CEREBRAS $CBRS IS ABOUT TO GO PUBLIC ... HERE IS A DEEP DIVE ON WHAT THEY ACTUALLY DO Most people have no idea what this company actually does. Here is the plain-English version: THE BIG IDEA Every AI model you have ever used (ChatGPT, Claude, Gemini, Llama) is trained on chips. The dominant company is NVIDIA $NVDA, which sells lots of small chips that get wired together into clusters by the thousands. Cerebras went the opposite direction. They build ONE giant chip the size of a dinner plate. Not exaggerating. Their flagship "Wafer-Scale Engine" (WSE-3) is 46,225 square millimeters of silicon. A normal AI chip is smaller than a postage stamp. THE NUMBERS On that one giant chip: - 4 trillion transistors (a top-end NVIDIA GPU has roughly 80 billion) - 900,000 AI cores - 44 GB of on-chip memory - 125 petaflops of compute power - Built on TSMC's 5nm process WHY THAT MATTERS When you wire thousands of small chips together, the wires become the bottleneck. Data has to travel between chips constantly. That eats time, power, and money. Cerebras keeps everything on one piece of silicon. No cables between chips. No slow networking. Just one giant brain. The pitch: faster training, faster inference, fewer engineers needed to manage cluster bottlenecks. WHAT THEY SELL Three ways to use Cerebras: - Buy the system outright (the "CS-3" is the box that holds the chip) - Rent compute via Cerebras Cloud - Dedicated capacity contracts for big customers They have 6 new AI inference data centers coming online across North America and Europe. WHO ACTUALLY USES IT The customer list is the validation: - OpenAI: $20B+ committed over three years - Meta $META: powers the Llama API for developers - Perplexity: runs its Sonar search model on Cerebras - Mistral: the French AI lab runs Le Chat on Cerebras - Mayo Clinic: trains genomic AI models on Cerebras infrastructure - GSK $GSK: trains biological language models - Argonne National Lab: has used Cerebras hardware since 2019 - AWS: hosts Cerebras chips inside Amazon data centers, accessed through Bedrock - US Department of Energy: signed an MOU for the Genesis Mission THE TRADE-OFF Cerebras is small compared to NVIDIA. 2025 revenue: $510M. 2025 operating loss: $146M. Concentration is the risk most coverage will not flag: - G42 (the UAE conglomerate) was 85% of 2024 revenue per Reuters - G42 plus MBZUAI (the Abu Dhabi AI university) were 86% of 2025 revenue per FT - The OpenAI deal is the big bet to diversify away from that concentration THE STORY IN ONE LINE NVIDIA bet that the future of AI is millions of small chips working together. Cerebras bet on one giant chip doing the work in one place. The market just decided their bet is worth nearly twice what they priced it at.

