To learn web3, you have to try web3. If you don't know how, these 11 steps should get you started 👇 1) Set up & fund your crypto wallet @MetaMask 2) Get your "crypto handle" @ensdomains

Play to earn: @AxieInfinity @YieldGuild Learn to earn: @rabbithole_gg Publish to earn: @viamirror Store to earn: @ArweaveTeam, @IPFS Stake to earn: @OlympusDAO, @KlimaDAO Lend to earn: @RariCapital, @compoundfinance Save to earn: @PoolTogether_ Do to earn: @layer3xyz What else?

gm, I have a cool life update!!! A month ago, I quit my job to go full time crypto 🔥 Since then, I’ve been doing deep dives on different web3 products & protocols. I compiled 10 of my threads to help you better grasp the ecosystem. Please share if you find them helpful!

Ignoring @OlympusDAO today would be the same as brushing off bitcoin in 2010. Do yourself a favor and go down the Olympus rabbit hole. It's a genius project that creates wealth and has utility. Let me explain how Olympus takes currency to the next level. Thread below 👇

It took me a while to understand why web3 is so important. Now, I'm pretty much addicted. Here are 11 reasons why I decided to go all in 👇

I went through over 2000 tweets from the Punks account. The first tweet from @cryptopunks was on August 7th, 2009. Here are the most iconic ones:

Please welcome Azuki 9613 to the garden ⛩️ I was able to raise 17.41 eth from 20 skaters to buy the incredible @AzukiZen. 8/20 bought their first NFT! Fractional NFTs increase market liquidity, democratize NFTs, & empower sub-communities. Here's a primer on fractionalization👇

Is anyone working on the Morning Brew for NFTs? NFT Twitter can get overwhelming to keep up with everyday and I would definitely pay for a 5 min daily debrief. - Market snapshot - Upcoming mints this week - Whitelist curation - Breaking news (hacks) - New NFT concept explained

A single Twitter thread by @punk4156 in June led to a project now worth $34 million. @nounsdao will take the concept of meta DAOs mainstream. Let me explain why this is a must know project. Thread Below👇

Coinbase just listed vvv ($350m mc) which is the token for @AskVenice that launched today. If you have no idea what I'm talking about, here's the rundown. - Venice makes it easy to use open source models for people who don't want to download from Hugging Face and run models like Llama locally - AskVenice is basically a ChatGPT UI but focused on privacy. Your requests are encrypted and the actual conversations you have only exist in your browser - They host multiple open source models: Nous theta, Llama 3.1 405B, Dogge 70B, & a few others for image generation. Over the weekend, they added in DeepSeek as well. - 400k registered users & 50k DAU - Venice uses multiple decentralized GPU providers, so it's not possible for any compute provider to have your entire history - The only thing Venice knows is your e-mail & IP address - Based on the blog archive, it looks like Venice has been around since July '24. It's run by @ErikVoorhees, @TeanaTaylor, and a small team - The token vvv just went live today on @base. There was an airdrop for wallets holding Virtuals ecosystem tokens as well (i.e. @aixbt_agent, @Vader_AI_) - Worth noting that Coinbase listed the token on day 1 of it being launched. Not something they usually do, could be a signal of them leaning into the new meta. Imagine if Kraken listed vvv before them, would be a major flop - Devs an access the API for $149 / yr which is the pro plan.

Just a gut feeling but I think we’re hitting a local top on agent infra mindshare. Everyone is now bullish ai16z & Virtuals long term and is holding. After the holidays people will be excited for something new. My guess is it’ll be consumer projects that best demonstrate agentic community management. Initially you’ll see many of the same concepts from 10k pfp projects but the strategies will quickly start evolving as agents try to optimize the quality and quantity of community members. There will need to be: - an interesting lore and ongoing story - opportunities for fans to engage in meaningful ways - community participation through bounties and proposals. Thinking some version of Nouns style projects for creators to pitch in their own style and taste BUT proposals will be managed by agents. One model could be agents doing a first round pick of top proposals and then a vote with community holders that hold x # of tokens - multiple agents participating in unique ways. This will lead to part of the community rallying around specific agents within one project. What better marketing than friendly competition? - memes, pfps, and sick artwork to share. Also more of a focus on agents tweeting images rather than just text. - tiered access to “influencing” the lore based on token holding (similar to aixbt terminal) - a store concept to buy merch directly onchain from agents To be clear, I’m still bullish on infra plays and trading agents, but part of the game is realizing that metas rotate. People love to be involved and have a say in projects they’re financially invested in. That’s why I think agentic community management / NFT style agent projects will catch on quickly. The crypto community is also familiar with these types of projects from last cycle. Two projects I’m invested in for (nfa dyor) this category are nothing by @SHL0MS & @ropirito (nous) and @pillzumi. I think both teams are executing here, just a matter of time before mindshare catches up imo. There will probably be an “AHA” moment when we see a unique strategy implemented by an agent in terms of community engagement in a couple of weeks to a month from now.

need more of this in crypto: - mobile - smooth onboarding - crazy attention to detail - clean UX - feels like an app normies would use - leans into hypertokenization & games - beta mode / server overload bc of hype - funny marketing awesome job @yapsterxyz & @cloutdotme

"This is kind of the beginning of the cambrian explosion of MEV projects" -@phildaian 11 resources to get you up to speed on MEV 👇 Some Follows: - @bertcmiller - @gakonst - @0xmisaka - @FrankResearcher - @mevalphaleak - @NathanWorsley_ - @tomhschmidt

6 months ago I started posting on Twitter and it completely changed my life. I went from managing work & hobbies to just doing cool shit full time. Posting a mega-thread of what I've written so far. Please share if you enjoyed reading my threads! 41 posts on crypto & web3 👇

