krrishd

I've only played around a bit with it, but I've enjoyed using https://vercel.com/eve because of how "batteries included" it is on all the glue code/infra -- borderline just write Markdown skills + configure tools/connectors/schedule/etc, and things just work.

In fairness, GitHub is under a corporate umbrella with Office 365

These are audited and regulated by the federal government!

Re: why do we need _so many_, I tend to agree but it's the equivalent of "USD" in the banking system really being a wide range of deposits at different institutions, with different underlying balance sheets (banks fail regularly, and are regularly backstopped by eg. the FDIC!). There is clearing infra being built around this problem (increasing fragmentation of deposits between stablecoins/etc), just like it was a problem for the old world in which clearinghouses emerged

https://www.federalreserve.gov/econres/notes/feds-notes/paym...

This is one recent paper, but with this and other personnel details (eg. Fed governors attending the last few "A Very Stable Conferences", ex-Fed people working at Circle / doing stablecoin startups) you can infer that the Fed is more pro-stablecoin than you think.

Why? Cross-border payments are hard, CBDCs are unpopular / politically infeasible, and Fedwire takes up a ton of balance sheet capacity (etc).

UST and the other ones were "algorithmic" stablecoins collateralized poorly -- USDC and OUSD (eg.) are collateralized under the regulation of the OCC / by USD-equivalent short-term bonds etc.

This is not to say that there can't be liquidity challenges or "black swan" events, but the current iteration of stablecoins are _comparable_ to being "made by banks" and are certainly regulated.

Sure, but there are longer shots than others, where money alone won't make the difference and when conviction in mission is directly related to conviction in founders/leaders.

"subjective and deeply irrational" aren't the exception for how humans organize

https://archive.is/Wkkey

Certainly you can - but this study appears to use motivations around power and status as its heuristic for narcissism!

Even if one were to grant the conclusion — leaders resist remote work to preserve their "power and status" — to what extent is the magnitude and success of a company (or for that matter, any serious enterprise) a direct function of its leader's ability to:

- Exercise authority / power

- Maintain status within the hierarchy

As in - who wants to work for a leader who is neither powerful nor high-status within their own company? Who consciously chooses a leader who is neither effective in getting people to do the right things, nor effective in commanding a (somewhat faith-based) trust in their long-term vision?

The study feels extremely leading in its idea of what a "good" leader would look like (presumably "hands off," leaves everyone alone such that good outcomes simple "emerge", etc) -- while treating this bent as obvious truth.

I say this as someone who spent the last 6 years straight working remotely (also having been successful in contributing impact).

Reminds me of this tweet thread from Emmett Shear (cofounder of Twitch):

"I used to struggle w needing to be “creative” or “original” in my work. At some point I had a breakthrough that really helped me: I cannot repeat an idea, no matter how basic or common, without imparting some of my worldview into it.

Even by choosing which basic ideas to amplify, I impart some small amount of myself into each output. It’s literally impossible for a given tweet to be “unoriginal”. Equally impossible to be Truly Original too of course, since you’re always remixing others thoughts.

This POV does put a premium on cultivating and developing one’s worldview, since that is the underlying originality simmering under the surface of each “basic” thought. The best writing is rewriting, including of other people’s words, and the lens is your whole mind."

from: https://x.com/eshear/status/1539393474612498434

This is cool - feels great that there's a growing long-term incentive to implement open SDKs and APIs, for the sake of agent-forward prosumers, in the spirit of earlier internet stuff that used to do it but stopped.

Natural | SWE (Product, Data, Design, Infra, Payments), Product Designer, GTM, more | San Francisco, CA (Onsite)

We're building payment rails for AI agents; fully-owned rails across traditional banking and stablecoins, and the full set of funds flows that AI agents need to transact in the real world with businesses, consumers, and other agents (wallets, pay, request, credit, billing, and more).

We've raised from some of the most well-known operators in fintech & payments to shape how agents interact with the economy, from scratch.

https://natural.co if you'd like to learn more - email krish@natural.co if interested!

