---
title: '$1,000,000 in One Year, or I Shut Everything Down'
source: 'https://youtube.com/watch?v=011eW6-S5RA'
video_id: '011eW6-S5RA'
date: 2026-07-31
duration_sec: 1026
---

# $1,000,000 in One Year, or I Shut Everything Down

> Source: [$1,000,000 in One Year, or I Shut Everything Down](https://youtube.com/watch?v=011eW6-S5RA)

## Summary

A crypto YouTuber publicly commits to making $1 million within one year or shutting down all his projects, including his company and YouTube channel. He outlines three distinct paths: crypto altcoin investments, systematic Polymarket backtesting, and a custom AI model to predict Bitcoin's hourly direction. The video is a high-stakes strategy breakdown with a hard deadline of May 1st, 2027.

### Key Points

- **The Million-Dollar Ultimatum** [00:02] — The creator commits to making $1 million in one year or shutting down everything, including Council and all projects, with no room for failure.
- **Starting from Zero** [00:40] — The new million-dollar goal is on top of any past earnings, effectively starting from scratch and pushing hard for the next 12 months.
- **Path One: Crypto Investments** [01:06] — The first route is investing in the crypto projects he has discussed over the past year, expecting a market bounce in late 2026 and 2027.
- **Una Launches Records** [02:40] — Una, a project flying under the radar, announced Records, the first AI infrastructure for IP growth built on Base, with partners like Pudgy Penguins, Doodles, and Moonbirds.
- **Persona IP's Explosive Growth** [03:05] — Persona IP saw 774% growth on Giphy over six months, dominating all other brands and demonstrating the potential of linking NFTs to real-world popularity.
- **NFT Provenance via Records** [03:18] — A Pudgy Penguin NFT with over 2 billion Giphy views would appear ordinary unless Records history is enabled, making its proven fame visible to buyers.
- **Polymarket as a Hedge** [05:31] — The second leg of the strategy is Polymarket because it is agnostic to crypto markets, providing a fallback if the crypto thesis fails.
- **Latency Edge on Polymarket** [06:25] — Most life-changing Polymarket winners exploited latency issues, such as receiving sports broadcasts a few seconds before the market updated.
- **Event Market Mispricing** [07:33] — Odds for four or more rate cuts in 2026 are at 3% due to Iran conflict fear; as fear fades, odds should rise, allowing a buy-low/sell-high exit before settlement.
- **Backtesting Polymarket History** [08:45] — He downloaded the entire history of all Polymarket markets to build a backtesting database, using AI to systematically identify profitable strategies.
- **The Accuracy Threshold** [10:05] — 65% accuracy on Bitcoin's hourly direction is enough to profit; 75% is a money printer; 85% is mind-blowing, while most investors sit at 53–55%.
- **Custom Non-Transformer Model** [12:40] — Instead of fine-tuning expensive transformer models like DeepSeek, he chose to build a custom state-based model from scratch.
- **Transformers vs. State-Based Models** [13:21] — Transformers maintain a giant notepad, using exponentially more compute as context grows; state-based models are fuzzier but closer to human memory.
- **Fixing State-Based Model Flaws** [14:42] — He worked for 1.5 years to give state-based models perfect memory and fix in-context learning (ICL), then needed a real-world test.
- **The Ultimate Test** [15:22] — The proof of concept is beating 65% accuracy on predicting Bitcoin's price direction hour by hour — an impossible task where his architecture could outshine transformers.

### Conclusion

The video lays out a high-stakes, three-pronged strategy to reach $1M in a year: altcoin investments, Polymarket trading with systematic backtesting, and a custom AI model aimed at predicting Bitcoin. It's a blend of practical tactics and a long-shot Hail Mary, all driven by the pressure of an all-or-nothing deadline.