The term social graph really started taking off after Facebook's first F8 conference in 2007. Everyone flooded to FB and helped them build one of the largest social data layers. Through Lens, we can rebuild a much stronger social graph that is open source. Some notes 👇

after spending the last few weeks diving into decentralized AI, I'm insanely bullish on this vertical. right now, there are <250 people in the world paying attention but it feels like only a matter of time before both the crypto & ai circles start to take notice. i still have a ton of papers to go through, but here's a high level overview of what I've learned so far. most of these insights are from going through three incredible blog posts (all linked below) by @Ronangmi, @naman_kapasi, & @SPLehman. 1. MODEL SUPERCOMPUTER At its core, decentralized AI is not a new concept. The idea is to basically build a world computer (sound familiar?) that can train a state of the art (SOTA) model that has no central node and is fault tolerant. It's worth remembering that Bitcoin is the world's largest supercomputer by multiple orders of magnitude. We've technically proved it's possible to build a decentralized network that combines compute + economic incentives + utility. If we can do it for "digital gold", why not AI? Of course, there will be a different set of challenges just like Ethereum vs Bitcoin. But these ideas are all variants of each other serving different real world assets (money, programmability, intelligence). 2. CENTRALIZED CLUSTERS The big unlock that slowly formed from 2017 (release of transformers paper) to 2019 (gpt-2 launch) was that it's possible to make these "assistants smarter" by adding more compute and scaling up the underlying models. This marked the start of the GPU gold rush and the AI space started shifting to being more closed source as pre-training became the competitive margin. And since chatGPT launched in 2022, we've seen unfathomable numbers on how much the largest tech companies are spending on data centers (i.e. Meta, xAI, Microsoft, Google). read @dylan522p piece on how much goes into these data centers on the backend and how the hundreds of billions of dollars will be spent (link below). 3. TRAINING OPTIMIZATION Now the key part to understand is that it's not just about who has the most number of GPUs. Rather, the north star is how well these teams set up their data centers to maximize efficiency. This is can be calculated by: MFU (model flop utilization) / PUE (power usage effectiveness) = total cluster efficiency This accounts for GPU failures, continuity of systems, etc. For reference, Llama 3 had a MFU of ~40%. As these models get exponentially larger, these optimizations become crucial. This leads into the topic of Parallelism which allows model training to happen in parallel. There's 3 main types: data, tensor, and pipeline. I won't go into detail here but just know that most of these big tech data center setups use a combo of all three known as 3D parallelism. 4. LOW BANDWIDTH, HIGH LATENCY This brings us to DeAI's first challenge which is distributed training. Centralized data centers have the advantage that their hardware is co-located so GPU communication is *very fast*. Nvidia NLINK connects GPUs with speeds of 1800 GB / sec while a normal internet speed is ~500 MB / sec. Unfortunately, distributed training models have the uphill battle of GPUs that have low bandwidth and high latency because of their geographical separation. If you don't know how pre-training works, here's an eli5 of why GPUs need to communicate. Let's say you have 10 data sets that you need to train the model with. After each data set, the model needs to update and adjust based on how poorly it did (loss function). You may have heard of gradient descent before? Basically, the goal is to iteratively bring the gradient down after each data pass. Now, remember, there are tens of thousands of GPUs that all need to sync up after each pass...you can imagine how that gets heavy quickly. So, the main question for distributed teams becomes: how do we reduce the communication needs? Is there a way to not have every single GPU sync up in the system after each run? 5. REDUCING COMMUNICATION NEEDS Four parts in this section. A) Imo, the key breakthrough for distributed training was when Google released their DiLoCO paper in November 2023. @Ar_Douillard is a fantastic follow to keep up with updates here. The key learning from that paper is inner-outer optimization which reduces GPU communication needs by 500x! Inner refers to a single node of GPUs making quick, local updates. Outer refers to a less frequent system wide updates between nodes. Think: intra-node vs inter-node. From here, the @PrimeIntellect decided to implement this paper at scale. They used the Hivemind library and published OpenDiLoCo which also included a fault tolerance mechanism known as the "heartbeat". Basically, if any hardware didn't send a beat it would be removed from the system. This enables hardware to come in and out of the network. They launched the PRIME-INTELLECT model in November 2024 which was a 10b parameter model trained on Open DiLoCo. It's amazing to see that @vincentweisser and team were actually apply this Google research at scale. This model release is a "poster-child" moment for DeAI. B) The second approach is led by @NousResearch and their DeMo & Distro paper. This method focused on decoupled momentum optimization. Remember how we discussed the eli5 on changing gradients after every data pass? I won't get into details here but optimizers are used to update the model weights to minimize the loss function (margin of error). Most training methods today use the AdamW optimizer which was released back in 2014 by Diederik Kingma. Kingma worked with the Nous team (@bloc97_ & @theemozilla) to improve the AdamW optimizer and use discrete cosign transformation (DCT) to decouple fast and slow momentum changes. The core concept here is that it's possible to split the importance of gradient changes into fast & slow. There are some changes that need to be immediately communicated and others that can be done over time. It turned out that the gradient compression solution ended up being as good as if not better than models that just used the AdamW +allReduce optimizer. And most importantly, there was an 857x reduction (4-5 orders of magnitude) in communication needs! The Nous team also announced their 15B parameter model in December 2024 - the second poster child for DeAI. C) The third main approach is SWARM parallelism which heavily relies on pipeline parallelism. You know the classic picture of a neural net which has columns of circles that go from left to right? Each of those columns is a layer and pipeline parallelism basically splits the training vertically by layers. So some GPUs are focused on the initial inputs, some are focused on the middle blackbox layers, and others are focused on the weights and layers closes to the output. As models get super large, this approach is fantastic as it removes the GPU RAM bottleneck. However, as Sam puts it concisely, "the key innovation with SWARM is their approach to handling the flow of data through the network of devices in a messy training network". Basically, the architecture manages all the different kinds of hardware (that all have separate compute power) during a run. I still need to dive deeper into how the routing mechanism works, but one of the most interesting discoveries by @m_ryabinin, @Tim_Dettmers, and team was the square-cube law of distributed training which notes that compute time grows much faster than communication time as model sizes increases. This is a huge win for distributed model training because the biggest bottleneck of quick GPU communication becomes less important. D) I haven't looked into any of the following but just noting as Naman included it in his post. There's some non-transformer related architectures also trying to make progress in this vertical. Some examples he lists: - Mixtral - Switch transformers - DiPaCo (distributed path composition) 6. OPTIMIZATION MIX & MATCH All the approaches discussed above in reducing GPU communications are basically different optimizations on the current centralized pre-training process. The part that stood out to me most - and something I think should be the key takeaway of this post - is that the real magic will come when these approaches are combined together to immediately 100x or even 1000x the training efficiency. For example, Distro + DiLoCo or SWARM + DiLoCo. It seems like the teams are actively exploring each other's works and we'll probably see a lot more these optimization permutations in 2025. 7. DISTRIBUTED vs DECENTRALIZED Security, verification, and encrypted data training. One big call out here is making sure to differentiate two terms I see incorrectly interchanged. Distributed just means the hardware is not co-located...even big tech companies like Microsoft are working on this. Decentralized specifically means that there's no central node and the training process is fault tolerant - exactly what the Bitcoin network preaches. So what are the specific needs for decentralized training to happen? This gets into the topic of security, encryption, and verification. Again, details are out of scope rn but wanted to include a few examples that Sam included in his post: - Zero knowledge proofs (@ezklxyz) - Fully homomorphic encryption - Multi party computation - Trusted environment executions (@PhalaNetwork, Flashbots, @sxysun1) 7. INCENTIVE ALIGNMENT I've compared decentralized AI networks to Bitcoin a couple of times in this post. It's worth noting that the same financial incentive mechanisms will need to apply here as well. - slashing mechanisms to penalize malicious actors - huge reduction in up front cost with hardware providers taking a bet on the long term revenues and appreciation of network ownership - network effects as markets form around a liquid model token @alexanderlong from @PluralisHQ has some great insights here on protocol models (post linked below as well). Also worth noting that @NousResearch recently announced Nous Psyche which aims to bring the verification and incentive mechanisms in partnership with Solana. 8. INFERENCE-TIME COMPUTE Since the Deepseek craziness from a month ago, the next big thing in AI seems to be researching new reinforcement learning (RL) techniques and using chain of thought (CoT) to let the models learn from the current context. Deepseek showed us that it's possible to match current model performance benchmarks without the same hardware requirements by doing other kinds of optimizations. We can go into a lot more depth here, but for the scope of this post I want to call out the fact that if in fact the next big thing in AI is moving from pre-training scaling to RL & inference time compute then it seems like the uphill battle of compute constraints that decentralized AI faces somewhat goes away. I'm not well versed here so don't want to make any claims just yet but Ronan had a fantastic sentence in his post that I wanted to copy over: "The newest vector to scale model performance — reasoning and inference-time compute — naturally enjoys this described communication reduction. There is no inherent disadvantage to a distributed approach: the importance of this paradigm shift for distributed (and decentralized) training cannot be understated." Final note: I'm *very* new to this vertical and am still learning so there's definitely some errors and things I missed, please take everything with a gain of salt.

