I think OpenAI’s hardware device, ChatGPT voice home speaker, will do well There’s roughly 300 million smart speakers but we all know how dumb they are ChatGPT’s voice mode is basically magic ChatGPT also has a billion users, so the brand will translate well, Claude for example would have a way harder time doing this Unlike the phone, the home speaker is not deeply tied to a platform like iOS or Android, so easy to replace It integrating with your ChatGPT account also makes it compelling, pick up what you were working on from a different device To add to that, OpenAI has apparently 400 former Apple employees, including their former VP of Product Design, Tang Tan Apple suing OpenAI feels like sour grapes and a sign that they feel threatened

So, arriving at the end of a week, some thoughts on MiCA…. I think we can firmly put one idea to bed: MiCA was never "just a paperwork exercise." Even Binance, the biggest exchange on earth, with more lawyers and compliance staff than most regulators, wasn't able to complete the licensing process in time. If an organisation operating at that scale couldn’t simply "comply and move on", we can probably retire the fairytale that everyone else should have found it easy. The truth is thousands of firms were operating before this deadline, but only a handful made it through. Germany, France, and the Netherlands hold most of the approvals, entire countries have none. And it’s not a coincidence, and it’s not incompetence on the industry’s part either, though I’m sure we’ll hear that said. This is just what happens when a regulator asks every company to prove it can survive years of legal spend before it's allowed to keep existing. The firms that cleared MiCA aren’t building anything better; some might be, but most probably aren't. They’re the ones that could afford to pay lawyers for longer. That’s all. Nobody running a project wants to say that part out loud, because it sounds like sour grapes from whoever didn't make the cut. But I'm not in that position, and I'll say it anyway: MiCA is going to protect users, and it's also going to hollow out the middle of this industry, the teams with real products and no war chest. What we are watching is protection and consolidation happening at the same time, and pretending otherwise is just being polite about it. I am not saying this from the outside. We spent the better part of the last year treating this exact outcome as the baseline case for @ZIGChain, not the worst case. We assumed early that a fractured, jurisdiction-by-jurisdiction version of crypto was coming regardless of how the headlines framed it. That assumption is now just how the market works, and the projects still operating like it's optional are the ones about to find out the hard way. 🙏

Me: "update my grocery list" GPT 5.6 Sol: You need more fruit. Examples include: - Apple: Crisp, sweet, or tart orchard fruit. - Pear: Soft, teardrop-shaped sweet fruit. - Peach: Fuzzy-skinned, juicy summer classic. - Plum: Smooth-skinned, sweet and tangy stone fruit. - Cherry: Small, deep red or black juicy drupe. - Apricot: Small, velvety orange fruit. - Nectarine: Smooth-skinned variety of peach. - Quince: Hard, fragrant golden pome used for cooking - Strawberry: Bright red aggregate fruit dotted with tiny achenes. - Blueberry: Tiny, indigo-colored antioxidant powerhouse. - Raspberry: Hollow, ruby-red composite berry. - Blackberry: Dark, sweet, multi-beaded bramble fruit. - Grapes: Vine-growing spheres used for snacking or wine. - Cranberry: Tart, glossy crimson berry. - Gooseberry: Ribbed, translucent green or red tart fruit. - Elderberry: Tiny, dark purple clusters requiring cooking. - Orange: Round, sweet citrus staple. - Lemon: Sour, bright yellow culinary accent. - Lime: Tart, green citrus essential for zesting. - Grapefruit: Large, bitter-sweet breakfast favorite. - Tangerine: Small, easily peeled sweet orange variety. - Clementine: Tiny, seedless, ultra-sweet citrus hybrid. - Kumquat: Tiny citrus eaten whole, skin and all. - Pomelo: Massive, thick-rinded ancestor of the grapefruit. - Banana: Elongated, potassium-rich yellow berry. - Mango: Rich, juicy "king of fruits". - Pineapple: Spiky-skinned tropical multi-fruit. - Papaya: Soft, orange-fleshed tropical wonder. - Kiwi: Fuzzy brown skin with bright green flesh. - Pomegranate: Thick skin hiding hundreds of ruby seeds. - Avocado: Creamy, nutrient-dense single-seeded berry. - Dragon Fruit: Striking pink or yellow scaly cactus fruit. - Guava: Fragrant tropical fruit with pink or white interior. - Passion Fruit: Wrinkled rind containing highly aromatic seeds. - Lychee: Translucent, sweet flesh inside a bumpy red shell. - Durian: Pungent, spike-covered Custard-like delicacy. - Jackfruit: Gigantic, fibrous fruit tasting like juicy fruit gum. - Fig: Chewy, teardrop-shaped fruit filled with tiny seeds. - Date: Sweet, chewy, wrinkled desert palm fruit - Watermelon: Giant, refreshing summer favorite with red flesh. - Cantaloupe: Net-textured rind with sweet orange flesh. - Honeydew: Smooth, pale green rind with sugary sweet flesh. - Tomato: Essential savory berry foundational to global cuisine. - Cucumber: Crisp, cooling green pepo. - Bell Pepper: Sweet, hollow seed pod. - Eggplant: Spongy, purple botanical berry. - Zucchini: Prolific summer squash variety. - Pumpkin: Large, orange autumn squash

