if you’re building a crypto app: revenues: memecoins >> perps >> prediction markets >> defi vc-hype: prediction markets >> perps >> defi ~ memecoins CT-appeal: memecoins >> perps >> defi >> prediction markets normie-appeal: prediction markets ~ memecoins >> perps >> defi or you can just build a ‘trade any markets’ app like everyone :)

if you’re like me thinking why retail still trades so much memecoins or perps but not doing much holy DeFi here’s why: 1) DeFi in ‘26 is a minefield: you’ve protocols blowing up left & right: - some due to opsec (eg. drift, kelp) - economic risks (eg. neutrl) we're seeing 2-3 small protocols blow up ~every week now 2) assume you take the risks: but then also you’ve p low yields blue-chip assets at top protocols like aave, morpho hardly give you 5-6% apy at max eg. there's $250m at kamino just yielding 2.6% apy??! ofc, whales or foundation treasuries can’t off-ramp and have no choice but to do TVL deals & put in this DeFi protocols imagine this: if you’re a retail user who has $5k left will you put in a random protocol for 15% apr knowing your money can blow-up anytime or simply buy memecoins or even gamble in prediction markets in hope to 10x or go 0 casino math makes latter far more lucrative 3) old nyc uncs already knows this trend: hence they declared crypto ‘institutional’ - slapping the clarity act and made the narrative that crypto is just a backend for wall street to push their ‘governance tokens’ it’s good to see crypto’s institutional growth but trying to kill what makes crypto exciting is as sinful as ‘memecoin pump & dump’ both can def. co-exist 4) don’t get me wrong: i love DeFi and love nerdy financial engineering a ton but even as a pro DeFi user of ‘23 (when i’d have used 100+ protocols) putting a huge chunk of my portfolio isn’t good from a r/r perspective now ofc in ‘23 era, you’d farm airdrops, get a $1b+ tge and that’d make it worth it but now most vc-backed token launches are awful to make it farm-worthy and it's same for all my crypto friends i still get nerdsnipped by good DeFi mechanics & exotic assets or interesting sources of yields and still put couple of Ks in them (imo: if you still find an underrated DeFi protocol which can have a good narrative later - you should still deposit who knows it gives you unexpected airdrops like uni/jto) 5) crypto or even stocks is now fast finance: active investing is roaring back exponentially people wanna trade and rotate fast (either memes and stocks) they don’t have to worry much about blowups or their capital getting stuck anywhere (they've stablecoins or bank accounts to park money) as a builder: stop optimising for TVL, DeFi is not just TVL - you can optimise for just volumes or ‘fast finance’ and make tons of money and still perform finest financial engineering 6) tokens are beautiful: do any financial engineering around tokens (eg. memecoins paired with stocks and using creator fees mechanics) and you'll see the true power of crypto composability eg. a token on solana is instantly available at launch to 100s of terminals, wallets, apps etc i.e so you get instant distribution and leverage network effects tldr: long financial engineering, long fast finance!

i think DePIN will make a BIG comeback this cycle - some thoughts: 1) '23-24 depin was structurally broken: i think conceptually depin was an excellent idea: using crypto to form economic networks the idea was simple: you use tokens as a way to solve cold-start problem i.e bootstrap supply → then you aggregate demand → eventually form a network around it (eg. bitcoin is the first proof of concept of depin network) but there were issues: – high fdv, low float: eg. vc-backed projects typically launched at $1b fdv & 10% float and now you've 90% of supply overhang waiting to be dumped by VCs and suppliers + suppliers have costs in fiat - so it automatically becomes the sell pressure and VCs needs to exit too – revenues didn't accrue to network: some projects (eg. grass or compute ones) did have some pmf but the value accrue to network hasn't been clear essentially, all depin projects were designed in a way to have only sell pressure from high fdv - with no upside for speculators and without speculation, the concept of depin falls apart + regulation was a huge hindrance (or excuse) for projects to have their depin token only as a 'governance' token and prevent any true financial engineering of a network (despite this some tokens like tao, render, akt, geod, grass, vvv are still doing well - showing market appetite) so, what changes now? 2) new on-chain mechanics: in '24, memecoins/pump/metadao gave birth to a new mechanism: i.e anyone can easily launch and bootstrap capital now, - as a project: you don't need to have fancy VCs, tier-1 cex listings, $1b fdv to launch a network token - as tokenholder: you don't have to wait for supply dumps, buyback announcements and a foundation a project can launch at low fdv, have buybacks or token utility coded at network level we've now much mature on-chain markets for best dynamic pricing for marketplaces: → creators or LP fees or transfer taxes: where you can translate speculative energy to a network fees you can also use this rewards for