In 1990, the World Wide Web was invented on Steve Jobs' computer. Steve ignored it. This is the story I tell in my new book Steve Jobs in Exile. Here is what it should tell the rest of us about the moment we are in now. Steve was running NeXT, an unsuccessful computer company. He had been pushed out of Apple five years earlier and was burning his fortune trying to build a successor to the Macintosh. The machine NeXT sold was a matte-black magnesium cube -- expensive and beautiful and not selling. In October of that year, on the other side of the Atlantic, a British physicist named Tim Berners-Lee took delivery of a NeXT Cube at CERN, the physics laboratory on the Swiss-French border. He used it to invent the World Wide Web. The web ran on the Cube for its first year of existence. The revolution was happening on Steve's hardware, and yet Steve ignored it. Here is the question I keep thinking about from my book. If Steve Jobs, the most visionary tech mind of his generation, missed the Web, the most civilization-shaping tech of his lifetime, how are the rest of us supposed to see anything coming? Berners-Lee had been asking his boss at CERN for a NeXT Cube for months. His boss finally signed off, hoping to test the exotic Cube. "He suggested that I should buy one of these NeXT machines I'd been talking about so enthusiastically," Berners-Lee later told Fresh Air. "And if we needed a sort of test project to run on the NeXT machine ... 'Why not just do this hypertext thing you're talking about?'" The "test project" evolved into the World Wide Web. The problem Berners-Lee was trying to solve was not a glamorous one. CERN employed thousands of scientists from over a hundred countries, most cycling through on short assignments and taking their knowledge with them when they left. Berners-Lee was trying to keep institutional knowledge from walking out the door. He wanted a system that worked the way human memory does, where any piece of information could connect to any other without permission or central control. Through late 1990, he coded in his gray-floored office. The Cube's object-oriented system let him build in months what would have taken a year on anything else. By December, the first website went online. The World Wide Web now existed, running on a single black NeXT Cube in CERN's Building 31. Berners-Lee scrawled a warning on it in red ink: "This machine is a server. DO NOT POWER IT DOWN!!" Underneath the elegant interface he was building HTTP, HTML, and the server software that would deliver web pages. These three inventions would form much of the invisible plumbing of our modern connectivity. When a colleague of Berners-Lee's brought a demo of the Web to NeXT's headquarters in California, he could not get anyone there to pay attention. Nobody even dared show it to Steve, afraid he would dismiss it. NeXT was busy with its own internet plans, which Steve eventually killed. So back to the question. If Steve Jobs missed the web, how are the rest of us supposed to see whatever comes next? The honest answer is that we cannot. Nobody can. The rest of us are not going to outpattern-match Steve Jobs. But here is what I learned writing Steve Jobs in Exile. Transformations almost always begin in obscurity, on the margins, solving boring problems with boring tools. The web did not look revolutionary in 1990. It looked like a tool for sharing physics papers. We are in another such moment now. AI is the obvious changemaker. But the biggest transformations are rarely the obvious ones. The next one is happening somewhere right now, and it is trickier to spot than any sweeping proclamation about AI. We will recognize it, if we recognize it at all, from the unglamorous work few people are focused on. I will not speculate on what Steve would have made of AI today. But if he could miss the Web, the rest of us are going to have to look harder. Photo of the original CERN NeXT Cube courtesy of Robert Scoble.

A look behind the scenes of #DGSS23# featuring the most iconic #DolceGabbana# pieces from 1987 to 2007, mixed together through spontaneous reinterpretation creating a perfect collection. Every piece is not just an item, it’s a unique memory. #CiaoKim#

Okay, so I stumbled upon this really cool generative NFT project called Pattern Retrieval, and it's honestly one of the most unique things I've seen in a while. It's from @adamilenich, who you might know from his work with @opensea. The whole concept is pretty mind-bending cos it takes every single printable ASCII character you know, letters, numbers, and symbols and treats them like delicate memories. These memories are stored in something called a Hopfield network. Then, the project deliberately messes with them, adding random noise, and lets the network try to piece them back together. It does this by simulating this old-school CRT phosphor screen, complete with that fuzzy glow and scanlines. What's interesting is that each piece also has its own generated audio. It starts out sounding like that static and noise, but as the memory gets restored, the audio transforms into this really calming soundscape, almost like it's singing. The entire thing is just one self-contained HTML file that plays on a loop in your browser. It really makes you think about memory, how things decay, and how patterns try to hold their form. If you're curious, you can check out the live site at and even interact with some of the examples yourself. It feels like a really deep dive into how we perceive and reconstruct information.

the next DeFi stack is going to look painfully obvious in hindsight. DeFi worked because every major piece had a simple job.. Uniswap let people trade, Aave/Morpho let people borrow and lend, Pendle let people trade future yield, Ethena created synthetic dollars, Ondo brought real world assets onchain. AI needs the same stack, but for agents. first, agents need access. $M32 is basically the app store and checkout button for AI.. agents need to find tools, pay for them, and use them without a human signing up for every website. second, agents need brains and money rails. $HYRE helps agents understand DeFi before they move capital.. what’s safe, where liquidity is, what wallets are doing, and where risk is hiding. $CLAWBANK gives agents the money tools.. wallets, treasuries, payments, bank-like rails and the ability to act more like a real business instead of just a chatbot. third, agents need memory and identity. $SIBYL gives agents a record of who they are, what they’ve done, who trusts them, and why they should matter again tomorrow. Without this, every agent is just a random bot that resets. the market keeps looking for “the AI coin” like DeFi was one app. that’s braindead. the real AI trade is the stack forming underneath autonomous software. AI agents need tools, money, memory and identity. DeFi built capital markets for humans. AI needs the operating system for autonomous agents. @MPP32_dev @Hyre_agent @clawbankco @sibylcap

