Sayan Sinha Ray

Need some help in speaking to folks working in the cross border payments space either in India working in a PA-CB or a AD Cat 1. If you know anybody in your network please connect me.

@keith_analog @Dipper_pol Exactly

@Sayan91 @Dipper_pol Having followed Taleb on twitter over some years, I get the impression he'd not agree to talk to Fridman as a guest.

While a $170B hedge fund just put AI-researched signals into production, Nassim Taleb has spent 30 years warning this exact moment is coming he made $35M shorting Black Monday 1987 and built a career proving AI and ML break when markets break watch him sit down with Lex Fridman for 10 min on antifragility, risk and why every quant model fails at tail events this is what comes when the math meets reality

UPI’s distribution has been fundamentally built by foreign capital flowing in as VC money through the large TPAPs like Phonepe and Google pay. I would urge you to do a thought experiment. Imagine ‘0’ VC capital flowing in, even with the government subsidizing the banks for upi payments you'd have the infrastructure ready but I don’t think you'd find the risk taking capital in India to build the network. Frankly, w.r.t AA companies wanting to build high quality financial experiences face challenges finding FIUs which are again licensed entities with their large sales cycles, bad tech and unreliable AA pipes. If a founder in India can today raise money from YC, Antler etc they would much rather make an attempt in the US or EU than in India.

🚨🚨Must Read🚨🚨 Why India cannot build a world-class fintech champ yet, writes @guptajiten Opinion | India has built the world’s most envied digital financial infrastructure—the UPI, Aadhaar, the Account Aggregator (AA) and ONDC. And yet, as of 2026, it has not produced a single consumer fintech like Brazil’s Nubank or the UK’s Revolut.. In every market where fintechs have scaled, regulators carved out a licence category between an unregulated tech vendor and a full-fledged bank — allowing non-bank players to hold customer funds, issue cards or extend credit. India has no such middle ground. RBI’s hierarchy is unambiguous: banks first, NBFCs below, fintechs last. Indian fintech funding has fallen from roughly $9–10 billion in 2021 to under $2 billion in 2024–25. The US, EU, Singapore, and Hong Kong have legitimised stablecoins. But India remains hostile.

Just out of curiosity for payers already in your served corridors (US, UK, EU, etc) who could pay via ACH/SEPA into the standard Xflow receiving account, what is the use case where they prefer to pay in USDT/USDC instead? I wanred to understand whether you're seeing: (a) Payers with stablecoin treasuries who would otherwise have to off-ramp themselves to pay via ACH and for whom the all-in cost of paying you in USDT is lower than the all-in cost of off-ramping first? (basically Cost(USDT payment to merchant) < Cost(self off-ramp → ACH) ) (b) Payers actively migrating their treasury from fiat to stablecoin and renegotiating payment terms on that basis? (c) Some other driver I'm not seeing? And on the receiving side: given that the stablecoin route necessarily incurs an off-ramp cost that the fiat route doesn't, the all-in fee to the Indian merchant should be higher (right?). What's the merchant's strong reason to accept this higher fee in a corridor where Xflow's standard product already works? Is it always demand from a specific payer, or are you seeing other adoption drivers?

@xflowpay The right structure was always clear to us: Stablecoin accepted outside India by a licensed partner → converted to fiat USD offshore → fiat enters India via AD Category I bank → INR settled to Indian business with full documentation Clean and regulated.

Stablecoins payments now live on @Xflowpay. I'll be direct about why we built this. We heard from Indian businesses who lost deals because they couldn't accept stablecoins. Not a compliance grey area problem, but a compliance infrastructure gap.

With all of this chatter around orbital data centers, I have been having this thought since last night about SDT (Space Data Transfer). The general idea is that there has to be substantial innovation in data transfer between satellites to make some of these ideas cost efficient and thereby realistic. Funnily I started thinking of it as gossip protocol on steroids :) For a moment imagine every LEO satellite in a constellation to be a node that contains structured information (for ex orbital direction, closeness to ground terminal, heat information etc) about its nearest neighbours. This allows for efficient packet forwarding from one LEO node to the next. The structured information will be periodically updated from ground terminals. I will do a more structured write up on this in the coming week.

If you parse this, it soon becomes evident that such an asset class brought to the public in a regulatory first manner becomes a treasure trove of information and truth in society. At Kresper (kresper.com) we are making this ambitious attempt.

Prediction markets are an interesting asset class. Fundamentally it is a derivative contract where the underlying is not a priced instrument (like an index option or a standard commodity future) . The underlying is an “event”, the outcome of which determines the market price.