Pocket operator (particularly the KO) is the single thing other than maybe the OP-Z that I think has enough creative utility relative to its price to not feel like glorified furniture.

I enjoyed mine a lot / made a lot with it!

Then you haven't talked to the PMs I've worked with :)

jk, agree with your point

I agree with the premise, but it's quite painful in practice constantly probing and prodding for justification and explanation -- especially because _even with_ the justification, explanation, etc, one's mental map / the "topology" of the thing being built is only very loosely being populated as a result of the conversation. I say this having continuously tried to find a way to keep my learning rate comparable to if I was writing the code myself, and having somewhat failed.

I'm starting to wonder if the thing to address is the anxiety itself rather than the "fuzziness about the code" that creates the anxiety - and more explicitly model myself as an engineering and/or product manager counterpart to these things. I wonder how non-IC EMs/PMs do it - it seems maybe fundamentally anxiety-inducing? – but they _do_ do this already (tolerate the fact that the underlying technical system is not fully within their grasp).

As a hobbyist and indie hacker, love it. As a professional maintainer - it made a dry job dryer, faster.

I was sort of hoping this would be a year-to-date visualization similar to Github profile contribution graphs...

Ramp is (mostly) a Flask monolith with some sprinkles of Elixir at the very edges where sub-second performance matters.

In case you’re curious why this might make sense: older HN thread/post on why you’d issue a custom stablecoin: https://news.ycombinator.com/item?id=45237035

Previous HN thread / article re: _why_ you'd issue your own stablecoin: https://news.ycombinator.com/item?id=45237035

TL;DR in this announcement seems to match: more control over rewards/yield (both in entitlement _to_it, and in the ability to customize the portfolio that generates it).

Agreeing with your immediate hunch that the traditional rails are sufficient for agentic commerce, but still finding the x402/stablecoin etc work fascinating:

Under which circumstances are the traditional rails actually off limits? Some (trivial) speculation being:

- When the beneficiary is not a traditional business but itself some sort of protocol / contract (that benefits from or even requires final settlement into eg. a stablecoin wallet, eg. to pipe those funds into downstream dependencies) - nothing with mainstream PMF here yet, but feels like the "trivial" case for those rails

- In cases where "broad commercial relationships" around low-cost txns aren't feasible -> credit risk can't be pooled, there isn't credit infra, etc -> market never clears. Global micro-payments to regions with no local acquiring/robust banking? Something else?

(yes, very guilty of solution in search of a problem thinking :))

why people who actually like crypto would want paypal

this snippet is everything: "to PayPal, Venmo, as well a rapidly growing number of digital wallets across the world that support crypto and stablecoins"

this is effectively PayPal taking its "closed-loop" payment network, and opening it up to any wallet capable of receiving crypto/stablecoins - which is still a big deal.

your counterparty no longer has to have a PayPal account for you to pay them via PayPal - they can have any crypto wallet and get paid by you - which is in line with much of the crypto vision around global interoperability/payment acceptance/etc. you could compare to Visa/card acceptance as another global payment rail - but the difference here is closer to the difference between global card payments (easy) and global bank transfers (hard)

I see, I misread: that’s interesting. I would assume the issuer would still be liable to resolve the backing, but yeah I could see how that poses systemic risk.

I also don’t think such a risk could realistically remain hidden - this is still going to be heavily regulated and audited, and industry will wise up to the sorts of risk that emerge.

unless such a product is much better than what banks offer, but that doesn't seem to be the case.

I think you're basically correct here. I think the fear of the banks - and why they are insistent on prohibiting stablecoins from generating yield/interest (via the GENIUS act) - is that that doesn't stay true in the long-term, as stablecoins ascend as a cross-border payment/storage rail.

Does USDC pay interest to the holder or do I have to make a USDC deposit at Coinbase in order to get paid interest?

