Tagged · 21 notes
Everything tagged LLMs, wherever it's filed, newest first.
The tells that make people ask "did AI write this?" in short emails and messages, a paste-in style block to hand the model, and a 30-second checklist to run before you hit send.
Thread: Changing how the model answers ↗Absolute Mode is a widely shared prompt that strips a chatbot down to blunt, unpadded answers, and this is my copy with notes on when it's the wrong tool.
Thread: Changing how the model answers ↗When a long back-and-forth with an AI finally lands somewhere good, the method is worth more than the answer. How I pull the method out and turn it into a prompt I can run again.
Thread: Building with Claude and MCP ↗A grounded take on the "tell me what you know about me" prompt that makes the model show its evidence and admit when it's guessing.
Thread: Prompts that hold up a mirror ↗God Mode is a widely shared prompt dressed up as a cognition protocol, and underneath the theater it reliably gets deeper, more structured thinking out of a model.
Thread: Prompts for plans, apps and side hustles ↗Eight short prompts for everyday jobs like sorting screenshots, planning dinner from a fridge photo and writing a resale listing, with a note on where each one needs a human check.
Thread: Prompts for plans, apps and side hustles ↗The Truth Prompt pushes a model to stick to what it can support, say "I cannot confirm this" out loud, and check itself for fabrication, which helps most when web search is on.
Thread: Changing how the model answers ↗API security, MCP security, an AI gateway and a verification engine, one at a time: what each one looks at and what it's there to catch.
Thread: Building G8KEPR ↗A prompt that points out where your stated values and your actual behavior don't match, and only flags a contradiction it can back up with both sides.
Thread: Prompts that hold up a mirror ↗A widely shared prompt that turns an app idea into a full spec you can hand to a no-code builder, plus the quick version for ideas you just saw in a post.
Thread: Prompts for plans, apps and side hustles ↗Execution Mode is a circulating prompt that cuts the pleasantries, apologies and hedging so every reply starts with the answer and stops when it's done.
Thread: Changing how the model answers ↗A pair of prompts that project where your current patterns lead in five years, then run the same projection assuming you fix your biggest blind spot.
Thread: Prompts that hold up a mirror ↗The widely shared "$10K a month" Growth Engine prompt walks a model through a four-phase side-income plan, and the structure is useful even if the headline number isn't.
Thread: Prompts for plans, apps and side hustles ↗A three-phase reflection prompt for a pattern you keep falling into, built to offer its reads as guesses you can confirm or reject rather than pronouncements.
Thread: Prompts that hold up a mirror ↗My old notes had a pipeline that wrote and published WordPress posts on a schedule. What I'd keep from it, what I'd throw out, and how I'd use AI for a client's site copy now.
Thread: Building a site for someone ↗The Life OS prompt sets a model up as a chief of staff for work and life, and it's useful as long as you know what it can't see and what you shouldn't hand it.
Thread: Prompts for plans, apps and side hustles ↗A lot of AI security tools ask another AI model to grade the first one. G8KEPR doesn't. Here's why, what it costs, and what the published numbers do and don't mean.
Thread: Building G8KEPR ↗What prompt injection actually is, why it's so stubborn to fix, how it sneaks in through documents and tools, and where G8KEPR's guard sits in the path.
Thread: Building G8KEPR ↗Most AI security looks at what goes into a model. G8KEPR's Verification Engine also looks at what comes out, with four kinds of checks and grounding at the center.
Thread: Building G8KEPR ↗Runaway agents, retry loops and sudden spend spikes can turn an AI feature into a very expensive surprise. Here's why G8KEPR treats cost as a security problem.
Thread: Building G8KEPR ↗How G8KEPR's AI Gateway routes requests across many model providers, what circuit breakers and bring-your-own-key mean, and why one checkpoint beats security bolted onto every app.
Thread: Building G8KEPR ↗