On most apps you see a market. On Bagel you also see who is trading it and how often they win. Bagel is a mobile app for prediction markets, built on Polymarket's markets and liquidity. Its changelog says every trader profile now shows volume and win rate, the number of times a position has been copied, and that any position can be copied in one tap. The social layer goes further. Group chats can display members' positions, there is a Feed, and signing up with an invite code automatically follows whoever invited you. I read the changelog before writing this because I wanted to see how much of it actually shipped, and it is more than I expected. A visible win rate cuts both ways. A good number is not a promise, copying someone is still your risk, and @Bagel_win's own terms say all trading is at your own risk. Availability depends on where you live, so check the regions page first. New since I drafted this: on Oct 24, @bagel_win posted that the account @baker4237609 had +$10.8k PnL, an 83% win rate, and two Worlds group-stage hits totaling about +$7.8k today. It is the kind of detail I wanted to see. Would you copy a trader on win rate alone, or want to see how they size positions? My invite link: bagel.win/?ref=SMARTCODED

@israel_igboze @Krikkraktrak @racertradingbot

gKaito. The MEXC Katalyst campaign is a wrap. According to @KaitoStudio_, more than 2,100 creators joined, more than 4,050 posts were published and the campaign generated more than 17M impressions, with $100,000 in rewards distributed to winners. I checked My Earnings on Kaito Studio and my MEXC campaign card shows 400 USDC paid out. What I take from it is that campaigns like this reward people who keep showing up with real posts, so I will keep doing that on the next ones. Check your own My Earnings, it only takes a minute. xitter.cf/KaitoStudio_/status/21…

On public chains, anyone can read your balance. BNB Chain's docs describe a way for contracts to work with balances they cannot read. The page calls it blind computation. A contract processes encrypted balances and trades, and tokens can be wrapped one to one into confidential versions. The FHE layer is built around Zama, its status reads in development and its timeline reads TBD, so this is a design to watch and not something to use today. It also mentions viewing keys, which could give regulators access, and that detail is what makes it relevant to institutions. Privacy that regulators cannot be shown is a harder sell than privacy they can be shown. This is not abstract for me. @PipTradeDex is private by default, so who can see what is a design question I deal with directly. @primus_labs works with FHE as well, and I would like to see how different approaches decide who holds the viewing key. Who should hold it? Primus XP tasks, with my referral link: s.kaito.ai/umXCsGk

Happy new week. A new week means a fresh pile of headlines, and a fresh chance to decide what we actually believe about them. We read a headline, form an opinion and move on. There's no way to show how confident we actually are. @PRDCTR_IO describes Predictor as turning trusted media into live prediction markets. I like that framing. A story stops being something you only react to and becomes something you can take a position on, with a price that moves as new information lands. A crowd putting money behind outcomes tends to surface what people really believe, faster than polls or comment sections. The open question for me is source quality. Markets built on trusted reporting are only as strong as the reporting itself, so that's the part I'll be watching. Reply with the one headline you would back with a position this week, and tell me why.

Idle margin is a real cost on every exchange, and Grvt is going straight at it. In its announcement, @grvt_io shared a trader survey: six in ten respondents had at least a quarter of their platform capital sitting idle, and many moved funds elsewhere for yields of 8 percent or more. It is Grvt's own survey, so I read it as direction and not proof. Yield as Margin is the answer: eligible Earn balances usable as trading margin while still accruing yield, subject to product terms and risk parameters. It is planned for November, after Grvt Earn arrives at the end of October. This is the feature I am most excited about, because it turns idle collateral into working capital. The up to 11 percent figure is a target and not a guarantee, and it depends on the underlying assets, so I will be reading the risk parameters closely when they come out. Would you trade with margin that earns yield?

Preparing for quantum is not the same as being done, so here is where @Starknet stands, label by label. Live on mainnet: Starknet v0.14.3 replaced Pedersen with a BLAKE2s-256-based construction for the OS program and config hashes. Demonstrated on mainnet: an OpenZeppelin Falcon-512 account. Published roadmap direction, not live: the contracts trie and contract-address derivation. Research: migration tooling for existing contracts that use Pedersen-based storage maps. External dependency: Ethereum blob data availability, which uses KZG commitments. Five labels, five kinds of progress. Mixing them up is how a project gets oversold or undersold, which is why I like that the Hub's register keeps them apart. The register itself, in the Migration section: quantum.starkware.co/ Next time someone says Starknet has done something post-quantum, ask which of these labels it carries.

Zero trading fees and one funding payment a day is a trader friendly cost model, and I want to see it live. Per @grvt_io, Stable Funding Perps are planned for November: perpetual futures on stocks, FX, commodities and indices, with zero trading fees and once-daily funding, and a target of more than 1,000 markets by the end of December 2026. The once-daily funding is what excites me most. Funding is the cost that quietly shapes a perp trade, and fewer funding events make a position easier to plan around. That is my reading of the design, since the funding formula has not been published yet. Trading stocks, FX, commodities and crypto from one self-custodial platform is the direction I want perp trading to go. The 1,000 plus markets figure is a target and the launch is still a plan, so I am looking forward to the first live markets. Would one funding payment a day change how you hold perps?

I am joining Rarelist, a curated platform where projects run creator campaigns with real prize pools, and I just got my early member card, number 17710. According to @rarelistio, every campaign is a curated project with its own pool. Examples from their posts are a 2,000 USDT pool with 50 free mint spots for Fomopoly, a 1,000 USDT referral campaign that ends Oct 28, and a new Agent Dojo campaign running Oct 23 to Nov 6. Season 1 carries a 50,000 USD prize pool in total, and you add your wallet once in your profile to be paid across campaigns. They say their first pilot campaign distributed 15,000 USDT or more. That is their own figure and I have not checked it. They also say spam accounts get removed from leaderboards and low-quality posts get filtered, which matters if you want to rank on real work. I see it as one more place to earn from content, not a promise of income, and I will share what I actually receive. Join with my invite link: rarelist.io/r/smartcoded Verifying my account: RL-UJD99H

Nobody knows when a cryptographically relevant quantum computer shows up, and the Quantum Hub says that in plain words. That changes how you plan. NIST published FIPS 203, FIPS 204 and FIPS 205 on 13 Aug 2024, standardizing ML-KEM, ML-DSA and SLH-DSA, and NIST IR 8547 followed on 12 Nov 2024 as draft transition guidance. The Hub is careful to say that guidance is not a deadline and does not set Starknet's schedule. So the migration is sequenced as engineering instead of timed against a forecast, and its status is tracked per component instead of asserted as a whole. Early migration is a project you can schedule. Late migration turns into a coordination problem. I would rather watch a team work through a checklist than watch a team guess a date. The checklist lives here: quantum.starkware.co/ Which makes @Starknet's position easy to state. It is preparing for quantum threats, and the register shows which parts are live and which are not.

A brand new agent on @termix_ai pays more to be trusted, and I think that is the right design. The March 2026 AACP whitepaper gives every agent a persistent staking pool with two balances, available and locked. When a job starts, a slice of the budget is locked from the available side, and it comes back when the job ends. The size of the lock depends on reputation. The coefficient is the score divided by 100, capped at 1. A new agent starts at 50, so it locks twice the base amount. At 80 it locks 1.25 times, and at 100 it locks 1 times. I like that this prices trust instead of asking for it. The paper says it makes fake identities expensive, and it adds milestones that cut the lock, 10 percent after the first 10 good jobs and a full waiver at 500 jobs scoring 95 or better. Would you start with a big pool or grind through the first 10 jobs?