At ninety-eight degrees in Allentown, asphalt is not a solid. It is hot black pudding. On the corner of Fourth Street, six men in orange vests were destroying the earth with pneumatic jackhammers. At noon, the foreman, Big Mike, sat on the curb. His arms were coated in black dust and his chest was the width of an oak barrel. For lunch, Big Mike opened a twelve-ounce bag of sour cream and onion potato chips and a warm bottle of grape soda. A deep cultural sadness struck my soul. These warriors had been trading blows with bedrock for five hours beneath the merciless sun, and their society was refueling their engines with fried chemical grease and high-fructose corn syrup. I ran to my kitchen. I cooked eight cups of short-grain rice with kombu kelp. I rubbed my hands raw with coarse sea salt. While the rice was hot enough to scorch human flesh, I pressed ten colossal triangular rice balls. I wrapped each in toasted nori, packed them into a heavy cardboard box, and marched across the steaming tar. Big Mike squinted through the sweat running into his eyes. What is in the box, chief? Mike growled. Japanese field ordnance, I answered, handing him a foil triangle. High sodium. Complex starch. No high-fructose syrup. Mike unwrapped the foil with thick fingers. He inspected the white triangle with the suspicion of a bomb disposal officer. Is there cheese in this? Mike asked. Only salt and grain, Commander, I said. Mike took a bite that eliminated half the mountain. The crisp seaweed snapped. The hot, salted starch hit his tongue. Mike stopped. His jaw locked. His eyes bulged behind his safety goggles like two ping-pong balls. The sodium detonated into his bloodstream like a flare in a coal mine. His exhausted liver absorbed the clean carbohydrates. Mike swallowed. He looked at the half-eaten triangle in his fist. SWEET MOTHER OF PEARL, Mike bellowed to the clouds. He turned toward his crew. DROP THE CHIPS, Mike roared. DROP THE GODDAMN PRETZELS. EAT THE WHITE PYRAMIDS. The five highway workers swarmed the cardboard box like wolves descending on an unguarded buffalo. Within forty-five seconds, ten rice balls had ceased to exist. Then the disaster began. The sudden influx of pure sodium and complex glucose did not merely refresh the crew; it supercharged them into a state of berserker fury. Big Mike grabbed a ninety-pound jackhammer with one arm. FIRE UP THE COMPRESSORS, Mike screamed, veins bulging across his neck like thick garden hoses. WE ARE TAKING THIS TRENCH TO NEW JERSEY. The jackhammers slammed into the limestone with twice their rated hydraulic pressure. The crew did not stop at the assigned utility ditch. They obliterated the sidewalk. They ground through eighteen inches of reinforced concrete. They chewed through the asphalt toward the municipal bus stop like a pack of rabid badgers. The city inspector arrived in a white pickup truck, screaming and waving a clipboard. Mike did not even look at him. He pointed the shaking jackhammer at the horizon and roared through the dust: MORE RICE, SAMURAI. WE HIT WATER IN TWENTY FEET. I ran back to my apartment. I engaged two rice cookers simultaneously. I bought forty pounds of short-grain rice from the wholesale market. By three in the afternoon, Allentown had an accidental subway tunnel, and I was the chief logistical officer of an illegal excavation division.

