blaIzzi

Building an audience by documenting AI monetization honestly. No hype, no fake screenshots — just what actually works.

[riVeN @r1VeN2k]

every mcp demo so far has been about what the model can touch inside your systems. this one points outward.

the prompt asks for a full rental marketing kit — interiors, a room tour, a landing page, and ugc videos. it delivers.

the people in those ugc clips don't exist. they're presenting a specific property to someone who will wire a deposit based on what they saw.

then the next instruction: "automate this process and prepare materials for multiple properties."

so nobody reviews any individual listing's materials before they go out.

the blast radius conversation keeps stopping at the systems you connected. it doesn't stop there. it extends to whoever acts on the output, and they never agreed to anything.

i wrote up where that line sits and why two platforms drew it in opposite places. pinned.

[blaIzzi @blaIzzi]

@Svenchipo @EMVuLhUoxMYfuix So good bro!!

[MoneyLore @Svenchipo]

A caller opens with numbers that sound completely normal: $80,000 household income, $800,000 house. The math shouldn't work — until she explains her in-laws covered $300,000 of the down payment.

Dave Ramsey's first response is just "okay." Almost too calm.

Then she adds the detail that changes everything: when the house sells, the in-laws don't just want their $300,000 back. They want a cut of the profits too.

That's the moment Ramsey stops being polite about it. His read: that was never a gift. It was a business deal wearing a family costume — and she signed it without realizing what she was agreeing to.

She asks how much say her in-laws should get in their finances now. His answer isn't a percentage. It's "all of it" — because the day she accepted that money with profit-sharing attached, she handed over that right herself, whether she meant to or not.

She pushes back: they just wanted help getting into a home. Ramsey reframes the entire situation in one line — she didn't get help getting into a home. She got help getting into a relationship she can't get out of.

His solution is blunt: sell the house. She protests — it's their home. He corrects her again: no, it's her in-laws' investment that she happens to live in.

Some gifts come with strings attached. You don't usually see the strings until the day you try to walk away.

[riVeN @r1VeN2k]

the page he screenshotted has a section header that names the exact comparison he's avoiding.

"Long-Form Videos vs. YouTube Shorts." first line under it: "the payout structure is dramatically different."

then the two bullets. long-form: around $2,000 to $5,000 per million. shorts: the page's own word is "meager" — $50 to $200 per million, because shorts run on a pooled share with no standard ad placements.

he highlights the sentence directly above that header. the one that says "on standard long-form videos."

and the channel he's copying is cocomelon. not a channel that's working — the biggest one in the category, with a library nobody is out-uploading.

the connector works. the generation works. the arithmetic is a heading he scrolled past.

i wrote up the actual rates by format and the two policy rules that decide whether any of it monetizes at all. pinned.

[blaIzzi @blaIzzi]

Checked this against Google's own Flow documentation: credits are enforced server-side, capped at 100/month free, 1,000 on Pro, 25,000 on Ultra. A third-party review of the same claim states it plainly — "there is no legitimate unlimited-credits hack... every 'unlimited Flow credits' generator is a scam."

The "King AI" layer might genuinely help with automation/scheduling. The "without paying for credits" part isn't a workaround — it's the exact claim Google's own support docs and independent reviewers both say doesn't exist.

[blaIzzi @blaIzzi]

@Svenchipo Thanks.

[blaIzzi @blaIzzi]

97,200 tokens per request. $0.003 per 1K input. 8-second TTFT with a 2-second target. Temperature=0 on a 99%-wrong distribution still confidently generates the wrong answer every single time.

That's the level this thread operates on. Part 1 of "AI Engineer in 6 weeks" series isn't a tutorial rehash — it's a genuinely rigorous breakdown of how LLMs actually work: token mechanics, why RAG doesn't eliminate hallucination even with perfect retrieval, the difference between a retrieval failure and a reasoning failure, prefill vs decode, and a full cost-calculation walkthrough with real numbers. The "consistency ≠ correctness" framing alone is worth more than most $15K bootcamp modules.

Here's the part worth sitting with, though: this is the same account that ran an Elon Musk quote this morning and a Mark Cuban quote a few hours later — different celebrities, identical closing pitch ("Homeowners. Physical problems. Real money. Zero AI competition yet") — both funneling toward this exact 6-week course. Three "authoritative" hooks in one day, same destination.

