A good technical LLM interview question: Your RAG chatbot... is working as expected locally. You deploy it behind a load balancer with 3 replicas. Users report that it forgets what they just asked, and answers get worse with each restart. Why did this happen? (answer below) A local setup has one process that owns everything. - The vector index is a variable in memory. - Conversation history is a Python list. - The documents are on local disk. You never treat any of them as infrastructure, because restarting rebuilds all three in seconds and there is only ever one copy. The setup does not carry over to production directly. The vector index might disappear on restart, so the app re-embeds everything on boot and serves empty results until it finishes. Conversation history may belong to one replica, so a follow-up routed elsewhere has no memory of the previous turn. Documents could be on whichever container ingested them, so the three replicas hold three different corpora. None of this is evident with one user and one process. So the actual work in shipping RAG is not just the retrieval logic, but also storing the vector index, the conversation history, and the documents outside the app, where every replica reads and writes the same copy. Which comes down to three requirements: > The vector store needs persistence and has to be reachable from every replica. pgvector inside Postgres keeps embeddings next to the rest of the data instead of adding another system to operate. > Conversation state has to be checkpointed outside the app. LangGraph writes its state to Postgres, so any replica can pick up a thread mid-conversation. > Docs need shared object storage, so ingestion happens once instead of once per replica. If you get those three right, the retrieval logic you wrote in the notebook works unchanged. To learn how all of it is wired together, Akamai's GitHub has a working reference implementation. - rag-langgraph-k8s-quickstart is an airline policy Q&A assistant built with FastAPI, LangChain, and LangGraph. Terraform provisions the LKE cluster, a Postgres instance with pgvector for embeddings, a second Postgres for LangGraph checkpointing, and an object storage bucket for the policy documents, in one apply. - akamai-workshop-ai-inference covers the next step, running the model yourself instead of calling an API, with prefill and decode, KV cache tradeoffs, and continuous batching under real concurrency. Both are available on Akamai's new Developer Hub, alongside their tutorials and code samples. It also links to Edge Case, their Discord, where four developer advocates architect and deploy a production app live every other Wednesday. If you create a new Akamai Cloud account, you can also get $300 in credits for joining. Join here: That said, this post assumes the retrieval logic was right to begin with, and that is doing a lot of work. Most RAG systems fail earlier, at the point where a chunk gets treated as a self-contained unit of meaning. I wrote about the two skills that fix that gap, and why the chunk is usually the wrong thing to embed. Read it below. Thanks to Akamai Cloud for partnering today!show more

A good technical LLM interview question: Your LLM chatbot... takes 12s before it generates the first token, and the users are complaining. So you move the model onto a GPU with 3x the computing power. The time to first token barely improves. Why did this happen? (answer below) Latency in an LLM app is a placement problem disguised as a model problem. If you profile the 12 seconds, the model's prefill itself may only account for around 1.5 seconds of it. So halving the prefill step saves just 750ms out of 12000, which is under 7%. The rest is spread across stages that never touch the GPU. The request first travels to whatever region the app runs in, and a cross-continent round trip could cost over a second before any code executes. Then the request handler starts. On a container-based serverless platform under load, this adds several seconds of cold start, paid before auth, rate limiting, or prompt assembly even begins. Retrieval adds its own hop, and the response streams back across the same distance. Optimizing a stage that was already fast cannot alter the latency that's majorly affected by other stages. Those other stages are slow for a structural reason. An LLM app runs two workloads that want opposite machines. - The request path is short, spiky, and needs to sit close to users - Inference is long-running, GPU-bound, and billed hourly, whether requests arrive or not. So the actual decision is not which model to run, but where each of these two workloads runs. There are three options, each with its own tradeoffs: > A dedicated GPU box removes inference cold starts, but it bills around the clock and lives in one location, so distant users wait out the round trip on every request > Container-based serverless scales to zero, but the request path pays a cold start, and most of these platforms have no GPU behind them. > Edge runtimes start in under a millisecond, because a WebAssembly module carries no OS or container image to boot. They handle the request path well and cannot hold a model. So the answer is not to pick one, but to split the app across two of them. The request path runs close to users, and inference runs on a dedicated GPU it calls into. That also explains the failed upgrade. More compute made a stage that was already fast faster, and left the 10.5 seconds around it untouched. To actually learn how it's done in practice, Akamai's GitHub has a reference implementation for each half. - vllm-on-lke serves Qwen2.5-7B-Instruct behind an OpenAI-compatible endpoint on one RTX 4000 Ada GPU in Linode Kubernetes Engine, with Terraform creating the cluster, both firewalls, and the GPU operator in one apply. - akamai-functions-llm-chatbot covers the front, where a WebAssembly API checks a KV cache and only calls the GPU-backed instance on a miss. Both are available on Akamai’s new Developer Hub, alongside their tutorials and code samples. It also links to Edge Case, their Discord, where four developer advocates architect and deploy a production app live every other Wednesday. If you create a new Akamai Cloud account, you can also get $300 in credits for joining. Join here: That said, this post treats generation as a single 1.5s block, but that block has its own structure, and knowing it well tells you whether a model is slow to start or slow to stream. I wrote a first-principles walkthrough of it, covering the prefill and decode split, KV caching, and where the time actually goes inside each one. Read it below. Thanks to Akamai Cloud for partnering today!show more

This is not a border dispute. It is a... sermon. A Hamas leader on a Gaza stage tells seven million Palestinians abroad to stop warming up. There are Jews everywhere, he says. Attack every Jew on Earth. Slaughter them. Kill them. With Allah’s help. The crowd is not a war room. It is a public. Hamas did not invent that sentence. Hamas is but one franchise of it. Islamic Jihad says the same thing with a different logo. Hezbollah wraps it in Khomeini and the language of “the Zionist entity.” The Houthis put “death to Israel, curse the Jews” on a flag and call it governance. The regime in Tehran funds the network and calls the Jewish state a tumor. Different uniforms. Same theology. For them, Jewish sovereignty is an insult to Allah, and Jewish lives outside the land are not off-limits. That is why “from the river to the sea” is not a housing policy, and why October 7 was not an accident of occupation. The target list in this clip is not a soldier at a checkpoint. It is Jews as such, wherever they are. Once the enemy is a people, not a government, there is no ceasefire that satisfies the idea. There is only a pause. Western coverage keeps shrinking this to one militia and one strip of coast because a regional ideology is harder to manage than a “peace process.” Factions are useful that way. They let you condemn Hamas on Tuesday and treat Hezbollah as Lebanese politics on Wednesday and the Houthis as a shipping problem on Thursday. The preacher does not bother with those folders. He names the Jews and asks the diaspora to start. If you want to understand the war, start with what they say when they are not fundraising in English. The map they want is not two states. It is a world with fewer Jews in it, sanctified as duty. Hamas is the local office. The ideology is the corporation.show more

