Episode 88 · Fintech · 32 min

Coordination, not currency

JediSwap is a decentralised exchange on StarkNet with 100,000 people in its Discord and not one employee anywhere. The founder's case is that crypto's real payload was never the token: it is a points ledger nobody can quietly edit, capable of paying the ten-thousandth person in a workflow chain for work no legal contract can see.

RG
Rohit Goyal
Founder, JediSwap · with Vishal Krishna
Coordination, not currency — episode thumbnail
31:55
Said in this episode
▶ 1:20
0
Employees in the JediSwap organisation
It is run as a DAO — a permissionless organisation where contributors coordinate on incentives rather than the obligation of legal paperwork.
▶ 17:29
100,000
People in JediSwap's Discord
Of those, five to ten thousand ever discuss JediSwap and roughly 500 put in meaningful time — as little as one to five hours a month.
▶ 16:43
100K → 5M
DeFi users, 2018-20 to today
Goyal's own on-air estimate, hedged as 'we might have'; he sets it against first using the internet himself around 2007, well after the West.
▶ 19:34
$10 on $100
Ethereum fee that killed the yield product
Paying ten dollars in gas to deposit a hundred made the low-risk tier pointless, and sent the team looking for layer-two scaling.
▶ 8:47
4-5M
Users on his pre-crypto mobile games
Hyper-casual, roughly one megabyte, low data consumption — built with no VC money and no full-time employees, just people who wanted to do it.
The brief

The argument in sixty seconds

Goyal's claim is that crypto was mis-sold. The interesting artefact was never the currency but the coordination: he read Satoshi's white paper and Sapiens in the same 2017 stretch, decided that one had built trust and the other had described how humans organise, and recognised what he had already been doing since 2010 — assembling companies out of friends who shared an imagination rather than a payroll. He runs the lens over a retail chain where 10,000 people do nothing but put a price on a package. Give each step a point, let the next person approve it, write the points somewhere nobody can quietly change, and you get a percentile ranking of who actually did the work plus a way to pay them without a contract. JediSwap is that thesis run as an organisation: a decentralised exchange on StarkNet, a Discord of 100,000, roughly 500 people who put in an hour or five a month, and zero employees. The route there cost him three bear-market years at Mudrex — a crypto investing platform he took into Y Combinator — before the end-of-2020 admission that he had built the centralised thing he came to replace. A yield product with three buttons, low, medium and high risk, then died on Ethereum gas, because paying $10 to deposit $100 is not a product. The stakes he sets: while crypto reads as tokens and Ponzi schemes, the coordination layer stays unbuilt — and the five people who can out-build 300 keep getting paid like employees.

Worth your time if you are

Corporate leaders who still file web3 under Ponzi scheme
Founders designing incentives instead of job descriptions
Engineers weighing a DAO contributor role against a payroll
Twenty-year-olds tempted to trade crypto rather than build it
Episode map

Where the conversation travels

Every block is a chapter, coloured by what it's about. Click any of it to jump straight to that minute on YouTube.

01Cold open: a DEX and a DAO 0:00 Recorded at the BUILD web3 summit, the Jedi jokes give way to a two-part definition — JediSwap is a decentralised exchange on the StarkNet layer two, and a permissionless organisation with not a single employee, coordinating on incentives instead of legal paperwork. 02Ten thousand people pricing a package 1:38 Asked for a use case a CEO would recognise, Goyal builds one out of a retail chain where 10,000 people do a single standardised step, and shows how a point awarded at each handoff and approved by the next person produces a percentile view of who actually did the work. 03Points on a ledger nobody can edit 4:26 Attach incentives to those points by writing them to a smart contract — in his phrasing, a database nobody can change, with visible rules about who may change what — and the contract, he argues, stops being necessary because the record is there to see. 04Vegetables in 2010, four million gamers 6:12 The backstory: engineering at IIT Bombay finishing 2012, companies built from his second year onward with no VC in sight, a fruit-and-vegetable venture two years before BigBasket, then hyper-casual mobile games scaled to four or five million users with no funding and no full-time staff. 05Satoshi and Sapiens in the same year 9:20 After hearing about Bitcoin twice on price spikes, he finally read the white paper in 2017 alongside Sapiens — currency as trust on one side, human coordination on the other — and concluded that coordination was what he had been practising all along. 06Mudrex, YC, and the centralisation itch 10:51 To get more people into crypto he co-founded an investing platform with three college friends, took it into Y Combinator as what he believes was the first Indian crypto company there, then spent three bear-market years before asking at the end of 2020 why he was running a centralised company. 07Mesh Finance: community before product 12:22 DeFi's habit of distributing tokens to early users returned him to the incentive problem, and with only one to two lakh people using DeFi he started Mesh Finance as a community and a blog post rather than a product. 08Lehman, printed money and a neutral currency 13:39 The plain-English detour: 2008 leverage, bailed-out banks and wiped-out American pensions, the argument that printing money decides who gets rich and who gets poor, and DeFi as the removal of the central intermediary at each layer of currency, banking, exchange and yield. 09100,000 in Discord, 500 who show up 16:43 Adoption tracks the internet's own lag — he first used it around 2007 — and today's honest DAO funnel is 100,000 people in Discord, five to ten thousand who ever discuss JediSwap, and about five hundred who put in an hour or five a month. 10Three buttons, then a gas-fee pivot 18:15 The first product replaced unreadable token tickers with three buttons — low, medium and high risk, with the upside and the downside spelled out — but Ethereum fees made a $100 deposit absurd, pushing the team to StarkNet and, to make it work there, to building a decentralised exchange. 11Humans as a collective computer 20:18 His frame for why transparency pays: humans are computers and an organisation is a collective one, so its output depends on how fast information moves and how much computational bandwidth is shared — illustrated by the startup that rebuilt a 300-person corporate platform with five people. 12The Ponzi problem, and trust as throughput 22:14 The bottleneck is perception — crypto read as tokens and Ponzi schemes — and the throughput objection dissolves into a trust question: a transfer from your brother needs fewer nodes than one from a stranger, so networks fractal into layer twos and threes that older stacks can integrate with rather than be replaced by. 13Kings, corporates and starting at twenty 24:49 Distribution of decision-making is evolution, he argues — you already live better than a king — so while JediSwap the product is not yet relevant to corporates, he is open to their coordination problems, and tells twenty-year-olds to start with friends and a community rather than a trade. 14A brick shop, Dalio and Renaissance 28:32 He traces the instinct not to technology but to a tier-four town and a family construction shop selling bricks, thanks the BUILD organisers, and closes on Ray Dalio, the secrecy of Renaissance Technologies, Outliers and Thinking, Fast and Slow.
Takeaways

