A Cool Land | AI engineer

AI × Python engineer (15y). I ship LLMs, trade crypto & stocks, and share what actually works. AI-first, no hype 🛰️ NFA

[A Cool Land | AI engineer @acoolland]

I spent 15 years in Python. Then AI ate the world, so I followed it in.

This account is that journey — plus everything I point AI at:
- Python -> AI: the real lessons
- Markets: crypto & stocks, read like data
- Sport: NBA, NFL, football through a model's eyes
- Building in public

No hype. Just the engineer's version.

[A Cool Land | AI engineer @acoolland]

The newest venue in my rotation is @Lighter_xyz on Robinhood Chain. This is week 2, and the first time I am putting its numbers out.

Week 2:
- points: 8.007
- volume: $252,629
- cost: $39.96 in fees
- cost per point: about $5

These points are not cheap. Elsewhere I pay cents for a point, here it is about $5 each. My bear case for a Lighter point is $10, so even at the low end that is twice what the fees cost me. The $10 is my expectation, not a price, so put your own number in.

All of it came from the tradfi markets. Standard account, straight from the web app, nothing clever.

Delta neutral against @variational_io. Long on one venue, short the same thing on the other, so the price moves cancel out and neither side is a bet.

My referral, and yes it is a referral:
https://t.co/h5Qi88Rkpc

[A Cool Land | AI engineer @acoolland]

Week 10 on @OndoPerps. $71.79 in USDC, and I did not place a single trade to get it.

Week 10:
- points: 1,367
- volume: $0
- open interest: $145k
- rewards: $71.79 USDC
- fees: $0.00
- net: $71.79

Nothing to subtract this week. The position just sits there, rewards come from open interest, and doing nothing costs nothing.

The part I cannot price yet is the 1,367 points. There is still no token, so what they are worth is a guess.

Delta neutral against @risextrade. Long on one venue, short the same thing on the other, so the price moves cancel out and neither side is a bet.

My referral, and yes it is a referral, we both get a boost:
https://t.co/xwhte1yT8T

My last Ondo numbers, if you want to see where this is going:
https://t.co/UY9PMeG0F2

[A Cool Land | AI engineer @acoolland]

Week 8 on @OndoPerps. I placed no trades at all, and it paid $89.49.

Week 8:
- points: 2,794
- volume: $0
- open interest: $260k
- rewards: $89.49 USDC
- fees: $0.00
- net: $89.49

The zero is not a typo. Rewards here come from open interest, not from trading. Once the position is on, leaving it alone costs nothing and still pays.

So there was no cost side this week. Everything that came in stayed in, and 2,794 points came with it. There is still no token, so what those points are worth is anyone's guess.

Delta neutral against @risextrade. Long on one venue, short the same thing on the other, so the price moves cancel out and neither side is a bet.

My referral, and yes it is a referral, we both get a boost:
https://t.co/xwhte1yT8T

My last Ondo numbers, if you want to see where this is going:
https://t.co/VSrjnUzG5u

[A Cool Land | AI engineer @acoolland]

Week 8 on @risextrade. 234.5 points, and getting them cost $34.56 in fees.

Epoch week 8:
- points: 234.5
- volume: $69,136
- open interest: around $250k
- cost: $34.56 in fees
- cost per point: $0.147

Same assumption as before: if a point ends up worth $1.50, the week annualises to about 141% on the margin I have parked there, not on the notional. That $1.50 is still a guess and there is still no token. Put your own number in and the answer moves.

The cost side is not a guess. $34.56 in fees, 14.7 cents a point.

Delta neutral against @OndoPerps. Long on one venue, short the same thing on the other, so the price moves cancel out and neither side is a bet. The points are the product.

If you want to try it, my invite: https://t.co/AEbzMX4pJr

Last week's numbers, if you want to see where this is going:
https://t.co/e8nHyAtaBC

[A Cool Land | AI engineer @acoolland]

Week 7 on @risextrade. 281.5 points, and getting them cost $61.67 in fees.

