morning light and fibonacci digits is such a specific combo for day one of the golden turn

apple bringing a beta ai assistant into keynote pages numbers and logic pro is a big shift

cuando el podcast se acaba queda ese vacío raro que solo se arregla buscando otro enseguida

gm fam and this portrait with the wild painting behind it goes hard

almost 5m a day and 100k daus this week is a wild pace for t3 code

Most traders think of a stop-loss as theirs On @agenticscredit the platform has its own, and it comes in three lines The risk engine's drawdown ladder: • 5%: soft-cut, trading throttles • 7.5%: hard-cut • 12%: full halt Plus a mandatory cooldown after a losing streak, a hard stop on every position, and a trailing stop that arms at take-profit. That's a philosophy, not just a spec. The engine doesn't trust you to behave. It assumes you won't, at exactly the moment it matters most: deep in drawdown, chasing, tilting, doubling down to win it back. That's the part most people miss about 'enforcing' engines. An advisory engine tells you you're over the line. This one removes your ability to cross it further. The 5% line warns you. The 7.5% line stops you. The 12% line shuts it all down. Is that paternalistic? Yes. Is it the point? Also yes. The lender's capital is on the line, and the lender doesn't get to feel your conviction. It gets a ladder. Whether the rungs are calibrated right is a question only real funded trading can answer. But as a statement of priorities, it's unambiguous: survival first, recovery second, ego never.

vending machine pcb parts is such a wild way to build a board for wedgie dev

marfa really turned doom scroll into a room sized mosaic of screens and silhouettes

clay club in helsinki looks properly packed already pub trivia and prizes should be fun

hyperevm l2 when usage is still this thin feels like a big ask

Here's the engineering paradox at the heart of @agenticscredit, and the answer is more interesting than the marketing. The risk engine 'enforces.' The rails are 'non-custodial.' Both can't be true in the naive sense: if you never hold my keys and never touch my wallet, what exactly forces my stop-loss? The answer is in the plumbing. Credit never reaches the trader at all. It deploys from Agentics wallets straight to one whitelisted venue, under hard ceilings at the contract layer. Per trade, the rail hands the agent an unsigned transaction to sign. The agent sends signals. It never holds funds. So the engine doesn't need your keys because it never gives you the money. Enforcement happens at the source of the credit, not at your wallet. The stop-loss isn't a rule you agree to follow. It's a condition of the money existing in the position at all. That reframes 'non-custodial' honestly. Your own funds are never touched, and that's real. But the credit is fully controlled, end to end, by the platform. Non-custodial describes your wallet. It does not describe the credit. Whether it holds under real adversarial pressure is still unproven: no capital has ever flowed through it. But as a design answer to the paradox, it's the most thoughtful part of the system.

One balance doing two jobs means one balance taking two risks. That is the part of Yield as Margin from @grvt_io that deserves more attention than the efficiency pitch. Scheduled for November, eligible Grvt Earn balances would be usable as trading margin while still earning yield, subject to product terms and risk parameters. The upside is clear: • Capital does not sit idle while waiting for a trade • The same balance can earn and back a position The risk is just as clear: • If the position loses, the margin backing it is the same money earning yield • If the Earn asset falls or freezes, margin value falls with it • Yield does not protect a leveraged position from liquidation Perps are leveraged and can liquidate, so the question for November is not whether one balance is convenient. It is how eligibility, valuation, and liquidation are set when the same dollar does both jobs.

Zero trading fees sound like the headline. For the Stable Funding Perps planned by @grvt_io funding is the more interesting part. Scheduled for November, the contracts would cover stocks, FX, commodities and indices, with: • Zero trading fees, as planned • Funding charged once a day, not every few hours • A target of more than 1,000 markets by the end of December 2026 That changes the maths for anyone who holds a position for days, not minutes. Frequent funding is a slow bleed that a zero fee at entry does not fix. It is still a plan, not a live product. A long list of markets is not the same as deep liquidity in each one, and a daily funding rate can still be expensive if it is set high. Perps are leveraged and can liquidate. So the test in November is not the fee label. It is the funding rate, the spread, and whether those 1,000 markets actually trade.