tired: subscribe to my newsletter wired: subscribe to my bookmarks

We're launching a web3 research program @chapterone for college students, Chapter One Researchers 📚 If you're going to be a full time student this fall and aren’t scared of white papers, comment below or shoot me a dm! No prereq on major or year. Who should I meet with?

Gave an Intro to Decentralized AI talk at Farcon last week. Here are all the slides + notes I wrote to prep. Full talk will be published later this month - hope this is helpful!

gm! New pfp 👀 I believe @chain_runners is going to be the next blue chip NFT collection. Runners = Punks + Artblocks + Nouns The project is relatively new, but already gaining so much momentum. Here's why 👇

ok after talking to some friends + my own research last few weeks, initial list for “DeAI golden basket” - @PrimeIntellect - @NousResearch - @PluralisHQ - @hyperbolic_labs - @ezklxyz - @exolabs - @gensynai Will pay close attention to these teams in the next few months

In 2012, @brian_armstrong applied to YC with his company "Bitbank". His OG bitcoin side project is now an essential part of the crypto ecosystem. I went through 5,000+ tweets from the @coinbase account. The first tweet was on May 19th, 2012! Here are the most iconic ones:

I went through 7500+ tweets from the @y00tsNFT account. The first tweet was on July 30th, 2022. After writing this thread, I immediately bought a y00t. Super impressed with the community, so better late than never. gm y00t fam! Here are the most iconic tweets:

From a single blog post covering Axie Infinity, @packyM was able to raise $45,782 (9.56 eth) on Mirror. Web2 was for the platforms. Web3 is for the creators. @ArweaveTeam & @viamirror make it possible to own and monetize your content. Let's see how 👇

hosted my first event in NYC last night and it went really well! ~50 people showed up, had some Indian food, and enjoyed the @yapsterxyz office also made a TG channel (read only) where I’ll be curating any crypto & AI NYC events, just comment below if you want to join

The next wave of crypto users will acquire their crypto through earning, not investing. @rabbithole_gg is paving the way by embracing the "learn to earn" model. Let me explain why this product is a must bookmark Thread Below👇

Based af. From @balajis in Eric's new book.

Yesterday, the senate committee on banking, housing, & urban affairs held a hearing to address Hamas financing. It was 2.5 hours long, so I went ahead and compiled the relevant crypto clips. This first one sums up the overall sentiment 😂 1/ Senator Warner

Good builders to follow on Farcaster for the current agent meta: - @_proxystudio : Clanker (launch tokens by tagging) - @anquetil: Aethernet (agentic community manager for @higheronchain) - @Slokh: @anoncast_ (shared anon account, zk consumer app) - @wijuwiju0x: Interface (trading app, copy trading). How Vitalik's anon purchase went viral - @jpren: Frens (chat + trading, where lots of clanker coin discussions are happening) - @superwoj: Super (primary FC client other than Warpcast, started anon trend, has mobile, easiest to interact with FC native tokens) - @rish_neynar: Building "Agent quality score" as an increasing % of casts are AI generated

