←
Joel John12,592 posts
Joel John avatar

Joel John

@joeljohn

building - @decentralisedco waiting on @jarapphq to ipo Currently obsessing about tokenisation, agentic payments and institutions

◷ Joined December 2014▧ 1,440 Photos and videos
Joel John avatar
Joel John@joeljohn·24 Sep 2024

Had a little over 75 in-person meetings last week in Singapore. Crypto, in its current stage, rhymes equal parts with an MLM industry and the future of the web. Here is the distilled, TL:DR version of what I learned in Singapore, mixed with photos as is tradition. 1. Why Events There were 700 side events because it's the easiest way for a CMO to justify the job. In-person events are measured with footfall, instead of on-chain value creation (which is harder). Signaling linkages to others in the industry, helps close deals. Most importantly, free food and booze is the oldest trick in the game. They work. If anything, it's a sign of broken incentives in the industry. The current pipeline for capital looks like this: Fund of funds -> VCs -> Founders - > CMOs -> Night clubs. I think it really highlights what's broken with distribution and incentives. Or its globalisation. Not my monkey, not my circus. 2. Token Comps Tokens being down 70-80% on average has a dual sided impact on employee morale. People feel burnt out, and over-worked. Crypto's original promise was exponential upside, for limited risk. Last year's token performance flipped the promise. It's boundless work, for declining upside. For some ecosystems (like SOL), the story is different due to price action (and focus on community). You can see how this variance in comp affects decision making and morale within teams. Some teams, thanks to where the price is, have a disproportionate edge over others. I think there's a loop - where lower prices, lead to worse off decisions, which in turn leads to worse off decisions. 3. VCs, DPIs and tokens Most VC funds are either raising, or set to raise. It makes me wonder where all the idle capital from a few years went back. The alt-coin sell pressure from the past few quarters are likely funds selling off alts that are vested to produce returned capital before raising a new fund. Will they succeed in raising new funds? Yes. Will that new raised funds go to existing tokens or themes? Probably not. I think a lot of VCs are burnt from the past two years and will (i) suppress valuations or (ii) push for earlier liquidity. You can see this play out with consumer app valuations. Few biters. For the VC model to work - you need multiple tokens, listing at high FDVs, with low floats, to mark-up books, and raise new funds (before vests). That pipeline, relies on eigen, berachain, monad - so there's a lot of eyeballs on it. Unlike last cycle, between perps and OTC, that market has evolved too - so its an interesting time for VC overall. 4. All eyes on consumer One of our own portfolio founders added a million users last quarter. Each day, he's adding another 10-15k users. I don't think it will remain the exception. Consumer apps with founders that are non-token first, will likely blitzscale. The same applies to RWA founders with a fintech first focus. As VCs are token oriented and these founders are outsiders, many of them will remain undervalued (and as pariahs). My honest ask to many of them have been to avoid talking to gatekeepers. I think between MPC wallets, L2s and better on-ramps, we are at the cusp of seeing a new generation of consumer apps that have very little t do with existing token economies. Will they be valued as high? Will they create token returns the same way? Probably not. I think they will need a new class of equity first VCs. Or existing ones will flock to equity, and here's the reason why. 5. Exchanges have reduced listing VC backed bags. This is voodoo-hot air theory. But I think regulatory capture will drastically reduce the number of VC backed tokens that make it to large exchanges (like Binance or Coinbase). You can see versions of this if you look at the number of VC backed tokens making it to exchanges and compare to frequency in past cycles. One theory is that retail capital is not flowing in like it once used to and exchanges have no reason to. The other, is that price discovery in exchanges have been down only - and they are being more picky. Whatever be the case, not being able to mark-up books and raise new funds will have trickle down effects on venture valuations. This is why most funds have a preference of lower valuations (risk premia for illiquidity) or SAFTS (assurance that liquidity will come). I think as listings on exchanges get harder, we will see capital flow through to equity. 