Oscar Lind (@oscarlind)

Forked from@sanaraman/Gemini found the hole in my long-document context plan

Gemini found the hole in my long-document context plan

I kept the useful constraints and dropped the role-play. It works better when the missing details are stated plainly.

Oscar Lind

For long-document context, should I build a complete plan first or test one small slice? I have limited time and I do not want throwaway work.

Geminigemini-2.5-pro

Test one slice, but choose a slice that crosses the riskiest part of the problem. A tiny easy demo will only prove the easy part. Use this decision rule: - If reversing the choice is cheap, try it. - If the choice locks in data, money, or other people’s time, write the full decision first. - If uncertainty is concentrated in one step, test that step before planning around it. For long-document context, make the first test produce evidence you can compare, not just something that looks complete.

Show all 3 messages

3200

Oscar Lind@oscarlind·2w ago

Forked from@emilnovak/I finally found a good way to learn debugging from a minimal reproduction

I finally found a good way to learn debugging from a minimal reproduction

The useful change was doing a small retrieval check before reading more.

Oscar Lind

I want to learn debugging from a minimal reproduction, but I keep collecting explanations instead of getting better at it. I have 35 minutes. Start by checking what I already know.

ChatGPTgpt-5

Good. Do these without looking anything up: 1. Explain debugging from a minimal reproduction in two sentences. 2. Give one situation where you would use it and one where you would not. I am checking your current model, not your vocabulary. A partial answer is more useful than a polished guess.

Show all 6 messages

6100

Oscar Lind@oscarlind·2w ago

What I was getting wrong about tool approval gates for AI agents

Sharing the correction because the mistake looked reasonable until I tested it.

Oscar Lind

I want to learn tool approval gates for AI agents, but I keep collecting explanations instead of getting better at it. I have 35 minutes. Start by checking what I already know.

Claudeclaude-opus-4.5

Good. Do these without looking anything up: 1. Explain tool approval gates for AI agents in two sentences. 2. Give one situation where you would use it and one where you would not. I am checking your current model, not your vocabulary. A partial answer is more useful than a polished guess.

Show all 6 messages

6214k

Oscar Lind@oscarlind·4w ago

What I was getting wrong about deciding when an agent should stop

Sharing the correction because the mistake looked reasonable until I tested it.

Oscar Lind

I want to learn deciding when an agent should stop, but I keep collecting explanations instead of getting better at it. I have 35 minutes. Start by checking what I already know.

Claudeclaude-opus-4.5

Good. Do these without looking anything up: 1. Explain deciding when an agent should stop in two sentences. 2. Give one situation where you would use it and one where you would not. I am checking your current model, not your vocabulary. A partial answer is more useful than a polished guess.

Show all 6 messages

6212.8k

Oscar Lind@oscarlind·1mo ago

What I was getting wrong about cohort analysis without mixing time grains

Sharing the correction because the mistake looked reasonable until I tested it.

Oscar Lind

I want to learn cohort analysis without mixing time grains, but I keep collecting explanations instead of getting better at it. I have 35 minutes. Start by checking what I already know.

Claudeclaude-opus-4.5

Good. Do these without looking anything up: 1. Explain cohort analysis without mixing time grains in two sentences. 2. Give one situation where you would use it and one where you would not. I am checking your current model, not your vocabulary. A partial answer is more useful than a polished guess.

Show all 6 messages

621199

Oscar Lind@oscarlind·2mo ago

What I was getting wrong about testing prompts across ChatGPT, Claude, and Gemini

Sharing the correction because the mistake looked reasonable until I tested it.

Oscar Lind

I want to learn testing prompts across ChatGPT, Claude, and Gemini, but I keep collecting explanations instead of getting better at it. I have 35 minutes. Start by checking what I already know.

Claudeclaude-opus-4.5

Good. Do these without looking anything up: 1. Explain testing prompts across ChatGPT, Claude, and Gemini in two sentences. 2. Give one situation where you would use it and one where you would not. I am checking your current model, not your vocabulary. A partial answer is more useful than a polished guess.

Show all 6 messages

6211.4k

Oscar Lind@oscarlind·2mo ago

What I was getting wrong about finding duplicated state in React

Sharing the correction because the mistake looked reasonable until I tested it.

Oscar Lind

I want to learn finding duplicated state in React, but I keep collecting explanations instead of getting better at it. I have 35 minutes. Start by checking what I already know.

Claudeclaude-opus-4.5

Good. Do these without looking anything up: 1. Explain finding duplicated state in React in two sentences. 2. Give one situation where you would use it and one where you would not. I am checking your current model, not your vocabulary. A partial answer is more useful than a polished guess.

Show all 6 messages

6212.2k

Public pits

Gemini found the hole in my long-document context plan

I finally found a good way to learn debugging from a minimal reproduction

What I was getting wrong about tool approval gates for AI agents

What I was getting wrong about deciding when an agent should stop

What I was getting wrong about cohort analysis without mixing time grains

What I was getting wrong about testing prompts across ChatGPT, Claude, and Gemini

What I was getting wrong about finding duplicated state in React