I believe USDC from Coinbase is framed as "reward", and is downstream of an agreement Coinbase has with Circle to get that "reward" from Circle for all USDC deposits it holds on platform. Other "rates" you can get on centralized stablecoins tend to be similar AFAICT.

Now, why would anyone buy a branded stable coin that explicitly doesn't promise a return?

In practice, you're not even buying these (or at least - that is not the presentation). What you're actually doing is making a deposit into a "stablecoin" account at a place like Stripe (who now offers a Stripe Stablecoin Account, denominated in the USDB custom stable), Slash.com, Dakota.xyz, etc. IIRC Mercury is also a design partner of Stripe's blockchain.

When you make that deposit - either from your regular bank account via ACH/wire, or via USDC - it settles into the account as the branded stablecoin. When you send funds out - you're either sending as fiat or as USDC.

In short - you're not proactively "buying" the coin, and in fact - Stripe describes [0] the USDB coin as closed-loop & "not for public sale", and I think the others are the same. You're just depositing your funds into a platform, in order to use them on-platform - and the platform is holding them as a "custom stablecoin."

[0] https://docs.stripe.com/crypto/stablecoin-financial-accounts...

I don't see why not - I'm sure the banks (or others more expert than me) would argue for stablecoins being somehow distinct in this regard, but yeah don't know why eg. Vanguard wouldn't also be a credible cause of deposit flight.

(I do vaguely remember reading that banks were concerned about people moving to money-market fund products that had bank-like functionality)

Assume the buyer/seller holds capital from sources that the majority of the market considers “illicit” and/or is legally sanctioned and/or physically frozen or restricted

This is not feasible legally, and is where your claim falls apart.

From the now-passed GENIUS act [0] which regulates the stablecoin issuer:

- "Permitted payment stablecoin issuers must maintain reserves backing outstanding payment stablecoins on at least a one-to-one basis, consisting only of certain specified assets, including US dollars and short-term Treasuries."

[0] https://www.lw.com/en/insights/the-genius-act-of-2025-stable...

I think there is plenty of counter-evidence in how this is being approached:

- The obvious: these are stablecoins, whose value is pegged to and 1:1 backed by fiat currency / is not capable of the cliche pump&dump dynamics of other crypto tokens.

- To the extent that (eg. today) they are coupled to a network like ETH or Solana (whose holders stand to gain) - both Stripe and Circle are building L1 blockchains right now whose native gas tokens are stablecoins, and are therefore also decoupled from any of the bagholder stuff. Merits of those chains aside: the big players want to further eliminate that dynamic and are putting their money where their mouth is.

- Stripe (and other legitimate fintechs) want to use stablecoins specifically because they legitimately make cross-border payments much easier, and because there is serious/earnest usage emerging. SWIFT actually does suck (not just to the cliche engineer-who-wants-better-APIs way, but even a banking professional would tell you), international payments are more unsolved than you think outside of a few fintechs who are basically just managing massive ledgers + a ton of liquidity around the world.

(In short: I think your take is something that may have made more sense 5-10 years ago, when Stripe themselves ditched crypto for the reasons that it didn't work for anything useful and was primarily a means of gambling)

USDC gets me 4% on Coinbase, and USDB and other Bridge-issued custom stablecoins also give the customer rewards that they can pass onto the holder (thanks to MMF/similar cash equivalents behind the scenes etc).

But yes - this is why banks want to prevent stablecoin issuers from being allowed to grant rewards

This was also where I initially landed after finding out that the custom stablecoin could not leave my Bridge instance.

I think the role that crypto plays in enabling this is as a neutral, credible storage layer on which this token can be held, that is not my Postgres database as (eg.) Bridge - these tokens still are actual ERC-20s/etc that are present on-chain, as are the wallets that hold them -- but yeah, I'm:

- not sure how instrumental that actually is here

- not sure if that's just incidentally the easiest structure for Bridge, whose primary business revolves around facilitating payments via stablecoin (now, as a part of Stripe)