HeyJaredNeural Stack - Software | AI | Open Source
Welcome to Neural Stack - a channel about building software, exploring AI, and figuring out where technology is headed. I make videos for developers, builders, and curious people who want to understand modern tech without all the hype.
- Indexed videos, last 90 days
- 78
- Latest publication
- Sep 29, 2026
- Audience
- ~1.6K subscribers
- Earliest in this view
- Jul 3, 2026
Latest videos
How to Manually Test an AI-Built Signup Form: 3 Checks to Record (opens the original)
Read excerpt
Learn a repeatable way to manually check an AI-built signup form: submit it blank, enter an invalid email, then correct the rejected value. Define expected behavior, record what you observe, and decide what to check next. All form visuals are hypothetical mockups. This walkthrough explains the testing process; it does not report actual test results. 00:00 Three checks you can repeat 00:45 Get the practice form ready 01:30 Define expected behavior 02:20 Set up your test record 03:05 Check 1: Subm
How to Build Your First Agent Evaluation Set From 20 Real Failures (opens the original)
Read excerpt
Turn 20 real agent failures into your first usable evaluation set. This walkthrough covers documenting what failed, preserving the inputs and context, writing standalone tasks, and defining observable success conditions with evidence you can inspect. You’ll also learn how to start collecting failures, keep unfinished entries separate, and run your first set—with a clear understanding of what the results can and can’t tell you. The tutorial distinguishes Anthropic’s starting guidance from the wor
How to Cut Redundant Examples from an Agent Prompt (opens the original)
Read excerpt
Trim repetitive examples from an agent prompt while keeping the distinct behaviors each one teaches. This walkthrough uses a hypothetical six-example prompt to show how to label behaviors, group overlapping examples, and choose what to keep. Then check for missing behaviors and prepare realistic tasks with executable acceptance checks to test the revised prompt. Finish with a before-and-after comparison and a map of the behaviors retained. #PromptEngineering #AIAgents #PromptDesign
How to Compare Claude Opus 5.5 on Your Coding Tasks (opens the original)
Read excerpt
Is Claude Opus 5.5 better and cheaper for your coding work? We unpack the reported benchmark and published API pricing, then outline a comparison you can run on your own backlog tasks using /model and /usage. This is a testing walkthrough, not recorded hands-on results. Learn to define success, compare the work produced, and read usage without treating it as proof of dollar savings. We also cover Claude and API access, plus the announced GitHub Copilot rollout. 00:00 What's changed? 00:40 What t
How to Ask Claude Code for Edge-Case Tests That Fit Your Repo (opens the original)
Read excerpt
Learn how to ask Claude Code for edge-case tests, review whether they follow your repository’s conventions, and check that the new tests were actually included in a test run. Start with one clearly identified function and an existing test as a reference. Follow the request and inspection workflow, keeping convention matching and execution as two separate checks. This is an instructional workflow guide, not a recorded repository demonstration. No generated tests or completed test-run results are
Publishing over time
Last 90 days. Choose a month to open its work.
Recurring subjects
Named in the text we hold. One piece can cover several.
Audience
~1.6K subscribers
Measured Sep 19, 2026
Source's subscribers, not the number who saw an individual piece.
About this data
Counts cover the work we have indexed. Tone needs enough text and a confident classification. Excerpts and episode notes are not full articles or transcripts.
Identity or attribution wrong? Suggest a correction.
See coverage about Neural Stack - Software | AI | Open Source