Epoch week 7:
- points: 281.5
- volume: $123,357
- cost: $61.67 in fees
- cost per point: $0.219

Same assumption as before: if a point ends up worth $1.50, the week annualises to about 169% on the margin I have parked there, not on the notional. That $1.50 is still a guess and there is still no token. Put your own number in and the answer moves.

The cost side is not a guess. $61.67 in fees, nearly 22 cents a point. That is the most I have paid for a point on this venue.

Delta neutral against @OndoPerps. Long on one venue, short the same thing on the other, so the price moves cancel out and neither side is a bet. The points are the product.

Invite codes. First come first served:

E3T9WOH9
BOEPWI3I
IJJJ79N7
OTEX8KLR
ASPJ20K8

Last week's numbers, if you want to see where this is going:
https://t.co/NxvNjxizb7

[A Cool Land | AI engineer @acoolland]

Week 7 on @risextrade. 281.5 points, and getting them cost $61.67 in fees.

Epoch week 7:
- points: 281.5
- volume: $123,357
- cost: $61.67 in fees
- cost per point: $0.219

Same assumption as before: if a point ends up worth $1.50, the week annualises to about 169% on the margin I have parked there, not on the notional. That $1.50 is still a guess and there is still no token. Put your own number in and the answer moves.

The cost side is not a guess. $61.67 in fees, nearly 22 cents a point. That is the most I have paid for a point on this venue.

Delta neutral against @OndoPerps. Long on one venue, short the same thing on the other, so the price moves cancel out and neither side is a bet. The points are the product.

Invite codes. First come first served:

E3T9WOH9
BOEPWI3I
IJJJ79N7
OTEX8KLR
ASPJ20K8

Last week's numbers, if you want to see where this is going:
https://t.co/NxvNjxizb7

[A Cool Land | AI engineer @acoolland]

Week 6 on @risextrade. 171.4 points, and getting them cost $22.97 in fees.

Epoch week 6:
- points: 171.4
- volume: $76,585
- open interest: between $100k and $200k
- cost: $22.97 in fees
- cost per point: $0.134

Same assumption as before: if a point ends up worth $1.50, the week annualises to about 103% on the margin I have parked there, not on the notional. That $1.50 is still a guess. There is still no token. Put your own number in and the answer moves.

The cost side is not a guess. $22.97 in fees, 13.4 cents a point. I open positions and leave them alone, so no slippage, and the fee stays at the bare 0.03%.

Delta neutral against @variational_io. Long on RISEx, short the same thing on Variational, so the price moves cancel out and neither side is a bet. The points are the product.

Invite codes, refreshed. First come first served:

E3T9WOH9
BOEPWI3I
IJJJ79N7
OTEX8KLR

My last RISEx numbers, if you want to see where this is going:
https://t.co/giOWvVxAn1

[A Cool Land | AI engineer @acoolland]

Week 8 on @OndoPerps. I placed no trades at all, and it paid $89.49.

Week 8:
- points: 2,794
- volume: $0
- open interest: $260k
- rewards: $89.49 USDC
- fees: $0.00
- net: $89.49

The zero is not a typo. Rewards here come from open interest, not from trading. Once the position is on, leaving it alone costs nothing and still pays.

So there was no cost side this week. Everything that came in stayed in, and 2,794 points came with it. There is still no token, so what those points are worth is anyone's guess.

Delta neutral against @risextrade. Long on one venue, short the same thing on the other, so the price moves cancel out and neither side is a bet.

My referral, and yes it is a referral, we both get a boost:
https://t.co/xwhte1yT8T

My last Ondo numbers, if you want to see where this is going:
https://t.co/VSrjnUzG5u

[A Cool Land | AI engineer @acoolland]

Week 4 on @OndoPerps. $47.85 in USDC came in, and $45.68 of it went straight back out in fees.

Week 4:
- points: 1,378
- volume: $152,250
- rewards: $47.85 USDC
- fees: $45.68
- net: $2.17

That is the entire cash result. Two dollars and seventeen cents. The rewards covered the costs and almost nothing beyond them.

So what did the week actually buy? 1,378 points, for nothing. The USDC paid for them. Whether that was worth doing depends on what a point turns out to be worth, and there is still no token, so nobody can tell you that.

