Soprano'

crypto, fintech & growth
usually somewhere between the people building, funding & scaling things
@polymarket trader | ads + web

[Iagon @IagonOfficial]

Every AI rollout builds a honey pot.

@NaVi_GaT0R, CEO at Iagon, opened with it at the @Cardano_CF stand at @EBlockchainCon in Barcelona. Data pulled into one place for AI becomes the target.

Iagon spreads it across a sovereign network.
Nothing whole to take.

Thanks to all who came by 🙏

[Soprano' @S0pranoo]

prediction markets had a pretty insane weekend

Polymarket U.S. cleared 1B+ contracts in total volume, while printing consecutive daily ATHs:

saturday -> 510m+
sunday -> 525m+
so we re launching Polymarket data a public source for U.S. + international exchange data

the numbers are getting harder to ignore

and imo, we're nowhere near the ceiling yet

[Soprano' @S0pranoo]

https://t.co/hFv1EtOqse

4/ AI should absolutely be part of the workflow

Use it for research
Use it to pressure test assumptions
Use it to structure messy thinking
Use it to generate alternatives
Use it to move 10x faster

But the final strategy still needs judgement, taste and actual understanding of the product + market

AI can produce the strategy document

It shouldn't be the reason the strategy exists

Because if all you re selling is the first polished answer ChatGPT gave you..

the founder can generate that PDF too :DD

[Soprano' @S0pranoo]

https://t.co/jMGpqLNBPM

3/ And founders trying to "outsmart" marketers with AI are missing the same point

"Lets see what ChatGPT says about our marketing"

Sure. The strategist has ChatGPT too 😂

The difference starts AFTER the model gives you an answer:

why this audience?
why this channel?
why now?
what assumption are we testing?
what makes this fail?
what do we learn if it fails?
what context does the model simply not have?

Thats the part you cant outsource to a prompt

[Soprano' @S0pranoo]

https://t.co/jMGpqLNBPM

3/ And founders trying to "outsmart" marketers with AI are missing the same point

"Lets see what ChatGPT says about our marketing"

Sure. The strategist has ChatGPT too 😂

The difference starts AFTER the model gives you an answer:

why this audience?
why this channel?
why now?
what assumption are we testing?
what makes this fail?
what do we learn if it fails?
what context does the model simply not have?

Thats the part you cant outsource to a prompt

[Soprano' @S0pranoo]

https://t.co/dfIpNnAubP

2/ chatGPT can generate 50 decent marketing ideas before you've finished ur coffee

But it doesnt automatically know which one makes sense for THIS product, THIS audience and THIS moment

I've seen teams push KOL campaigns before they even had clear positioning

Founders asking for community growth when users still couldn't explain what the product actually does

Campaigns proposed because they look good on a deck, without anyone asking what user behavior they're supposed to change

Thats where judgement + experience start mattering

[Soprano' @S0pranoo]

https://t.co/dfIpNnAubP

2/ chatGPT can generate 50 decent marketing ideas before you've finished ur coffee

But it doesnt automatically know which one makes sense for THIS product, THIS audience and THIS moment

I've seen teams push KOL campaigns before they even had clear positioning

Founders asking for community growth when users still couldn't explain what the product actually does

Campaigns proposed because they look good on a deck, without anyone asking what user behavior they're supposed to change

Thats where judgement + experience start mattering

[Soprano' @S0pranoo]

1/ One day founders are gonna realize their "marketing strategist" just sent them the same ChatGPT strategy they generated themselves 10 minutes earlier 😂

Market research
Positioning
Content
KOLs
Community.
Partnership
KPI
30/90-day roadmap

Clean it up, throw it into a nice PDF and suddenly its a "strategy"

The problem isnt using AI

The problem is confusing AI output with actual strategic thinking

[Soprano' @S0pranoo]

1/ One day founders are gonna realize their "marketing strategist" just sent them the same ChatGPT strategy they generated themselves 10 minutes earlier 😂

Market research
Positioning
Content
KOLs
Community.
Partnership
KPI
30/90-day roadmap

Clean it up, throw it into a nice PDF and suddenly its a "strategy"

The problem isnt using AI

The problem is confusing AI output with actual strategic thinking

[Soprano' @S0pranoo]

GM fam :3

I think theres a pretty reasonable argument that $GRAB taking a controlling stake in Atome is actually bullish for $SE and Monee

1) if Grab’s fintech stack was scaling as well as they wanted, there would be much less reason to buy control of an outside platform like Atome. meanwhile Monee has already built significantly more scale across the region

