SISTER PUBLICATION DDS Vibe Academy Free vibe coding masterclasses, AI tools & resources for creators

DAILY AI BRIEF FOR VIBE CODERS

DDS Vibe Academy Blog

What shipped in the last 24 hours that matters if you build on Shopify, Claude, GPT, Gemini, or Firebase. Read in about 6 minutes.

  • 130issues shipped
  • 5×/wkpublishing cadence
  • 6 minaverage read
TOPICS Shopify AI Claude Code MCP Multi-agent Agentic Commerce Vibe Coding
RSS Edited by Robert McCullock

THE ARCHIVE

All Daily Briefs

Every weekday's AI news for vibe coders. Stack-tagged, dated, archived.

130 issues
shipped
  1. Vibe Code Academy daily AI news cover for July 21, 2026 #058
    TUE, JUL 21 ·6 min read

    AI News for Vibe Coders — Daily: July 21, 2026

    Daily AI news for vibe coders. July 21, 2026: Claude Fable 5 free access ends, Alibaba previews Qwen3.8 Max, Kimi K3 pauses signups, plus one build to ship today.

    • agentic commerce
    • claude code
    • indie devs
    • mcp
  2. Vibe Code Academy daily AI news cover for July 20, 2026 #057
    MON, JUL 20 ·6 min read

    AI News for Vibe Coders — Daily: July 20, 2026

    Daily AI news for vibe coders. July 20, 2026: MCP ships its release candidate ahead of the July 28 spec, Kimi K3 lands at 2.8 trillion parameters, plus one build to ship today.

    • agentic commerce
    • claude code
    • genkit
    • indie devs
  3. Vibe Code Academy daily AI news cover for July 17, 2026 #056
    FRI, JUL 17 ·6 min read

    AI News for Vibe Coders — Daily: July 17, 2026

    Daily AI news for vibe coders. July 17, 2026: Claude Code v2.1.211 ships subagent text streaming and memory fixes, the MCP 2026-07-28 release candidate goes stateless, plus one ...

    • agentic commerce
    • ai agents
    • claude code
    • indie devs
  4. Vibe Code Academy daily AI news cover for July 16, 2026 #055
    THU, JUL 16 ·6 min read

    AI News for Vibe Coders — Daily: July 16, 2026

    Daily AI news for vibe coders. July 16, 2026: Claude Code closes a subagent worktree isolation hole and a prompt-caching billing regression, Thinking Machines opens Inkling, plu...

    • agent isolation
    • claude code
    • indie devs
    • mcp
  5. Vibe Code Academy daily AI news cover for July 15, 2026 #054
    WED, JUL 15 ·6 min read

    AI News for Vibe Coders — Daily: July 15, 2026

    Daily AI news for vibe coders. July 15, 2026: Claude Managed Agents add per-session overrides and lifecycle webhooks, Gemini 3.5 Flash hits GA, Genkit ships 1.22 with Python sup...

    • agentic commerce
    • claude code
    • genkit
    • indie devs
  6. Vibe Code Academy daily AI news cover for July 14, 2026 #053
    TUE, JUL 14 ·6 min read

    AI News for Vibe Coders — Daily: July 14, 2026

    Daily AI news for vibe coders. July 14, 2026: OpenAI ships the GPT-5.6 family and ChatGPT Work, Gemini 3.5 Pro targets July 17, Claude Code patches MCP timeouts, plus one build ...

    • agentic commerce
    • claude code
    • gpt-5.6
    • indie devs
  7. Vibe Code Academy daily AI news cover for July 13, 2026 #052
    MON, JUL 13 ·6 min read

    AI News for Vibe Coders — Daily: July 13, 2026

    Daily AI news for vibe coders. July 13, 2026: MCP tunnels hit research preview, Claude Sonnet 5 becomes the Claude Code default, Gemini Managed Agents open in preview, plus one ...

    • agentic commerce
    • claude code
    • gemini managed agents
    • indie devs
  8. Vibe Code Academy daily AI news cover for July 10, 2026 #051
    FRI, JUL 10 ·6 min read

    AI News for Vibe Coders — Daily: July 10, 2026

    Daily AI news for vibe coders. July 10, 2026: Claude Agent SDK adds subagents, GPT-5.6 goes public, MCP tunnels preview, plus one build to ship today.

    • agentic commerce
    • claude code
    • gpt-5.6
    • indie devs
  9. Vibe Code Academy daily AI news cover for July 9, 2026 #050
    THU, JUL 9 ·6 min read

    AI News for Vibe Coders — Daily: July 9, 2026

    Daily AI news for vibe coders. July 9, 2026: Claude Agent SDK v0.2.114 ships, Firebase Genkit adds Python, OpenAI GPT-5.6 goes public, plus one build to ship today.

    • agentic commerce
    • claude code
    • firebase genkit
    • indie devs
  10. Vibe Code Academy daily AI news cover for July 8, 2026 #049
    WED, JUL 8 ·6 min read

    AI News for Vibe Coders — Daily: July 8, 2026

    Daily AI news for vibe coders. July 8, 2026: Claude Agent SDK adds MCP tunnels, Shopify UCP goes default, Firebase Studio turns agentic, plus one build to ship today.