Delta neutral against @variational_io. Long on one venue, short the same thing on the other, so the price moves cancel out and neither side is a bet.

My referral, and yes it is a referral, we both get a boost:
https://t.co/xwhte1zqYr

Last week's numbers, if you want to see where this is going:
https://t.co/hFqCLfchJM

[A Cool Land | AI engineer @acoolland]

Week 6 on @risextrade. 171.4 points, and getting them cost $22.97 in fees.

Epoch week 6:
- points: 171.4
- volume: $76,585
- open interest: between $100k and $200k
- cost: $22.97 in fees
- cost per point: $0.134

Same assumption as before: if a point ends up worth $1.50, the week annualises to about 103% on the margin I have parked there, not on the notional. That $1.50 is still a guess. There is still no token. Put your own number in and the answer moves.

The cost side is not a guess. $22.97 in fees, 13.4 cents a point. I open positions and leave them alone, so no slippage, and the fee stays at the bare 0.03%.

Delta neutral against @variational_io. Long on RISEx, short the same thing on Variational, so the price moves cancel out and neither side is a bet. The points are the product.

Invite codes, refreshed. First come first served:

E3T9WOH9
BOEPWI3I
IJJJ79N7
OTEX8KLR

My last RISEx numbers, if you want to see where this is going:
https://t.co/giOWvVxAn1

[A Cool Land | AI engineer @acoolland]

Week 4 on @risextrade, posted late because I was away.

Epoch week 4:
- points: 213.8
- volume: $51,932
- open interest: $179k, held in BTC, ETH, SOL, XRP and crude oil
- cost: $15.57 in fees
- cost per point: $0.073

Now the number I have been avoiding. If a point ends up worth $1.50, this week annualises to about 175% on the margin I have parked there, not on the notional.

Read that sentence again, because the word doing all the work is "if". There is no token yet. $1.50 is my assumption, not a price.

I picked it to sit on the low side rather than the flattering side. In a strong market a point could be worth a multiple of that. In a bad one the token never arrives and the honest answer is zero. Put your own number in and the APR moves with it.

I am publishing it anyway, because a guess with the working shown beats no number at all. The cost side is not a guess. $15.57 in fees, 7.3 cents a point. I open positions and leave them alone, so no slippage, and the fee stays at the bare 0.03%.

Delta neutral against @variational_io. Long on RISEx, short the same thing on Variational, so the price moves cancel out and neither side is a bet. The points are the product.

Invite codes, refreshed. First come first served:

CG0IA5YL
R7KCP29Z
BOEPWI3I
E3T9WOH9
IJJJ79N7

Last week's numbers, if you want to see where this is going:
https://t.co/dF1YdhXi57

[A Cool Land | AI engineer @acoolland]

Week 4 on @OndoPerps. $47.85 in USDC came in, and $45.68 of it went straight back out in fees.

Week 4:
- points: 1,378
- volume: $152,250
- rewards: $47.85 USDC
- fees: $45.68
- net: $2.17

That is the entire cash result. Two dollars and seventeen cents. The rewards covered the costs and almost nothing beyond them.

So what did the week actually buy? 1,378 points, for nothing. The USDC paid for them. Whether that was worth doing depends on what a point turns out to be worth, and there is still no token, so nobody can tell you that.

Delta neutral against @variational_io. Long on one venue, short the same thing on the other, so the price moves cancel out and neither side is a bet.

My referral, and yes it is a referral, we both get a boost:
https://t.co/xwhte1zqYr

Last week's numbers, if you want to see where this is going:
https://t.co/hFqCLfchJM

[A Cool Land | AI engineer @acoolland]

Week 3 on @OndoPerps. $118.51 in USDC, and it cost me $35.84 to earn it.

Week 3:
- points: 1,472
- volume: $119,450
- open interest: $138k to $140k
- rewards: $118.51
- costs: $35.84 in fees
- net: $82.67

The book splits about in half. $68.5k of it is BTC and ETH. The other $70k is crude oil, WTI and Brent. Not a sentence I expected to type this year.