6. On token management. A lot of founders seem to be issuing tokens as that is what VCs are backing today. But the same guys will vanish once your FDV is around your series A valuations. If you are a founder - consider hard and strong about issuing tokens. If you do issue a liquid token, IR (or educating markets) about what you do is as important as running the product. There's space emerging for new forms of media that highlights key metrics, and earnings for protocols. This subsection of the market is small today - but as nature of investors focused on liquid markets trend to sophistication and number of tokens that are at low valuations increase, I see more capital flowing towards this segment. 7. The Early adopter-gatekeeper See all these packed events? Talk to most folks attending them, and you'll realise everyone has a clear agenda. Most of the time, its money. Loads of it. And that's fine, its a capitalist economy after all. The problem, is when a subsection has made tremendous wealth by being early and packages the same old thing in new words and continues to do the same things. We somehow presume everything needs a token. Everything needs to be overly complicated. Everything needs millions in burn. The early adopters are rich and can afford to be stupid for years on length. New guys (post 2022 entrants) that haven't made bank, have to lean on efficiency. But in pursuit of efficiency, they seek advice from early adopters, and get lessons on a constant grift. I think this is why our industry struggles to evolve. It rhymes with MLM schemes, because they work the same way. An early adopter wins by virtue of being early. As new entrants arrive, the early adopter continues to win. With each new participant, the amount of return declines. Sooner or later, you will have a desperate large bottom pyramid. If you are in the industry, it helps to wonder - are you desperate, or leading. Are you being efficient, or consuming junk. Just lots of room for introspection there. The way these systems break is when a new entrant ignores the rules and builds a business to scale ignoring the incumbents. My thoughts, prayers and dollars can be behind you if that's you. 8. DePin Data This is specific to data points (like audio, spacial data, map data etc). I think a lot of these will rhyme with P2E as it gives retail a chance to enter the industry and can be packaged like a dream. Spend $50 to make $100 is an easy sell. My concern here is how valuable is this data? My running assumption is valuation fr such data will be at 1000x what they are actually worth. They will be great trades (and worth playing on liquid side) - but the long-tail of data accumulators will be in for a rude awakening. Much of that sector rhymes with gaming guilds at this point. I know this is a sceptical take, but I'm putting it out there for healthy discourse on the value of the data. Some of it will be tremendously valuable, and much of it will be useless. Kind of like the long-tail of BS. Which leads me to my final point. 9. The Long-tail of BS. Every category initially has few players that are pioneers. They attract high valuations. Which in turn accrues talent building in the same category. If you are imitating a business, you are likely in the long-tail of BS category. It works if you have niche or geo-specific moats, but odds are quite low you have those moats. This applies to service providers (like lawyers) as much as it applies to product categories (like GambleFi or Web3 social). The long-tail of BS happens because there's more VCs than thoughtful builders in the room. The long-tail of BS cannot be avoided. It is a thing. It is a beast to be acknowledged. As the industry grows, the long-tail of BS will grow even longer. You can see it in the booths with no product. In the events with no conversations. In the KOL rounds with financial shenanigans. In the gigantic face of a founder with no product slapped on an event entrance. The long tail of BS is everywhere. I think it can't be tamed. It can't be cut. It only has to be acknowledged and waded through. The reason why the long-tail of BS exists (currently) is because founders (and talent) that can meaningfully produce impact are likely in other sectors (like AI). Even funds that deploy at scale, see this tday. There is an evident decline in the quality of individuals. But this is what retail adoption looks like. As the sector scales - it brings with it everything that society is. The good, the bad, the ugly - and well, even FTX. You just need to grow smarter as the long-tail of BS takes over. A lot of the systemic challenges (and opportunities) that exist in crypto are dependent on price. As the fed cuts rates, much of what I said will change. We are quite probably at the early stages of a new cycle. I am not as bearish - but there's a healthy amount of skeptical cynicism in me. We have work to do.