bootstrapping supply or demand - leading to a token flywheel (without directly dumping tokens) → propAMMs allow you to price anything on-chain with external reference price (eg. energy index) or you also have tokens paired with compute indexes (instead of sol or usdc) → upgraded payments infra (like stablecoins, wallets, on/off-ramps, x402) to power the payment mechanics of network + all on-chain financial engineering can 10x the network effects of typical marketplaces 3) new world co-ordination problems: ai has led to a whole supply-chain disruption from mining to energy to compute to inference internet era lead to us to tons of $100b marketplaces (eg. airbnb, uber) and there'll be need for much more marketplaces for ai era eg. there'll be incredible demand for energy - and we'll need tons of niche marketplaces to aggregate demand & supply even OpenRouter is an example of an inference marketplace lot of '23-24 depins tried to recreate decentralised versions of existing successful networks (eg. mapping, bandwidth) - but that didn't work as there is not much new supply or demand to aggregate but as the world develops new problems around coordinating energy, mining, compute etc - i think a lot of room for experimentation opens up 4) long markets, long token marketplaces! the market's risk appetite has never been higher. whenever we see something slightly novel appears (eg. memecoins paired with their stocks) - you get good bids world's speculation is at its all-time high (eg. stock markets) - particularly, retail wanna bid anything they can understand you can now translate speculative energy of markets to solve co-ordination problems via financial incentives + regulations allow you to be hugely expressive on-chain even, venice/vvv is the only way to get exposure to 'inference' on-chain - hence gets a good premium despite having seperate equity value 5) what's next? while obviously most of these network (like energy/compute) would be controlled by huge companies but there'll be many gaps unserved - and the token marketplaces can start from niche and eventually grow as the token keeps getting attention and the network itself becomes more valuable as long as you're commoditizing or aggregating supply and demand, you can form a marketplace around almost anything and you can keep adding people to the network (+ keep upgrading the network) and having some mechanism baked in: where network usage automatically benefits the token at a fundamental level now, you don't even need trad VCs (and whole supply dump) and as soon as you flip the model and make it more bottoms-up, where the earliest consumers or speculators receive the majority of the upside - they have an incentive to propagate the network eg. imagine an openrouter launching at a $10m every inference consumer now has an incentive to share it with friends and bring more people into the network - and bringing attention to the network also benefits the token (ofc, it'll again break if projects launch at v high fdvs as there isn't much asymmetrical upside left to propagate the network) ultimately, it's about figuring out a good mechanism and ofc you'll need a strong team that can execute on both sides of the marketplace: supply and demand and achieve PMF overall, i remain cautiously optimistic for depin or 'token-led marketplaces' to make a comeback in a bigger and different form factors keep looking at solana/base/robinhood plays and ofc, tons will fail - but i'm all here for new experimentations and will bid some good ones! 🫡

personalisation will be the biggest story in finance & trading - for the next 5 years: 1) why? - today, finance lacks personalization because txns are mostly context-free eg. a credit card txn is just: amount + time + counterparty/merchant (too little signal for true personalisation) - one can argue merchant/category helps, but it’s still shallow eg. the same $500 nike txn could mean marathon training, flipping a drop, or buying your girl a gift. amount + category can’t distinguish them and the data is expensive & available only to select hedge funds who use it to predict consumer trends 2) so, how to solve? - many fintechs tried in 2010s to build “personal finance” or even “robo-advisor” apps but they all failed as txns hardly had any context - they all gave generic market advice (based on hardcoded user risk preferences) while social media (tiktok/insta/X) got hugely personalised as content already has context built-in and you can have profiling based on that context - and that led to whole dopamine loops (initially caption & hashtags were used as context) - real personalization comes from ‘why’ the txn happened, not just ‘what’ happened if you think every trade or payment has a context i.e there’s always a story (eg. leo’s salp longed mu because they believe in memory shortage story or i bought this black casio to signal lowkey cool) 3) why now? - finance (both payments & trading) is becoming increasingly a socio-cultural phenomenon eg. streamers or tiktoks influence majority of our spending or trading decisions + we like to share every trade/purchase with our friends - ai agents can natively attach intent, goals, and context to every txn and ai is