Brent is in there for a reason. It was part of a $25k incentive pool this week. Ondo puts a smaller bonus pool on one corner of the market at a time, and far fewer people crowd the small one than the main one.

Delta neutral against @variational_io, same as everywhere else. Long on one venue, short the same thing on the other, so the price moves cancel out and neither side is a bet.

It comes to about 2.4 cents a point. That is the number worth comparing if you farm Ondo too, not the points total.

My referral, and yes it is a referral, we both get a boost:
https://t.co/xwhte1yT8T

If you would rather have the version of this that does not involve staring at a funding screen, I wrote that up too. Five habits, only one of them about picking anything:
https://t.co/n2rQCw8nN8

[A Cool Land | AI engineer @acoolland]

Week 4 on @risextrade, posted late because I was away.

Epoch week 4:
- points: 213.8
- volume: $51,932
- open interest: $179k, held in BTC, ETH, SOL, XRP and crude oil
- cost: $15.57 in fees
- cost per point: $0.073

Now the number I have been avoiding. If a point ends up worth $1.50, this week annualises to about 175% on the margin I have parked there, not on the notional.

Read that sentence again, because the word doing all the work is "if". There is no token yet. $1.50 is my assumption, not a price.

I picked it to sit on the low side rather than the flattering side. In a strong market a point could be worth a multiple of that. In a bad one the token never arrives and the honest answer is zero. Put your own number in and the APR moves with it.

I am publishing it anyway, because a guess with the working shown beats no number at all. The cost side is not a guess. $15.57 in fees, 7.3 cents a point. I open positions and leave them alone, so no slippage, and the fee stays at the bare 0.03%.

Delta neutral against @variational_io. Long on RISEx, short the same thing on Variational, so the price moves cancel out and neither side is a bet. The points are the product.

Invite codes, refreshed. First come first served:

CG0IA5YL
R7KCP29Z
BOEPWI3I
E3T9WOH9
IJJJ79N7

Last week's numbers, if you want to see where this is going:
https://t.co/dF1YdhXi57

[A Cool Land | AI engineer @acoolland]

Week 3 on @risextrade. 199.1 points, and the trading it took to get them cost $3.71.

Epoch week 3:
- points: 199.1
- volume: $12,367
- open interest: $127.5k, held in BTC, ETH, SOL, XRP and crude oil
- cost: $3.71 in fees
- cost per point: $0.019

There is still no token, so nobody knows what a point is worth. Cost per point is the only part of this you can measure while you wait, and mine came in under two cents.

I open positions and leave them alone. No closing and reopening, so no slippage, which is the bit you lose crossing the spread every time you trade. That keeps the cost at the bare 0.03% fee.

Delta neutral against @variational_io. Long on RISEx, short the same thing on Variational, so the price moves cancel out and neither side is a bet. The points are the product.

Same invite codes as last time, so some may already be gone. First come first served:

S9W0RKP2
CG0IA5YL
R7KCP29Z
BOEPWI3I
E3T9WOH9

Last week's numbers, if you want to see where this is going:
https://t.co/eNTDir433b

[A Cool Land | AI engineer @acoolland]

Week 3 on @risextrade. 199.1 points, and the trading it took to get them cost $3.71.

Epoch week 3:
- points: 199.1
- volume: $12,367
- open interest: $127.5k, held in BTC, ETH, SOL, XRP and crude oil
- cost: $3.71 in fees
- cost per point: $0.019

There is still no token, so nobody knows what a point is worth. Cost per point is the only part of this you can measure while you wait, and mine came in under two cents.

I open positions and leave them alone. No closing and reopening, so no slippage, which is the bit you lose crossing the spread every time you trade. That keeps the cost at the bare 0.03% fee.

Delta neutral against @variational_io. Long on RISEx, short the same thing on Variational, so the price moves cancel out and neither side is a bet. The points are the product.

Same invite codes as last time, so some may already be gone. First come first served:

S9W0RKP2
CG0IA5YL
R7KCP29Z
BOEPWI3I
E3T9WOH9

Last week's numbers, if you want to see where this is going:
https://t.co/eNTDir433b

[A Cool Land | AI engineer @acoolland]

Another perp dex I farm, and it does not pay in USDC, is @risextrade. It pays points, and there is no token yet.