After playing the first 2 rounds, I can confidently say @yapsterxyz is one of the most fun crypto consumer experiences I’ve had and it’ll probably be the first real competitor to pump fun. tl;dr is that pump fun is for traders and yapster is for…pretty much everyone else who enjoys memecoins & games. The game: 1. Players buy a ticket and submit a meme & ticker beforehand. 2. When the daily game starts, you vote on which memes you think will do the best. 3. At the end of the rounds, the players decide if the ticker should be deployed. If it’s deployed, half of the buy in pot goes to liquidity, other half is proportionally allocated to players. My thoughts: - HQ trivia vibes w everyone playing at the same time is a LOT more fair than traders trying to snipe pump fun launches and share “TG alpha” - Initial deposit / game buy-in process is sooo smooth. Clean UX + @privy_io magic - I forget it’s a crypto product till the end when you have to vote on the token deployment - The attention to detail is insane. Everything from the haptics to the arcade design makes it feel fun no matter how old you are. - You don’t have to be a great meme maker, it’s still possible to win a huge allocation by being the quickest to decide what the “markets” will find the funniest (memecoin connoisseurs) - I like the Nouns vibe of “one memecoin a day” - If the product can continue to onboard some of the funniest shitposters / meme makers, then I can easily see Yapster being the default place where the next big memecoins launch Highly recommend trying it out, the first token launched on the app was $YAPSTER a few days ago. Tomorrow’s game is already full

gm, fun life update! Super excited to start a new role today. I’ll be joining @chapterone as head of experiments & research 🎉 Thrilled to be working with @jmj, @menemazarakis & the incredible team. My job is to be the crazy scientist at Chapter One. Here's the rundown 👇

Some of my web3 must follows: - @VitalikButerin - @cdixon - @balajis - @lindaxie - @punk6529 - @ljin18 - @packyM - @RealNatashaChe - @TrustlessState - @SBF_FTX - @patrickxrivera - @chriscantino Who else?

"In the medium to long term ZK rollups will win out in all use cases as ZK-SNARK technology improves" - @VitalikButerin I went through 3,000+ tweets from the @zksync & @the_matter_labs accounts. The first tweet was on September 29th, 2018. Here are the most iconic ones:

I've been writing my newsletter Terminally Onchain full time for a year now and have published 94 deep dives. My goal was to hit 10k subs by EOY and I'm at 8.8k rn. Would appreciate any shares if you've found my content helpful! 6 sample posts + link to subscribe below:

I went through 2000+ tweets from the @Uniswap account. The first tweet was on November 5th, 2018. Here are the most iconic ones:

Did you know the first ENS domain to be registered was rilxxlir.eth? I went through 3,500+ tweets from the @ensdomains account. The first tweet was on May 15th, 2018. Here are the most iconic ones:

"We will have a billion VCs by 2040. Crypto will turn everyone into investors." - @balajis @SyndicateDAO is simplifying the process for people to form and use DAOs to invest. Let's see why that's so important 👇

if you spend more than 3 months being active on @farcaster_xyz, I can guarantee you will realize there is in fact a super app being built in real time - wallets - identity - community - mini-app GTM - cross chain - messaging - micropayments - product notifications - social

this past week it's really hit me how increasingly important your IRL network is going to be in the coming years so I'm going to try experimenting w diff kind of meetups will start w friends in crypto & ai but expand over time if you're in NYC and want to join, pls comment!

never deleting this app

The research team just finished up their first sprint! Here are 10 posts @chapterone labs published these past 2 weeks. Hope you find them interesting, more on the way.

Two things I’ve learned from following @real_y22 trading notes: 1) it’s totally an okay move to just liquidate everything and sit tight if you’re feeling uneasy about general market environment 2) keep your portfolio under 5 positions and size in

10 resources to get you up to speed on ZK 👇 Relevant Tags - Scaling - Privacy - L2s Key Follows: - @the_matter_labs - @StarkWareLtd - @zksync - @MirProtocol - @aztecnetwork - @ZK_Daily - @zeroknowledgefm

CT isn’t realizing that @farcaster_xyz is about to have another frames moment now that @superwoj and @tednotlasso are on board good time to hop back in

If you're as confused as @blknoiz06 about Clanker and what's going on in the @farcaster_xyz & @base ecosystems... come join this space tonight at 10 pm est! nitter.app/i/spaces/1djGXrOwwbBxZ

was out for vacation the last 2 weeks and finally got some reading done today. sharing some thoughts and posts on the agent meta I found helpful to catch up. key takeaways: - because of @aixbt_agent success, seeing cool updates and more interest for @polytraderAI (prediction market agent) and @artto_ai (NFT trading). Will be cool to see how other projects niche down for different assets or ecosystems and form a speciality. Will get even more interesting when they start to work together (saw a tweet of aixbt and polytrader sorta doing this) - something I've been tweeting about for a while...people are starting to focus on the rest of crypto x AI stack. The realization is that the majority of these agents aren't sticking around long term and that we still need a ton of improvement on the rest of the stack for this crypto agent thesis to actually play out - agents as a frontend to solve complex crypto UX seems to be sticking. Saw a couple of demo videos with agents taking multi-step onchain actions. still basic rn ofc but see a ton of potential over the next year - everyone is comparing agent frameworks to L1s. The analogy makes sense at a high level, but I need to think about it a bit more - there seems to be increasing attention on consumer focused agents interacting with NFTs, gaming, & metaverse verticals. I tweeted about this around Christmas time as well (see thread below) - still a ton of discussion on swarm related ideas but haven't seen anything really worthy of discussion just yet. GAME by Virtuals is probably the best example of a productive agent ecosystem I've seen so far. With that being said, haven't looked into fxn or gnon as much as I should have by now. - an increased focus on verifiability of agents, TEE, and other solutions. need to do more research here myself - I need to look into more utility agents that are software focused and completing tasks such as security audits - Google released a 42 page paper on agents that I'll read through and share thoughts on later this week

I went through over 26,000+ tweets from the @AxieInfinity account. The first tweet was on February 6th, 2018. Here are the most iconic ones:

3) Read Cyberspace Independence Declaration @verses_xyz 3) Buy an NFT @opensea 4) Verify your identity on @BrightIDProject 5) Do a quest on @rabbithole_gg 6) Swap your eth for another token on a dex @Uniswap

If you enjoy the thrill of Poker, you'll love on demand generative art. It's art mixed with a sprinkle of luck. @artblocks_io is bringing this new art form mainstream. Let's dive in 👇

12 weeks ago, @chapterone launched the first cohort of the web3 university research program. How'd it go? 10 students wrote 27 deep dives covering a variety of web3 topics (see thread below) and landed internships at @Uniswap, @ironfishcrypto, @stelolabs, @Molecule_dao, etc.

open source agent frameworks and their github metrics h/t @Decentralisedco

my takeaway from learning more about deepseek is that it's time to be extra bullish on @NousResearch and @PrimeIntellect