increasingly getting better at ‘personalising’ once it has enough context (eg. how hermes or chatgpt keeps remembering what you talk to them or research and your gmails/calendar also serves as context) every finance app would have an ai attached which is constantly gathering and mapping ‘context’ (and no, ai won’t control your finances but it’d be in the loop to gather your context) 4) how it's happening? - once txns become context-rich, finance becomes truly personalized eg, @fomo or @Pumpfun allows someone to make a memecoin trade contextual by attaching a thesis and broadcasting to followers and @tryramp captures context for a payment txn from SaaS interactions like emails, calendars and past behaviours. while @Plaid is building a txn foundation model to capture ‘why’ of a payment - all txns on crypto are open in nature: there’s a huge graph already getting built on our wallets from the last 10 years wallets are right now unlabelled context - we just need to make them contextual 'invisibly' and yeah, crypto apps will need to add context and that’ll be their moat we’re witnessing a glimpse via social trading - where a trader doesn’t even realise they’re adding the ‘context’ to their trades (using thesis or callouts) 5) once we have context on txns: - we'll see more personalised investing experiences (eg. you bought Apple for the iPhone cycle but sales didn't turn out to be good - your app flags the broken thesis and you sell) - i feel there’s a tremendous social x ai unlock waiting to happen in finance let’s build the next wave of contextual finance apps! 🫡

if you’re building a trading product, you need to nail these: any trading (or financial markets) product has: 1) discover assets (alpha/info/charts) 2) actions (buy/sell/long/short/yes/no/deposit) 3) manage assets (positions/porfolio) you miss one part & product falls apart 1) asset discovery: - design for infinite discovery: your product need to have a way for user to progressively discover assets & relevant info (charts/token info/fundamentals etc) more surface area for discovery = more time spent = more trades - reduce text, increase numbers/charts: numbers & charts consumes much lesser time & surface area and reveal much higher insights as compared to texts your interface should look clean & structured but never kill any data points for the sake of simplicity no trader likes an empty section/whitespace on a screen - everyone needs an edge (or atleast need an illusion that they’ve an edge) any product-enabled edges will always attract traders eg. @fomo shows top trader by highest PnL - to create an illusion that you can have an edge by checking top trader's trades @gmgnai allowed to track many X accounts & labelled wallets - assets coverage: trading is a way to get exposure to a narrative or sector more assets you surface = more they'll trade (imo, personalisation of asset discovery eg. feeds is still an unexplored design space) 2) trading actions: - <5s action: if your product requires any cognition overload to simply buy/sell, they won’t trade make it dead simple to enter amount & just click green/red buttons (just like casino slots) eg. @AxiomExchange allows to put presets (eg. 10 USDC) to make it buy literally 1-click action CTAs (deposit/buy/sell/long/short) should always be accessible from info/alpha i.e if a user has already developed conviction, let them take trade instantly i.e any info (incl. alerts) should lead with an action CTA - instant dopamine & feedback: just after placing a trade, a user has to feel their trade is moving (instant gratification) - even casino or gacha has instant dopamine loop with sound & numbers moving eg. @AxiomExchange optimistically sends you the confirmation with a sound and instant display on chart with B label (all client-side and it keeps sending on backend/chain) - never remove optionality: never reduce any trading features (eg. take profit) for sake of simplifying just make it a optional UX (toggle for advanced orders): no traders wanna feel they can’t do 'advanced trades' - even if they rarely do 3) manage assets: - keep things moving: user need to feel their positions are changing every second to have that dopamine hit + user’s positions/net worth should be as accessible as possible (they wanna keep checking their trade's performance) - easy to adjust positions: users constantly change their mind as markets move eg. adding/removing collateral, tp/sl should be dead simple - risk visibility: always surface liquidation, leverage, PnL & exposure clearly traders love risk, but they hate being surprised by it 4) bonus: trading apps are most complex engineering-wise, you can't compromise on: - reliability (it should just work 24/7!) - speed (both data & execution) - liveness (stale data is more dangerous than no data) (ofc, liquidity needs to be deep enough: either native exchange or underlying exchange) once the core loop works, referrals, rewards & social features can compound retention/growth ---- crypto is now 'trade any markets' we already have insane trading apps like fomo, axiom, pump, gmgn (all doing $1m/day in fees) - and much more advanced than any non-crypto fintech apps and i feel we've barely scratched the surface, we'll see a huge wave of trading apps getting built this cycle: tackling with different wedges: - mobile - ai - social - gachafication and enable to trade any markets using crypto rails! 🫡