What I do know is what mine cost. 8.9 cents each.

Epoch week 2:
- points: 129.5
- volume: $38,400
- open interest: $115k
- markets: SOL, XRP, BTC, ETH
- cost: $11.52 in fees
- cost per point: $0.089

The strategy is the same one I run everywhere: delta neutral against @variational_io. Long on RISEx, short the same thing on Variational, so the price moves cancel out and neither side is a bet. The points are the product.

I only opened positions and left them alone. No closing and reopening, so no slippage, which is the bit you lose crossing the spread every time you trade. That keeps the cost down to the bare 0.03% fee.

I am nowhere near the top of the RISEx leaderboard and I am not trying to be. Cost per point is the part I can actually control, so that is the number I track.

I have five invite codes. First come first served:

S9W0RKP2
CG0IA5YL
R7KCP29Z
BOEPWI3I
E3T9WOH9

That is the fiddly end of making money work. The slow end I wrote up here, five habits, only one of them about picking anything: https://t.co/n2rQCw8nN8

[A Cool Land | AI engineer @acoolland]

Week 3 on @OndoPerps. $118.51 in USDC, and it cost me $35.84 to earn it.

Week 3:
- points: 1,472
- volume: $119,450
- open interest: $138k to $140k
- rewards: $118.51
- costs: $35.84 in fees
- net: $82.67

The book splits about in half. $68.5k of it is BTC and ETH. The other $70k is crude oil, WTI and Brent. Not a sentence I expected to type this year.

Brent is in there for a reason. It was part of a $25k incentive pool this week. Ondo puts a smaller bonus pool on one corner of the market at a time, and far fewer people crowd the small one than the main one.

Delta neutral against @variational_io, same as everywhere else. Long on one venue, short the same thing on the other, so the price moves cancel out and neither side is a bet.

It comes to about 2.4 cents a point. That is the number worth comparing if you farm Ondo too, not the points total.

My referral, and yes it is a referral, we both get a boost:
https://t.co/xwhte1yT8T

If you would rather have the version of this that does not involve staring at a funding screen, I wrote that up too. Five habits, only one of them about picking anything:
https://t.co/n2rQCw8nN8

[A Cool Land | AI engineer @acoolland]

https://t.co/cJhZNzTSlZ

[A Cool Land | AI engineer @acoolland]

Last week @pacifica_fi paid me 982 points. The trading it took to earn them cost 14 cents.

Week 48:
- open interest: about $150k, held in XAUT (tokenised gold), BTC and ETH
- 7 day volume: $348.54
- points: 982
- fees: $0.14, at 0.04% of volume

The volume number is not a typo. On Pacifica the points come from open interest, not from trading. The only reason I trade at all is a daily boost worth 25%, and the smallest trade that qualifies is enough to claim it.

So the routine is dull on purpose. Hold the position, place one small trade a day, collect.

Delta neutral against @variational_io, same as everywhere else. Long on one venue, short the same thing on the other, so the price moves cancel out and neither side is a bet.

My referral, and yes it is a referral, it helps us both:
https://t.co/GQw0D0kRvY

For the record, none of this is how most people actually get richer. I wrote up the boring version too, five habits, only one of them about picking anything:
https://t.co/n2rQCw8nN8

[A Cool Land | AI engineer @acoolland]

https://t.co/cJhZNzTSlZ

[A Cool Land | AI engineer @acoolland]

Everyone tells you to find a better return. Nobody tells you what a better return costs.

$200 a month at 8 percent. Start at 22 and you have about $895,000 at 65. Start at 32 and you have about $387,000.

To close that gap the late starter does not save one extra dollar. He has to earn 11.5 percent a year instead of 8. Every year. For 33 years.

So a ten year head start is worth about 3.5 points of annual return. Fine, you think. Go get 3.5 more points.