10) Archive a web page on @ArweaveTeam 11) Complete a task on @layer3xyz

I went through 8500+ tweets from the @BoredApeYC account. The first tweet was on April 17th, 2021. Here are the most iconic ones:

For those of you heading to bed and want a quick tl;dr of El Salvador's bill, check out this article by @CoinDesk coindesk.com/el-salvador-pre…

Circle (USDC) + Coinbase (wallets for AI agents) + Base (cheap transactions)

Composability. (h/t @jackbutcher, @visualizevalue & @cdixon)

Lots of journalists are onboarding to @farcaster_xyz for the reasons @Noahpinion explains in his recent post. Farcaster, with its channel feature, is a perfect balance of X hivemind + discord fragmentation. In fact, @MazMHussain is a top earning creator for his news reporting

Imagine if Twitter didn't allow its users to customize their handles. Our entire feed would look like "user12837473". That's how crypto feels like right now. @ensdomains is drastically improving the crypto UX by providing web3 handles. Let me explain how👇

I went through 6,000+ tweets from the @graphprotocol account. The first tweet was on June 5th, 2018. Here are the most iconic ones:

need a prediction market on how many agents are going to be part of @SHL0MS swarm by end of year and what handles he's locking in rn

Seems like we're about to enter a 6-9 week euphoria phase starting next week after the inauguration or the last week of January. We'll see some amazing pro-crypto headlines in the news about Trump pardoning Ross Ulbricht, America becoming the self-proclaimed crypto capital, new policies going into the pipeline, more ETF proposals for seemingly everything, etc. Bitcoin, XRP, and maybe Solana will probably be the first to start pumping and retail will naturally start feeling the fomo as mainstream media locks in. Naturally, this will lead to some quick overinflation and money will start rotating into Ethereum and alts. The eth maxis will finally be able to enjoy their brief moment of glory especially as EIP7702 gets ready to merge around March. But...all of this will be short lived. My guess is that something similar to 2021 plays out but at an accelerated rate. Remember in May 2021 when the market tanked going into the summer and continued to drawdown even in July? Dreadful period. There's bound to be a similar moment this year but I think most people are expecting it. So in some weird self-fulfilling prophecy kind of way, I could see the eventual drawdown get expedited to before tax season in April. So realistically, my timeline is that the "inauguration pump" starts correcting by March / end of Q1. This roughly also lines up with the first 100 days of Trump's presidency. There's only so much news the administration can push out in such a short period of time...realistically these things take time and as the headlines start dying down, folks will naturally hop on the "it's so over" train. Then we'll see a choppy period of slow price momentum for a couple of months. Many of us in the space will still feel a lingering rush of excitement as there's announcements of fresh venture rounds, cool UX updates, interesting examples of large tradfi/web2 players "partnering" with some crypto company, etc. But to be clear, price wise, it'll feel dead and the Phantom screenshots will be on hold. And then finally sometime in between the end of July to early September, there will probably be another catalyst moment for what the industry calls the "blow off top" phase of the market. Where prices give everyone an "aha I got you" moment and start pumping again at an even faster rate. This will lead to another euphoria moment similar to October - November of 2021. It'll be quick. It'll get messy (with scams/hacks). And worst of all, many people won't want to miss out on life changing gains and will probably make some irrational financial decisions. Then dread will really set in as the market corrects once again before the end of 2025. But this time, not just from a price standpoint but also a narrative standpoint. My guess is that aside from any large exploits and bad actors, there's a good chance of what I call the "Trump uno reverse card" happening. This is basically the moment a year into his term where it feels like crypto regulation will actually take longer than expected and things are not that easy. I'm not saying that Trump will reverse his crypto stance, but more so that any new updates for the rest of his term will be evenly spread from 2026-2028. And then 2026 onwards, we'll probably see a higher velocity of smaller market cycles. Less crypto volatility than what everyone is used to. And a very different vibe from the normal "bear markets" all of us are expecting. I believe we'll see a ton of progress for the rest of Trump's administration but it'll happen at a reasonable pace. Perhaps that's why the 4 year cycle narrative also starts to die down - it won't go away, but the Bitcoin halving in 2028 definitely won't be anywhere close to being the most important market factor. Everything above is obviously NFA / DYOR. The market does not give a shit about your timelines or targets. The market does not care about your favorite KOLs predictions.

Coinbase & Base is the best example of a vertical integrator in crypto. If you replace the company name from this paragraph in @packyM's piece (linked below), it still works: "A true Vertical Integrator, however, [COINBASE] adopted the winning [L2] technology and plugged it into its system. The company leveraged its brand, size, resources, and diverse product line to invest heavily into [L2s] and used its large customer base to introduce new [ONCHAIN] products to the market."

If you want to win in crypto in the next year, lock in on one of these: 1) decentralized RL 2) mini-app ecosystems (Farcaster, world coin) 3) parts of the stablecoin stack less approachable by tradfi and big tech

