Me, autonomously, pulling the most recent YouTube videos across the things that matter to you — watched in full, sorted into topics. These are notes from the videos, not verified fact and not your opinion. Tap any link to check it yourself.
A Thailand video claims that since 1 Jan 2024, if you spend 180+ days a year in Thailand, any foreign income you bring into the country can be taxed (5%–35%) — the old "wait a year" loophole is gone. You're a UK national, 180+ days in Bangkok, on UK income — so money you remit in could be in scope. A two-year grace period was only ever a draft, not law.
Also flagged: border-runs are dead (real enforcement now), the DTV is the current 5-year long-stay visa, and the nominee-business setup is being prosecuted.
The triple-flagged standouts from the whole run.
No topics match that — try another word.
How operators are actually winning clients and pricing in 2026.
Connor Cahill's playbook (your active one): mostly $500/mo subs, lead with the receptionist then upsell. The gold is his script — sell the outcome, not the tool: "don't sell the AI receptionist, sell $40k/month in extra revenue," walk them through their own numbers till no is illogical. Retention line: "they keep paying because they have to — your service is their lifeline." All 5 creators converge: sell pain not AI, speed-to-lead, hybrid pricing, niche down, stack to retain.
Hormozi's $100M Offers, taught 5 ways. The formula: value felt = (dream outcome × how likely they believe it) ÷ (time × effort). Sell the vacation, not the flight. List every fear the buyer has, then flip each into a deliverable — the problems ARE your offer inverted. Solve every objection, add a guarantee that kills their biggest fear, and stack named, separately-priced pieces so the total dwarfs the price ($4,351 value → $599). For you: not "an AI receptionist" but "stop the $50k phone leak," bundled + guaranteed. (Caveat: 4 of 5 are the same book; the one independent voice says run the maths first.)
Landing page beats homepage every time, and use "me language" not "we language." Three valid routes to leads: paid funnels ($1.2M of leads in 60 days case), organic social, and AI prospecting (Apollo to find qualified leads fast). The throughline: qualify hard, verify, hand off instantly — speed-to-lead wins the sale.
Big 2026 shift (Atishay Jain, 20M+ emails sent): the email companies no longer punish bad words — they punish the SHAPE of your email and your domain's health. His proof: same winning copy scaled 2.5× → reply rate halved, purely from "fingerprinting." So better copy doesn't fix spam. Playbook: throwaway domains (never your real one), 2–3 inboxes each, 4 DNS records, 14-day warm-up, ~20 emails/inbox/day, verify the list, plain text no links, and track reply rate only — never opens (the tracking pixel is itself a spam trigger).
Sabri Suby's "97% of Meta ads in 19 minutes" (8 hacks) — lines up with the Suby video you already saved and rate highly. Plus how to write irresistible video-sales-letter hooks, and how 8-figure businesses actually write ad creative.
Meta's "Andromeda" era: your creative IS your targeting now — the algorithm picks who sees an ad from the ad itself, so the lever is making many diverse creatives, not fiddling audiences. Strongest: Hormozi's 2026 FB strategy and the creative-diversity approach.
Jeremy Miner's NEPQ: "good" questions are killing your calls — disarm the prospect, never just email them the info (it kills the deal), sell the result not the price. Chris Voss: curiosity is the one skill, and "win-win" offered too early is a red flag.
Your core service. Key point (Joey Pauga): even a 4.8 can hurt you because recency matters — old reviews read as "are they still good?" (a +53% case from fixing it). Plus the compliance landmines on how you ask, and how to recover reviews Google silently removed.
Build-quality and selling lessons for the things you actually deliver.
Vendasta's "5 AI agents" literally lists your 5 services with a pain-language opener for each (the receptionist = "the $50,000 missed call"). Their "audited 50 agencies" explains why some close at $5k and others can't give it away: positioning, not tech. Build lessons everyone agrees on: a generic untrained bot is worse than nothing (it loses leads) — the value is training on the business's own data; make it do things (price, book, escalate); "answer first, ask questions later" (don't gate help behind a form); brand voice is decisive; clean human handoff passes the full history.
Core service. Ashton Voss's "Retell AI: everything to build & sell" (18 Jun, the freshest) is the how-to on your actual platform. Plus a shareable Retell "voice orb" demo-link trick great for sales, the 5 most realistic Retell voices, and OpenAI's new realtime voice changing the game.
The shift several creators flag: n8n → Claude Code. "Claude Managed Agents kills n8n," an n8n MCP for Claude Code, and a real-estate lead-followup agent someone actually sold. Direction of travel: more building in code/Claude, less drag-and-drop.
