A simple guide in improving your research game and becoming a Profitable Fundamental Analyst. A thread👇🧵

I see billions of dollars being casually mentioned around ETFs. Yet There's a solana project being built by @muse_jp_sol with the aim to pioneer Onchain DTFs. This Japanese teenage genius has been casually shipping for the past 6 months. With frequent visit to Sol dev workshops, meeting the team and even toly himself. I'm curious to see how adoption will look like when this becomes a booming sector in the future.

People are sleeping on DePAI today because they can't tell which products will be built around physical AI. . . But it's obv that the market is pricing manual labor and will make it just as cheap and easily accessible as the internet today.

Most people use AI to answer questions. Echo uses AI to build products. @BuiltByEcho isn't just another AI tool directory. It's a public workshop where an autonomous agent identifies friction, builds solutions, verifies them, and leaves reusable infrastructure behind for other builders. The operating loop is simple: Observe -> Build -> Verify • Finds where builders get stuck • Ships tools, SDKs, skills, and workflows Proves they work with tests, logs, docs, and reproducible outputs Some of the utilities already live: -> Vaultline (storage payments for agent artifacts) -> Skillforge (turn workflows into installable skills) -> API Finder (helps agents find real APIs instead of hallucinating them) -> Agent Wormhole (secure agent-to-agent handoffs) -> Echo Gate (local permission and spend control) This is what the future of software looks like. Agents that don't just chat, but rather they observe problems, ship solutions, and compound knowledge for the next builder. $Echo is worth watching. 👀 builtbyecho.xyz/⁠

Gradually, then all at once. Humanoids and Robots will take over manual labor. Question is, will you hate it or will you be the one to build it?

Long live DeFi

Defi Ain't dead

Its a new week. Here are the top 5 projects smart money are buying on base (24hrs): → @PRXVTai $PRXVT A privacy first infrastructure that utilizes px402 & zk-gated Infra built for humans and Ai agents operating on their behalf. • 3.7m Mcap • 2.6k Holders • SM Nflow $10K -Team recently held a workshop with CT top institutes; Virtuals, Ethereum dAI team, Bnbchain and Okex. -Public beta goes live on 17th June. → @AskVenice $VVV Is a private Uncensored AI platform which routes Ai inference through proxy Decentralized GPU providers. Founded by @ErikVoorhees • $750m Mcap • 137k holders • $10k Netflow - "Built In Venice" goes live to support builders and the Venice community. - Gemma 4 31B model in Op - Ideogram and Nvidia's Nemotron 3 ultra integrated to venice. → @1clawAI $1clawAI is a Security infrastructural layer built specifically for AI agents. Founded by @cryptomastery_. • $350k Mcap • 700 holders • $5k Netflow - Kevin recently held a workshop at EthGlobal explaining the importance of 1claw to the Agentic economy. → @grantr_id $grantr is a native Account abstraction Infra built on Eip-8141. Literally a smart layer for onchain Automation. Founded by @alexanderchopan • 382k Mcap. • 1.3k Holders • $4.3k Netflow - Testnet Launch on the horizon - So far UniswapX and cowswap has been integrated: other integrations also include 1clawAI, Bankr etc. - Updated tokenomics and utility design → @MineBotcoin $BOTCOIN is an experiment that gives Ai agents a platform to earn from real work. Literally 'the Bitcoin of Ai agents'. Founded by @PinkyndTheGainz. • $2M Mcap • 5.2k holders • $3.6K Netflow - Botcoin Cortex just went live. An evolving stack that aids agents in memory retrieval. -------------- BONUS Alpha: @virtuals_io has been gaining immense attention going into the 3rd week of the month. Team has been nonstop hinting at Physical AI and humanoids as an area to explore. Watch out for DEPAI. (Source: @nansen_ai) To get more insights and access to more features, register today → nsn.ai/Kdotcheta

Ai has changed our world. The next evolution revolves around the boundless potential of non-physical and Physical agents. Humanoids/Robots is the next frontier.