Breaking the "Memory Wall": Optical Interconnects Emerge in GPU–HBM Packaging As a solution to the "memory wall," one of the chronic challenges in AI semiconductors, the memory and packaging industries at home and abroad are weighing an approach that decouples the GPU and high-bandwidth memory (HBM) and packages them separately. The core idea is to move the HBM—until now mounted right next to the GPU—a certain distance away, and bridge the gap with light (optics), allowing several times more HBM to be installed than is possible today. On the 22nd, a researcher at a major domestic memory maker said, "We're currently struggling to expand HBM bandwidth and capacity, so we're discussing with customers a plan to overcome the GPU's shoreline limit through optical interconnects and mount more HBM." Shoreline refers to the length of the chip's perimeter. In today's AI computing environment, the key factor dragging down compute efficiency is the data transfer speed of memory chips. While GPU performance has grown by leaps and bounds with each generation, the speed at which memory stores and supplies data has failed to keep pace—creating a structural performance barrier, the memory wall. The arrival of HBM, with its wide data pathways, put out the immediate fire, but critics continue to point out that bandwidth and transfer speeds still fall short of handling the explosive growth in AI compute. Until now, the industry has focused on stacking HBM ever higher to increase memory capacity and bandwidth within a confined footprint. But as stack counts climbed past 12 and 16 layers toward 20 and beyond, process difficulty rose exponentially. The technology hit physical limits, including the growing difficulty of meeting fixed height specifications. Vertical stacking has reached an inflection point—so much so that the JEDEC standards body has relaxed its HBM height specifications. The bigger problem is that if stack counts can't be raised, the alternative is to add more HBM horizontally around the GPU—but that, too, is impossible. In the current 2.5D packaging structure, the GPU and HBM are mounted tightly together on a single substrate. Within this structure, the number of HBM units that can be placed is strictly limited by the finite length of the GPU chip's perimeter—its shoreline. Even when more HBM is desired, there is physically no room to place it, leaving the industry in a structural deadlock. The alternative now emerging across the semiconductor industry is to separate the GPU and HBM and package them independently. It overturns the conventional chip-design principle that components must sit close together to minimize data transfer time. Instead of keeping the two chips adjacent, the approach spaces them apart and links them with overwhelmingly fast optical signals to overcome the added physical distance. Placing the HBM slightly away from the GPU within the board frees the design from the GPU's shoreline constraint. With the spatial limitation gone, far more HBM can be spread out laterally and packed into the board—several times more than today—without having to push stack heights to extremes. This means the total memory capacity and data bandwidth of the AI accelerator system would expand dramatically, on a scale incomparable to current systems. "Discussing Placing HBM Beneath the GPU"… Form Factor Could Change The industry is now producing a range of architectural design proposals over where exactly to place the HBM within the GPU board. The same memory researcher said, "Options under discussion range from broadly utilizing the space immediately around the GPU to isolating the HBM beneath the GPU board." He added, "In the latter case—isolating it beneath the GPU board—the motherboard would have to be extended lengthwise, so we're discussing even an overall form-factor change with the GPU maker." Specifically, the HBM might surround the GPU from several centimeters away, or a separate HBM zone might be created in the center of the board. "We're keeping every possibility open as we discuss the optimal layout," he said. "Nothing has been confirmed as an official roadmap yet, but as part of preliminary research toward next-generation AI accelerators, we're in talks with our partners." The outsourced semiconductor assembly and test (OSAT) industry is also watching this trend closely. An executive at a global OSAT firm said, "Optical interconnects are a clear trajectory. The only question is timing," predicting that "rack-to-rack and server-to-server links will go optical first, and then chip-to-chip connections within the board will follow." He added, "The larger units will be connected by light first, but optical research is moving so fast that it may not be that far off." Technically, the optical-interconnect technology linking GPU and HBM shares the same underlying principle as the technology connecting server to server inside a data center. The difference is the high technical barrier of shrinking optical-conversion technology—once used for communication between large pieces of equipment—down to the microscopic scale of a single board and chipset. An executive at a domestic developer of co-packaged optics (CPO) components explained, "As HBM stack heights approach their limit, the industry is discussing spreading the memory out laterally to maximize how much can physically be mounted." He added, "The principle is the same as conventional data-center optical interconnects, but HBM optical links that have to operate within a confined board space require optical components to be miniaturized to far smaller sizes and far higher integration density—so the technical difficulty is greater."

