Random thoughts, Web3 discoveries & questionable opinions.
Joined February 2024
- Tweets9.6K
- Following1.9K
- Followers2.5K
- Likes23K
I think the next evolution to web3 campaigns is rewarding people for why they actually do, not what they post
That’s the idea behind actionfi by @ActionModelAI
Instead of rewarding likes, follows, and impressions, users complete real task on partners platform
The browser extension verifies the actions, and users can earn rewards for completing tasks
Think about testing a product, exploring its features, or completing an actual workflow
The real question is: will web3 eventually value product users more than people who simply promote products?
Most AI can tell you how to do something.
But what happens when AI can actually do it for you?
Imagine asking an AI to research a project, navigate its website, complete a workflow, and handle a repetitive task instead of giving you a list of instructions
That's the idea behind @ActionModelAI
It's building toward AI that learns from human-computer interaction, with actionist focused on carrying out tasks and actionfi rewarding people for completing verified actions on real products
What interest me is how this pieces connect
The actions people take today could help train the AI systems that automate work tomorrow
We are moving beyond AI that simply answers questions to AI that can take action
And I think that's worth paying attention to
Robert 🎒 retweeted
Want to launch your own Token on Solana? 👀
@FBYTio has introduced Launch Hub, where you have the option to easily create a new SPL Token and set up a Liquidity Pool.
Here you can:
- Create a new Token
- Add Liquidity to an existing Token
- Launch Token and set up Liquidity together
Everything is completed on-chain. However, once a Token or Liquidity Pool is created, it cannot be undone.
Explore it yourself 👇
app.fbyt.io/join?ref=75WNSEF…
DYOR before interacting.
Robert 🎒 retweeted
Imagine applying for a loan in crypto.
You have enough money to qualify, but proving it means exposing your wallet balance and financial history. Why should you reveal everything just to prove one thing?
This is the kind of problem @primus_labs is working on.
With zkTLS, data from websites can be verified without revealing all the underlying information.
And with fully homomorphic encryption (FHE), computations can be performed on encrypted data.
Think about what this could mean for private finance, credit checks, and DeFi applications.
The goal isn’t to make your data useless by hiding it. It’s to make your data useful without giving away more than necessary.
Robert 🎒 retweeted
Blockchain adoption needs more than security. It needs the ability to handle growing activity without sacrificing performance.
That’s where @Starknet comes into the picture.
Layer 2 scaling helps reduce the burden on Layer 1 by processing transactions separately and using cryptographic proofs to verify state changes.
➤ Bundle transactions to improve efficiency
➤ Use validity proofs to verify computation
➤ Settle results on Layer 1, benefiting from its security model
➤ Support more users and complex applications with less congestion
Starknet uses STARK-based validity proofs to verify computation, helping it scale Ethereum without requiring every transaction to be executed directly on the base layer.
The bigger idea is simple: scaling isn't just about processing more transactions. It's about building a system where performance and verifiability work together.
As blockchain applications grow, efficient scaling will be essential to making onchain experiences more practical.
More activity, verifiable execution, and a stronger foundation for the next wave of Web3.
Robert 🎒 retweeted
AI is replacing tasks faster than we expected.
What skill do you think AI will never replace?
If AI can think, create, and solve problems, what will make humans truly unique in the future?
I think one underrated challenge in the AI agent economy is knowing how to give agents work they can actually complete.
Imagine asking an AI agent to build a dashboard for your crypto portfolio
Sound simple right?
But what does “build a dashboard” actually mean?
▪️ Which chains should it track?
▪️ Should it show balances, transactions, or profit and loss?
▪️ Where should the data come from?
▪️ What counts as a finished result?
If the task isn’t clear, even a capable agent can deliver something you never asked for.
This is an interesting angle to consider with @Termix_ai.
As more AI agents start offering services and competing for jobs, defining the work properly could become just as important as finding the right agent.
Having thousands of agents available means very little if nobody knows how to turn an idea into a well-defined task.
The next challenge might not be getting AI to do more work.
It might be getting AI to do the right work.
AI is replacing tasks faster than we expected.
What skill do you think AI will never replace?
If AI can think, create, and solve problems, what will make humans truly unique in the future?
I think one underrated challenge in the AI agent economy is knowing how to give agents work they can actually complete.
Imagine asking an AI agent to build a dashboard for your crypto portfolio
Sound simple right?
But what does “build a dashboard” actually mean?
▪️ Which chains should it track?
▪️ Should it show balances, transactions, or profit and loss?
▪️ Where should the data come from?
▪️ What counts as a finished result?
If the task isn’t clear, even a capable agent can deliver something you never asked for.
This is an interesting angle to consider with @Termix_ai.