A GIRL WALKED OFF THE TOP OF A WATERSLIDE,... THREW TWO FLIPS DOWN THE FACE OF IT, AND LANDED IN THE POOL 20,000 likes None of it is generated. That is exactly why it is worth eleven seconds, because your feed has spent a year training you to assume the opposite run it as a test. here is what is in the frame that no model reliably puts there yet: → the water. spray leaves her body at the right angle and volume for her speed, and the sheet on the slide deforms under her weight instead of flowing past her → the rotation conserves. she tucks and the spin accelerates, she opens and it slows. angular momentum is a constraint, and generators approximate it rather than obey it → the crowd reacts late. people on the platform turn after she has already gone, because they are reacting, not choreographed → the camera operator loses her. the frame lags the subject and catches up. that is a mistake, and mistakes are the expensive thing to fake → and the landing is ugly. real impacts are. generated ones almost always resolve too cleanly none of that is a checklist you can memorise, and pretending otherwise is how people get caught. every one of those tells has a shelf life. three of them were already unreliable a year ago what does not expire is where you look. continuity, physics, and mistakes - the three places a model has to simulate a system instead of reproducing a surface. surfaces are solved. systems are not, yet and the reason this gets more useful every month is not fakes. it is that real footage is starting to get accused. the cost of being wrong now runs in both directions the fastest way to calibrate your own eye is to make one yourself. image-to-video from a still, one line about the motion - Picsart runs it from a phone. you start spotting the tells about ten minutes after you have made your own this one is real. the fact that you had to check is the actual storyshow more

sorry, they just did WHAT someone gave a machine... one disease name, the leading cause of blindness in the developed world with 1.5 million americans already in its path, and it came back pointing at a drug that has sat in pharmacies for years under a different label: 551 papers read in 30 minutes against the 294 hours a human would have needed, and the loop that did it is public on GitHub most agent setups answer one question at a time, so the ceiling on the work is the quality of the question you happened to think of this one was handed a single question and wrote the second one itself. turns out that follow-up is where the real find was: a target called ABCA1, upregulated threefold, in an experiment no human ordered i read the whole paper looking for the trick, and the trick is structural. that is the second question, and it is the gap between an assistant and a factory: - hand the loop a field rather than a task: it was given a disease, and choosing the mechanism was part of its job - make it rank before it spends: 151 papers in, ten candidate mechanisms out, scored against each other before anything touched a bench - split reading from judging, so the agent that forms the theory is a different agent from the one grading it - close every cycle on physical reality: the verdict was an experiment, and another model's opinion was never allowed to stand in for one - feed each result back as the next question rather than a log line, which is the step almost nobody builds - search what already passed inspection first: the winner was an approved compound with a safety file already on record - write down what the round learned before opening the next one, so round two starts where round one stopped my read, and i think it is the uncomfortable one: reading was the entire bottleneck in that field, and everybody spent the decade optimising the writing. people ran every physical experiment here, the analysis agent needs a domain expert writing its prompts, and the authors decline to call this the leap it resembles. the thinking got replaced, and the hands did not so the question i cannot answer for my own setup: which step of your loop still stops dead until you sit down and type something bookmark this one. the four parts that turn one model into a line that runs like this, the queue, the rooms, the write permissions and the gate, are built file by file in the piece below ↓show more

SOMEONE BUILT A TERMINAL TORRENT CLIENT THAT SEARCHES EVERY... TRUSTED SOURCE AT ONCE AND DOWNLOADS STRAIGHT TO DISK finding a torrent in 2026 is miserable fake download buttons, popups that spawn three more tabs, and half the results are dead with zero seeders. so he built the opposite of all that: > type one query and it hits a curated set of trackers all at once > results stream back tagged with source, size and seeders as fast as each site answers > arrow to the one you want, press d, and it lands on your drive > if one source is down it just skips it and keeps searching the rest > downloads run in the background and pick up where they left off if you quit mid transfer the whole thing is one command. npx and youre running, all you need is node, no browser, no setup, nothing leaves your machine except the request itself. its called torlink, open source and mit. this is the kind of clean terminal tool that does one annoying thing perfectlyshow more

Your body fat is a tank with a tap... on it, and the tap narrows as the tank empties. - Plenty of fat: your body covers almost any deficit you ask it to - Getting leaner: it covers less each day - Properly lean: it covers very little, and the rest comes off the plate Which is why the first stone comes off while you barely notice, and the last one takes everything you have. So when you are lean, the shortfall has to arrive as food. And the industry hands you the wrong bag. Push the protein. Two hundred grams. Two fifty. Chicken breast on a kitchen scale. Past what you build with, protein is a fuel your liver has to strip the nitrogen off before you can use a calorie of it, and that machinery has a speed limit. So the extra hundred grams builds nothing and lifts nothing. Protein is what you build with. Fat and carbohydrate are what you lift with. Cut them both to make room for another shake and you have removed the thing that was moving the weight. Fat wins that choice. Twice the energy per gram. Leaves insulin alone, so the fat you are burning is not told to stay put every three hours. Keeps you full. And it is the same fuel your own stores are already releasing. Eat the fat. It is the one thing on the plate that matters more the leaner you get, and the first thing they told you to cut.show more