When Andres Freund, Linux kernel contributor & Microsoft engineer was debugging slow SSH logins on his Debian machine in March 2024, he noticed something weird: liblzma (part of XZ Utils) was using way too much CPU power, so he kept digging, and what he uncovered was a multi-year supply-chain attack! An attacker using the name “Jia Tan” had spent two years slowly infiltrating the tiny XZ Utils project, a compression library used by virtually every major Linux distribution. The backdoor wasn’t in the source code. It was hidden deep inside the build scripts. It would have given the attacker remote root access on millions of servers the moment a specially crafted SSH key was used. Freund caught it days before it would have shipped in Debian, Fedora, Ubuntu and more. One man, one anomaly, one routine debug session saved the internet from a potential catastrophe. Respect!

@thenarrator In addition to what you said I think the capital efficiency problem needs to be solved for the broad market maker interest. Some innovative Margin techniques need to be brought in instead of full cash collateralization. I am attempting a regulated play in Europe.

prediction markets need to rethinking what market makers actually get in return for taking this structural risk and potentially rewarding MMs with something beyond fees if the industry solves market making incentives, everything else follows

Mostly time for complex work is unknown and veiled with uncertainties. Driven individuals create a timeline to accomplish complex workstreams and communicate this abstraction to the general public since humanity fears ambiguity and uncertainty. #musingssayanrdotcom

The outcome need not be a successful one but it should be a one that influences the invocation of the subsequent outcome. Humans need discrete slicing of the time continuum to interpret progress.

The sum total of all ups and downs in smaller time intervals seen on a larger timescale is deemed as progress. Hence it is important to reach an outcome in those smaller time intervals.

It was a tectonic shift in the way of thinking about things. Maybe there will come a time when an inquisitive mind will craft a new interpretive framework to reason matrix multiplications. Till then the hard work to understand, tinker and build continues.

Back in the day Newton was able to elegantly articulate the effect of gravitation through math and experiments but the cause remained elusive. Einstein uncovered deeper secrets of our universe by trying to reason the cause.

I have consumed a ton of elaborate ,extremely impressive youtube videos of implementing nano LLMs to understand under the hood. Amidst all of this there is an interesting analogy that keeps popping in my head. LLMs seem like “pre Einstein gravitational understanding”.

How long do you think HDFC Bank will take to transfer the money from your account to your nominee's after you die? Public sector banks get a bad rap, but it's been more than a year and we have not been able to get funds transferred from my now deceased husband's account in the country's largest private bank. The nominee is our adult daughter, there are no conflicting claims, the paperwork is clear, and yet for more than a year now HDFC has been giving us the run around. Get this paper, get that paper. Each time my daughter comes home for holidays, we have to take 2 days out to go and wait at the bank branch. Today, more than a year after the first application, they said a) we can't find the paperwork including the death certificate you submitted because the employee you submitted to is not here b) you have to get a stamp paper and notarise it for the auto debits that hit his account between the date of his death and the date the account was frozen (on the basis of the paperwork in point a, now non-traceable) c) you have to get HDFC Standard life to sign off because there is an auto debit for an investment from his account. (ME: they are the payee, he is the payer, why does the payee need to approve stopping a standing instruction? Bank: Hamare main aise hi hai). If you wait for 30 minutes, we'll get you the form. d) after 30 minutes. Sorry, we can't get the form, you have to go to their office and do it. By now it is 2.45 PM, HDFC Standard life office is at least 35 minutes away. They are open only till 3.30 PM. We are not sure we'll make it there on time. I call their number to ask if the branch is open tomorrow. Person: Sorry, I can't give you the information today, you will have to call and check tomorrow. (To repeat: HDFC Standard Life's call centre cannot tell you if a branch is working or not the following day) In all, we have spent many hours over many months trying to get this money in a case the bank officials themselves say "should not be complicated". What is worse is for my 20-year-old daughter to relive the trauma of losing her father each time we have to go and explain the case. Just today, we had to speak to three different people and start the story from the beginning. This is not just the bank's inefficiency, it is its utter callousness. (My husband also had money in Indusind Bank. It took one visit and two weeks for that to be transferred). The other, more important point, is that I have a job, we have the money to continue to pay for our needs without having to immediately access these funds. But I am certain that is not the case for many others. What about them? How can this be the system of a bank whose customers are dependent on it for giving them access to their own money?! It beggars belief. @HDFC_Bank @HDFCBank_Cares

@UIDAI @ceo_uidai @UIDAIBengaluru @UIDAI Have DM'd the details. Please check.

@UIDAI @ceo_uidai @UIDAIBengaluru Have DM'd the details. Please check.

@UIDAI @ceo_uidai @UIDAIBengaluru it's quite appalling. I was trying to log into the aadhar portal to check out the offline ekyc XML and to my surprise my aadhar is in a deceased status.