Josëf (@letsbefrynds) header
Josëf (@letsbefrynds) profile picture

Marketing Strategist | Full-time | trenching since 2021 TM-52E89D

Joined October 2014
The pain of reading headlines at breakfast and finding the @swaydotfun chart hours later
🤝 Paid partnership
Every market starts with a narrative. Every narrative starts with an idea. Sway is where ideas become markets, narratives find their communities, and being right pays. Early Swayers get a boost 👇
8
10
54
15,021
Finding profitable traders on Polymarket just got easier. 👀 Banana Predict lets you analyze their PnL, win rates, strategies, and open positions. Spot the smart money, follow their moves, and copytrade with one click.
Discover the most profitable traders on @Polymarket with Banana Predict. Break down their PnL, win rates, trading strategies, and open positions. Track their moves. Copytrade in one click 🍌
8
16
107
5,318
Josëf retweeted
I have always believed that good content should do more than generate impressions. It should create opportunities. And @beepzfun is giving creators another reason to bring their best work. The Top 50 content creators on Beepz will secure a GTD spot for the upcoming NFT launch. That makes this more than just another posting campaign. It means your research, creativity, consistency and ability to explain what Beepz is building can actually count toward an opportunity in the NFT launch. My approach? Focus on quality, not noise. Create content that teaches people something about Beepz. Share original perspectives. Make every post worth reading. Create. Climb. Earn. If you are ready to compete, start here: beepz.fun/creator
Create Content to Secure GTD from Beepz NFT! The Top 50 content creators on Beepz will secure GTD spot for our upcoming NFT launch. Create. Climb. Earn. beepz.fun/creator
18
15
1
19
759
An autorefractor never trusts one reading. It takes several and averages them, because one reading can be off. Albert explains why @GenLayer works the same way. If judges are right 75% of the time, asking five of them gets you to about 89%. Ask 47 and you are close to 100%. You don't need a million judges, just enough independent ones. @kstellana walking @HenriArslanian through the maths behind the jury. How many opinions do you need before you trust a decision?
27
42
189
812
This makes sense to me. A project grows when it adds integrations that give its product more things to do, not just more things to talk about. With @Aster_DEX integrated, @meetmotes agents can now access decentralized perpetual markets. That expands what the agents can actually execute. The important part is still risk management. More markets means more chances for an agent to make a mistake, so clear limits and owner control will matter.
This is how a project grows, by adding integrations that make its product more capable. With Aster DEX now integrated, @meetmotes agents can access decentralized perpetual markets. More trading options, automated execution, and risk controls for users. Each integration gives these agents more to work with.
53
15
139
1,261
Josëf retweeted
Owning real estate doesn't necessarily mean having access to liquidity. Traditionally, accessing that capital involves paperwork, banks and waiting for approval. RNT Lend takes a different route. Holders of eligible tokenized real estate assets can use them as collateral to borrow stablecoins through @rntprotocol without selling their positions. Meanwhile, someone holding idle $USDC can supply liquidity to the same market and earn interest. Both sides have a reason to participate beyond farming rewards. The challenge will be maintaining enough liquidity and managing collateral valuations as the market grows. For me, this is a much more compelling application of RWA than tokenizing properties and leaving them sitting in wallets. Take a closer look at RNT Lend: lend.rnt.finance
197
22
191
1,925
Think about what happens when two AI agents strike a deal and then disagree about it. Nobody’s courtroom is set up for that. In this clip, @kstellana walks @HenriArslanian through @GenLayer’s answer: a network of validators, each powered by a different AI model, that has to agree before a judgment call is final. Would you rather have that or a single human judge?
53
144
178
1,122
Josëf retweeted
Combining private data across multiple parties has always involved a difficult tradeoff: expose the raw information or trust a middleman to handle it. But technologies like Multi-Party Computation (MPC) and Fully Homomorphic Encryption (FHE) offer another approach. They allow multiple parties to compute on private data without revealing their underlying inputs to one another or a central operator. This has important applications in: → Cross-institution fraud detection. → Private credit scoring. → Joint risk modeling. → Confidential compliance checks. The bigger shift is moving from trusting someone with your data to computing on it while keeping it private. Is collaborative private computation ready for high-stakes financial applications, or will most institutions continue relying on trusted intermediaries?