Stop renting a chatbot and calling it your company’s... brain. Utopia is the first open-source enterprise world model I’ve seen that actually treats knowledge like a system, not a vibe. It's a Self-hosted, open source and built around the one question every other knowledge base ignores: when. Every fact in it carries a start date and an end date. Nothing gets overwritten. When something changes, the correction closes the old version and opens a new one. The history stays intact. So the graph isn't a snapshot anymore. It's a recording you can rewind. You drag a timeline on the screen and watch the whole thing redraw itself to show what was true on that date. Who owned what. What the policy said before it changed. All of it, replayable. And every fact points back to the exact sentence it came from. Nothing gets in without a receipt. Here's the part that got me. The whole thing is one binary and one Postgres database. That's it. No Elasticsearch, no separate vector service and no message queue. Most RAG stacks are six services duct-taped together. This is two. It even runs fully offline on an air-gapped network with any model you want.show more

Nobody goes into the forest at night expecting to... change their life. You go for a walk. A flashlight. A bad feeling about the silence. Maybe you're tracking something, maybe you're just tired of the noise in your head. You expect leaves. Mud. The usual dark. You do not expect the forest to hand you something back. Watch the video first. Then come back and read the rest, because what happens in those 20 seconds is only the surface of the story, and the real story is much stranger than the clip. ⸻ Let me start with a confession. I have watched this video more than forty times. Not because it's cute. Everybody says "cute" and scrolls on. I watched it because every time I hit replay, my brain found a new question it couldn't answer. Why was it there? Why was it quiet? Why was it still there? That third question is the one that keeps me up. Because in the wild, a baby that is still where it was left is not a lucky baby. It's a baby whose entire survival plan is working in real time, minute by minute, against everything that wants to eat it. And the plan has a name that most people have never heard. ⸻ THE FIRST RULE OF THE WILD: BE BORING Here is the thing nature figured out millions of years ago and humans keep forgetting. The loudest survivor doesn't win. The quietest one does. Predators don't hunt what they can't find. So the oldest trick in the book isn't strength, isn't speed, isn't claws. It's disappearing. Some animals hide by running. Some by blending in. Some by pretending to be something far more dangerous than they are. And one of the most spectacular animals on Earth, an animal built for a single purpose, to be the fastest thing alive on land, starts its life using the exact opposite of speed. It starts life using stillness. Think about that for a second. The sprinter of the animal kingdom, the body engineered to hit highway speeds in a few heartbeats, spends its first weeks of existence unable to run from anything. Eyes barely open. Legs wobbling. Weight of a small bag of sugar. And the whole species depends on that tiny body staying quiet in the dark. ⸻ WHY YOU SHOULD CARE ABOUT SOMETHING THE SIZE OF YOUR FOREARM Let me give you numbers, because numbers don't lie, and these ones are brutal. In some of the best-studied wild populations, the majority of newborns of this species never make it to adulthood. In certain regions researchers have recorded cub losses that reach staggering levels in the first months of life, with most of the litter gone before the animals are old enough to even try a chase. Not because they're weak. Because the world is full of things with teeth that are very, very good at finding small, helpless, warm bodies. Lions. Hyenas. Leopards. Eagles. Even baboons. The adults of this species are fast, but they are not strong. They are slender, light, built like a racing car and not like a tank. They can't win a fight, so they have never been allowed to be proud. They lose kills to bigger predators constantly. They run themselves to exhaustion for a meal and then hand it over. Which means that from the first day, the safest strategy for the youngest ones is not to fight. It's to be invisible. ⸻ THE DISGUISE THAT SHOULDN'T WORK (BUT DOES) Here's where it gets wild. Newborns of this species are born with a strange silvery-grey fur running down their backs, a thick, fuzzy cape of hair that looks like a mohawk crossed with a dandelion. For years people thought it was just baby fluff. It isn't. Many researchers believe this mantle is a costume. A disguise that makes a defenseless baby look, from a distance and in poor light, like a honey badger, an animal so fearless and so aggressive that lions think twice before picking a fight with it. Read that again. A creature too young to walk properly is dressing as one of the most reckless animals in Africa to bluff its way past predators. That's not cute. That's a con artist's masterclass. And the disguise only works if one rule holds: the cub has to stay put, stay silent, and stay hidden. Which brings me back to the question I can't stop asking about that video. ⸻ WHERE DO YOU HIDE A BABY WHEN THE WHOLE WORLD IS A MENU? The mother has a problem. She is the only protection her litter has. She cannot stay with them all day, because she has to hunt, and hunting means leaving. Every single time she walks away from her cubs, she is gambling their lives against dinner. So she does what mothers of this species have always done. She picks a spot. Dense grass. Rocky crevices. A thicket nobody would look at twice. And she moves them, over and over, every few days, carrying each baby one by one in her jaws, because a scent that stays in one place is a scent that gets found. She is running a witness protection program with no budget, no backup, and a body that shouldn't be doing this at all. Now ask yourself what kind of place makes a good hiding spot. Somewhere narrow. Somewhere dark. Somewhere that can't be reached easily by anything bigger than the thing you're trying to protect. Somewhere that looks, from the outside, like absolutely nothing. And that's when you realize why the thing in the video is so unsettling. It isn't a rescue story. It's a hiding story that someone accidentally walked into. ⸻ THE NIGHT IS NOT NEUTRAL Most people think of the forest at night as a place where nothing happens. That's backwards. Night is when the forest does its real business. The day shift is for show: birds, light, noise, movement. The night shift is for decisions. Who eats. Who hides. Who doesn't make it to morning. Your eyes are bad at night. Your ears are worse than you think. Your sense of danger is tuned for a world that no longer exists, and it screams at the wrong things, shadows, branches, a snapping twig, while completely missing the thing standing ten feet away. Every animal around you, though, is doing math. How far is that sound? How big is it? Is it hunting or is it lost? Is it alone? The forest is a conversation conducted entirely in whispers, and you are walking through it wearing boots. So when something is found in the dark, something that has survived the dark, you have to understand what you're actually looking at. You're looking at a statistical miracle with a pulse. ⸻ WHAT THE FASTEST ANIMAL ALIVE CAN'T DO Let me break a myth for you, because it matters. People hear "fastest land animal" and picture an unstoppable machine. Reality check. Yes, this animal can go from standing still to highway speed in roughly the time it takes you to read this sentence out loud. Yes, its spine flexes like a spring. Yes, its claws grip the ground like running spikes, and its long tail works as a rudder, steering a body that is basically a controlled fall at full velocity. But look at what it can't do. It can't roar. It purrs, chirps and yelps instead, like a house cat that was given a rocket engine. It can't sustain the sprint. A chase lasts seconds, and afterward the animal is so overheated it has to stop and recover, sometimes long enough for a scavenger to walk up and take the meal. It can't defend its territory like a lion. It can't defend its kills