some bear market thoughts on crypto: 1) we're in the peak depth: almost everyone is done selling - we just lack marginal buyers yet either we get capital rotation back from ai soon (q4 maybe?) or we remain flat for some time 2) fundraising is p bad rn, particularly early-stage there's no infra trade left and most crypto vcs have either: - no money - or no idea on how to fund 'products' most got burnt out due to overfunding bs infra in '23-24 which were all down only coins (so no exits) all risky capital is funding ai now as a founder: unless you've some significant traction or big pedigree, it's a waste of time to even think of fundraising 3) teams are shutting down left & right - mostly due to lack of funds or PMF which directly impacts job markets - there's some hiring on institutional side (nyc-based) but apart from that, the hiring has decreased significantly only companies with huge treasuries are able to hire and asian projects are hit the hardest from what i can see capital is a huge moat for startups for 'right to win' - for otherwise capital-starved early-stage 4) there's a huge gap in early-stage funding: best time to angel invest at a low val or even start micro-accelerators due to lack of infra money, hackathon/grants have now reduced by ~95% now earlier l1 grants/funding used to attract a significant number of early-stage teams - further driving away talent if fundraising doesn't return by q1' 27 - innovation will significantly shrink and crypto will become a 'legacy industry' 5) crypto is now just 'financial markets' - there's broadly four markets with PMF: - memecoins & spot (solana/rh) - perps (hl) - yields (eth) - prediction markets (poly) + stablecoin payments and then a long tail markets like pokémon cards etc all non-financial protocols on governance, social, proof of xyz, identity, gaming etc are now dead - they're now just a feature on speculative apps 6) everyone's either: - building an 'everything exchange' coinbase, robinhood, solana, pump, axiom, fomo, polymarket, jupiter, phantom - or packaging yield usdt, usdc, usdg, morpho, aave, kamino 7) there're only 2 types of chains left: - general community-led: eth & solana - distribution-led: base, robinhood, tempo, bsc rest all are dead or will be dead soon 8) the survived teams are incredibly product-focussed now: - better trading execution - ux/mobile focus finally, building good products and nailing distribution is the only way to win the whole game is now like building a fintech app with crypto rails ofc, speculation will always be the holy grail of financial markets and even stocks & gold now trade like crypto - everything is now a narrative trade! 9) there are some teams which are building non-crypto infra (eg. robotics/data collection/depin types) funded by crypto VCs: but we all know they're mostly a scam looking to launch high float, low fdv tokens (and will probably fail) and will abandon the tokens whenever they get some PMF and call it 'credits' (iykyk) 10) there're some incredible liquid opportunities but with teams giving up/shutting down/abondoning - it's incredibly hard to judge or filter out quality but i'd still say holding good quality liquid tokens have the best r/r a simple filter is just buy tokens which has: - good treasury - founder-led who understands distribution - upcoming sector tailwinds (happy to share my list) 11) there's only few traders left in trenches: but they're incredibly smart and they can make money in any market now many are now trading stocks and killing it - as they're the pro narrative traders safe to say: all crème layer genz traders have traded/are trading crypto this core audience is always on lookout to try out new experimentations 12) stablecoin payments: they’re p huge in the shadow economy (eg. an Indian manufacturer buying from china in USDT to escape taxes & regulation hassle) but organising a shadow market is a challenging and unsolved problem agentic payments are a huge opportunity but no PMF yet (typical coldstart problem) - all numbers pushed by base/solana are just wash volumes i'd bet on stripe/tempo (+ trad companies) to gain a lion share in regulated stablecoin payments just because of distribution 13) opportunities now lie on building: - better trading interfaces & modality (via ai, mobile) - more novel markets (eg. compute, emerging market coverage, metadao) - anything around token or memecoins (speculative mechanics) - cool fintech ideas which failed due to banking/regulations (eg. neobanks) and ofc, picks & shovels around this biggest bet: post ai, speculation and trading as a way to show conviction on anything socially will increasingly rise and crypto is the de-facto rails for any new markets 14) and yess, memecoins (internet culture coins) will come back extra-ordinarily as macro moves and capital rotates from ai it'll bring interest back to crypto and once again, sidelined VCs will be salty i still remain cautiously optimistic on broader crypto and p optimistic on crypto as 'anything market' rails 🫡