Veo UGC Workflows, Codex Migration, 1Password .env — AI Daily May 02

780 messages · 70 active members

780
messages
70
active members
@jasonakatiff, @tounano, @Kieran
top contributors

Overview

Saturday's 780-message marathon centered on AI video production and agent tooling. Kieran walked the room through a complete UGC ad workflow — 8-second Veo scenes chained with ffmpeg, same reference image across cuts, submagic for captions, Kling for silent b-roll, and Mandarin prompts to bypass Seedance's English content filters. Kie.ai emerged as the go-to onramp for non-technical partners (yoga teachers, vet techs) who want Veo/Kling/Seedance/Nanobanana behind one credit-based UI. Separately, a new free Claude skill (bradautomates/claude-video) lets the model 'watch' any video by piping yt-dlp frames plus Whisper/Groq transcripts for ~$1 per 45 minutes. On the agent side, @jasonakatiff began migrating from Claude Code to Codex and immediately hit rate limits on xhigh fast (15% of weekly usage in 2-3 hours) with 45+ min job times. Consensus: Codex needs explicit step-by-step prompting and should be treated as a tool, not a peer; symlink .claude→.codex and CLAUDE.md→AGENTS.md. Codex's hidden /goal beta got hands-on testing, Grok 4.3 hit #2 on the AA Intelligence Index, and Hermes is gaining real traction over OpenClaw for ecom automation with 50+ crons and native subagent spawning. @jcartu pushed an uncensored Qwen3.6-35B on a 5090 via vLLM as a 160tps local alternative. Security dominated the back half: Jason migrated hundreds of vars to 1Password's new .env environments beta while Sav310 countered with Infisical for machine-first workflows. A long SOC2 thread put real pen tests at ~$15k each, with @geilt arguing prompt-based audits only catch basics and AI tends to bolt on unnecessary layers like JWT over robust RBAC. The All-In pod's claim that all human-written software needs rewriting in 12 months made the rounds.

Topics

Kieran detailed his production stack: Veo for talking UGC, Kling for silent b-roll, 8-second scenes chained via ffmpeg with a consistent reference image, submagic for captions — skipping last-frame workflows due to high failure rates. Seedance 2 wins on realism for ecomm trust signals but distracts from proven Veo ad volume. Mandarin prompts bypass English content filters, and kie.ai is the best non-technical onramp now that Google Flow's UI is 'absolute horseshit'.

@jasonakatiff hit Codex rate limits fast (15% of weekly xhigh in 2-3 hours) with 45+ min job times vs 10-15 min on Claude Code. Codex needs explicit instructions and should be treated as a tool, not a peer; symlink .claude→.codex and CLAUDE.md→AGENTS.md is the standard bridge. The hidden /goal beta lets you fire-and-forget with a success definition, and Grok 4.3 jumped to #2 on AA Intelligence Index (53) but with an 8-point dip in non-hallucination rate.

Jason migrated hundreds of vars across 15+ files to 1Password's new environments feature (public beta), which replaces .env.local with a 1Password-backed mount. Sav310 pushed back recommending Infisical for machine-first/agent workflows, successfully mapping 262 sprawled keys into Docker and Cloudflare via Agent Vault. Frequent auth prompts when Claude accesses secrets remain a UX pain.

@jasonakatiff runs 12 pen-test libs and is budgeting $15k each for real pen testing and SOC2 audits. @geilt argued prompt-based audits cover only basics and that AI tends to bolt on unnecessary layers like JWT over already-robust RBAC — infra architecture matters more than code. Separately, @jcartu advocated an uncensored Qwen3.6-35B (NVFP4) on a 5090 via vLLM hitting 160tps as a cloud-Claude-Code replacement.

A new free skill (bradautomates/claude-video) splits any YouTube/Loom/Instagram video into frames + Whisper/Groq transcripts so Claude can analyze it for ~$1 per 45min. @tounano and @samtome converged on running most 'E2E' tests in happy-dom with React mounted in the test runner — reserving real browser E2E for 2-3 happy paths and drag-and-drop. @jasonakatiff also shared a Notion writeup on the agent improvement loop using OpenTelemetry traces, scoring, and offline evals.

Key Takeaways

  • For UGC video ads, chain 8-second Veo scenes with ffmpeg using a consistent reference image and skip last-frame workflows; Mandarin prompts bypass Seedance's English content filters.
  • Codex needs different prompting than Claude Code — explicit step-by-step instructions, treat it as a tool not a peer; symlink .claude→.codex and CLAUDE.md→AGENTS.md.
  • 1Password's .env environments beta is production-ready for local dev; use Infisical for machine/agent workflows.
  • Real SOC2 + third-party pen testing runs ~$15k each, and AI-generated code often adds unnecessary security layers — 'works' ≠ 'needed' or 'secure at scale'.
  • Run most E2E tests in happy-dom with React mounted in the test runner (95%+ speed gains); reserve Chromium for 2-3 happy paths and drag-and-drop.

Hot Threads

@tounanostarted

Video model recommendations and Veo workflow — Kie.ai, 8-second scene chaining, Seedance vs Veo for UGC

60 replies5 participants
@jasonakatiffstarted

1Password .env migration, Infisical, AWS vs GCP vs Vercel for prod secrets

35 replies6 participants
@jasonakatiffstarted

AI security, SOC2 compliance, and whether all human-written software needs rewriting

30 replies5 participants

Linked Items

Veo UGC Workflows, Codex Migration, 1Password .env — AI Daily May 02 | Built with AI