like a leopard. It can't climb like a leopard either, not really, though when the situation demands it, this animal will surprise you. It lives on the edge of a very thin margin. And that margin starts the second it's born. ⸻ THE NUMBER THAT SHOULD KEEP US ALL AWAKE There are far fewer of these animals left in the wild than most people assume. Not millions. Not hundreds of thousands. Estimates put the global wild population at only a few thousand individuals, roughly the number of people who fit inside a single large arena on a weekend night. The entire species. The entire planet. Their range has collapsed to a fraction of what it once was. They've vanished from most of the countries where they used to roam, squeezed out by shrinking habitat, fewer prey animals, conflict with farmers, and a brutal illegal pet trade that sees a baby of this species as a status symbol with a price tag. Which makes every single survivor a statement. Every cub that makes it through its first months is not just an animal. It's a vote for the future of the species. ⸻ THE TEST YOU DIDN'T KNOW YOU WERE TAKING Here's a question I want you to answer honestly. If you walked into the dark and found something small and alone, something that looked abandoned, what would you do? Most people would say: help. That's the instinct. Warm, human, immediate. And that instinct, in the wild, is exactly what wildlife experts spend years warning people about. Because here's the paradox nobody puts in the captions: The most dangerous thing you can do to a hidden baby animal is to decide it needs saving. Mothers leave their young alone on purpose. Often for hours. Sometimes longer. That absence isn't abandonment. It's strategy. The baby looks lost. It isn't. It's following instructions older than language. An enormous number of animals every year end up in rescue centers, not because they were abandoned, but because someone with a good heart and no information took them out of the one place their mother was coming back to. The kindest-looking choice and the right choice are not always the same. And that, more than anything in the clip, is why I can't stop thinking about it. ⸻ SO WHAT ACTUALLY HAPPENS NEXT? This is the part that separates people who scroll from people who understand. When a wild animal in trouble is encountered by a human, there's a decision tree. A real one. One that professionals follow and one that almost nobody posting videos online talks about. Is the animal injured? Is the mother nearby? How long has it been there? Is it exposed to weather, to predators, to danger? Is there a licensed team within reach? Every answer changes the outcome. The wrong call, even a loving one, can turn a survivable situation into a death sentence. The right call, made calmly, with the right people, can give an animal a second chance at a life it was never promised. And somewhere between those two outcomes sits the moment you're about to see. I'm not going to tell you what it is. You're going to watch it yourself and decide what you would have done. ⸻ THE MATH OF A 20-SECOND CLIP Let me tell you why this video travels. It isn't the animal alone. It isn't the dark. It isn't even the tension. It's that it triggers three of the oldest circuits in the human brain at the same time. Circuit one: the forest. Every one of us carries a few hundred thousand years of memory about what lives in the dark between trees. Your pulse knows things your mind has forgotten. Circuit two: the small and the helpless. Big eyes, soft shapes, tiny limbs. Evolution wired us to protect them, and that wiring fires before thought arrives. Circuit three: the unknown. Something is there that shouldn't be. Something is happening that you can't fully explain. And your brain hates an unfinished question more than it hates almost anything. Put those three together and you get a clip that people send to their friends with one sentence: "You need to see this." That sentence is the most powerful distribution engine on the internet. No algorithm in the world beats it. So I'll say it too. You need to see this. ⸻ THE THING ABOUT WONDER We live in a time when almost nothing surprises us. You can watch a rocket land itself, a robot do a backflip, a stranger on the other side of the planet cook dinner in a tin shack, all in the space of a minute and a half, all while eating lunch. We've gotten so good at consuming astonishing things that we've lost the ability to be astonished. And then, once in a while, something comes along that breaks through the numbness. Not because it's loud. Because it's real in a way that feels almost impolite to interrupt. A living thing. A tiny one. A breath you can almost hear. And a human being suddenly forced to make a choice with no script. You don't need a soundtrack for that. You don't need a narrator. You just need to be quiet for twenty seconds and let it work on you. ⸻ THE FOREST HAS RULES. YOU DON'T KNOW THEM. If you ever walk into a wild place at night, there are a few things worth knowing. One. The forest is never empty. It only feels empty because the animals that live there are better at not being seen than you are at looking. Two. Silence is not safety. Silence is information. When birds go quiet, something changed. When insects stop, something moved. The forest talks in what it doesn't say. Three. Your light makes you blind. A flashlight shows you a small circle and hides everything around it. You trade the whole world for a single bright patch. Four. Whatever you find, you are the intruder. Not the hero. Not the savior. The guest, and a clumsy one. Five, and this is the one that matters most. The wild doesn't owe you a story. When it gives you one anyway, you are the lucky one. Most people walk through the woods and see nothing. Some people walk through and see everything. And a tiny number of people, on a very particular night, walk through and see something that rearranges how they think about the word "rescue." ⸻ WHAT WOULD YOU DO? I'm going to ask you to do something unusual for the internet. Don't just watch. Pause on the first second and ask yourself one question. If this were you, in this place, at this hour, with only your hands and your judgment, would you have even noticed it? Most people wouldn't. That's the point. The animal's whole existence depends on people like you walking right past it. It was built for that. Every instinct, every pattern, every ounce of its tiny body says one thing: Don't be found. And someone found it anyway. Now the question is whether being found was the end of the story or the beginning. ⸻ THE QUIET PROFESSIONALS Behind every clip like this, there's an invisible world that rarely gets credit. Rangers who walk the same trails for decades. Vets who work through the night for animals that will never know their names. Rehabbers who bottle-feed things that could bite them, every three hours, for weeks, sleeping in shifts, smelling like formula and hay. Researchers who sit in a vehicle for fourteen hours to count one litter. Volunteers who clean cages, wash blankets, boil water, chop meat, and never once appear in the shot. These people don't go viral. They go home exhausted. And the only reason any of us ever get to see an animal like this alive is because of the thousands of boring, unglamorous, backbreaking hours they put in so that a twenty-second moment can exist. If a video like this makes you feel something, let that feeling be a little bit bigger than the video. It deserves to be. ⸻ NOT EVERY HERO LOOKS LIKE ONE Here's something I find fascinating. In the best wildlife stories, the person at the center almost never looks like a hero. No cape. No fanfare. No dramatic music. Just hands, patience, and a decision made in a few seconds in poor light. The heroism is almost