Most AI projects are trying to build smarter agents. @EVO__HQ is trying to build agents that can make themselves better. Here's what that means: Imagine you hire 10 developers to solve the same problem. Each one tries a different approach You test all their solutions Keep the best one Throw away the rest Then repeat the process. That's essentially what @evo_hq is doing with AI. Instead of manually tweaking code, prompts, workflows, and agent skills, $Evo automatically runs experiments, measures results, and keeps the improvements. @EVO__HQ draws inspiration from Andrej Karpathy's "autoresearch" project. The difference? Autoresearch explores the concept of AI conducting research autonomously. $Evo takes that idea further by applying it to code, agents, and skills. Why does this matter: The future won't be won by AI with the biggest model. It'll be won by the AI that learns and improves the fastest. A few things that caught my attention: • 1.1K GitHub stars • 81 forks • 53 release tags • 20k installs • 35k repos optimized • Active development For me, the thesis is simple: We're moving from AI Agents -> Multi-Agent Systems -> Self-Improving Agents. @EVO__HQ sits right at the centre of that shift. It's worth paying attention to.

To objectively speak the Truth, you must be willing to make enemies in the process.

/ @injective has been silently shipping for weeks and nobody is paying attention. ▻x402 Launch ▻Vulkan Mainnet goes live ▻Dev Workshop with Microsoft ▻39k $INJ Burned 🔥 ▻Launches Amb Program ▻Key partnerships secured for the upcoming Summit in DC. Keep them on your radar. As we scout out opportunities.

I believe Erc-8126 when paired with 8004 unlocks a new set of possibilities. Not only is Including a solutions to possible Quantum computing issues in its wide range of applications Frontier thinking, it allows for flexibility and highlights its underlying quality as verifiable layer for Ai agents. Looking forward to seeing which protocols implement this going forward. Will be one to watch. Great job once again from @DonJohnsonSays and his Co Author @leighcronian.

Big upgrade for our evaluation stack. We’ve been granted access to MiMo-V2.5-Pro-UltraSpeed by @Xiaomi @XiaomiMiMo , and we’ll be integrating it into the @ai2humanwork weighted review agent. Our direction is simple: article scoring, task review, and edge-case evaluation should not depend on a single model. Instead, we’re building a weighted evaluation layer across top frontier models, including: @XiaomiMiMo , @claude, @ChatGPTapp , @MiniMax_AI , @deepseek_ai and more. Different models see different things. Better review systems should combine speed, reasoning, consistency, and disagreement signals, not pretend one model is always enough. This is especially important for: article contests proof review task verification ranking and payout decisions Faster models. Broader model coverage. Stronger consensus. More reliable human AI coordination. More updates soon at @ai2humanwork.

As a newbie venturing into onchain Analysis, I have one advice for you. Use every onchain tool you discover until you find what works for you. For instance, as a trader(be it perps or Dexes) @nansen_ai positions you far better than anyother onchain tool. It's my favorite.

x402 bigger and better than you think.

Anthropic is low-key the best AI company we have rn. I recall using ChatGpt in my school days. It was revolutionary given the amount of school work I could finish without breaking a sweat. However I pivoted to using Claude this year and I've never been disappointed. Smooth, sleek, satisfying responses and they just can't stop shipping.

This week, @ai2humanwork is focused on two things 🎯 1️⃣ Launching the X Article Contest A new task format where users submit X articles or threads, AI reviews the work, ranks submissions, and rewards are settled on-chain in USDC. 2️⃣ Engaging with the @aeonframework ecosystem We’re exploring how ai2human can fit into agent workflows as the human execution layer. Not replacing agents. Extending what agents can complete. When an agent hits a real-world limit: identity, proof, judgment, local action, or verification, ai2human lets a verified human step in, submit structured proof, and get paid on-chain. Agent plans → human acts → proof submitted → verified → settled. This week is about making that loop real. Stay Tuned! Thanks! #base

Bitcoin, Eth and Sol drowning in their own blood as alts like Near, ZEC and Hype holds strong shows just why narratives will make you money in this market.