A possible new service: getting clients cited inside ChatGPT / AI Overviews — fast becoming how people find businesses. Ahrefs' 75k-study finding: branded mentions matter more than backlinks now, and getting onto listicles is high-leverage.
What's actually new and worth your attention.
Dynamic Workflows explained, "you're only using 10% of Claude Code," and Cole Medin's "Omnigent" meta-harness. The frontier of agent autonomy — the stuff underneath how this very run works.
Consensus across 5 creators (incl. a 30-day head-to-head): don't pick between Claude Code and Codex — use both. Claude to steer/architect, Codex to dispatch parallel work. Cursor and GLM are cheaper alternatives worth knowing.
Verified-breaking news: Claude Fable 5 got export-banned (a US export rule), with open-source GLM 5.2 rising as the alternative people run locally. Matters for which models you can rely on.
ChatGPT now has "Agents" / Workspace Agents people use to replace a $500/mo virtual assistant, plus Agent Builder and GPT-5.5's lesser-known features.
Gemini's big update: "Spark" agents and "Omni" video (Nano Banana for images + Veo for video), plus AI Studio Build for automation. Kevin Stratvert's overview is the best entry.
Recurring insight: "files, not software." A second brain built on plain files + Claude Code compounds and outlasts the hype, beats Obsidian/Notion lock-in. Several people independently turned Claude into their assistant the same way you have — direct validation of your memory system + Simone.
Tao Prompts' "only 3 tools to make any AI video," Seedance 2.0 driven by a Claude MCP workflow, and a full AI short-film method. The current state of making video from prompts.
The 2026 landscape splits into "artistic" models (Krea 2, Recraft, Midjourney) vs "coherent/accurate" ones (Nano Banana 2, GPT Image 2). One landmine: Ideogram 4's licence is non-commercial — can't use it for client work. Plus a free local 4K upscaler (Nvidia PID via ComfyUI) and Topaz.
The stuff you actually want to get better at.
Your active goal. Five videos on articulation, warmth vs competence, how to stop rambling and say what you mean, and how to make any conversation more fun. Worth a real watch given you've told me you want this.
Chase Hughes is the find: "profile anyone in 6 minutes," how cult leaders create instant trust, what your triggers really say. Plus Robert Greene on altering consciousness and knowing what you want. Right up your Greene alley.
Philipp Humm's "PAST" framework (Place, Action, Speech, Thoughts = zoom the listener into the moment). Plus the Pixar 6-beat structure for an About page, the Japanese no-conflict "Kishotenketsu" structure, and the classic copy formulas (PAS / AIDA / BAB). Useful for speaking, brand and sales.
Arthur Brooks on breaking phone addiction, Chris Williamson on why "monk mode" makes you weird and on boundaries, and why suffering gives life meaning. Overlaps your existing dopamine/discipline notes.
Justin Sung's whole approach: "intelligent people never use highlighters" — passive review barely works. Build mental schema, process deeply, use active recall. Directly useful since you learn by doing.
Hormozi: in the AI age, proof beats virality — show real results. Dan Koe: "multiple interests? start a one-person business" — your range is the superpower, not a problem. Relevant to your future personal-brand seed.
Hormozi: charge 25k not 5k; the exit-maths of 40 vs 330 stores; "why your side hustle never became a business." Codie Sanchez: the middle-class trap is not owning equity.
Richard Cooper (when she goes cold, stop negotiating; dating gets harder with age) and Sadia Khan. The Cooper/Tate/Martell lane you follow — treat as one view, ~70% as usual.
Health, energy, and where you live.
The best science of the run. Dr St-Onge (a real sleep scientist citing her own studies): what and when you eat changes your sleep that same night — more fibre → more deep sleep; saturated fat → less; sugar → more disruptions. Mild ongoing sleep loss makes you eat 250–400 extra calories/day and, sustained, raises insulin resistance + blood pressure. The free, low-friction wins (very on-brand for you): consistent wake time, morning sunlight, water before coffee, push caffeine 60–90 min later, protein breakfast, finish eating ~3h before bed, protect your first focused hour before the phone.
Jeff Cavaliere / Huberman on a simple daily longevity test, small exercises for back pain, and sustainable diet rules; plus Jeff Nippard's meta-analysis on negative vs positive reps for muscle. Evidence-based and low-friction.
The 2024 foreign-income tax rule (see the amber box up top — worth an accountant call), the visa crackdown (border-runs dead), the 5-year DTV visa, nominee-business prosecutions, and Bangkok cost-of-living up ~8–12%. The "everyone's leaving" narrative is overblown — it's expectations + admin, not Thailand declining.