Ideas to carry out of this hour

01

Crypto's real product is coordination, not currency

Goyal's central claim is that the token is the least interesting thing crypto produced. Reading the Bitcoin white paper against Sapiens in 2017, he took Satoshi to have built trust and Harari to have explained that humans win by coordinating — which is what he had been doing since 2010, assembling companies from people who shared an imagination rather than a salary. The reframe is the whole argument: as long as crypto is understood as tokens and bad actors running Ponzi schemes, the coordination layer nobody has built stays unbuilt.

02

A points ledger prices work a contract cannot see

His worked example is deliberately unglamorous: a retail or manufacturing chain where 10,000 people do one tightly specified step, each handing off to a checker. Award a point per completed step, let the next person in the chain approve it, and within a week you know on a percentile basis who did a thousand points of work and who did a hundred. Write those points to a smart contract — a database with visible rules about who may change what, and levels above which nothing can be changed — and the incentive attaches directly to the record, which is why he says the paperwork stops being necessary.

03

The honest number for a DAO is 500, not 100,000

JediSwap has 100,000 people in its Discord, but Goyal breaks the funnel down without flattering it: five to ten thousand ever talk about JediSwap at all, and roughly five hundred invest meaningful time — which he defines as low as an hour, or five hours, in a month. That is the working population of an organisation with zero employees, and the reason he insists on community first and product second: Mesh Finance began as a telegram group and a blog post, and the people who joined it were the asset.

04

He spent three years building the thing he came to replace

Mudrex was the sensible move — get more people into crypto by helping them invest in it — and it took three college friends into Y Combinator as, he believes, the first Indian crypto company to get in. Then came three bear-market years to 2020 and a harder audit: he was running a centralised company where everyone deposited their money, which was not what he came for. His verdict on those years is unusually blunt for a founder on camera: he lost the progress on his actual problem statement while raising funds and hiring people.

05

Gas fees killed the product before the market did

The first Mesh Finance product attacked comprehension: instead of tickers no human can parse, three buttons — low, medium, high risk — each showing both the yield you might earn and the share you might lose. It shipped on Ethereum and the economics collapsed, because paying $10 in fees to deposit $100 makes the low-risk tier meaningless. That single unit-economics failure drove the search for layer twos, the move to StarkNet, and the discovery that they would first have to build the exchange underneath — which is how a yield app became JediSwap.

06

Trust, not throughput, decides which chain you need

Asked the standard enterprise objection — blockchains are too slow, and we have already bought the cloud — Goyal reframes throughput as a trust budget. A transfer from your brother needs no settlement guarantees; the same $10 from a stranger might need Ethereum's hundreds of thousands of nodes. Networks therefore fractal: some work belongs on one server, some on five, some on a base chain, which is what layer twos and threes are for, and why he says the old stack can integrate the new rather than be replaced by it.