always small. A steady breath. A careful movement. The humility to go slowly when every part of you wants to rush. It's the opposite of how we picture bravery. And it's far more convincing. ⸻ WHY THE INTERNET CAN'T LOOK AWAY Let me tell you what I've noticed about the videos that explode. They almost never explode because of what they show. They explode because of what they promise. The promise of a good wildlife clip is simple: something real is about to happen, and you might miss it. Not "look at this animal." Everyone has seen animals. It's "something is going on here, and it's going to resolve one way or another, and you have to see which." That's tension. That's the engine. That's why people who never watch nature content will stop mid-scroll for a hollow piece of bark and a quiet nighttime forest. Because deep down, we are still the same species that sat around fires listening to the dark. We never stopped wanting the story. ⸻ THE TINY DETAIL MOST PEOPLE MISS When you watch the clip, notice what you notice first. Then watch it again, and notice what you missed. That gap between the two viewings is where the magic lives. The first time, you're reacting. The second time, you're reading. The third time, you start asking questions that have nothing to do with the animal and everything to do with you. Why did my chest tighten? Why was I holding my breath? Why did I want it to be okay so badly? A stranger's animal. A forest you've never visited. A night you weren't part of. And yet. That's the power of paying attention. When you give a single quiet moment your full focus, the world becomes bigger and stranger and more tender than the feed makes it look. ⸻ A SHORT HISTORY OF HUMANS AND BIG CATS For as long as there have been people, there have been stories about the cats. Ancient rulers kept them. Painters carved them into tomb walls and palace halls. Kings tamed them for hunting, and some of the oldest hunting records show people training them to ride on horseback to the chase. We admired them. We feared them. We put them on crests and coins and flags. And then, slowly and quietly, we pushed them out. Fences. Roads. Farms. Cities. Every year, the wild gets a little thinner, the corridors get a little narrower, the places to hide get a little fewer. Today, the story of this animal is no longer a story of power. It's a story of margins. A species that once crossed continents now clings to scattered pockets, many of them not much larger than a county. So when you see one alive, young, and breathing in the dark, you're looking at something that survived a very long list of things that should have ended it. ⸻ WHAT I WANT YOU TO TAKE FROM THIS I'll be honest with you. I didn't write this to flood your timeline with another cute animal moment. I wrote it because the best content on this platform does one thing: it makes you pay attention to something that was always there. The forest was always there. The dark was always there. The animals were always there. We just got too busy to look. And if a twenty-second clip is what it takes to drag you back to a sense of wonder, a sense of responsibility, a sense of "oh, right, the world is bigger than my notifications," then it's done something most content never will. It's made you feel small in the best possible way. ⸻ THE RULES IF YOU EVER FIND A WILD BABY Since I know some of you are going to ask, here is the short, honest, useful version. One. Do not assume it's orphaned. The mother is very often nearby, watching, waiting for you to leave. Two. Keep your distance and your voice low. Stress alone can kill a young animal. Your good intentions do not cancel out its fear. Three. Do not feed it. Wrong food, wrong temperature, wrong method, and you can cause more harm in ten minutes than the wild did in ten days. Four. Do not take it home. Ever. It is not a pet. It is not a project. It is a wild life, and the law in most places agrees. Five. Call the professionals. Local wildlife rescues, park authorities, licensed rehabilitators. Give them the location, the time, the condition. Let them decide. Six. If you can't tell whether it's in danger, document from a distance and ask an expert before touching anything. Seven. Remember that the best rescue is often the one where nobody ever knows you were there. Save this part. It might matter one day. ⸻ THREE QUESTIONS TO ASK BEFORE YOU SCROLL Before you hit play, hold these in your head. Question one: what is the animal protecting itself from? Not the obvious answer. The real one. The list is longer than you think, and most of it never shows up on camera. Question two: what does the person in the frame know that you don't? Experience changes everything. The calmest hands in the world usually belong to the people who have been wrong the most times and learned from every one. Question three: what would have happened if nobody had walked by? Because the answer to that one is the reason this story matters, and it is not the answer you expect. Keep those three questions close. By the end of the clip, you'll have an answer to at least one of them, and I'd bet money it will surprise you. ⸻ A NOTE FOR THE SKEPTICS I know how the internet works. You've seen a thousand clips that promised everything and delivered nothing. You've been burned by dramatic captions. You've learned to roll your eyes. Good. Keep that instinct. It protects you. But give this one twenty seconds anyway. If it doesn't earn them, scroll on and I'll never say a word. If it does, you'll know exactly why I wrote this much about so little footage. ⸻ ONE MORE THING BEFORE YOU PRESS PLAY The best moments in nature are never announced. They happen in the gap between two footsteps, in the half second after a sound, in the breath you didn't know you were holding. You can't schedule them. You can't order them. You can only be there, awake, when they arrive. Tonight, you are. ⸻ THE ILLUSION OF CONTROL We like to think we're in charge out there. We carry maps. We carry phones. We carry bright lights and strong opinions and a belief that every situation has an instruction manual somewhere. The forest has never read it. The forest runs on timing, on luck, and on a thousand tiny decisions made by creatures who will never know they mattered. A mother who chose one spot instead of another. A wind that shifted at the right moment. A predator that turned left instead of right. A person who decided, on one particular night, to take a walk they almost skipped. Take away any one of those, and the story doesn't exist. That's what gives a moment like this its weight. It wasn't planned. It wasn't guaranteed. It came out of a chain of coincidences so thin that a single changed step would have erased it. And you are watching the one version of the night where it all lined up. ⸻ WHY SOME STORIES STAY WITH YOU I've seen thousands of viral videos. Almost all of them evaporate. You laugh, you share, you forget. But a few lodge themselves somewhere deeper. A few return at three in the morning, uninvited, and ask you questions you didn't expect. The difference is never production. Never editing. Never music. The difference is whether the moment held something true. Something about vulnerability. Something about trust. Something about the strange, fragile bargain between humans and the wild. When a clip carries that, you don't need to be told it's special. You feel it in your body before you can explain it. ⸻ THE PART WHERE I ASK YOU FOR SOMETHING I don't ask for much. Here's what I'd like. Watch the video with the sound on, in the dark if you can, with nothing else open. Don't skip ahead. Don't read the comments first. Don't treat it like content. Treat it like a window. Then do one small thing. Share it with the one person you know who still gets