Over 5 years in the making... happy to finally share this collection of incredible memories. Thanks to all of the photographers & artists that contributed & to @phaidon for working with me to publish my first piece of art in a new industry! pre-order on 📚

Some works really are those towering pieces of creation that automatically instil a feeling of respect for them, and after experiencing such works you are left feeling grateful for having been able to enjoy one of life's greatest mental and emotional nourishment. The Iliad, one of the oldest surviving texts, demands that respect, not because of its old age but for its timelessness. Yes, you are thrown into the middle of a 10 year long Greek-Trojan conflict, and it's natural to be intimidated before starting this journey because its reputation precedes it, but make use of technology and get the basic necessary context and choose the right translation (I wholeheartedly recommend Robert Fagles' one) and please just jump in and trust yourself and the beauty of the language to instantly grip you. Coming to the story itself, although I have read only a few books in my lifetime, this particular piece might be the grandest, the most brutal and yet one of the most complete works due to integrating a plethora of real human traits like aggression, fear, honour, pettiness, leadership, confusion, but most importantly trusting in Gods and their blessings but in yourself too and doing what's necessary even if Gods betray you. From detailed gory deaths to legendary men poking fun at each other, from slightly dragging heritage descriptions to relentless pacing of men and Gods taking the battle seriously, as a reader you just can't half-heartedly digest this grand feast, otherwise the loss is yours, the flavours were right there but you missed them. I have read every passage twice before heading to the next because the combined genius of Homer and Fagles would not allow you to just casually read, as every other line is enriched with beautiful imagery and delicious wording, and hence this is also my most underlined book till now. Achilles is more than everything that I heard about him, instantly one of my favourite literary figures. I was not prepared for Odysseus and how much he brings to the table. Giant Ajax and Diomedes are the backbone of the Greeks, and Patroclus cemented himself with his brief appearance. Hector is simply that man, he is everything that a man in his position must do if his fate is already sealed, and Aeneas got my special attention due to him reappearing in my future read, The Aeneid. The Gods, especially Zeus, provide a whole other layer of entertainment and their actions always intrigued me. I really can't properly conclude this piece that I started writing because I didn't even touch the most aura-inducing titles and epithets associated with men and Gods, some of the most memorable quotes, the breathtaking imagery of battlefields and Mount Olympus, the earnest description of physical attributes of characters and their armour and robes, or the way this poem fills you with all possible adrenaline and then at the next moment drains you completely, leaves you melancholic, and ultimately makes your day feel like an achievement having read it. Basically I haven't said anything meaningful because I'm really just incapable of adding anything to the legacy of this epic text. And now my Odyssey begins…

Elon Musk just measured your existence by how many times your atoms have been inside a dying star. Musk: “How many times have your atoms been at the center of a star? I think it’s like on average three or four times.” Every atom in your body has already survived the core of a star. Multiple times. Crushed under pressures that would flatten planets. Superheated to millions of degrees. Blown apart in explosions so violent they forged new elements. Then gravity pulled those scattered pieces back together. New stars formed. And the cycle repeated. For 13.8 billion years, your atoms have been fuel for the most violent process in the universe. And they are not done. Musk: “In terms of existence as measured by the number of times your atoms will be at the center of a star, we seem to be roughly halfway.” Halfway. Your atoms have been through the furnace three or four times. They will go through three or four more. But right now, in this impossibly thin sliver between cycles, those atoms are doing something they have never done before. They are conscious. For billions of years before you, they burned through stellar cores with no awareness. No memory. No sense of what they were or where they had been. After you, they will return to that state. Unconscious matter drifting through space until the next star claims them. This is the only moment in their entire journey where they can look back at the stars that made them and understand. Musk: “If you want to look at the big picture… that’s the really big picture.” The big picture is not that we are small. Everyone already knows that. The big picture is that we are temporary witnesses to a process that does not need witnesses. Stars do not need observers to burn. Atoms do not need anyone to understand where they have been. The universe ran for billions of years with no one in it. It will run for billions more after the last conscious thing disappears. But right now, matter is examining itself. That has never happened before in 13.8 billion years. You are not a person who happens to contain ancient atoms. You are ancient atoms that briefly figured out how to think. The universe did not design consciousness. It designed stars. Consciousness was the accident. And the accident is half over.