07

If five people can out-build 300, pay them like it

Vishal offers the case of a health platform a large tech corporate staffed with 300 people that a startup rebuilt with five, and Goyal takes the implication further than the efficiency point: the five should be rewarded far more than the 300, and the network gets better still if the five and the 300 share what they know. His model is that an organisation is a collective computer whose output depends on how fast information travels and how much computational bandwidth is pooled — which makes transparency a performance argument, not an ethical one.

The numbers, drawn

What the episode measures

Every figure below was said on air — timestamps included, caveats kept.

Conversation share

portion of the hour spent on each theme
Payments & fintech · 24%Leadership & org · 20%Founder journey · 18%Product strategy · 13%Data & digitisation · 11%Savings & wealth · 8%
Payments & fintech24%
Leadership & org20%
Founder journey18%
Product strategy13%
Data & digitisation11%
Savings & wealth8%
Computed from the chapter map of this episode.

What 100,000 Discord members actually amount to

people
In the Discord100,000Ever discuss JediSwa10,000Meaningful contribut500
As stated in conversation: 100,000 on Discord, 'in the range of 5 to 10,000' who talk about JediSwap (upper bound shown), and about 500 who invest meaningful time — an hour or five hours in a month.▶ 17:29

DeFi's user base, by the guest's own count

users
02,500,0005,000,0002018-20At tapin
Goyal's approximate figures, hedged on air as 'we might have 5 million'; the earlier number is given variously as one lakh to two lakh users. Not independently sourced in the conversation.▶ 16:43
Worth keeping

Lines that stay

It's a DAO — a permissionless organisation, which means we do not have a single employee in the whole organisation. We are creating a framework of coordination where you coordinate based on incentives instead of legal paperwork, where you have obligation.

— Rohit Goyal ▶ 1:20

Satoshi was creating a currency, and currency is just trust. Sapiens was talking about coordination — and I realised that throughout my life I was just doing coordination. I didn't raise money, I didn't pay anyone, but we were coordinating.

— Rohit Goyal ▶ 10:04

I try to think about humans as computers. All of us are computers, and we build as a collective computer.

— Rohit Goyal ▶ 20:40

The whole vision of what crypto can push for is not understood — it's limited to tokens, sometimes bad actors doing Ponzi schemes. If you start thinking about crypto as a coordination tool, the whole thing changes.

— Rohit Goyal ▶ 22:30

I want to work on what I like doing, at the time I prefer, from a place I like. All of those are freedoms I want to choose — and everyone wants the same thing.

— Rohit Goyal ▶ 27:20
Clips that travel

Short on time? Start here

Operators who suspect their org chart is a coordination tax

The 10,000 people who price a package

The full worked example — a point per handoff, approval by the next person, percentile rankings, and a ledger nobody can quietly edit.

2:09 → 6:12 · 4 min ▶ Watch clip
Founders who suspect they built the wrong thing

Satoshi, Sapiens, and quitting his own company

The two-book coincidence of 2017, an Indian crypto startup into Y Combinator, and the end-of-2020 admission that he had built a centralised company.

9:20 → 12:38 · 3 min ▶ Watch clip
CXOs who want DeFi explained without a single token

Why 2008 explains decentralised finance

Lehman, leverage and bailouts, the argument that printing money picks winners, and the four-layer stack DeFi is trying to disintermediate.

13:39 → 16:43 · 3 min ▶ Watch clip
Anyone running a community as a growth channel

100,000 in Discord, 500 who actually show up

An unflattering contributor funnel, the three-button risk product, and the $10 gas fee that forced the move to StarkNet.

16:43 → 20:18 · 4 min ▶ Watch clip
Enterprise architects weighing blockchain throughput

Trust decides how many servers you need

The brother-versus-stranger test for settlement, fractal layer twos and threes, and the case that old stacks integrate rather than get replaced.

23:02 → 25:21 · 2 min ▶ Watch clip
Glossary

The jargon, unpacked

DEX (decentralised exchange)
A venue where you swap one asset for another without ever handing custody to an operator — Goyal's framing: at no point does he control your assets.
DAO
A permissionless organisation run on incentives rather than employment contracts; JediSwap's version has 100,000 Discord members and zero employees.
StarkNet
The layer-two network JediSwap is built on — a chain that settles to Ethereum while keeping fees low enough for a $100 deposit to make sense.
DeFi
Decentralised finance: rebuilding currency, lending, exchange and yield without the central intermediary who can lose your money at each layer.
Smart contract
In his 'noob language' definition, a database nobody can unilaterally change, with visible rules about who is permitted to change or read what.
Gas fees
The per-transaction cost of using a blockchain; on Ethereum these were high enough that a $10 fee on a $100 deposit killed his first product.
Yield protocol
A platform that pays a return on deposited crypto — the category Mesh Finance tried to make legible with three risk buttons instead of token tickers.
Connections

If this resonated, go here next

Full transcript

The whole conversation, searchable

125 segments

Auto-generated captions, lightly cleaned. Click a timestamp to open that moment on YouTube.