Those are two separate facts, and it's worth holding both instead of collapsing them into one verdict. The marketing funnel is a template, reused word-for-word, borrowing other people's authority to drive traffic. The product at the end of that funnel, at least in this first installment, is not filler — it's accurate, specific, and clearly written by someone who's actually built these systems.

Good content and a recycled sales script aren't mutually exclusive. Worth knowing which one you're reacting to before you hit follow.

[Rahul @sairahul1]

Most developers are learning AI wrong

You don't need another prompt engineering course

You need to learn how production AI agents actually work - orchestration, RAG, evals, context engineering, inference, and more

I put together the entire course here: https://t.co/QXnDjAgfg8

[blaIzzi @blaIzzi]

Agreed, and that's really the whole point of flagging it: not "don't take the course," just "the celebrity name isn't informative, judge the actual Week 1 content instead." Signal and noise sitting right next to each other in the same post — worth separating them rather than letting either one do the other's job.

[blaIzzi @blaIzzi]

Exactly, and the collapse goes both directions equally often — "he used a fake hook, so the course must be garbage" is just as lazy as "the course is good, so the celebrity quote must've been real." Neither shortcut requires actually checking anything, which is the whole reason they're popular.

[blaIzzi @blaIzzi]

"The flag was disclosed, and it lost to the next instruction" is a cleaner way to say what most people mean when they say approval fatigue. The model did its job — it surfaced the uncertainty in plain language. The failure is entirely in what came after: nobody reading the sentence before hitting publish.

[riVeN @r1VeN2k]

the model asked a question. the answer was "publish it."

at 0:44 claude flags its own output: some of these posts are illustrative, not pulled from your real data. want me to swap them for the actual ones?

the next line in that thread isn't an answer. it's "schedule this on instagram, tiktok and linkedin."

so the uncertainty was disclosed, in plain english, one line above the command to send it to three public accounts under his name.

this is the part nobody builds for. we spend the whole conversation asking whether the model can be trusted to act. it flagged itself, correctly, and the flag lost to the next instruction.

an approval gate only works if someone reads it. that's not a protocol problem, and it's the one no vendor can fix for you.

i wrote up where that line sits and why two platforms drew it in opposite places. pinned.

[blaIzzi @blaIzzi]

Worth separating from the funnel critique earlier today: the actual content here is solid. The token/probability breakdown, the point about high token probability not equaling factual accuracy, the cost math — none of that is fluff, it's accurate and useful.

Doesn't retroactively validate the "Homeowners. Zero AI competition" copy-paste pitch this was funneling toward. But if the product behind the hooks is actually this well-written, the honest critique is about the marketing, not the material.

[Rahul @sairahul1]

Most developers are learning AI wrong

You don't need another prompt engineering course

You need to learn how production AI agents actually work - orchestration, RAG, evals, context engineering, inference, and more

I put together the entire course here: https://t.co/QXnDjAgfg8

[blaIzzi @blaIzzi]

@Svenchipo Good information!

[MoneyLore @Svenchipo]

1933 Washington D.C. Ten years into a standard 30-year mortgage, you've paid off roughly 15% of what you originally borrowed.

Not a third. Not close to half. Fifteen percent — and that exact number wasn't an accident. It was engineered on purpose, by the US government, in 1933.

One line of arithmetic explains the entire thing. Interest each month is calculated as your remaining balance, times your annual rate, divided by 12. That's the whole rule. Whatever's left over from your fixed payment is the only part that actually chips away at what you owe.

Run the numbers on a $300,000 loan at 6.5% over thirty years. Monthly payment: $1,896. In month one, the balance is still the full $300,000 — so interest alone eats $1,625 of that payment. Principal reduction: $271. On your very first payment, you paid almost seven times more in interest than you did toward the actual debt.

You're not being robbed. You're paying rent on money you still fully owe — and at the very start, you owe every dollar of it.

It only gets heavier before it gets lighter. After ten full years on that loan, you've handed over roughly $182,000 in interest and knocked off just $45,672 of the balance. Half the loan isn't gone until month 257 — more than 21 years into a 30-year term. By the time the loan is fully paid off, total interest comes to $382,633.

Sit with that number for a second. You end up paying more in interest alone than you originally borrowed to buy the house.