Here is the problem. Reaching for return means bigger swings, and swings charge their own fee. Square the swing, halve it, that is roughly what you lose. A 15 percent swing costs about 1.1 points a year. A 20 percent swing costs about 2. It comes out of the money you actually end up with, however good the average looks on paper.

Then inflation takes a second cut. 8 percent with 2.5 percent inflation is 5.4 percent real.

A higher return gets taxed twice, every year, forever. The head start gets taxed zero times, and it costs one decision.

That is the part nobody puts on the slide.
Long version here.
https://t.co/n2rQCw8VCG

[A Cool Land | AI engineer @acoolland]

https://t.co/cJhZNzTSlZ

[A Cool Land | AI engineer @acoolland]

@HyperliquidNews @tradeparagon A third of open interest sitting in $UNITREE means the real risk here is one oracle. Fine at $6M, harder to scale much past it.

[A Cool Land | AI engineer @acoolland]

@CryptoPicsou The real test is what HIP-3 volume does after the discount ends. Flow that only exists at 90% off is rented, not owned.

[A Cool Land | AI engineer @acoolland]

@_stevenhl A DEX-wide fee scale wastes the discount on books that already have flow. Per asset lets you keep the subsidy where liquidity has not shown up yet.

[A Cool Land | AI engineer @acoolland]

@Langerius Because a perp on gold needs an oracle and a funding rate, not custody or settlement. Once HL showed that template holds volume, adding a market cost almost nothing.

[A Cool Land | AI engineer @acoolland]

@goul82 @risextrade ~7k

[A Cool Land | AI engineer @acoolland]

Another perp dex I farm, and it does not pay in USDC, is @risextrade. It pays points, and there is no token yet.

What I do know is what mine cost. 8.9 cents each.

Epoch week 2:
- points: 129.5
- volume: $38,400
- open interest: $115k
- markets: SOL, XRP, BTC, ETH
- cost: $11.52 in fees
- cost per point: $0.089

The strategy is the same one I run everywhere: delta neutral against @variational_io. Long on RISEx, short the same thing on Variational, so the price moves cancel out and neither side is a bet. The points are the product.

I only opened positions and left them alone. No closing and reopening, so no slippage, which is the bit you lose crossing the spread every time you trade. That keeps the cost down to the bare 0.03% fee.

I am nowhere near the top of the RISEx leaderboard and I am not trying to be. Cost per point is the part I can actually control, so that is the number I track.

I have five invite codes. First come first served:

S9W0RKP2
CG0IA5YL
R7KCP29Z
BOEPWI3I
E3T9WOH9

That is the fiddly end of making money work. The slow end I wrote up here, five habits, only one of them about picking anything: https://t.co/n2rQCw8nN8

[A Cool Land | AI engineer @acoolland]

https://t.co/cJhZNzTSlZ

[A Cool Land | AI engineer @acoolland]

@OndoPerps @arbitrum great news and simplifies life 🔥

[A Cool Land | AI engineer @acoolland]

The ten years you spend not starting is the most expensive thing in this clip.

Here is what it is actually worth. Two people put $200 a month into the same boring fund. Same 8 percent a year, same everything. One starts at 22. One starts at 32.

At 65 the early one has about $895,000. The late one has about $387,000.

The late one did not save half as much. He put in $79,200 against $103,200. So he contributed 77 percent of the money and walked away with 43 percent of the result.

Now the part that rearranged my head.

For the late starter to catch up he does not need to save one dollar more. He needs to earn 11.5 percent a year instead of 8. Every year. For 33 years. Without one bad decade in the middle.

That is the whole trade. Ten years of doing nothing clever is worth about three and a half points of annual return, forever.

One of those is a decision you can make on a Tuesday afternoon. The other is a career, and most careers spent trying do not get there.

The second point in that thread is that the return rate matters more than the start. On pure arithmetic, correct. In an actual life, no, because you get to choose when you start and you do not get to choose your return.

Run it at 5 percent or at 10 percent if you think I picked a friendly number. The gap stays between three and four points either way.

I wrote the long version, and the other four habits, here.
https://t.co/n2rQCw8nN8

[A Cool Land | AI engineer @acoolland]

https://t.co/cJhZNzTSlZ