## Transcript

million dollars, and if I fail, I'm shutting down everything. That includes Council, all of it. All of it gets shut down if in a year I haven't made at least a million dollars. I'm giving
myself zero room for failure over the next 12 months to ensure I push as hard as physically possible on every opportunity currently available. And that's because I believe right now we're at one of the most exciting periods in
that there's ever been. And I want to make sure that I don't waste that opportunity by being lazy or just kind of counting on my past successes to carry me through. And to be very specific, I have until May 1st, 2027 to
pull this off, and I have a couple different routes that I'm planning on million-dollar goal. And this million-dollar goal is above and beyond anything I've made up to this point. Okay, so I, you know, back in the day I
had my 1 1 million goal, I made a million dollars, but none of that money actually counts towards this goal. It's as if effectively I'm starting from zero and really pushing that over the next year I'm going to make another million
dollars given the opportunities that are at hand. And I'm approaching this from is the one that you guys are the most familiar with, and that's investing in in all the projects that I've been talking about over the last year and
bounce back, especially later into this year and into 2027. And in this video I'm holding, where I'm kind of placing my bets when it comes to this of a brief update on some of these projects. But I'm not just putting all
my eggs in that one basket because this is obviously really high stakes for me. to try to get me to this million-dollar goal is Polymarket and various Polymarket. And over the past couple weeks I actually went and downloaded the
entire history of like all markets on Polymarket from the very beginning of I've been back-testing a lot of different ideas and strategies based on that data on on what works and what doesn't when it comes to making money on
Polymarket. And the third route that I'm taking is probably the one that excites that's disappointed me the most and proven to be so far the most difficult to actually get anything done with, and that's a project that I've been working
going to get into a little bit later in this video. Okay, first off, I'm going on currently, and I'm going to make this really really brief because I talk about this already a ton. That's going to be projects like ResearchHub RSC, Infra,
which is just a really tiny AI bet, Ava, which is another really tiny AI bet, and dive into a little bit today and highlight what they're doing is Una. Una has been really flying under the radar and doing quite a bit of building in
announced an expansion called Records. And as they describe it, Records is the first AI infrastructure for IP growth built on Base. And Una has some top-tier partners when it comes to launching Records, including Pudgy Penguins,
Doodles, Moonbirds, Clone X Sords, and some of the top brands in the NFT sector. And this is Giphy growth for the Persona IP over the last 6 months, and you can see Persona is absolutely dominating all other brands across Giphy
with 774% growth over the last 6 months combined. And they're doing this starting with NFTs, which is why they have a lot of these top-tier NFT for me and kind of walked me through the platform, and one of the things they
showed me was this NFT, this Pudgy Penguin NFT, that on Giphy has over 2 billion views. But traditionally, typically, if you were to go buy this You'd have no way to connect that this NFT is actually a pretty famous NFT. It
to you would just kind of appear like any other Pudgy Penguins NFT. Records these NFTs where somebody going to buy that NFT with Records history enabled would see that this NFT has over 2 billion views on Giphy, and that would
add to the value or the story around that NFT. And that's just kind kind the surface, and NFTs is really just the starting point for what records could be. You can imagine in a world where it's really easy to make this top-tier
cinematic content using AI-generated, you know, videos, etc. A lot of brands will want a way to tie in their brand into this user-generated content universe. For example, imagine if Pokémon wanted to run a big campaign for
encourage fans to go make user-generated live-action trailers or or, you know, whatever. Records provides all the infrastructure to be able to do that. It uses AI to categorize the top creators,
the top content, the content that's getting the most reach or achieving that brand's goals or for whatever they're looking for. And basically allows these IPs to run these campaigns with user-generated content universes outside
of the normal brands' IP. There's a lot more to it than that and in a future I mostly just want to highlight that during the last 5 months when prices are are out there still building and they're still pushing the frontier. And I think
Unagi is one of those teams that I just feel like it's inevitable at some point succeed. And when that happens, I think you're going to see a violent repricing of the UNA token because it is valued so, so lowly. Obviously, I hold some
all the details in that because I said I was going to make it quick. But that's kind of in a nutshell my altcoin strategy really is determined by macro and altcoin conditions going into the later half of 2026 and 2027. A lot of
I believe. I'm happy to dive into that in a future video if people want more my strategy. The second part of my strategy, as I mentioned, is Polymarket. And I like Polymarket as a second leg of my strategy because Polymarket is