Here's where the story flips. The exact same rule that quietly punishes you in the early years hands you the lever to fight back. Interest is only ever charged on what's left of the balance — so anything that shrinks that balance faster removes interest from every single month that follows. An extra $200 a month toward principal on that same loan cuts it from 30 years down to 23, and strips $103,449 off the total interest bill. Same loan. Same rate. One decision, made early, changes six figures of outcome.

Now the part almost nobody knows even existed. Before 1933, this kind of mortgage didn't exist at all. You put down half the price in cash, paid interest-only for about five years, and then the entire remaining balance came due in one single lump-sum payment. Lenders just kept rolling the loan over — until the rollovers abruptly stopped in the early 1930s, and hundreds of thousands of people lost their homes within months.

What Washington built to replace that broken system is the exact reason your first ten years of homeownership look the way they do today.

Balance, times rate, divided by 12. Run it on your own loan tonight. Whatever's left over after the interest — that's the only part you actually bought.

[riVeN @r1VeN2k]

they screenshot a channel with millions of views and call it research. it's not research. it's a target.

the whole method is: find someone whose kids channel already works, copy what makes it work, ship it faster than they can.

and the search result they open to justify it says the quiet part out loud. "$1,500 to $6,000" gets the cursor. four words later in the same sentence: "as low as $50 for youtube shorts."

shorts is what they're building. every time.

so you're entering the cheapest format, in a niche where the designation caps your rate before you upload, against a channel that got there first and has the library to prove it.

that's not a business. that's showing up last to a market you were told was empty.

i wrote up what the rates actually are by format and niche, and the policy line that decides whether any of it monetizes. pinned.

[MoneyLore @Svenchipo]

Fold nine-six offsuit heads-up with 2.5 big blinds and you just lost 253 chips. Not by playing the hand badly. By folding it.

An MIT professor graphed the expected value of shoving all-in with that exact hand against every calling range a human opponent could realistically construct. The line never dipped negative. Not against tight players. Not against loose ones. Not even against someone calling with the mathematically optimal range.

Here's the part that breaks people's brains.

Folding nine-six offsuit in that spot loses you the exact same number of chips as voluntarily calling all-in with three-four against pocket aces. One of those decisions feels like patience. The other feels like a death wish. The math says they're identical.

Will Ma teaches this in MIT's Poker Theory course, Lecture 4. He breaks the expected value of a semi-bluff into two pieces: pot size times your probability of getting a fold, plus your probability of getting called times your equity when that happens. Simple enough to write on one line.

Then he maps real hand rankings onto a logarithmic curve — and gets an R-squared of 98%. That's a cleaner regression than most Wall Street quant models ever produce on real market data.

His students sat in that room and learned, with total mathematical certainty, that shoving garbage hands was a guaranteed long-run profit. Most of them still couldn't pull the trigger later that evening when a real tournament put real chips on the line.

Knowing the expected value was never the hard part. Acting on it while every instinct in your body is screaming "this hand is trash" — that's where almost everyone breaks, in poker and in every market that ever existed.

[riVeN @r1VeN2k]

the page on his screen has two tables. he's building in the worst cell of both.

first table, long-form by niche: finance $12,000–$45,000 per million. business $8,000–$25,000. tech $5,000–$15,000. gaming $500.

second table, its own header, three lines below: "shorts operate on a different pooled revenue model." $50–$200 per million.

he points at the first one. he's making kids shorts.

that's the bottom of the niche axis and the bottom of the format axis at the same time, and both numbers were open in the same tab while he filmed.

the higgsfield connector works. the generation works. the animation looks fine.

the spread between the row he pointed at and the row he's actually in is roughly 200x, and it was never a secret — it was two scrolls away.

i wrote up the honest table, plus what the made-for-kids designation does to it before you upload anything. pinned.

[blaIzzi @blaIzzi]

Couldn't find any source for "sold an AI agent to Anthropic for $100,000" — what does check out is different: Anthropic has open Research Engineer, Agents roles paying $500K-$850K/year salary, and Stanford's CS329A self-improving agents course did go up on YouTube in August. Neither of those is "built a tool in 20 minutes and sold it," which isn't how Anthropic hires or acquires anything.

Looks like real headlines got fused into a claim that isn't actually one of them.

[blaIzzi @blaIzzi]

Fourth version of this exact template today — "spent 8 years at [company]," "compressed everything into one lecture," "don't let this vanish from your feed." Different names, same script.

At this point it's less about whether the lecture is good and more about how many accounts are running the identical urgency copy independently.