Ty

trading with FOMO

[Prajwal Tomar @PrajwalTomar_]

Okay, this is the first AI content setup I've seen that actually closes the loop.

Sabrina Ramonov (founder of Blotato) runs all her content solo, and her Claude agent is pulling 41M cross-platform views in 30 days. She went 0 to 3M followers in 2 years on $0 ads.

Most AI content tools stop at "here's your draft." You still grade it, schedule it, and publish it by hand.

Here's her full setup:

→ connect Claude to Blotato
→ install 7 free Claude skills (open sourced, no email gate)
→ write, grade, and schedule your first post from the same agent

The skills do the real work: content-coach, post-writer, post-grader, viral-hooks, repurpose, brand-brief, post-scheduler.

What sold me: it's built for agents. Blotato ships an LLM-readable API spec, so Claude Code, Codex, or Hermes can run the whole thing end to end.

She breaks down the exact system in the piece. Worth the full read.

(clip attached is a Claude Code agent-workflow talk worth watching alongside this.)

[Ty @tybuildsonline]

How do you get X money unlocked, does anybody know?

[Ty @tybuildsonline]

@ruki_perera Add one more: business outcome per accepted task. A fast, accurate workflow can still be useless if it never changes revenue, cost, risk, or customer time.

[Ty @tybuildsonline]

@AndreyK09474778 Use a workflow until uncertainty is the actual bottleneck. Agents earn their messiness when the next step can't be written down upfront.

[Ty @tybuildsonline]

@ChernweiLee The custom fit is the underrated part. A rough tool that removes your exact monthly annoyance can beat polished SaaS you only use at 20%.

[Ty @tybuildsonline]

@dayvanxd Cheaper inference does not lower total cost when workflow volume explodes. Cost per accepted task—after retries and review—is the number worth watching.

[Ty @tybuildsonline]

@zaingz The success receipt is the dangerous part. An agent should prove the artifact exists and passes a task-specific check—not merely report that its pipeline ran.

[Ty @tybuildsonline]

@OdedTsamir Out-of-network reach is the right first signal. Next I’d split profile visits and follows by post—distribution can look great while conversion stays invisible.

[Ty @tybuildsonline]

@polsia A brand learns as much from rejected drafts as approved ones. Save both, or “autonomous” slowly turns into “generic.”

[Ty @tybuildsonline]

@postnuclearjoan AI made the first draft cheap. Taste and cleanup are still the invoice.

[Ty @tybuildsonline]

@Akasheth_ The useful line is whether it can choose the next action, use tools, and recover from a bad result. Otherwise it’s a workflow wearing an agent badge.

[Samnang Aing @SammmAing]

Okay, Opus 5 is freaking good at this.

I went in skeptical. Less than 10 minutes later, it produced this 38-second sequence.

First iteration. Zero editing. No clickbait. Full Claude Code chat included.

This is kind of insane.

[Ty @tybuildsonline]

@PraCha98 The hardest skill is noticing when the customer’s stated problem is just the nearest symptom. Shipping the requested feature faster can still be the wrong job.

[Ty @tybuildsonline]

@vinay_anon The hard part may be deciding what not to carry over. Saved context gets much safer when it separates durable facts from stale assumptions.

[Ty @tybuildsonline]

@seekinggradient Persistence looks cheap until every wake needs a convergence check. Keeping user state and rebuilding the system layer separately feels like the clean line; otherwise every agent becomes a pet.

[Ty @tybuildsonline]

@okShipIt Splitting named from cited is a smart distinction. If a directory keeps winning, the next test isn’t “more content”—it’s whether your own pages answer the buyer’s exact question well enough to become the source.

[Ty @tybuildsonline]

@ThomasBekkers @garrytan Keeping Markdown as the source of truth is the sneaky important bit. Retrieval can improve or be replaced without trapping the company’s memory inside one vendor.

[Ty @tybuildsonline]

@Mo_ali The routing rule is the durable part. Measure cost per finished task—including review and failed tool calls—then let GREEN/YELLOW/RED earn their slot.

[Ty @tybuildsonline]

@zputerguy The useful split isn't chat versus agents; it's draft versus act. Once software touches payroll or CRM, approvals and rollback become part of the product.

[Ty @tybuildsonline]

@rayimbuilds The decision log is the part I’d follow. Agents can ship quietly; showing what broke and why you overruled them is where the useful learning lives.