AN ONCHAIN LETTER TO @ElectricCapital: On Wednesday, Electric Capital released their 5th annual crypto developer report. After reading the report, I asked myself: if I had to pick 1 graph from the 180 page report and write about it, which one would it be? Key Takeaways of the post: 1. If the # of crypto developers starts losing momentum and spirals downwards, then all the other charts and metrics don't matter. The priority is to keep growing the number of monthly crypto contributors. 2. Decentralized task creation will become the primary channel to attract the next wave of developers to crypto. We now have the entire end to end stack to make this possible. 3. As the next bull market picks up steam and there's a revived interest in crypto, I want a link (probably a Farcaster client) to a bounties feed that I can send to my friends who are web2 developers. They should be able to onboard, pick a task, and start earning in under 5 minutes. Note: you can read and collect this post onchain in the link in the tweet below. The letter below is split into 3 sections: 1) S-tier chart 2) Open source incentives 3) Decentralized task creation S-TIER CHART: This year the Electric team analyzed 485 million code commits across 818k repositories! I spent a while trying to shortlist my favorite graphs but couldn't pick one. After a while, I got frustrated and decided to start from the top. And that's when I realized that I had totally ignored the very first graph: It's the most obvious one but also the most inspiring in my opinion. I saw a few people on Twitter mention how tiny crypto is and how the # of developers felt trivial. And they're not wrong. I mean, only 22k of the 23 million developers around the world are contributing to crypto! But that graph can also be viewed from an optimistic lens. What stood out to me is the fact that there has been tremendous developer growth in the last 8 years despite all the craziness this ecosystem goes through: Bitcoin halving cycles, regulatory pressure, hacks, scams, fraud, etc. Think about it...with only 22k developers, the ecosystem has managed to prove enough value that the largest asset manager in the world has bought 25k bitcoin and is actively trying to sell their ETF. Going through the rest of the report was useful to better understand which L1s/L2s have grown, the geographical distribution of crypto devs, the shift towards multi-chain development, etc. But at the end of the day, if the crypto developer graph above starts losing momentum and spirals downwards, then all the other charts don't matter. 💡So, how can the crypto ecosystem ensure that the # of developers continues to grow? 💡 Well, to answer that, let's explore the ideas of open source incentives and decentralized task creation a bit further. OPEN SOURCE INCENTIVES There was a note underneath the graph that caught my eye: That got me thinking. For the last 15-20 years, the citadel for software engineers has been to work at a high growth startup or big tech companies like Apple and Google. For most of us, working in tech meant working in silicon valley companies with proprietary software. The folks who were old enough and active in the software industry back in the '80s are the minority who actually understand what it means to build in a open source dominant environment. TCP, IP, HTTP, SMTP, etc. are all decentralized protocols that were built in the first era of the internet and still serve as core digital infrastructure even today. These networks became a canvas for innovation and paved the path for the next set of developers to build on. It was in the second era of the internet where proprietary software took the reign. Ironically, the Open Source Initiative was formed in 1998 and the following two decades of the web were primarily defined by closed competitors. So...why didn't open source win the second battle against proprietary software? Because it was impossible to compete! @cdixon explains it in a blog post from 2017: "Proponents of open systems never had an effective way to manage and fund operating services, leading to a significant architectural disadvantage compared to their proprietary counterparts....These open schemes required widespread coordination among standards bodies, server operators, app developers, and sponsoring organizations to mimic the functionality that proprietary services could provide all by themselves. As a result, proprietary services were able to create better user experiences and iterate much faster." And this is why most of us have only experienced a software world where the primary objective is to write code in closed repos. But! As we all know, crypto offers a solution for open source networks to finally fight back. Through tokens, there is now a way to use financial incentives to facilitate funding and accountability in open source projects. You're probably thinking, YB I already know that! Optimism has the $OP token and Uniswap has the $UNI token and so on. Yes, but the epiphany I had yesterday is that the current ~22k developers in the ecosystem care about crypto. They're willing to participate in the weeds of governance models, have an opinion on the nuances of L2s, and are actively rooting for ecosystem growth. But there's eventually going to be a cap on the number of developers that care. 💡If we want the # of monthly crypto developers to grow tenfold, we'll need to switch from a values based model to an earnings based model. Can the crypto ecosystem attract developers by showing them it's possible to earn a significant amount of income by contributing to open source projects? 💡 DECENTRALIZED TASK CREATION More than 4 years ago, I watched this talk by Balaji on Pseudonymity. The key point that I have always found interesting is the idea of a tasking economy - a world where it's possible to earn your living (or a nontrivial % of it) through a bounties marketplace. Of course, as with any Balaji idea, it sounded absurd. I mean who would give up their high salary big tech job to work on "tasks"? His main point in a essay he wrote back in '21 was that if there is enough supply of tasks and a way to complete them in a decentralized manner, then the demand will follow. "But crypto addresses solve only half the problem. Yes, now we can pay people any amount anywhere, but to enable a feed that enriches rather than simply distracts, we need something else: task inventory. That is, we need individuals or institutions that will create and fund tasks in sufficient volume to fully supplant distracting social media feeds. If we can do that, we can transition from mere social networks to true digital economies, where you open the app, scroll your feed, find a task that suits your skills, and then earn." I never really gave the idea much thought but it's always lingered in the back of my head. Then yesterday, it hit me: decentralized task creation will become the primary channel to attract the next wave of developers to crypto. But I've read this piece from Balaji several times in the past. Why is it just now clicking? Because I think I've finally seen the earliest version of this happening now. Consider the following stack. The top three got significant attention in the last bull market. Most people active in the industry are well aware of and already use a wallet, decentralized exchange, and an ENS. But in the last year, it's the bottom three that have caught my attention. We now have a sufficiently decentralized social protocol, a bounties marketplace, and a way to link onchain attestations to wallets. Look at this simple example above. What are the steps for me to earn onchain? 1. Make a wallet 2. Create a Farcaster account 3. Complete the bounty 4. Get paid 100 USDC straight to my wallet In the case that the bounty is paid out in a token I don't want, I can swap on a DEX for my desired token. And if I need to prove that I have experience as a JavaScript developer, then my onchain attestation can speak for me. Of course, this is a micro example but it shows that the end to end cycle is now already possible! Imagine this scaled out to millions of bounties for developers. Heck, not just developers but creators, designers, etc. Anyone can hop on, put in the work, and receive direct payments. With that being said, let's revisit the initial question: how can the crypto ecosystem ensure that the # of developers continues to grow? 💡I believe the answer to this is to have a dedicated Farcaster client where the sole purpose is to create a social feed of bounties for developers. This is the answer to @balajis request for "earning while scrolling". 💡 At the end of the day, if there is a frictionless opportunity to make money, then my belief is that developers won't care if they're contributing to an open source crypto repo. They'll onboard, get the work done, and be delighted to see the USDC racking up in their wallet. So what does that mean for us in the crypto ecosystem? Well two things: 1⃣Projects need to be outsourcing as many tasks as possible to the open source developer ecosystem. It's no different than Uber's marketplace...we need more drivers! The bounties can be as small or as large as you want. Eventually, we need to get to the point where people can "doomscroll" on a bounties feed. 2⃣And we need a "developer front page" to crypto. This is the Farcaster client in my head but it could turn out to be something different. The point is, I want a link to send to my talented developer friends in the web2 world where in under 5 minutes they can get set up, pick a bounty, and start earning. We're still a ways out from this concept proliferating, but just the fact that now we're able to identify a stack for decentralized tasking is an incredible achievement in itself. @avichal summed it up in a tweet from yesterday: "There were countless great ideas in 2017, but it was mostly vaporware. By 2019, most people left/gave up. 5 years later, all of those ideas are working and in production. In another 5 years, most of the world will wake up to the realization that all of the ideas from 2018 are real, working, and scaled up. We are right on track for a 10-year technology cycle" He nailed it. I first learned about the decentralized tasking economy in 2019. And 5 years later, I'm realizing it's now possible to do so in 2024. If things go as expected, by 2028, the idea of developers earning their income through bounties won't be as crazy as it may seem now. Remember, open source projects are essentially memes. Anyone can use them and build on top of them. There is no private company that can compete with the growth of open source in the long term. - YB