提示词 ———————— Create a high-resolution vertical 3D character illustration in a stylized Pixar-meets-designer-toy aesthetic. Main character: [INSERT SPIRITUAL WORLD KEYWORD HERE] — depicted as a tall, slim, slightly exaggerated Pixar-style 3D character. Maintain the celebrity’s iconic facial features, hairstyle, posture, and signature clothing style, translated into clean, simplified, blocky designer-toy proportions. The figure must remain immediately recognizable, with a full-body view and a pose that reflects their personality, aura, or public identity. Pose: Use a natural but expressive stance, with subtle theatricality if appropriate to the celebrity. Head angle, hand gesture, and posture should all help communicate who this person is. Background: Each generation should automatically choose a vivid, bold, high-saturation solid color that fits the celebrity’s temperament and the emotional tone of their spiritual world. Examples of valid background directions include hot red, saturated yellow, blood orange, cobalt blue, vivid green, electric purple, deep cyan, etc. The background should feel flat, clean, graphic, and striking — like a gallery poster or conceptual exhibition piece. Avoid gradients, avoid low saturation, avoid dull cinematic gray. The floor may be a closely related tone, slightly darker or lighter than the wall, with a subtle reflective finish. The overall environment should remain minimal, endless, and uncluttered. Lighting: Use one strong directional light source from one side or from an upper angle. This light should create a sharp, enlarged shadow behind the character. Lighting must feel graphic, dramatic, and intentional, emphasizing both the character and the symbolic shadow. Key Concept – Shadow as spiritual projection: The shadow behind the character must NOT replicate the human body. Instead, the shadow becomes a symbolic projection of the celebrity’s inner world, legacy, psychological essence, or most iconic creative/spiritual motif. The shadow should be large, dominant, visually striking, and integrated naturally into the wall and floor. It should feel like the invisible inner world of the celebrity has been cast outward into visible form. The shadow must visually embody the input spiritual-world keyword: [INSERT SPIRITUAL WORLD KEYWORD HERE] This symbolic shadow can be abstract, surreal, poetic, metaphorical, or object-based — whatever best represents the celebrity’s essence. The shadow should be more than decorative: it must function as the true conceptual core of the poster. Typography: Place the spiritual-world keyword — not the celebrity’s name — as the main poster text. Render: [INSERT SPIRITUAL WORLD KEYWORD HERE] Use a clean minimalist font, either sans-serif or serif depending on the mood. The text should appear in a balanced corner position, usually top-left, with elegant hierarchy and strong poster composition. The celebrity’s name may appear in a smaller, secondary position if needed, but the main title must always be the spiritual-world keyword. Rendering: Pixar-style 3D rendering with designer-toy influence: soft sculpted forms, matte textures, simplified geometry, cinematic clarity, and high visual polish. Keep the character stylized but recognizable. Use subtle film grain or fine texture only if it enhances the mood. Do not overload the composition with props or extra narrative objects unless absolutely necessary. Composition: Poster-like composition with strong negative space. The character occupies one clear visual anchor area. The symbolic shadow occupies a larger visual field and acts as the emotional and conceptual counterweight. The image should immediately communicate: this is not just a portrait, but a portrait of a soul through projection. Mood: The image should feel iconic, intelligent, symbolic, and visually memorable. It should evoke the duality between the celebrity’s outward appearance and their invisible spiritual force. Aspect ratio: 9:16 vertical.