agnostic to the crypto markets, macro, etc. It is just overall a betting market on a lot of different random things that isn't necessarily just tied to crypto. So, if I was wrong about crypto and crypto didn't end up doing well going
into late 2026, 2027, this strategy could still pull through and help me get to my million-dollar goal. And obviously, there's a lot of stuff on on Polymarket. From this market with
over $58 million in betting volume, betting will Jesus return in 2027 to of course lots of sports markets, lots of markets on crypto, markets on macro finance, and everything else in between. There's just an absurd amount of
opportunity on Polymarket. And as you can see, some of these people have made $22 million, $16 million, $12 million trading against these markets. So, my opportunity. But from what I've seen, I've gone on X and I've gone through a
it's like, "Oh, I made I turned $4,000 into a couple million." Or I you know, "I did this to make a life-changing amount of money on Polymarket." I've kind of the commonality between all those people is they exploited typically
some sort of latency issue. And so, a lot of the people that have made those life-changing $4,000 to a couple million dollars have all been focusing on he noticed that he could get a sports broadcast
like a couple seconds quicker than the data updated on Polymarket. And he was where he ended up making a ton of money on Polymarket because he had that couple where he got a little bit faster, a little bit
And he was able to use that to exploit the market and basically make a ton of aren't really strategies that I feel like even feel like there's probably a lot of edge to exploit on just because there's
so many really sophisticated actors that are already doing strategies like that. different types of strategies on Polymarket that maybe people aren't break these things down from kind of a first principles
viewpoint of like what actually creates opportunity on Polymarket. This is one opportunity on Polymarket where right now the odds of a rate cut on in 2026 for four or more rate cuts is actually at 3%, which I think you'll see this
number go quite a bit higher later into the year. There's just a lot of fear conflict in Iran right now. But what a lot of people don't know about of these markets at any given time. Meaning, I can place a bet yes that
believing that we're going to hit four or more rate cuts in 2026. And then when the fear from all the oil, etc. goes away, you should see the odds of four away, you should see the odds of four cuts go up from 3% to maybe 9% or 12% or
don't think necessarily that's going to happen. But when it does go up, let's say it went up to 15%, I could then sell at 15% and take a profit before this market actually ends. And so you can jump in and out of markets before they
finalize and that's another opportunity to make money on Polymarket. And that's probably countless, probably thousands of different ways and strategies that what I ended up doing is I downloaded the entire history of Polymarket's
markets. So like from the start all the way until today, every single market and downloaded that, created the entire database, like backtesting database, and have been systematically going through that data using AI to try to identify
the top opportunities by backtesting that data and then shadow market, liquidity conditions, executability, etc. And then once I find a strategy that passes all of that, I
it'll work in practice before actually deploying any capital. And so far I have found quite a few really interesting and intriguing strategies that gets me excited about where this strategy could go over the next 12 months. Obviously,
of this. I started this just a couple weeks ago. But but I'm pretty excited about what it could potentially yield. And my final Hail Mary wildcard bet is year and a half. And as I've mentioned, this has been the number one cause of
and a half, all my headaches, and just so much work. And it's just been by and far one of the single hardest things I've ever done. And that is a year and a half ago, I set out to try to build my
own custom architecture to try to better predict markets. And specifically to try to better predict the price of Bitcoin and whether any given hour it would go up or down. The reality is you actually don't have to have that high of accuracy
the markets. In fact, if you could achieve just 65% accuracy, you could investing in the market. If you could achieve 75% accuracy, you literally have a money printer on your hands. 85%
accuracy is like the most mind-blowing thing you could possibly imagine. A lot indicators and these these people on X and stuff, you're assuming they're like 90% accurate. But in reality, like most people are like 53, 55% accurate when it
comes to their investing. Like it's just really hard to get above that threshold. traditional transformer models to try to kind of break past this number. And my this was, you know, the the best of the best have already tried to use
market, to try to find edge on the market. And so coming at that problem the same way is very unlikely to yield any sort of like positive result. And I this. I I'd taken Claude, I'd taken GPT, and all these different models. I've
grand for access to like all the top-tier data in the market. I'd already run through it, and I couldn't really find any edge that would hold. I I built my own custom light GBM model, and it did yield surprisingly decent results,
wanted to get to. And that's when I came up with kind of a crazy sort of side project, side quest of what if I built my own custom model like from scratch? What if I came up with my own custom architecture that was non-transformer