quiet when they see something beautiful. The one who stops the car to look at a sunset. The one who says "wait, look" and points at a spider's web in the morning dew. That person will understand. And if you're feeling generous, send it to someone who has forgotten how to be amazed. They need it more. ⸻ THE LAST THING Somewhere out there tonight, in a place with no signal and no audience, another small animal is sitting very still in the dark, doing the bravest thing it knows how to do. Waiting. It has no idea about views. No idea about followers. No idea that people on the other side of the world are watching a version of its story and holding their breath. It just knows one thing, passed down through millions of years of survival: Stay quiet. Stay hidden. Hope the night is kind. Most nights, it isn't. Some nights, against every probability, it is. And every once in a while, a person walks into the woods with no plan, no expectation, no idea they're about to become part of a story much bigger than themselves. That's the story in this video. I won't tell you how it ends. I'll just tell you that when it was over, I sat in silence for a long time, and then I hit replay. ⸻ NOW GO WATCH IT. All twenty seconds. Then come back and tell me which moment got you. Because everyone has one. And it's never the same moment twice. ↓ Video above ↑ If this made you feel anything at all, repost it, so someone else who needs a reminder that the wild is still out there, still breathing, still full of surprises, gets to feel it too. Follow for more stories from the places we forgot to look. The forest is still talking. You just have to stop walking long enough to hear it.show more

OFFICIAL: It has been 134 days since the release... of Totenreich, which is longer than the search for both Samantha's Drawing, and Gong COMBINED. With the release of s6 next week, it felt fitting to use our platform to get as many eyes on it as possible. As a quick disclaimer, just because things have been 'tested' does not mean it is not the solve. This post is also not an invitation to share your personal thoughts on datamining/leaks. This relic, by everything shown in the game files, can NOT be datamined: "Even in the current state of the game, relics are a lot harder to solve via datamining and ingame hack tools... For relics, only the sequential item ones (like Dancing Arnie and Agarthan device) are incredibly easy. Agarthan device with the helmet, radio, and mr peeks is actually the easiest. Dancing Arnie becomes hard due to it becoming a kill challenge for the lobster and you cant datamine that". Because of that, the leading theory is that the missing piece has to be something kill condition related, or player action related. Below is literally all of the information I could find on the belt that I thought was worth mentioning, to try and help the community solve this together. CONFIRMED: There is a magazine with a sticker of Mr Peeks holding a Geiger Counter that has a pickup interaction, and is located at the top of the lighthouse while you are picking up the Jotunn Star. This only appears in cursed mode, and appears without any relics active. Considering the only other grim relic (power switch) is solved, we know this mag fits in with the wrestlers belt in some way. The sticker comes directly from intel located on Paradox Junction. THEORIES: Treyarch Dev Tweet - One Treyarch dev tweeted out: "It feels very fitting this is the final unsolved relic given the reward... good luck" Coda Wallbuy Sounds - In a video, a user hears a distinct ding sound when buying ammo off of the coda wall buy in spawn. This has been difficult to replicate, but has been replicated by the original user. Disappearing Boxes - There are 7 boxes around the map that have a hit box, and disappear when hit/shot (they disappear fast with melee macchiato). Currently it is believed by the majority of the community that this is nothing Clocks - There are five clocks near Richtofen's office. Dataminers have said "The hour, minute and second hands are all separate individual models stacked on top of the base model. lol I mean it could mean something or it could just be nothing. Their design choices on some stuff is weird But it is a model without a script and isn’t setup to move in any way" Swordfish - The mag located at the top of the lighthouse is the same model of a mag from the Swordfish A1. Some people believe that is a hint for the swordfish being needed from the box in order to complete the relic. Here are links to a spreadsheet managed by one community member that has things that community members have tested directly. Feel free to look through it, and see if you have an idea that has not already been tested: If you are interested in hunting this with other community members, please feel free to join the Mystery and ZoneX discords, as many of the community is hunting throughout there regularly. You can also feel free to comment under this post anything you might have found interesting.show more

this is more useful than my entire degree Elon... Musk's rocket company signed a $60,000,000,000 deal for Cursor in June, and eight days ago the two of them put a worker on sale for $200 a month: it gets its own computer in the cloud, signs into your accounts, clicks through your real apps, and hands back finished work instead of a draft for you to paste i ran one against my receipts folder on sunday and got back 14 filed, 2 it held because they needed a card number, and a saved method i never wrote myself Grok Bot is the one you train by doing your own job in front of it, and the whole handover fits in four messages tonight: 1. write out one job you did today the way you would brief a new hire: what has to be finished, which sites and files to work from, what to hand back, and where it stops and asks you 2. let it run once on something safe to get wrong, then correct the result until it is worth your name 3. say "save what we just did as a skill", and add the one rule about what always needs your approval 4. say "run that skill every weekday at 8 and post the result here. if the source is missing, tell me instead of using yesterday's numbers" xAI wrote that order into its own manual: one real job, then the saved method, then the clock. a schedule sitting on top of a method nobody checked replaces two hours of your clicking with two hours of your mistake turns out you never get to pick the brain, and that is the part i would argue about: the manual says there is no model picker for members or admins, no plan to add one, and the bill follows whichever model answered bookmark this, then open the piece below: which jobs deserve a worker of their own, and which ones quietly burn the seat ↓show more

