Not a "ChatGPT is bad" video. The actual complaint is narrower: a model can give a polished answer, follow individual instructions, and still lose sight of what the work was supposed to accomplish, so you end up supervising the direction of the work, not just checking the result. That's been Nick's experience with the newer OpenAI models on longer Codex tasks specifically, stated as his experience, not a universal claim. • The one question that actually matters: not "which model is smarter," but did it preserve what you asked for, did it finish the work, and how much attention did you spend getting it there. An impressive first answer doesn't settle that if the next few turns keep missing the point. • Public reports are treated as anecdotes, not proof: a Reddit thread about GPT-6 running extensive checks and coming back with the task unfinished is "their report, not a run I've independently reproduced," and positive reports (a successful overnight task with lower usage) get the same treatment, not cherry-picked out. • The fair test, so you don't have to take anyone's word for it: take an actual unfinished task, give an alternative model the same information and the same goal, and check whether the important requirements survived and whether the result is usable, separate from supervision cost. My take: this is one of the more carefully-hedged opinion videos on the channel, and worth noting exactly where it slipped. The opening line adds "and you probably want to be as well," which is a stronger universal claim than anything else in the video argues for. And near the end, Nick says directly on camera that the video carries some bias and fallacies of its own while trying to be as reasonable as possible, that's not scripted hedging, that's him saying it plainly. 📎 Full guide PDF pinned below: the three judging questions, how every public report is framed (and its limits), the usage-reset nuance, Opus 5.5 vs Astra's actual split benchmark result, and the five-check fair test to run on your own work.

Build and sell an AI service a business pays you $1,000/month for, built and pitched in under 24 hours. It's Hermes Agent, an open-source AI employee with over 250,000 stars on GitHub, hosted on a managed Cloudways server so you're not the one installing security patches at 2am. • Pick a payer, not a problem: a niche passes three checks, a five-figure customer value, 10,000+ of them, and one channel you'll actually use. IT managed service providers pass clean: 25,000-40,000 in the US, and one client is worth $50K-$70K to them over time. • The offer fits in one sentence, no "AI," no tool names: "we find the firms, write and send the emails, and put the calls on your calendar." $500-$1,000/month plus a few hundred per call he shows up to, and if he doesn't sign 5 new clients in 90 days of sending, he gets the month back. • Cloudways is what makes the delivery side actually run: deploy Hermes in a few minutes from the dashboard, pick a region and a model provider, and the server tests every Hermes update before installing it and backs the agent up daily, so a broken release is a one-click rollback instead of a 2am panic. Scout plan runs about $9.90/month, plenty for one nightly job. Code BFCM802 gets 80% off for 2 months plus free migrations, through December 4. My take: the build and the hosting are the easy, cheap part (about $20/month all in). The actual hard part, same as every video like this, is getting a stranger to say yes to the sample. That's most of the video, down to the exact two-sentence message. 📎 Full guide PDF pinned below: all three niche checks, the one-sentence offer formula, the exact Telegram prompts used to set up both jobs, and the two-sentence message that lands the first client. 👉 Which part would actually stop you from shipping this: picking the niche, writing the offer, or the Cloudways setup? Get Cloudways (80% off 2 months + free migrations, code BFCM802): https://www.cloudways.com/en/?id=2202934&coupon=BFCM802

It's brand new, just a handful of subscribers so far, and it's not a replacement for the main channel. It's shorter, more practical videos (8 to 10 minutes) on the part that usually gets skipped: actually getting a team to use the AI system once it's built, not just building it. The newest one, "You Need to Make Your Team Insanely Good at AI," walks through the rollout plan for a recruiting system Nick's team built for a client, candidate search, a scoring queue, contact details filled in automatically. The build demo looked clean. It didn't show whether the recruiters actually knew how to use every result, which is the real point of the video. • The practice has to be real, not a demo: watch how someone already builds a shortlist today, then have them practice on an ordinary candidate, then one with a missing phone number on purpose, then one where the fields contradict each other. Finish by actually handing the shortlist to the next person, not just showing the screen. • One team's results don't predict another's: a study of more than 5,000 support agents found an AI assistant lifted average productivity, but the effect varied a lot by each agent's existing skill and experience. Same system, different outcomes per person, which is why you check instead of assuming. • What actually gets measured: not logins or how many prompts someone ran, but finished jobs and how often the output needs correcting. The rollout only widens once the work holds up and checking it doesn't mean redoing it. My take: this is the half of "AI adoption" nobody wants to film, since there's no slick build moment in rolling out training. Worth a subscribe if you're the one who has to get a team to actually use what gets built for them. 👉 Watch it here ▶️ https://youtu.be/HsMkVIuunS0

A summary of every update across your business isn't the useful job for AI. It can shorten a long report, but it still leaves you to find what matters. This video covers the three places AI actually helps a decision maker: thinking a choice through before you commit, spotting a change in the numbers that needs you, and making your direction to the team clear enough nobody has to guess. • The obvious fix isn't always the fix: a surgical practice client had a missed-call rate already below 1%. Instead of an AI receptionist, the real win was fixing how inquiries got routed to the right staff member. Before building anything, put your proposal and the real facts into an AI chat and ask it to argue against your own plan, then use what it raises to find what still needs checking. • One number alone doesn't tell you why: lower sales could mean fewer orders or smaller ones, and those point to different problems. A healthcare consultancy client gets a daily executive email built from comparing current and earlier billing data, flagging the change so a human can go ask about it, not just summarizing everything. • "Look into it" isn't a decision: a request like "look into a new booking system, get it running next month" leaves out who approves the price and whether next month is a deadline to decide or to be using it. A marketing agency's finished build sat unused because the client still needed to pick an email address and invite their team, a build finishing isn't the same as the work being done. My take: none of these three replace the decision. AI raises the objection or spots the gap, you still have to make the call and say it clearly enough that your team isn't guessing. That's a smaller ask than most "AI for CEOs" pitches make it sound, and that's probably why it's usable. 📎 Full guide PDF pinned below: all three jobs with the real client examples, the exact prompt pattern for each, and a quick checklist to try on something you already own. 👉 Which of the three do you actually need right now: thinking a choice through, spotting a change, or making your direction clearer?

Right now your business runs on stuff that only lives in one person's head — usually yours. Every decision waits on them, and if they leave, it leaves too. So I got it out: every price, process, and judgment call into one folder the company owns, that the whole team works out of. One shared brain that doesn't resign, take weekends off, or walk out with 15 years of knowledge. No code, one afternoon. - It's just a folder + skills. The folder is what the company knows (prices, customers, processes); skills are how it behaves (the rules). Gather your scattered docs into one folder and it reads them like a sharp ops manager on day one — then flags what's missing, usually the pricing that only lives in your head. One interview prompt gets that out: the AI asks, you talk, it writes your pricing rules into the folder in ~20 minutes. Turn it into a skill in one plain-English paragraph and anyone runs the job the same way, every time. - The quit test. Ask the brain what your most-depended-on person knows. What comes back is safe forever; what's missing is your homework. In the demo, the volume-pricing question that sat unanswered all weekend now comes back in seconds — routine 90% done, the one real judgment call escalated with the thinking finished. My take: it drafts, you approve (~90% automated, a person finishing the last 10%). The hard part was never the tech — it's deciding to get the knowledge out of people's heads. 📎 Full guide PDF pinned below — the folder + skills model, every prompt, and the quit test. 👉 Which person in your business would you run the quit test on first? Drop it below. [Watch the video here ▶️]