OG crypto since 2016 | Crypto founder since 2020 | Support $15M+ raised Gamefi projects | CertiK partner | SEA × MENA network | DM open. Building @lookout_xyz
The Ledger / CryptoBilis story raises an uncomfortable founder question:
If a customer follows your recommended buying route, how much responsibility stays with you?
My weekend article looks at what self-custody still depends on 👇
#Crypto#Business https://t.co/DjvRzqLeY8
I think crypto sometimes asks memes to prove the wrong thing.
A meme doesn't need to pretend it's a software business to be interesting.
Culture, identity and speculation can explain why people participate. I'd rather hear that honestly than read a roadmap invented to make a joke look like a startup.
But “it's culture” can't answer every market question either.
If most distribution comes from holders who need another buyer, I'd want to know whether the audience would still care without the ticker.
And a community can be culturally alive while the token becomes a worse trade. People enjoying the joke doesn't tell me the price makes sense.
That's where I'd separate three questions:
Is the meme spreading beyond existing holders?
Is there fresh demand for the asset?
What does the current price already assume?
Those questions can produce three different answers.
The uncomfortable bit: something can be a good meme, a weak business and an overcrowded trade at the same time.
I'd still research it. I just wouldn't borrow evidence from one category to defend another.
Where do you draw the line between a meme people want to belong to and a market people want to exit through? #Crypto #Memes
A bank making crypto easier to use can be good for adoption and uncomfortable for some crypto businesses at the same time.
Coinbase's September 28 announcement with Citi is a useful example.
The setup described includes stablecoin acceptance where the merchant receives fiat settlement without having to hold or manage the stablecoin directly.
From a merchant's point of view, that can be a feature.
They want to get paid, reconcile the payment and deal with the occasional problem. Becoming a crypto treasury operator may not be on their to-do list.
But here's the part I'd debate:
If the bank or payment platform owns the customer relationship, the chain may become easier to replace than the interface.
That doesn't mean the chain captures no value. It means I'd want to know how it captures value before treating payment volume as an investment thesis.
A dollar can move through several layers:
The issuer may earn from reserves.
The payment provider may charge for the service.
The bank may retain the commercial relationship.
The chain may collect fees.
Those are different businesses. More activity doesn't reward all of them equally.
And reducing fees is great for adoption, but it can also make transaction volume a less useful shortcut for estimating revenue.
As a founder, I'd look for the part customers would struggle to replace: distribution, reconciliation, local payout access, support, or integration with the way they already work.
As a crypto researcher, I'd ask a second question: does the asset I'm evaluating actually have a mechanism to benefit from that business?
My hypothesis is that some of the strongest stablecoin businesses will look fairly boring from the outside. Customers may barely notice they're using crypto.
I'd change my mind about who captures the value if users consistently choose the underlying network, pay a premium for it, and make that preference matter to the provider.
Until then, “stablecoins are growing” and “this token benefits” are two separate arguments I'd want someone to connect. #Crypto #Business
The most expensive AI model can be the wrong hire for your business.
Coinbase's fraud-model results this week make that worth discussing.
Here's how I'd choose: which mistakes cost money, what the benchmark misses, and when a smaller model makes sense 👇 #AI#Crypto
My view: the valuable AI capability inside a business may be knowing when an answer is good enough to act on.
That includes knowing when to stop, ask for more evidence or hand the case to someone else.
For something like Lookout, generating a convincing market story is different from establishing that the evidence supports it.
For payments, explaining a transaction is different from deciding to approve it.
I'd spend on stronger models where they improve those decisions. I wouldn't use their reputation as a substitute for checking the result. #AI #Crypto
There's a countercase to all this.
A narrow model can look excellent until the input changes. Fraud patterns change. A new customer group arrives. The workflow grows beyond the task it was trained for.
And owning the model means owning more of the maintenance.
For a small team with limited data and irregular volume, a frontier API may still be the sensible starting point.
I'd move toward specialization when the task repeats, the outcome can be checked and the savings justify the extra work.
Small isn't a strategy by itself.
A token paying people to use a product doesn't automatically make the usage fake.
But I'd separate the people who want the service from the people who want the reward.
The useful question is what they keep doing when the reward gets smaller. #Crypto
If AI makes a business cheaper to run, the owner has a choice:
Keep the margin, lower the price, or spend more on service.
I'd think about which one earns repeat customers before calling the savings a moat. Competitors can use the tools too. #AI#Business
For @lookout_xyz, “why did this appear?” matters as much to me as “what appeared?”
A wallet lead, a narrative lead and a repeated update should send me into different kinds of research.
If they all look like buy signals, the product has failed me. #Lookout
A meme can be funny in ten countries and still need the same small group to buy every dip.
I'd want to know whether the joke travels further than the trading circle.
People sharing it because it's funny is a different signal from holders promoting their bags. #Memes#Crypto
The AI business I'd find interesting: a tool that helps a company get paid sooner.
Not just drafting an invoice. Finding the missing approval, wrong detail or disputed delivery holding it up.
I'd pay attention to days saved, not messages generated. #AI
As a non-technical founder, I want AI to explain which part of a product is expensive to change later.
Button colour? Fine.
How we identify a customer or store their history? I want to understand that before clicking “build” 😂 #AI#Founder
AI can help me do more jobs. It doesn't mean every job deserves equal time.
If I'm tweaking a landing page while nobody is talking to customers, being a one-person team suddenly looks less impressive 😂
I still need to choose where the day goes. #AI#Founder