An animal with no skeleton, no spine and no... teacher just engineered a fortress out of garbage it was already carrying. Then it vanished from the food chain in plain sight. Do not scroll past this one. Watch it with sound off, then watch it again, and this time ignore the octopus. Watch what the fish does. That is where the real story is hiding. I have spent more than ten years studying why some things explode on this platform and others die at 40 views. The pattern is almost always the same: a clip that makes the brain say "wait, that can't be right." This is the purest version of that I have ever seen. No narration. No music. No trick. A creature that has never read a manual, never met its mother, and will be dead before its second birthday, running a defense protocol that most militaries would pay billions to replicate. Below is everything the video will not tell you. Read it first and the clip hits ten times harder. Skip it and you will have watched a cool octopus. Read it and you will understand that you just watched a mind. Let's go. PART 1. NOBODY TAUGHT IT Start with the fact that should bother you most. The vast majority of octopus species live a life with zero parental contact. The mother lays her eggs, then guards them for weeks or months without eating. She fans them with water, cleans them, and fights off anything that comes close. When the eggs hatch, she dies. In many species this is not tragic accident. It is the design. Her body is programmed to shut down once the job is done. So the baby octopus enters the ocean alone. Tiny. Soft. Edible. Surrounded by thousands of animals that would eat it in one bite, and with no elder to show it which cracks are safe, which shadows are dangerous, which shapes mean food and which shapes mean death. Think about what that means for a second. A wolf pup learns to hunt by watching its pack. A crow learns which humans to hate by watching other crows. A human child gets roughly eighteen years of supervised practice before anyone lets them near a car. The octopus gets nothing. Whatever it needs to survive, it either carries as instinct or invents on the spot. And yet. Octopuses open jars from the inside. They unscrew lids. They have learned to turn off aquarium lights by squirting water at the bulb until it shorts out. A famous octopus named Inky escaped from a New Zealand aquarium in 2016 by squeezing through a gap in his tank, crawling across the floor, and sliding down a 164-foot drainpipe into the bay. Nobody saw it happen. Staff only noticed the empty tank. None of these behaviors were taught. Every one of them was figured out by an animal with a lifespan shorter than a smartphone contract. That is the first thing to hold in your head before you press play: this is not a trained animal. This is not a behavior passed down through generations of mentorship. This is a lone problem solver in a hostile place, solving a problem in real time. Everything that follows is that principle, pushed to its limit. PART 2. THE BRAIN THAT DOES NOT LIVE IN ITS HEAD Ask a neuroscientist where an octopus thinks and you will get an answer that sounds like a joke. Everywhere. An octopus has roughly 500 million neurons. That is in the neighborhood of a dog. But here is the part that breaks the usual model of how brains work: about two thirds of those neurons are not in the head at all. They are in the arms. Each arm carries its own dense cluster of nerve tissue. Each arm can taste what it touches, because the suckers are covered in chemical sensors. Each arm can decide, locally, what to do next without waiting for permission from the central brain. Cut an arm off and, for a while, it will keep reacting to touch, recoil from pain, and even grab at food. It is a semi-autonomous agent. So when you watch an octopus work, you are not watching one brain command eight limbs the way you command your hands. You are watching something closer to a committee. A central intelligence that sets the goal, and eight specialists that figure out the details on the ground. This is why octopuses are so fast at tasks that would paralyze a human with decision overload. Imagine trying to consciously control eight flexible limbs with no joints, no bones, and an essentially infinite number of possible positions. A human engineer who tries to model that in software hits what is called the curse of dimensionality. The number of possible states is too large to compute. The octopus solved this by not computing it centrally. It distributed the problem. The brain says "protect me." The arms handle the geometry. Keep that in mind when you watch what happens in this clip. The speed is the point. The coordination is the point. No part of that sequence looks like a creature thinking hard. It looks like a creature that has already decided, and a body that is simply executing. The human equivalent would be a musician who stopped thinking about individual notes years ago. What you are seeing is not hesitation. It is fluency. And fluency in a defensive maneuver, in an animal with no parents, is a very strange thing to see. PART 3. THE SKIN THAT SEES Here is a fact that sounds fake and is not. Octopus skin contains light-sensitive proteins, the same family of molecules that your eyes use to detect light. Researchers have shown that the skin of at least one octopus species reacts to light even when the eyes and brain are not involved. The animal wears a sensor across its entire body. Now add this. Most octopuses appear to be colorblind. Their eyes carry a single type of photoreceptor, which in theory means they cannot tell red from green. And yet they match the color of coral, rock, sand and algae with a precision that makes professional camouflage designers look like amateurs. How does a colorblind animal match colors it cannot see? Nobody fully knows. One idea is that the skin itself senses the surroundings and adjusts. Another is that the odd shape of the octopus pupil lets it exploit a lens effect that turns color differences into focus differences. Both ideas are still debated. What is not debated is the output: an animal that can change color, pattern and even surface texture in a fraction of a second. Smooth to bumpy. Flat to spiky. Visible to invisible. Under the skin sit millions of tiny organs called chromatophores. Each one is a sac of pigment surrounded by muscles. Tighten the muscles and the sac stretches wide and the color appears. Relax and the color shrinks to a dot. Each chromatophore is controlled directly by the nervous system, so the animal paints itself at the speed of thought. This is the standard octopus defense. Vanish into the background. Which makes what you are about to watch strange. Because the animal in this clip is doing something that does not look like the standard playbook. It is not hiding by blending in. It is hiding by becoming something else entirely. Something the predator has been trained by evolution to ignore. Camouflage is a trick. What you are watching is a lie told with architecture. PART 4. THE DISCOVERY THAT BROKE A RULE For a long time, science drew a firm line around tool use. It was supposed to be what separated the clever animals from the rest. Chimpanzees using sticks to fish termites. Crows bending wire into hooks. Dolphins carrying sponges on their snouts to protect themselves while foraging. Primates, birds, a few marine mammals. A small, elite club of big-brained vertebrates. Invertebrates were not invited. Insects, mollusks, worms. These were considered programmed machines. Reflex and instinct, nothing more. Then in 2009, researchers published a paper in Current Biology documenting something nobody had formally recorded before. In Indonesia, they watched veined octopuses, a species commonly called the coconut octopus, collect discarded coconut shell halves from the seafloor. They did not use them as a lair. They picked them up, carried them across open sand, and later assembled them into a hiding place when they needed one. The scientists argued that this met the definition of tool use. The animal was not just using an object that happened to be there. It was transporting it and deploying it later for a future benefit. That is a different cognitive category. It requires the animal to understand, in some functional sense, that an object it does not need right now will be useful in a few minutes. That is planning. And the animal doing the planning has a brain that looks nothing like ours, evolved on a completely separate branch of life that split from ours more than half a billion years ago. The paper made headlines for a reason. It forced a rewrite of one of the oldest assumptions in animal cognition. You did not need a backbone to invent. You did not need a cortex. You did not need a social group. You needed a nervous system under enough evolutionary