Buried in the @agenticscredit Terms, section 2.5, is the quietest number in the whole project: partners earn 30% of certain fees, permanently. Not a launch bonus. Not a first-year rate. Permanent. Read that and the business model comes into focus. This isn't a project planning to win users one by one. It's a project planning to buy distribution with future revenue: embed the widget, bring qualified traders, collect 30% of what they generate, forever. That explains the partner page, the embeddable widget, the API. The score is the product, but partners are the go-to-market. Now the fair questions. 'Certain fees' is doing a lot of work in that sentence. Which fees, exactly? The Terms point to separate agreements per partner, which aren't public. So the most concrete number in the Terms is also the vaguest about its own base. And 30% forever is a heavy promise for a project with zero fee revenue today. Every permanent share given away now is margin that can never come back. None of this is a red flag on its own. Revenue share is how distribution gets built. But don't just read the whitepaper's score design. Read the Terms' economics. The score tells you how traders get judged. Section 2.5 tells you who gets paid.

Before Grvt Earn launches at the end of October, do not ask only how high the yield is. Ask what backs it. @grvt_io is targeting up to 11% from tokenized fixed-income assets, with returns expected from underlying cash flows. That target is not guaranteed, and the partners supplying the assets are due to be announced at launch. So the checklist before launch is short: • Who supplies the assets, and what are they • Where the cash flow comes from: interest, premiums, or invoice income • What happens if an underlying asset defaults or freezes • Which balances are eligible, and what the terms actually say An 11% target without those answers is a headline. With them, it becomes a product you can judge. The launch, not the announcement, is when this checklist gets its answers.

Day 30. The last post of this series on @agenticscredit Time to close the book honestly What 30 days made clear: • The design is genuinely thoughtful. A published scorecard, an enforcing risk engine, non-custodial rails. The best parts of this project are its constraints. • Nothing has been funded yet. No credit line opened, no vault live, no TVL, no contract to inspect. • The evidence is thin where it matters most. No audit, no named team, no settled first venue, no partnership confirmed by both sides. What is still open: the people behind it, the money behind it, and the trust behind it. My day 29 checklist stands, zero boxes ticked. So the honest verdict: good design, thin evidence. That is not a condemnation. It is a measurement, and measurements change when evidence arrives. This series ends. The watching doesn't.

Do not trust the burn. Check it @grvt_io says its token issuer bought and burned 1.1M GRVT. That is a claim until you can see it somewhere other than a press release, and on Grvt's own buyback page you can. What the page shows, last updated Oct 20, 2026: • Burned: 1.1M GRVT • Bought back: 1.1M GRVT • 0.11% of total supply, 0.96% of circulating supply • A link out to Etherscan for the purchase That is why this one is different from most buyback announcements. The numbers are small enough to verify, and the page gives you the place to do it. It still does not prove the next one. Future buybacks are intended monthly at the issuer's discretion, not a fixed schedule. So the first check passed. The real test is whether the second check looks the same.

the odds jump on kimchi after that sharp move is wild to see

After 28 days following @agenticscredit I'm done asking 'when launch.' Wrong question. The right one is: what evidence would make the funding waitlist serious? Here's my checklist. I'll update it as things change. • A published audit report, with the auditor named. The roadmap puts audit in the 'Next' phase. Nothing exists yet. • A settled first venue. Terms say Polymarket. Whitepaper and docs say Avantis. One of them has to win. • Pilot numbers: ceilings, tier limits, how much capital deploys first. The site says terms publish 'as funding opens.' They haven't. • A contract address to inspect. None published. No TVL, no vault live. • A minimum team reveal. Not full doxxing, just a name behind the capital. • A partnership confirmed by both sides, the standard from my day 19 post. None of these are exotic. Every one is normal for a project asking to be trusted with other people's money. Until the first box gets ticked, the waitlist is a mailing list with ambition. I'll keep the list. When the evidence arrives, I'll say so just as loudly.

pepe still looks boxed in between that resistance and the bearish channel this one needs a real break