different direction than anyone else is going to try to build a model that would you guys who don't know me, this is just something that I like to do for fun, like an entire year trying to crack one of the millennial problems,
problems. There's seven of them, they're really hard. I tried to crack this one called Navier-Stokes. I used AI and GPT Pro and all this different stuff to try process being that, you know, at some point AI is going to solve this thing
just need a little bit of hand-holding to to get there. So, I spent a year learned a lot about physics and fluid dynamics and a lot of stuff that that. Obviously, I had my one to a million journey where I tried to make a
just about every side hustle that, you know, known to man at this point. I I in general like experimenting and trying out different things like this to learn architecture was probably the most exciting idea that I'd ever come up with
because it just seems like literally an impossible challenge. One that I like 99% chance that was going to fail at, but if I succeeded at was obviously like started. I decided to go an entirely different direction again than most
is they'll take an open-source model like DeepSeek, etc., and then they'll to build their own model. The reason they do that is because it's insanely expensive to train a transformer-based model. You need absurd amounts of
route. So, I had all these limitations of where, you know, I don't I don't have building a custom DeepSeek model wasn't going to get me to my goal. I had to try something that nobody else was trying in order to do what nobody else could do.
attention-based model like a transformer, I decided to build a models are not a new thing. This is There's tons of examples of current state-based models like Mamba, etc., they're out in the wild, but they work
wildly different from attention-based models. Attention-based models are have like a giant notepad in front of them of everything that is in their memory. And it takes exponentially more and more compute in order to keep track
of that notepad, which is why, you know, if you run if you ever ever run an AI goes kind of psycho. And that's because it's trying to maintain its attention across a massive amount of context on this giant notepad. And I always like to
model like a transformer like Claude, GPT, etc. Is like if I tried to tell you my number and you wrote that number down on a piece of paper. It's basically because it's sitting on a piece of paper in front of you. But that's how you
patient who can't remember anything in your own brain. Everything you remember things you try to remember on that paper, the fuller that paper gets to room to write down on that paper. State-based models are a lot more like
how we think and how we remember. It's a lot more fuzzy and a lot more blurry to might memorize it in your brain, but then later when you go to recall that phone number, you might not remember it exactly. You don't have that pristine
paper. It's actually in your brain and and your brain's a little bit more fuzzy that's why a lot of people don't pay attention to state-based models. It's because of that kind of blurry function that limits it in a lot of capacities
do with AI today. And so basically a year and a half ago, I went and set out first one being trying to fix the memory issue, give a state-based model perfect memory. The second one being how do I fix uh the second big flaw of
state-based models, which is this thing called ICL or in-context learning where transformers excel, but state-based models really really suck at it. And so I spent the last year and a half trying to fix those problems. And I got a lot
would be able to get. And basically came to the point where on paper like everything was perfect. Everything looked good. I'd proven a lot. I'd done and accomplished a lot, but I I to prove that in reality. I I needed some sort of
test or some sort of thing uh that was like undeniable that I'd built something awesome. And for me, that was trying to build a better market prediction mechanism and trying to build something that could get above that 65% accuracy
threshold on predicting Bitcoin's direction any given hour, which is a a it sounded like the exact kind of right impossible task where my architecture could excel versus a transformer and really kind of prove itself as something
more to share on that, but as that is my third like wild card Hail Mary, obviously crazy, and I'm not saying it's not crazy, guys. I'm like that's an insane insane gamble bet, but the one that gives me the most energy and joy
and motivation and excitement is that one cuz it's so dang fun. And I learning and all this kind of stuff that I never knew before actually trying to the setup. That's those are my three paths that I'm hoping to over the next
million dollar. I actually want to make a lot more than that, but at minimum a across that threshold so I don't have to shut off my YouTube channel and the OC you're curious about seeing my entire portfolio or you want to see every time
I buy and sell various tokens as well as different weekly video market updates, to new members, but you can sign up for the wait list in the description of this of this is me telling you to do anything
financial advisor and you should always do your own research. If this video is guys think. Is this insane, you know, am I stupid thoughts? And if you want to see more videos like this, make sure you hit that
next to it to be notified each time I release a new video. Thanks for release a new video. Thanks for watching, and I'll see you next week.