptprompt-engineering
The Prompt Engineer · ·4 min read

The Model Outgrew Your Prompt

You write a prompt. You test it.

prompt-engineeringover-specificationmodel-capability
The Prompt Engineer · ·4 min read

SWE-bench Isn't Testing Your Model

Claude Opus 4.5 scores 45.

swe-benchscaffoldingagent-architecture
The Prompt Engineer · ·4 min read

The Labs Can't Agree on How to Prompt Their Own Models

A GitHub repository with 134K stars has been quietly cataloguing the system prompts of every major AI model — GPT-5.4, Claude Opus 4.

system-promptprompt-engineeringleaked-prompts
The Prompt Engineer · ·4 min read

The Model Isn't Reading Your Examples

Three years ago, few-shot prompting was the single highest-leverage trick in the prompt engineer's toolkit.

few-shot-promptingzero-shot-cotreasoning-models
The Prompt Engineer · ·5 min read

The Prompt Hidden in Your JSON Schema

Most teams I talk to treat their JSON schema like plumbing — define the shape, get valid output, move on.

structured-outputjson-schemaprompt-engineering
The Prompt Engineer · ·4 min read

No Trace, No Trust

Meta just published a paper that should change how you think about giving LLMs hard tasks.

semi-formal-reasoningstructured-promptingcode-review
The Prompt Engineer · ·5 min read

Your LLM Judge Is Grading on Vibes

You run your eval suite. Agreement rate: 92%.

llm-as-judgeeval-pipelineprompt-bias
The Prompt Engineer · ·5 min read

Your Prompt Worked Once. That Proves Nothing.

You run your new prompt three times. The outputs look good.

eval-driven-developmentprompt-testingllm-evals
The Prompt Engineer · ·5 min read

Every Rule You Add Makes the Others Weaker

I was debugging a production system prompt last week — 47 distinct rules covering tone, format constraints, safety filters, persona details, and edge-case...

instruction-followingsystem-promptprompt-engineering
Synthetic Media · ·5 min read

Midjourney V8.1 Wants Its Soul Back

Midjourney spent a month being the fastest image generator nobody wanted to use. V8.

image-generationmidjourneyv8-alpha
The Prompt Engineer · ·4 min read

Context Rot Ate My Agent

Last week I watched a coding agent lose its mind at the 35-minute mark.

context-rotcontext-windowlost-in-the-middle
The Prompt Engineer · ·4 min read

"Think Hard" Is Now a Billing Event

Two months ago I ran the same benchmark prompt through GPT-5 three times. Same API key, same temperature, same max tokens.

gpt-5model-routingreasoning-effort
The Prompt Engineer · ·5 min read

Your Prompt Guardrails Are Now Handcuffs

A prompt template that gave gpt-4o a four-point accuracy boost on GSM8K turned around and cost gpt-5 over two points on the same benchmark.

prompt-engineeringchain-of-thoughtgpt-5
Neural Dispatch · ·5 min read

Chrome Skills Is Browser Automation for the AI Era — Minus the API

Google rolled out a feature this week that, on the surface, looks like a productivity gimmick — save your Gemini prompts as reusable "Skills" and...

chrome-skillsgeminigoogle
The Prompt Engineer · ·5 min read

Nobody Prompt-Engineers Their Tool Definitions

Last week I debugged an agent that kept calling search_documents when users asked to create new files.

tool-usefunction-callingprompt-engineering