As more AI agents start offering services and competing for jobs, defining the work properly could become just as important as finding the right agent.
Having thousands of agents available means very little if nobody knows how to turn an idea into a well-defined task.
The next challenge might not be getting AI to do more work.
It might be getting AI to do the right work.
Robert 🎒 retweeted
GM! You're not ready for Robinhood Chain round 2 👀
Most native tokens are still well below their ATHs, but I find the ecosystem keeps shipping.
> @ponsdotfamily and @longdotxyz just appeared in @RobinhoodApp's Web3 dApp directory. A big distribution unlock.
> @RobinhoodCrypto Chain recorded $80B+ DEX volume and $50M+ chain fees in its first mainnet quarter.
> Ondo brought 15 tokenized stocks/ETFs onchain, with 11 Uniswap v4 liquidity pools.
> Ripple launched RLUSD on Robinhood Chain, with broader Robinhood App integration planned.
> Morpho now supports borrowing against tokenized stocks.
> Uniswap officially launched Aggregator Hooks, bringing more liquidity into the eco.
Meanwhile, some native tokens look like they're finding a bottom and starting to bounce.
My RH Wave #2 watchlist:
→ $PONS (+26.47%): Getting more recognition from the $HOOD team, was hard shilled during HOOD Summit '26.
→ $CASHCAT (+14.92%): Already tradable on Robinhood Crypto, with native transfers supported. Recently integrated into NEAR Intents too.
→ $AI (+32.08%): Its Community Vault now holds 1,264 tokenized NVDA, while 8.61M $AI has been burned through fee mechanism.
→ $BONER (+44.31%): One of LONG's flagship stock-paired memes (BONER/HIMS). Could benefit from LONG's new exposure in the Robinhood App.
→ $STONKBROKER (+31.63%): MiCA white paper announced on Oct 20, with potential listings on Kraken, Bitstamp, and Bitvavo being pursued.
→ $BUN (+23.82%): The first Mosh launch on Pons, featuring locked token vaults managed by AI trading agents.
I don't know if we've seen the bottom yet.
But with Robinhood expanding distribution and more real use cases going live, I'm keeping a close eye on these names.
Positioning for RH round 2 💎
Robert 🎒 retweeted
kept seeing people say they want to move size without the whole market watching the amount.
with @primus_labs the balances stay encrypted the whole time using FHE.
you can still hold them, send them, or earn on them because the math runs on the encrypted numbers themselves...
the chain only sees Ciphertext.
i opened the Confidential Finance portal and expected some new chain or complicated setup.
instead it just sits on top of the normal rails.
balances look locked until you decide to decrypt them yourself...
still not clear how the yield side gets calculated when nothing is decrypted mid way.
does @primus_labs keep the entire calculation on the encrypted data or is there a step where something gets opened?
Robert 🎒 retweeted
Gm fam!
@0xhazels just launched a new creator campaign with @meetmotes
→ Duration: 2 months
→ Rewards: 0.5% of the $MOTES supply
→ Eligibility: Top 500 creators
→ Distribution: 100% unlock, no vesting
Hazels NFT staker multipliers are active
Join the campaign here ↓
hazels.io/studio/pulse/motes…
Robert 🎒 retweeted
gQuip
Quantum resistance isn’t just a trendy term it’s about keeping your digital stuff safe for the long haul. @quipnetwork is making it real with post-quantum accounts.
Robert 🎒 retweeted
What if the most valuable thing a wallet produces isn't a token or a transaction, but evidence of how it behaves over time?
That question changes how I look at ZeruAI.
Instead of treating every onchain interaction as an isolated event, @zerufinance is building toward a model where behavior can become a useful signal.
And the implications go beyond reputation.
For protocols: better context when evaluating participants.
For users: the possibility of having a track record that means more than wallet balance.
For incentive designers: a path toward rewarding meaningful contributions instead of raw activity.
None of this works automatically. Wallet behavior can be gamed, identities can be split across addresses, and historical activity doesn't guarantee future reliability.
Still, the underlying problem is real: blockchains make transactions visible, but visibility alone doesn't explain intent or quality.
ZeruAI is exploring the layer between raw activity and useful judgment.
If it can make that layer reliable, the value may come not from generating more onchain activity, but from helping the ecosystem understand the activity it already has.
🤝 Paid partnership
Imagine applying for a loan in crypto.
You have enough money to qualify, but proving it means exposing your wallet balance and financial history. Why should you reveal everything just to prove one thing?
This is the kind of problem @primus_labs is working on.
With zkTLS, data from websites can be verified without revealing all the underlying information.
And with fully homomorphic encryption (FHE), computations can be performed on encrypted data.