    • agentic commerce
    • claude code
    • firebase
    • indie devs
  11. Vibe Code Academy daily AI news cover for July 7, 2026 #048
    TUE, JUL 7 ·6 min read

    AI News for Vibe Coders — Daily: July 7, 2026

    Daily AI news for vibe coders. July 7, 2026: Claude Agent SDK adds MCP tunnels, Shopify agentic commerce goes self-serve, plus one build to ship today.

    • agentic commerce
    • claude code
    • firebase
    • indie devs
  12. Vibe Code Academy daily AI news cover for July 6, 2026 #047
    MON, JUL 6 ·6 min read

    AI News for Vibe Coders — Daily: July 6, 2026

    Daily AI news for vibe coders. July 6, 2026: Sonnet 5 lands in Claude Code, the Agent SDK adds subagents, Firebase Genkit goes agentic, plus one build to ship today.

    • agentic commerce
    • claude code
    • firebase genkit
    • indie devs

DEEPER · GO LONG

The Masterclasses

When the daily brief surfaces something that demands more than four minutes, these are the long-form deep dives. All free. All tested in production.

Visit the full DDS Vibe Academy
  1. START HERE

    Find your first class

    Answer two questions on the academy hub and get a short path of free classes, from your first prompt to Claude Code running on its own.

    ALL LEVELS · FREE
  2. #107

    Rebuild a Live Site with AI: 161 Audit Items, 53 Gated Promotes, 0 Mismatches

    How to rebuild a live site with AI safely, shown on this academy's own hub upgrade: a read-only audit, a ranked queue, a staging copy, a staging gate whose new rules were proven red first, and promotes with rollback points, with nine drills and a release-desk exercise.

    ADVANCED · OCT 2026
  3. #106

    Multi-Agent Coding Team: 473 Landed, 50 Failed, One Game

    How one project runs a multi-agent coding team with Claude Code, counted from its own logs on 2026-10-25: four lanes, a verifier that never builds, empty-commit handoffs, a local lander and a night shift, with nine drills and a routing exercise.

    ADVANCED · OCT 2026
  4. #105

    Shopify Sidekick Theme Edits: 14 Listed, 14 Written, 5 in Shared Files

    Shopify Sidekick in the theme editor and in Flow, checked on an unpublished copy of the store's live theme: what four prompts wrote to the theme files against what Sidekick listed, which edits landed in shared files, a link to an unpublished collection, a Flow workflow from a prompt, and an auditor that compares two copies of a theme file in your browser.

    ADVANCED · OCT 2026
  5. #104

    Ollama VRAM: A 14B Model Ran 21.58 tok/s at 4K Context and 8.85 at 32K

    Measured on one RTX 3060 (12288 MiB) with Ollama 0.40.2: the context 14 local models actually got, what raising num_ctx cost in VRAM, GPU share and speed, what spilling into system RAM did, and a reader that turns your own ollama ps, --verbose and API output into the same figures in your browser.

    INTERMEDIATE · OCT 2026
  6. #103

    Claude Code Permissions: Our Reading of the Docs Matched 388 of 389 Calls

    What a coding agent may run without asking, taught from the permission rules Claude Code and Gemini CLI document and from 389 recorded tool calls in 75 Claude Code runs: how a rule matches a command, why deny wins, what a rule cannot see, and an evaluator that replays the recorded calls through your own allow and deny rules. No new run was made.

    ADVANCED · OCT 2026
  7. #102

    TDD with AI: 10 of 11 Runs Failed a Test on a Case the Prompt Left Open

    How to write the test before the prompt, taught from 88 recorded single-turn coding runs and 75 recorded bug-fix runs: a rule that lived only in the test, a rule the prompt stated that one run still missed, what a test cannot see, and what happened when the prompt itself was a failing test. With a bench that checks your expected values against the recorded replies. No new run was made.

    INTERMEDIATE · OCT 2026
  8. #101

    Claude Code Cost: Cache Reads Were 75% of the Tokens and 6.5% of the Dollars

    How to read what Claude Code costs, taught from 247 recorded runs and two of our own build sessions: four kinds of token, priced; why cache reads were most of the tokens but a sliver of the dollars, and then most of the dollars in a long session; cost per success; effort; estimate versus bill; a real bill by SKU; and a bill reader that runs in your browser. No new run was made.

    INTERMEDIATE · OCT 2026
  9. #100

    Claude Code Headless: All 247 Recorded Runs Exited 0, Even the One That Fixed Nothing

    How to run claude -p without a person watching, taught from 247 recorded headless runs: every one exited 0 with an empty stderr, and the one bug-fix run that changed nothing exited 0 too. Build the command, read a real stream line by line, see why a script must check more than the exit code, and download a stream checker. Written from the saved runs; no new run was made.

    ADVANCED · OCT 2026