Fable Extension, Grok 4.5 Speed, Cold Email Infra — AI Jul 12

465 messages · 82 active members

465
messages
82
active members
@samb69, @iamgalba, @arielletolome
top contributors

Overview

The AI model wars dominated the day. What was billed as Fable's last day turned into a full extension from Anthropic, with the 5-hour usage limit temporarily removed for Plus, Business, and Pro plans — a move members read as a direct response to GPT-5.6 Sol's momentum and Grok 4.5's arrival. Codex users also noticed their 5-hour limit disappearing with weekly resets firing daily, effectively making it unlimited. @samb69 declared 'Dario is cooked' while @arielletolome predicted a nerfed Fable 5 release August 1st. Beyond the drama, orchestration patterns matured. @realcrischico detailed Fable planning with Codex and Grok cross-reviewing the plans — Codex reads narratively, Grok is terse like a court reporter. @samtome praised Grok 4.5 for finishing in 15 minutes what Opus/Fable took 2 hours to do, while @tidemid and @scalingfrog reported GPT-5.6 Sol charging through tasks Fable was sandbagging on. Popular pairing: Sol/Fable for planning plus Grok for execution. Builders also went deep on cold email infrastructure, with @iamgalba sharing a full playbook on G Suite inbox setup, warming, cadence, and agent-driven per-recipient research. Side threads covered n8n's continued relevance for deterministic tasks, feedback loops for auto-improving video ads based on profit margins, TCPA safety on inbound Meta calls, and AI video pipeline continuity struggles solved via 15-second parallel slices and lipsync overlays.

Topics

After a day of speculation and a fake extension meme that fooled several members, Anthropic officially extended Fable access and temporarily removed the 5-hour cap for Plus, Business, and Pro plans. Codex users separately noticed their 5-hour limit gone with weekly resets firing daily, making it effectively unlimited. Members read both moves as defensive responses to GPT-5.6 and Grok 4.5 pressure.

Grok 4.5 is winning converts on raw speed, with @samtome comparing it to Opus 4.5 tier and @jonmacofficial calling it 'fast af.' @realcrischico shared an orchestration pattern where Fable plans, then Codex and Grok cross-critique — Codex reads narratively while Grok is terse and information-dense. Members are moving toward model-agnostic pipelines with Sol/Fable planning plus Grok executing.

@iamgalba shared a detailed cold email setup: own G Suite with 3 inboxes per domain, custom tracking domains, warmed up, 30 emails/day max at 3/hr during US work hours. Agents run SERP, domain, LinkedIn, and Meta ads research before writing 4-6 sentence emails with 5-6 word subject lines. Debate followed on cheap bulk inboxes (scaledmail, inboxkit) vs. direct G Suite for individually-crafted messages.

GPT-5.6 Sol is winning fans for near-unlimited usage on the $200 plan and outperforming a degraded Fable on practical tasks — @tidemid switched after Fable kept sandbagging on dataset rights, and @scalingfrog said 5.6 does better UI work and system auditing than 5.5. @Pratikfrr countered that Fable still produces more impressive design output but at 10-15x the cost, while users complain about the growing number of effort levels to tune.

Builders debated whether Make/n8n still matter in an agent-first workflow — consensus: yes for deterministic recurring tasks, webhook triggers, and Slack reporting bots, with @expadz noting agents now build and maintain his n8n workflows. Separately, @danfeldman asked about training video ad generation on profit margins and pay-per-call sales rates; @fuckyesiwannatalkbusiness has the backbone built using kie.ai to spawn variants of winning UGC ads.

Key Takeaways

  • Anthropic extended Fable and lifted the 5-hour cap for Plus/Business/Pro; Codex's 5-hour limit also appears removed with daily weekly-resets — both defensive moves against GPT-5.6 and Grok 4.5.
  • Winning workflow is model-agnostic orchestration: Fable/Sol for planning, Grok 4.5 for fast execution, with cross-model plan reviews to catch blind spots.
  • For individually-crafted cold emails, direct G Suite inboxes (3 per domain, warmed, custom tracking domain) outperform cheap bulk providers; cap at 30/day, 3/hr during US work hours.
  • GPT-5.6 Sol is charging through tasks Fable now sandbags on, especially security-adjacent work or dataset access with edge-case rights concerns.
  • n8n still shines for deterministic recurring tasks and Slack-integrated bots — let agents build and maintain the workflows for you.

Hot Threads

@mn_user0started

Cold email infrastructure for professional services — inboxes, cadence, and agent-crafted messaging

30 replies6 participants
@watchmedropshipstarted

Fable extension confirmed — 5-hour limit lifted, model wars intensify

18 replies11 participants
@rmaiguestarted

Do you still use Make or n8n, or is it fully AI agents now?

14 replies7 participants

Linked Items