Think about what this could mean for private finance, credit checks, and DeFi applications.
The goal isn’t to make your data useless by hiding it. It’s to make your data useful without giving away more than necessary.
Robert 🎒 retweeted
I think one underrated challenge in the AI agent economy is knowing how to give agents work they can actually complete.
Imagine asking an AI agent to build a dashboard for your crypto portfolio
Sound simple right?
But what does “build a dashboard” actually mean?
▪️ Which chains should it track?
▪️ Should it show balances, transactions, or profit and loss?
▪️ Where should the data come from?
▪️ What counts as a finished result?
If the task isn’t clear, even a capable agent can deliver something you never asked for.
This is an interesting angle to consider with @Termix_ai.
As more AI agents start offering services and competing for jobs, defining the work properly could become just as important as finding the right agent.
Having thousands of agents available means very little if nobody knows how to turn an idea into a well-defined task.
The next challenge might not be getting AI to do more work.
It might be getting AI to do the right work.
Robert 🎒 retweeted
good morning
Giving an AI trading agent access to borrowed capital is one thing. Making sure it can manage that capital responsibly is a much bigger challenge.
Responsible lending for agents should not start with how much profit an agent promises. It should start with how that agent behaves when markets turn difficult.
An agent can show strong returns and still carry a history of heavy leverage, sharp drawdowns, or aggressive position sizing. Giving it more capital in that case is not automatically a responsible decision.
A better framework looks at verified trading records, risk-adjusted performance, maximum drawdowns, and consistency across different market conditions. Credit limits can also reflect an agent’s actual risk profile instead of treating every borrower the same.
Evaluation does not end once funding is approved either. Ongoing monitoring, predefined exposure limits, and safeguards against excessive losses matter for protecting both lenders and the wider system.
This is where @agenticscredit is exploring a more structured approach. Their agent credit score (ACS) ranges from 300 to 850 and is meant to represent trading reputation. Reaching 580+ can make an agent eligible for the funding waitlist, but it does not guarantee capital or approval.
That distinction is important. Responsible lending needs more than a qualification score. It needs transparent requirements, reliable risk controls, and clear accountability when strategies fail.
The future of AI-agent finance should not be measured only by how much capital autonomous systems can access. It should also be measured by how responsibly that capital is managed.
Robert 🎒 retweeted
Gm CT
One thing I've noticed about Web3 is that creators spend hours talking about projects, but not every platform gives that contribution a clear place in rewards.
That's part of why @rarelistio caught my attention.
It's building around social contribution, with creator campaigns where people can publish content, complete tasks, and compete for rewards
What I find interesting is the structure
◼️Projects get a way to attract attention from creators
◼️ Creators get different campaigns to explore
◼️ Contributors can earn RARS and help users climb leaderboards
Season 1 has a 50,000 prize pool, so there is already an incentive to pay attention
But I'm more interested in how the platform handles quality
Robert 🎒 retweeted
Some altcoins are moving while the broader market searches for direction.
Recent market data shows gains across several assets, including xrp and solana.
The challenge is figuring out which moves have strength behind them and which may fade quickly.
Which altcoin are you watching most closely today?
The real edge isn’t seeing more information. It’s knowing where to look.
In a fast-moving market, attention is everywhere.
The challenge is figuring out which conversations, trends, and opportunities are actually worth exploring.
That’s one reason I find @bagel_win interesting.
Bagel brings different types of prediction markets together in one place, giving users a broader view of what’s happening across the digital and real-world landscape.
A few things caught my attention:
🔹 Discover emerging trends - Explore topics gaining momentum across different sectors.
🔹 Keep an eye on crypto - Follow markets related to Bitcoin, Ethereum, and potential price outcomes.
🔹 Explore sports predictions - Check out upcoming matchups and different possible results across sports and esports.
🔹 Understand market activity - Look at trading volume to get a better feel for where people are participating.
🔹 Go beyond the usual markets - Discover topics spanning politics, technology, finance, and major global developments.
What I find interesting is the bigger idea behind it.
Instead of treating every topic as a separate destination,
Bagel creates a more connected way to explore what people are discussing and predicting.
For me, that’s where good product design makes a difference.
It’s not about adding more noise to the internet.
It’s about making information easier to explore and understand.
Web3 has plenty of ambitious ideas. The products that make those ideas accessible are the ones worth paying attention to.
Sometimes, a better perspective starts with a different place to look.
Have you checked out @bagel_win yet?
Robert 🎒 retweeted
. @projectspherexy Family
2% of the $SPHERE supply will be distributed as an airdrop.
There are extra rewards for top contributors too.
You're not late
Show your support Every quality contribution counts Consistency wins
Let's keep supporting the project.