2) from Atome’s side, selling control also says something. if they had very high conviction in becoming the dominant standalone fintech platform in SEA, giving up control at this stage would be a strange move

it reminds me a bit of the Tokopedia/GoTo situation. Tokopedia and TikTok Shop were competing, then TikTok ultimately took a controlling stake in Tokopedia which told you quite a lot about where the competitive pressure was heading

so while the headline is obviously about Grab expanding its fintech footprint, I think the deal indirectly strengthens the case for Sea’s position too

net positive for $SE imo

[Soprano' @S0pranoo]

This setup is kinda wild

Right now, someone would pay me about $3,930 just for committing to buy 100 shares of $META at $500

One put. Jan 21, 2028 expiration

If META stays above $500, the option expires worthless and I walk away with the entire $3,930 premium

If I get assigned, I buy 100 shares for $50,000 but after accounting for the premium, my effective entry is $460.70/share

And since my existing portfolio can cover the margin requirement, I wouldn't necessarily need to park the full $50k in cash while waiting

Thats the attractive part of a portfolio-secured put

The tradeoff is obvious tho

Im committing to that $500 purchase regardless of how far META might fall

If its trading at $400 at expiration, I'd still be buying at $500 and would be down roughly $6,070 after the premium

So basically

$3,930 upfront for taking on the obligation to buy META at a price I agreed to today

Interesting way to get paid for a limit order with very real downside attached

[Soprano' @S0pranoo]

https://t.co/MKISa6mZgq

imo the biggest variable isnt whether Hyperliquid keeps the best product

its whether perps themselves become commoditized

if Robinhood + other major platforms make onchain/perp exposure simple enough for normal users, distribution starts mattering just as much as execution

HL probably keeps the crypto-native power users either way

the real battle is who captures everyone that comes after them

[Flood @ThinkingUSD]

Hyperliquid bear case:

Robinhood UX is good enough to win retail perps and Hyperliquid starts declining in market share.

Hyperliquid metrics start stagnating around a reasonable level 3-5% of global volumes with a hardcore fanbase but gets capped at a price valuation due to easily forcastable revenues.

Hyperliquid base case:

Continued but slightly lower than expected growth, boring supply sink gains of 5-10% a week until blow off top around psychologically relevant numbers (77/88/99/111) etc.

Hyperliquid bull case:

Hyperliquid continues rapidly gaining in market share, it wins out developer attention and there is a new gold rush for people building out products and monetizing via builder codes. Hyperliquid token itself becomes one of the most valuable assets in crypto due to the cashflow it represents and the functionality needed for HIP-3 like implemenation. ETH Halves, Hyperliquid 10xs and Flips ETH.

[Soprano' @S0pranoo]

https://t.co/2H3fencfBG

gm all ^^

AI infrastructure is a compute problem is not the whole story

It's a problem of data ownership too

With the increase in size of AI comes an increase in the volume of sensitive data that passes through centralized systems and having all that data in a few locations is a clear vulnerability

That's why the way we're going at Iagon is important

Distribute the infrastructure. Maintain control nearer to the users. Create AI with a more sovereign base

I believe the side of AI infra will only get more significant for me

[Iagon @IagonOfficial]

Every AI rollout builds a honey pot.

@NaVi_GaT0R, CEO at Iagon, opened with it at the @Cardano_CF stand at @EBlockchainCon in Barcelona. Data pulled into one place for AI becomes the target.

Iagon spreads it across a sovereign network.
Nothing whole to take.

Thanks to all who came by 🙏

[Iagon @IagonOfficial]

The idea behind the talk, in one slide.

Your data stays readable only in your environment. Retrieval and policy run on your gateway. The model gets a query, returns an answer, and holds nothing of yours.

The model comes to the data. Not the other way round.

[Soprano' @S0pranoo]

@sjwolosz123 God Bless

[Soprano' @S0pranoo]

@antonia_mdprjct Amazing, can’t wait to hear more about this!

[Soprano' @S0pranoo]

@CohenSite @bobbyfijan I wouldn’t rebuild any home without redoing grading, soil recompaction before foundation/footings. Too much risk in vertical build without confirming soil is reliable after this disaster.

[Soprano' @S0pranoo]

@davegordon14 Amazing!

[Soprano' @S0pranoo]

@RobertMSterling Thank you RS!

[Soprano' @S0pranoo]

@jmf_5 Thank you so much

[Soprano' @S0pranoo]

@LA_Multi_Fam Like a war zone, can’t believe it.

[Soprano' @S0pranoo]

@JaywalkerCM A neighborhood in Pacific Palisades gone.

[Soprano' @S0pranoo]

@aussieflya Yes ok for now, these fires are popping up and spreading like crazy.