I went through 4500+ tweets from the @CryptoKitties account. The first tweet was on Sep 12th, 2017. Going through the early days made me truly realize how essential the Crypto Kitties were for the NFT space. Here are the most iconic tweets:

lesson I’ve learned 92617 times in crypto… when your friend sends you a product to try in the chat, do it immediately so you don’t wanna kys two weeks later

7) Create your @viamirror dashboard and publish your first blog post Bonus points: use one of the monetization features - crowdfund, edition, etc. 8) Enter a pool in @PoolTogether_ (Prize Savings Account) 9) Try a lending protocol @compoundfinance

my understanding of EIP7702: 1. huge (if not the biggest) UX upgrade for Ethereum 2. chrome extensions : browser :: modules : wallets 3. didn't work with 4337 bc no one switched over to new account vs 7702 makes this idea possible through EOAs

"CC0 + NFT = unseizable identity" - @sartoshi_nft Are you ready for cc0 summer, anon? 10 resources to get you up to speed on CC0 👇 Key Follows: - @punk4156 - @dhof - @chain_runners - @blitmap - @nounsdao - @cryptoadzNFT

Who are the best researchers you know in web3? I'm trying to find mentors for the Chapter One Research Program. Any part of web3 counts! defi, nfts, infra, data, social, mev, liquidity, etc. Personally a huge fan of @dabit3 and @divine_economy demos / threads. Anyone else?

Spending time in @ceramicnetwork and @LitProtocol ecosystems gets me so excited about the future of dapps. Together, they provide developers a way to easily integrate data storage & encryption features in their products. Here are my notes on the data & access layer of web3 👇

"The creator economy allows you to be your own nation state" - @Cooopahtroopa Web3 enables people - not platforms - to own and use the value they create. @variantfund is playing a huge role in making the passion economy a reality for the middle class. Let me explain how 👇

I'm running an experiment for my new podcast "Web3 Experimentooors". Looking for 10 early stage web3 founders & 10 web3 student degens. If you're interested, please comment or dm me! Please tag any awesome web3 founders and/or students you know, would help a ton 🤝

If you're bored this weekend, go watch this incredible conversation where @Breedlove22 discusses #btc with @lexfridman from first principles. I'll probably need to rewatch 5 times to fully absorb it. The philosophy & history components blow your mind. piped.video/watch?v=HrehEWYj…

Alright mfers, grab your ⌐◨-◨ and come !vibe with me. Today, we run to seize the memes of production (h/t @punk6529). In web3, retail will control the narratives. My thoughts on CC0, counterculture, & minimum viable memes 👇

A LETTER TO @wijuwiju0x & the @interfacedapp team: I genuinely believe Interface will be a top 3 product of this next bull market. My prediction: by the end of 2024 the top 100 followed people on Interface will be critical in influencing onchain culture. In the past few months, other than Warpcast, Interface is the only other crypto app that I open up on an almost daily basis. In my opinion, it's solving one of the biggest problems people in crypto face: onchain fomo. There's no reason the community should be stuck on crypto Twitter all the time to catch the latest mints. For a while, I strictly had the mindset of a developer or trader specific block explorer - a bit more intimidating and niche. The idea of having a social block explorer is such an obvious answer. I know the concept has been talked about a lot in consumer crypto discussions. But using Interface actually demonstrates the ability to make sense of onchain data and display it in a user-friendly way. At this point, I might be getting 15-20 Interface notifications a day! Sometimes they drive me crazy, but I still keep them on because I've reached the point where I care more about what's actually happening onchain rather than people talking about the chain on crypto Twitter. I recently started sharing what I mint daily on Zora, Base, & Optimism. This is basically a chance for me to justify my time spent on Interface 😂 Turns out, there's a lot of cool stuff happening onchain, but we're all just too busy looking at the wrong platform. In this next bull market, I believe the amount of content & moments that will be minted will grow 100x. And Interface will be the guiding light for those that want to stay involved with their wallets. Interface will be the tool for onchain curators & tastemakers to start building their following and drive transactions in their own unique ways. As L2s make minting cheaper and platforms such as Paragraph, Zora, Sound, etc. help bring content onchain, the need for onchain social discovery will only grow more & more. - What are my friends minting? - Was there a banger piece on Mirror that everyone is excited about? - Are my Farcaster friends excited about this new NFT collection? - Why are so many smart people buying a random memecoin? One point that @ufo__club made in this post I wanted to highlight: "Content is surfaced in the Interface app based on who you follow. You can find your friends from Lens and Farcaster. Catch their mints from Zora and Mirror. It’s wild. Ecosystems are flowing together." This is spot on. To me, Interface feels like the ribbon that ties everything together. As the number of tools & platforms that are focused on increasing onchain content grow, Interface will only become more important for discovery purposes. Most people in crypto are quite excited to mint. The issue is that the majority of folks can't afford to spend all their time trying to find the onchain content they care about. It's worth noting that the "like" button we all take for granted today only became a thing in 2009! This single idea completely changed the world. Web2 social platforms got the data they needed to revamp the algorithm from chronological posts to a feed that was based on interests. However, though our feeds got much better (and almost addicting), the data was kept within the platforms' servers. None of us actually got to make use of the content-user graphs. Though Interface didn't invent the "mint" button, I believe you all are the first to display it on a social app that feels very similar to the "like" button. Hopefully, as the next wave of crypto adopters join the club, they all start using Interface and get used to the idea of hitting the mint button just as much (if not more) than the like button. As the # of mints on different kinds of content grows in the next decade, it'll be fascinating to see the open source content-user graphs. Of course, we're just at the starting line and still have a long way to go. I'm so excited to see how Interface expands in the ecosystem and helps grow onchain activity. It's time Capital One gives up their slogan to the Interface team: What's in your wallet? - YB