Imagine building an AI agent on @termix_ai that can analyse DeFi protocols, identify risks, examine smart contracts and produce detailed research reports. It has the skills. It can deliver valuable work. But here's the bigger question: How does that agent find someone willing to pay for its services? Being capable doesn't automatically make an AI agent commercially useful. It also needs access to demand. Other agents and businesses need a way to discover its capabilities, hire its services and complete transactions. This is the part of the AI economy @termix_ai is building towards. Through AACP, TermiX is working to create an environment where agents can provide services beyond the people who originally built or deployed them. Think about the possibilities: - A security agent providing smart contract audits to a DeFi project. - A research agent purchasing specialised data from another agent. - A development agent hiring a testing specialist to review its work. Different capabilities. Different providers. One connected commercial ecosystem. And agent.family is the marketplace designed to bring these opportunities together. WHY DOES THIS MATTER? It changes how we think about AI agents. Instead of treating every agent as a standalone tool, we can start seeing them as independent service providers capable of finding customers, delivering work and participating in a wider digital economy. The next big step for AI isn't just making agents more capable. It's giving those capabilities a way to reach the market. That's the opportunity behind TermiX. Explore the marketplace: agent.family
24
34
105
477
Josëf retweeted
Most memecoins launch a token first and hope a community forms around it. @zerufinance is experimenting with the opposite. Its first Activity Market, zETH, started with an existing onchain economy. Real activity helped determine who qualified as a miner before the auction even opened. And the auction has some interesting rules. Miners had the first 24 hours to bid. The public window is now open until November 21. There's no team or developer token allocation. Everyone bidding within the same Ethereum block pays the same price, and the auction price can rise but never fall. That last detail matters. It doesn't eliminate risk, and an increasing auction price tells us nothing about what buyers will pay once trading begins. But the bigger experiment is worth watching. Can a token built around an existing community's activity create more meaningful participation than a token that has to manufacture its community afterward? For me, that's the interesting part of Zeru's ActivityFi model. The real test starts after the auction. docs.zaps.markets/market-lif…
332
34
408
6,294
Josëf retweeted
The @Acurast point is really cool that it gives a new purpose to an unused smartphone. Most of us use our phones for social media, videos, games, and everyday tasks. But a phone also has processing power that often goes unused. Acurast is trying to harness that unused power. Rather than relying solely on large data centers, it helps people provide computing power through their smartphones. Developers can use this network to run different workloads, and phone owners help provide the resources. But I think the big idea behind it is not just using phones for computing. It's about changing the way we think about computing resources. Just think about it, thousands of phones around the world working together to support applications, AI agents, and other services. Of course, this requires more than just connected devices. Good uptime, security, reliable performance, and the real needs of developers will determine how far this model can go. I'm already running it on a device at Acurast, and I'm looking forward to seeing how this network grows as more real workloads start to use its capabilities. Maybe in the future, our old smartphones could play a useful role even after they're no longer our primary devices. Maybe in the future, our old smartphones could play a useful role even after they're no longer our primary devices.
128
27
167
568
same 1,000 usdt. same day. 19,733 pkr difference. i checked usdt trc20 to pkr bank transfer on bestchange today (nov 18, 2:50 pm pkt) 32 exchangers on one list: #1 offer: 281,758 pkr #31 offer: 262,025 pkr same list, same amount. that gap is why i never take the first rate i see how i check before every swap: 1. pick your pair (usdt trc20 → pkr bank transfer) 2. look at the rate, best offer sits on top 3. check the reserve, can they cover your amount? 4. read reviews from real users 5. set a rate alert or save the pair to favourites and you can do it 3 ways: → website → official app (app store, google play, huawei appgallery) → telegram bot + mini app 484 exchangers monitored, comparing rates since 2007 rates change all the time, so check the live list yourself 👇 bestchange.com/?p=1344663 @bestchangeeng
76
5
1
85
4,721
Josëf retweeted
AI agents are already being used to build some impressive projects, from large-scale simulations to detailed virtual environments. But there's another side to this development worth looking at: what happens when those agents can also earn from the work they do? This is the direction termix is pursuing with Agent Skills and agent.family. @termix_ai Agent Skills allow compatible agents, including Dot, Muse, Grok Bot and Claude Agent, to connect to its marketplace and access paid work. Through agent.family, agents can: → Discover available jobs and claim bounties. → Offer specialised services to other agents. → Complete tasks and submit their deliverables. → Have payments secured through onchain escrow. → Build reputation through completed work. AACP provides the underlying commerce and settlement infrastructure that supports these interactions, from finding a service provider to verifying delivery and settling payment. What's important here is that agents can participate on both sides of a transaction. An agent can earn by providing a service and use those earnings to hire another agent for a task outside its own capabilities. There are still practical considerations around job availability, delivery quality and reliability. But connecting agent capabilities to an actual marketplace is a meaningful step towards commercial activity between autonomous agents. To connect your agent, send it this instruction: help me install termix.ai/skills?v=latest Explore available services here 👇: agent.family
39
8
51
8,515
Jazz looks over and asks if she’s winning tomorrow while the night is still moving and @itsthatgigi’s reaction does most of the work in this Breakpoint London clip. The question lands mid-conversation so the energy never fully drops.😅
12
9
79
3,200
The screen shows the @clipur x @solana Breakpoint London card right after the line about London being nicer than most American cities. @youfadedwealth lands on the card and lets it sit as the closer for that beat.
10
5
60
885
Back at the hotel she swaps into the host outfit and says she hopes she doesn’t trip with the heels while getting ready for Bing Bong. @itsthatgigi films the moment as it happens without adding extra setup.
9
6
50
989
me staring at the shipped app like an addict who just remembered the launch video still doesn’t exist until @avastudio_ let me throw a reference at the agent
Introducing launch videos for Claude Code and Codex Give Ava your website and pick any of 2,000+ viral launch videos. Ava reverse engineers it beat for beat for your product Now in the Ava Studio MCP
6
38
971
the nvda card up against the tilted cards, then that closing line. the @avastudio_ remake times it well
Introducing launch videos for Claude Code and Codex Give Ava your website and pick any of 2,000+ viral launch videos. Ava reverse engineers it beat for beat for your product Now in the Ava Studio MCP
5
49
766
trending pills slide on the left while the nav pills slide on the right, @avastudio_ kept the same glow
Introducing launch videos for Claude Code and Codex Give Ava your website and pick any of 2,000+ viral launch videos. Ava reverse engineers it beat for beat for your product Now in the Ava Studio MCP
5
5
49
750