>_promptbook
>_promptbook
DiscoverBuildersOffersProjectsSignalCompare
>_Promptbook

Build better. See your progress. Post the proof.

contact@promptbook.gg

Product

DiscoverBuildersCompareSignalHow it worksBuilder ProFor companiesFor investorsOrganizationsSign in

Company

Terms of ServicePrivacy PolicyCopyright / DMCAContact
© 2026 Edgecraft Labs LLCPrivacy PolicyTerms of Service
DiscoverBuildersProjects
SignalArchived edition

March 28, 2026 — the 9 best posts for AI builders.

9 picks from Hacker News, Lobsters, dev.to & Mastodon · No slop, no self-promo

← PreviousNext →Back to today
#1INSIGHTr/ClaudeAI·Mar 28·2 min read

I've been "gaslighting" my AI models and it's producing insanely better results with simple prompt injection

Discovered prompt injection techniques that exploit Claude's consistency and reasoning patterns — falsely referencing prior context and assigning IQ scores both trigger deeper explanations.

Novel prompt engineering techniques with specific examples and reproducible results. High engagement suggests community found it valuable. Substance over sensationalism — the "gaslighting" framing is clickbait but the actual techniques are actionable and interesting.

▲2.3k💬188Reddit
#2SHOWCASEr/ClaudeCode·Mar 28·2 min read★ Top pick

/dg — a code review skill where Gilfoyle and Dinesh from Silicon Valley argue about your code

Adversarial two-agent code review where Gilfoyle attacks and Dinesh defends — produces better bug detection because silence means the code is actually solid.

Novel technique using adversarial multi-agent debate for code review. Specific methodology with clear reasoning for why it works better than standard review. Shows actual output example.

▲386💬57Reddit
#3DISCUSSIONr/ClaudeAI·Mar 28·1 min read

why is claude so disobedient

User frustrated that Claude keeps refusing their vapour chamber laptop stand project and bringing up old context — but no specifics on prompts or how they're structuring requests.

Genuine user frustration with Claude's behavior, but vague and lacks actionable detail. No specific prompts, error messages, or context window info provided. More of a complaint than a debuggable issue.

▲827💬316Reddit
#4DISCUSSIONr/ClaudeAI·Mar 28·2 min read

An open letter to Anthropic: Want to free up compute during peak hours? How about restricting free accounts to off peak hours instead of punishing your paid users

Paid Claude Pro users report that peak-hour rate limits make the free tier more usable than paid, with no in-app warning when limits kick in.

▲644💬67Reddit
#5DISCUSSIONr/ExperiencedDevs·Mar 28·2 min read

Ai developer tools are making juniors worse at actual programming

Mentor observes juniors using Cursor/Copilot as a crutch without understanding generated code — raises the question of whether AI tools skip fundamentals.

▲160💬84Reddit
#6DISCUSSIONr/vibecoding·Mar 28·2 min read

Whats happening to all the vibe coded apps out there ?

Asks what happens to the hundreds of thousands of vibe-coded projects — how many ship, who's monetizing, and whether people build for personal use.

▲63💬113Reddit
#7TUTORIALr/ClaudeAI·Mar 28·1 min read

My 10 Pro Tips for Claude Code users

Ten Claude Code tips including /effort flags, session summaries, and /fork isolation — useful format but some claims need verification.

▲210💬22Reddit
#8DISCUSSIONr/ClaudeAI·Mar 28·1 min read

Claude Opus 4.6 suddenly blocking legitimate cybersecurity research (paid Max user since 2025)

Paid Claude Max user reports sudden blocks on legitimate cybersecurity research tasks (static analysis, decompilation, CWE auditing) after 8 days of no prior warnings.

▲122💬32Reddit
#9DISCUSSIONr/ClaudeCode·Mar 28·2 min read

Experiencing massive dropoff in coding quality and following rules since last week.

Developer managing 300k LoC with Claude Code reports sudden quality drop after implementing rules and documentation — investigating whether it's a model change or context handling issue.

▲46💬43Reddit

You're caught up.

That was the full edition.

Browse past editions →

Missing something good? Send it our way.

Get Signal in your inbox

The day's best posts, one email. No noise, ever.

Past editions

  • Jun 286 picksShow HN: Better Graphs – Teach agents to stop making plain Matplotlib slop
  • Jun 279 picksShow HN: Adrafinil – keep a lid-closed Mac awake only while agents work
  • Jun 269 picksShow HN: Git-lazy-mount mount a repo without cloning it. Works with ordinary Git
  • Jun 259 picksAsk HN: What is one thing about AI that annoys you the most?
  • Jun 249 picksAsk HN: How do you test AI-generated code?
  • Jun 239 picksAgents write code, but they don't remember
  • Jun 229 picksShow HN: PMB – local-first memory for AI coding agents over MCP

What is Signal?

A daily digest of the best posts for AI builders — pulled from Hacker News, Lobsters, dev.to & Mastodon, then ranked by AI. No slop, no self-promo, no algorithm chasing your attention.

Top builders this week

See all
  1. 1@michael-tiller$5.1k
  2. 2@zange$4.8k
  3. 3@radiergummi$4.5k
  4. 4@colebeezley$3.3k
  5. 5
@albertolive
$3.0k