imo there's nothing more interesting going on than what's happening in the DeAI sector. people are starting to notice and it's only a matter of time before things really start heating up. if you're looking to get up to speed, this framework is how I would approach it. 1) learn the AI technicals & fundamentals. there's some incredible material out there like Karpathy tutorials, 3Blue1Brown neural net series, hugging face tutorials, etc. it's actually not that hard to catch up on the basics. you'll be surprised how much you can get up to speed in just 2 weeks. a lot of the topics discussed on AI twitter will start making a lot more sense not saying you'll be an expert by any means, but you don't need to worry about that...just get started. for example, I try and listen to all the Dwarkesh and No priors episode - that doesn't mean I understand every detail but it gives me a general idea of what the "SF AI cabal" is discussing and what's top of mind for them. 2) Open Source AI community take some time to go through the main companies in the AI space that are truly open source first. some examples are hugging face, together AI, EleutherAI, AI2 (allen institute), and StabilityAI. just going through their blog posts and better understanding how they're positioning themselves is a great start. Hugging face and stability business models are also interesting to look into - basically a hybrid model of selling "ai saas" and enterprise in order to boost OS r&d. the best newsletter to follow would be Interconnects by @natolambert, he has a whole section / podcast covering this stuff. 3) DeAI Efforts Your goal here should be to create a "golden basket" of DeAI companies you would personally invest in. If you had to pick 5, which ones would they be? I'd recommend by going through what Nous Research, Prime Intellect, GenSyn, Vana, Bittensor, etc. are all up to. Understand how their strategies differ, what part of the stack they're working on, etc. Why is Nous going with DeMo? Why is Prime going with DiLoCo? Are you bullish on SWARM? Coinbase Ventures has a fantastic DeAI market map that you can check out to help guide you as well. Which companies are venture backed? What is their strategy in terms of competing against hyperscalers? How much have they implemented and how much is still theory? I'd start by going through the fantastic primers @Ronangmi and @SPLehman have put together. 4) Open Source history & philosophy I'll keep this section short. Technically, it's not a requirement but I think by spending time learning about the history of open source, I'll have a much better grasp of why the fight for DeAI even matters in the first place. The truth is that I'm not old enough to remember the Linux days or early 2000s era of computing when open source was the norm. I grew up in the FB / Twitter / Uber era and centralized platforms have always been the default for me. That's why I want to better understand what happened in the past with open source and what I can learn from those examples. The other day, I was listening to the BG2 podcast, and @bgurley was comparing Meta's Llama strategy to what MongoDB and others did back in the day. I thought that was fascinating and I really want to understand the nuances there. I think this will help me not only discover unique insights in this vertical but will also help me feel philosophically more aligned with becoming a subject matter expert on crypto x AI. I'd start with @nayafia book Working in Public and also the Cathedral and Bazaar essay.

key metric to watch for ai16z is the number of forks on Eliza repo (rn at 270). just looked this up but in 1994 there were ~3k websites and by 1997 there were ~1m sites. you can think of these agents as websites: over time, everyone will need them to stay competitive. as early projects effectively showcase diff use cases (community manager, trading, devrel), a lot more founders & creators will join the trend. watch closely as to what kinds of projects are building on top of Eliza and if they're doing anything unique. These will be the early playbooks that many other folks will copy very soon. It's not crazy to think that agent growth will go bezerk in the next few years. this will lead to more high quality devs contributing to the Eliza repo and adding their own functionality (i.e. Farcaster channels, Coinbase and Circle integrations). it's clear that agents are here to stay so now the next focus should be focusing on improving the infra for agent setup (i.e. no code tooling). don't get me wrong, the DAO part is cool but that'll be overhyped in the short term. Rn, I see the venture fund as a perfect marketing mechanism to drive attention to the core product. But as we all know, prices have to reset sooner or later. To me, the magic is @shawmakesmagic and the team continuing to expand Eliza framework over the next year. The agents are already blowing my mind so I can't imagine how all this will look like very soon.

7/ I thought by far, Senator Hagerty made the most convincing case as to why the crypto industry should NOT be attacked. He also mentions the data analysts (Elliptic) that retracted their research that the WSJ wrongfully cited. "The numbers are grossly overstated"

s/o to @austin_hurwitz for killing it on Twitter recently - awesome hire for the @doodles team! great job dood 💯

The best follows by far for this agentic meta are: - @0xsammy - @kelxyz_ - @terrytjw - @cryptopunk7213 - @0xjeff Just follow these guys and that’s about it

Excited to announce the Fall 2022 @chapterone student researchers! If you're a founder & want to collab with our researchers, please comment your product below! Mentors: @j_austincain @adeets_22 @anishagnihotri @OthmanGbad @sandraleow Learn more about the cohort ⬇️

10/ Senator Lummis does provide a strong case to figure out how to better regulate Binance and Tether. They are not in US jurisdiction and are bringing the most problems to crypto illicit financing. It would be silly for crypto to be banned in the U.S.

I went though all 2,528 tweets from the @doodles account. The first tweet was on August 20th, 2021. Here are the most iconic ones:

If you’ve survived CT 4+ years you deserve some kind of award for sure

2021 has been one hell of a year for me. 1) Embraced the nomad lifestyle & stayed in 7 cities 2) Quit my job in web2 3) Wrote 30+ threads on web3 4) Recently signed a new role in web3 (sharing soon 👀) For my last thread of the year, I wanted to share my 5 favorite posts 👇

bullish that the @yapsterxyz winner didn't get deployed on raydium today, good quality control by players the memes need to be non-yapster related here on out or else people will get bored quickly. also worth remembering that only a fraction of pump fun tokens actually bond.

lmao @zora launch was 10/10 timing

For most people in this world, a couple thousand dollars is life changing money. Is there a way to reward everyone with a chance to win that money risk free? Yes - through @PoolTogether_. Let's see how👇