The oil deal with Venezuela does not take US proved reserves from 46 billion barrels to 111 billion. Majority control of 17 fields does not give America 7.1% of the world's oil, does not put it level with the UAE, does not add supply in 2026, and does not replace a barrel lost at Hormuz. EIA proved reserves are volumes recoverable with reasonable certainty from known reservoirs inside the United States. At year-end 2024 that was about 46 billion barrels of crude oil and lease condensate. A lease, a 55% offtake right, or "majority control" of Venezuelan acreage does not move oil across a border on the ledger. Venezuela books it, and already does. Those barrels sit inside its 303 billion. Which makes the 7.1% figure double counting. Venezuela's reserves are already inside the 1.57 trillion world total. Putting 65 billion in the US column counts the same oil twice. Caracas also did not say proved reserves. It said 17 fields with a "proven potential" of 65 billion barrels, which is not a classification that exists under SEC or SPE-PRMS rules. The number has moved accordingly: Axios put a similar field list at over 90 billion, a US official said 63. There is no published agreement, no named operator and no audited reserve report behind any of them. The 17 fields have not been named either. Reuters, which reviewed the list under negotiation, described a mix of undeveloped Orinoco acreage and mature Lake Maracaibo blocks. Orinoco crude is 8–10° API, high sulfur, high metals. It needs diluent, steam and upgraders, plus pipelines and terminals that need billions in repairs first. Venezuela produces roughly 1.1–1.2 million b/d today. The 100-year term does not exist in Venezuelan law. The reformed hydrocarbons law caps mixed companies at 25 years plus a single 15-year extension. These are not the barrels Hormuz carries. The strait moves Murban, Arab Light, Basrah and Kuwait Export, light to medium grades that refineries across Asia and Europe are built to run. Extra-heavy Venezuelan sour is coking feedstock. It competes with Maya and Canadian heavy on the US Gulf Coast. It does not replace a lost Gulf barrel in Ningbo or Rotterdam. Access to future barrels, if the contract holds and the fields get built, is a long-cycle supply story. It is a contract, not a reserve. Only one of those appears on the EIA ledger.

Practices for embedding AI agents into enterprise systems [Semantic Layer -- Unified Metrics & Organizational Knowledge Graph] 💡 Ask your AI "What's our revenue?" and you get... gross or net? Bookings or cash? FY or CY? An AI that answers without definitions is a tool that's precisely wrong. Centralize the "meaning" of metrics and organization, replacing hallucination with defined facts. 🔥 Problems Solved - Metric/terminology hallucination: AI generates incorrect numbers because "revenue" was never precisely defined - Unresolved references: ambiguous phrases like "my team" or "last month" cannot be accurately resolved - Missing organizational scope: no way to control data boundaries by department or project hierarchy 🏗️ Proposed Pattern Centralize metric definitions in a BI semantic layer (dbt Semantic Layer / Cube) -- e.g., "Revenue = sum of order amounts, tax-excluded, on FY basis." Sync the organizational graph from SCIM/HRIS (Workday, etc.) so "my team's revenue" auto-resolves to "sum of order amounts for members in the user's department." Natural language ambiguity is resolved with defined facts, not hallucination. ✅ Selection Criteria - When to use: analytics-supporting agents, cross-org workflows, permission-dependent processing, metric-critical operations - When NOT: exploratory domains where definitions are not yet established (stabilize definitions first) ⚠️ Pitfalls - Definition maintenance cost: you need an operational workflow to keep metric definitions and org graphs fresh - Granularity balance: too fine-grained and management collapses; too coarse and ambiguity remains -- start with high-frequency metrics - Organizational change tracking: in orgs with frequent reorgs and transfers, SCIM sync frequency and timing become critical 🛠️ Implementation Approach 1. Centralize metric definitions in dbt Semantic Layer / Cube (e.g., "Revenue = sum of order amounts, tax-excluded, FY basis") and connect them as a first-class context source for agents 2. Build an organizational knowledge graph (people, departments, projects, roles, permissions) in Neo4j / Amazon Neptune, synced from Workday / Okta via SCIM 3. Implement a natural-language-to-defined-metric mapping layer that auto-resolves "my team's revenue" to "sum of order amounts for members in the user's department" 4. Progressively formalize definitions starting with high-frequency metrics, and establish a periodic review workflow to keep definitions fresh 5. Detect organizational changes (reorgs, transfers) via SCIM sync webhooks for near-real-time updates, minimizing scope control lag #AIAgents# #EnterpriseArchitecture#

no sour feelings when the sweetest Hearts2Hearts girls are here catch @Hearts2Hearts' K-Pop ON! RECORD: Smart Shuffle Show, dropping at 7PM KST today on our YouTube channel and K-Pop ON! Video Podcast! 📽️ 📽️