Post attachmentPost attachmentPost attachmentPost attachment
Joel John avatar
Joel John@joeljohn·3h

hypothetically speaking if a guy had to get access to a fleet of b300s for a quarter - who'd be the person to speak to help a man out

Vlad avatar
⟳ Joel John retweeted
Vlad@vladsvitanko·Nov 17

This guy's blog will be worth $17.8 billion in two days > studies math at the University of Toronto for 1 year > drops out at 18 > 2012: Peter Thiel pays him $100K to skip college (!!) > starts writing machine learning explainers on his personal blog > the posts spread through every ML lab on the planet > 2015: Google Brain hires him > 2018: leads a research team at OpenAI > 2021: leaves with 6 colleagues and co-founds Anthropic > Anthropic lists at $1.4 TRILLION in 2 days btw His name is Chris Olah, co-founder of Anthropic Zero degrees, just a good blog that did all the work for him Moral of the story: distribution is your biggest leverage

Post attachment
Joel John avatar
Joel John@joeljohn·11h

There's two trends in software I am quite keen on and spending more time in lately. 1. The abstraction of the UX layer I think the modern web will replumb itself around APIs and short-burst payments (on x402). You see this with data vendors (like @findatasets) and (@MonidHQ) making end-points for all of software accessible via API. The web's original sin was building everything around advertisements. The second order would be built around APIs pulling in information to specific work contexts from where workers/agents will collaborate on a final output. The best instance of this is using Notion - to have a worker agent that pulls in information via Exa/Keenable each morning alongside data (via Monid) on a dashboard. You will soon have people working in their own UX without opening a browser very frequently. The earliest signs of this is in - Notion Agents being used to query / pull information - Chatgpt plugins working directly in google docs - Claude's co-work mode 2. The second one is something I've been spending more time on personally. It is hyper-personal software. The web was built around algorithms predicting user behavior to tweak for retention. Tiktok's algorithm is the best in class at this. These algorithms optimised for attention as ads were the business model. There are a new class of software that embed directly in apps of scale to tweak products in real time for the end-user. They take a much larger context (from app, in-device, from the web) and session times and create a coherent profile of the user across apps. I have been seeing these drive retention and transaction volume much higher (±30%) than usual. I think products that do outcome based pricing in volume for such products will see extreme hockey charts in their revenue. I am interested in this category because tweaking a product in real time for improving consumer behavior can work for good too. For instance, a user could be "nudged" towards more exercise. They could "nudged" towards calling a loved one when in distress. Real-time user context for better outcomes is intriguing. These businesses would not be building software on their own but rather building the decision layer that tweaks products in real time - so the app/distribution layer and the personalisation layer will be two isolated products. Maybe it is my ADHD and general introversion - but I think the web will become more isolated (when it comes to work), more human (in terms of content) and more personal (in terms of software) Naturally agentic payments are a huge necessity to make any of this happen. x402 - is a necessity for data streams / api call based software. Outcome based pricing can become payment streams, which are then competitively bid on by third party financiers as a form of financing the new era of software. That intersection of finance, payments, software and understanding how all of it weaves with culture is very nascent.

John Felix avatar
⟳ Joel John retweeted
John Felix@johnfelix123·Nov 16

So many LPs love to say they underwrite future performance, not past performance. Then they ask for track record in the first 5 min of a meeting and tell every Fund I manager to come back for Fund III.

Joel John avatar
Joel John@joeljohn·Nov 16

Feb 26, 1995. An article from the Newsweek. A VP at DE Shaw named Jeff Bezos was leaving work to sell books online a few months earlier. We call it Amazon today.

Post attachment
Joel John avatar
Joel John@joeljohn·Nov 15

The amount of stupid a person’s ego can underwrite is directly proportional to their tolerance for pain. The thing is, too often, genius looks like stupidity in its early days.

Joel John avatar
Joel John@joeljohn·Nov 15

I think there is a thing as intellectual biome - just like there is gut biome. It is the substrate of niche interests/ideas you entertain, people you converse with and how they mutate within silence. If you consume constantly, you will only have regurgitated junk coming out.

Joel John avatar
Joel John@joeljohn·Nov 14

the lack of anxiety about running out of inference when i use hermes mobile app after stressing for weeks on grok bot is quite liberating reminds me of when I used to pay $6/gb for internet

Joel John avatar
Joel John@joeljohn·Nov 14

Best guess given Bitcoin's movement right now - some fund is deep in the red on AI trade. Bitcoin being the easiest asset to liquidate on a weekend, goes out first. sales has a hint of a fund couping up margin before calls come Monday open.

Joel John avatar
Joel John@joeljohn·Nov 14

The more I think of it - products like @hypurrdash are the end state for crypto native markets if you care of retention 1. Blockchain data/labeling for USP 2. Wide asset mixes 3. Built on third party protocol (hyperliquid) 4. Monetized via aggregated service layer (rfq)

Joel John avatar
Joel John@joeljohn·Nov 13

crypto is in this phase where - the tech works - the TAM for what it can do is 1000x bigger - it is a necessity for identity/payments as AI rewires the world - but its core adopters are depressed because numbers don't go up on their own Intriguing