https://www.youtube.com/watch?v=lQU5RX8EyWw Jack Roberts runs the same four real jobs on ChatGPT Dots, Grok Bot, and Hermes Agent — Skool image/video via Higgsfield, podcast → usable asset, Zapier tool connect + email draft, and competitor strategy — then compares cost, learning, control, and gaps. Course takeaway: use this as your Dot capability map. If your life already runs inside ChatGPT, steal his four-job test and wire Dots the same way before you add a second stack. What This Means For Dots Members 1. Dots is strongest when you stay ChatGPT-native — plugins, memory, and one personal assistant instead of a multi-bot OS 2. Image/video and deep tool chains are not automatic — plug Higgsfield (or similar) into the Dot path, same as Jack does for the other stacks 3. Zapier (or native plugins) is how a Dot sends real email and talks to your apps without paste-and-hope 4. Competitor strategy still needs a clear brief + a verify pass — treat the Dot like a researcher with a checklist, not a one-shot oracle 5. Watch the cost and control chapters before you bolt on Hermes or a second computer agent — many members only need a well-connected Dot Do This With Your Dot This Week 1. Create or name one Dot for “community content” and connect an image/video tool (Higgsfield or your preferred plugin) 2. Ask for one Skool image + one short video with almost no hand-holding — note rework vs Jack’s Job 1 results (~6:23) 3. Drop one podcast/long audio into the Dot and ask for a publish-ready asset pack (titles, cuts, posts) 4. Connect one Zapier (or native) path and draft a real email reply end-to-end 5. Run a one-page competitor strategy with a short verifier pass; save it where you act on research 6. After the four jobs, decide: stay Dot-first, or add Grok Bot / Hermes only for the jobs that clearly needed a shared computer or agentic OS Job Map (Timestamps) 1. Where to get Dots, Grok Bot, Hermes (~0:51) and what to know first (~1:48)

https://www.youtube.com/watch?v=sA3XDlU1udA First-hour hands-on with a ChatGPT Dot (named Marvin): create from Home, connect calendar and a computer, set daily recurring jobs, and use voice calls. Chris shows cloud computer takeover, a daily iOS bug-fix loop, read-only finances nudges, and calendar call prep — useful if you want concrete first jobs beyond the launch demos. Access And Create 1. At recording time he sees Dots on Pro ($100) after enterprise-only early access; $500 tier mentioned as new 2. Open ChatGPT Home — if you see Dot, you have access; if not, wait for rollout 3. Click Dot → name it → pick an avatar → connect only the apps/tools you need 4. For local computer control he updated Codex / the desktop stack so the Dot can run on his machine, not only OpenAI cloud Cloud Computer Vs Your PC 1. Every Dot has a hosted cloud computer — open it, take over the browser, type, browse 2. Cloud is safer for throwaway or risky tasks (he compares to paying for a DigitalOcean VPS for OpenClaw) 3. Local desktop access is optional and more convenient for real work files — use a spare/work machine if you can 4. Mac Mini / dedicated desktop pattern from OpenClaw days still applies for people who want isolation First Jobs He Actually Ran (Day One) 1. Daily app improvement loop — Dot proposes and lands small iOS fixes (scorecard save bug) toward 1% better every day 2. Finances plugin (read-only) — daily nudges on subscriptions to cancel, overspending, and savings ideas 3. Dual calendars — morning prep for personal + work meetings; deeper color-coded call lists from prior months 4. Voice call while working out — assign and refine jobs without sitting at the keyboard 5. Planned next: Telegram read-only on selected trading chats to summarize what he missed How To Think About Setup 1. Prefer recurring daily jobs over one-off chat prompts — that is the Dot advantage vs plain ChatGPT / Codex 2. Start with connections you already trust (calendar, read-only money) before broader messaging write access

https://www.youtube.com/watch?v=5JgKvZ6iAe4 Five-minute plain-English guide to ChatGPT Dots: what a Dot is versus a normal chat, who can access it today, how to create one, how memory and schedules work, cloud vs local computer, and which approvals to lock down before connecting apps or files. Best for members who want a clean mental model before they hand a Dot real work. What A Dot Is 1. Not a typing indicator — an always-on AI teammate that keeps working between conversations 2. Built for continuity: responsibility + desired result + tools + when to escalate to you 3. Tracks what changed, plans next steps, and runs background tasks while you keep chatting 4. Think ongoing teammate, not one giant prompt Who Can Access (As Of This Video) 1. Rolling out gradually — eligible accounts may still wait several days 2. OpenAI lists Pro $100, Pro $200, and Pro $500 for adults outside the EEA, UK, and Switzerland 3. Business Premium rolling out worldwide 4. Enterprise (including Edu and healthcare) beta starts off — admin must enable it 5. First Dot included with eligible Pro or Business Premium; deeper work allowances may change after launch How To Create One 1. Use the ChatGPT desktop app or ChatGPT in a desktop browser (Windows supported) 2. You cannot create a Dot on mobile or mobile web yet 3. Follow onboarding: name, look, review app/computer connections before enabling 4. After setup, the same Dot can appear in the mobile app once that update reaches you Good Assignment Shape 1. Responsibility — what it owns day to day 2. Sources — what it should read / watch 3. Boundaries — what it must never do without you 4. Updates — where and how often it reports 5. Examples: launch checklist, support-issue research, recurring reports, deadline risk watch 6. Exact-time work needs a saved schedule; open Activity for in-progress / scheduled / completed Memory And Research 1. Uses current conversation, relevant ChatGPT memory, and its own saved notes

https://www.youtube.com/watch?v=db1KRJQEt8w Igor (The AI Advantage) spends his first day inside ChatGPT Dots on the $200 plan: create a Dot, run a real browser job, set recurring monitors, wire results into Space, then wake up to proactive overnight work. Best for members who want a day-one operating pattern, not launch hype. Access Notes (As Of His Test) 1. He used the $200 plan; Business Premium ($100/seat) also had access in his window 2. Not available in Europe for him — he VPNed to the US 3. Cloud Dot first (not local computer); Skills did not work inside Dot, but Plugins did 4. Only one Dot at the time of filming — more Dots promised later Create + First Job 1. Create your Dot (he named his Mini Alfredo and gave it a character) 2. Ask what context it already has (recent chats summary + skills list) 3. Keep it on the cloud computer before granting local access 4. Give one concrete browser job (his demo: used camper vans in Lisbon, 20–50k EUR, top three fits) 5. Open the Dot computer view and watch the VM work — do not babysit every click Always-On Schedules 1. After the first search works, ask the Dot to refresh listings on a timer and notify you 2. Confirm the job appears under the Dot's Scheduled tab 3. Edit the prompt / interval there if needed (his ran hourly in the same Dot chat) 4. Expect wait time — Dots are slower than juggling several Grok Bot / OpenClaw agents in parallel Space As The Workbench 1. Open Spaces — markdown docs / Notion-style pages inside ChatGPT 2. Slash commands for headings, pages, images, AI inserts, comments, nested pages 3. Tell the Dot to create a tracker page (e.g. Camper Vans) and append new finds there 4. Treat Space pages like shared Google Docs the Dot reads and writes Overnight Proactive Pattern 1. Connect Plugins you actually use (Slack, calendar, Granola, etc.) after the first job succeeds 2. Let scheduled monitors run overnight 3. Review morning suggestions (his Dot flagged a stalled O-1 visa follow-up from calendar context)