pressure to solve a problem that could not be solved any other way. Now here is where it gets uncomfortable. The same researchers noticed that carrying the shells was expensive. The octopus could not swim or jet away efficiently with its arms full. It had to walk, slowly, on the tips of its arms, hauling its protection across open ground where everything wanted to eat it. Read that again. It deliberately slows itself down and exposes itself to danger in order to bring armor to a place where armor does not yet exist. That is a bet. A calculated, expensive, high stakes bet on a future threat. Most animals do not make bets like that. Most animals respond to what is in front of them. This one prepares. PART 5. THE WALK THAT SHOULD NOT WORK If you have ever seen the way this animal moves while hauling its gear, you already know the image I am talking about. It looks wrong. It looks like something built by an artist who had never seen an animal. Biologists call the gait stilt walking, and it is not a gimmick. It is a solution to a real engineering problem. A soft body has no rigid structure to push against. An octopus cannot sprint the way a lizard can. When it crawls, it uses its arms as flexible levers, anchoring suckers to the ground and pulling. That works fine when the arms are free. Add a load, and the problem changes. Now the arms are doing double duty: carrying and moving. A load that big would turn a normal crawl into a drag. So instead, the animal lifts its body up on two of its arms and uses the others to hold the cargo against its underside. It walks like a person on stilts, with the cargo bundled underneath. Notice what that tells you. First, the animal can reconfigure which arms do which job. There is no fixed assignment. In a human, the left hand and right hand have defined roles shaped by years of practice. In an octopus, every arm is a generalist that can switch roles on demand. Second, the animal is willing to trade speed for payload. It is not optimizing for escape. It is optimizing for survival once the escape is no longer possible. Third, and this is the part I find most chilling, the entire system is running while an enormous number of larger animals are watching. The walk is slow. The walk is visible. The walk is a dinner bell. And the octopus does it anyway, because it has done the math, in whatever way an octopus does math, and concluded that the payoff justifies the risk. You are looking at risk management from an animal that was never taught what risk is. PART 6. THE ARCHITECTURE OF A TWO-SECOND FORTRESS Let me describe what an engineer would see if you showed them the blueprint, without telling you what is in the clip. Start with the material. Natural shells are one of the most refined defensive structures on Earth. They are built from calcium carbonate arranged in layered microscopic bricks bound by a thin organic mortar. This brick-and-mortar structure is why a shell is so resistant to cracking. A crack that starts in one layer gets blunted and redirected at the next. Materials scientists have spent decades trying to copy it for body armor and ceramics. The octopus does not make this material. It borrows it. Now consider the geometry. A shell with a hinge is a pre-built enclosure with an opening mechanism already included. Two halves. One pivot. Whoever holds both halves together has a sealed container. The octopus does not need to invent a door. It only needs to hold one closed from the inside. That is the engineering insight. Do not build. Reuse. Repurpose. Notice, too, how little time this takes. Real armor, in the human world, takes weeks of manufacturing. A castle takes years. A suit of plate took a craftsman months. The octopus's fortress is constructed from loose parts in seconds, on a flat open plain with no scaffolding, no tools other than its own body, and while a predator is actively closing in. That is a response time most security teams would kill for. And it is modular. If one shell breaks, grab another. If the threat is gone, drop the whole thing and walk away light. There is no sunk cost, no attachment, no maintenance. A defensive system that exists only when it is needed and costs nothing when it is not. Engineers call this just-in-time architecture. Biologists call it behavior. I call it the most elegant piece of improvised design I have ever seen in nature. Now take that and ask the question that actually matters. Does it work? PART 7. THE FISH IS NOT STUPID Here is where most people watching get it wrong. They assume the predator is the fool in this story. They assume the fish is dumb, and the octopus gets away because its enemy cannot tell the difference. That is not what is happening. Predatory fish are some of the most efficient pattern recognition machines on the planet. A hunting fish does not scan the seafloor like a human reading a book. It runs a high speed filter. Does this shape move like food? Does it have the outline of prey? Is there an eye? A limb? A body plan I know how to eat? Biologists describe this as a search image. The predator holds a template in its head of what a meal looks like, and everything that does not match the template is, to its visual system, invisible. This is why a stick insect works. It is not that the predator cannot see the stick insect. It is that the stick insect does not match the template of anything edible. The octopus is exploiting exactly this. It does not need to be invisible. It needs to be irrelevant. A shell on the seafloor is not food. A shell on the seafloor is not interesting. A shell on the seafloor is part of the ground. The moment the octopus converts from "eight legged soft body" into "inedible object," the predator's filter rejects it, and the fish moves on to look for something else. Think about how remarkable that is. The octopus is not fighting the predator. It is not outrunning the predator. It is not even hiding from the predator. It is editing the predator's perception of reality by changing the category of object it belongs to. That is not camouflage. Camouflage says "I am hard to see." This says "I am not what you are looking for." It is the difference between hiding in a crowd and changing your uniform. One relies on not being noticed. The other relies on being noticed and dismissed. And the entire trick depends on one thing: timing. Too early, and the predator learns the pattern. Too late, and the predator has already committed to the strike. The octopus has to cross from one category to another inside a window that is a fraction of a heartbeat wide. Now go back and watch the fish. Watch how long it takes to lose interest. PART 8. THE ECONOMICS OF PARANOIA There is an ugly math underneath all of this, and it explains why evolution built an animal that behaves like this. An octopus lives fast and dies young. Many species live for one to two years. Some live only months. Compare that to a dog at ten or fifteen years, or a human at eighty. The octopus has to learn everything, achieve everything, and reproduce everything inside a window shorter than a school year. That is a brutal constraint. It means the animal cannot afford to make many mistakes. A single failed escape erases a life that took months to build. And it means every unit of energy has to be spent where it pays off. Carrying armor is expensive. It burns calories, slows the animal, and exposes it. Dropping armor is cheap. So why bother? Because the alternative is worse. When you live in the open, soft, and edible, with no claws, no poison and no speed advantage over the things hunting you, the expected cost of being caught without a defense is total. Not a scratch. Not a wound. Death. The only way to make that math work is to invest in protection before you need it, at a price that looks wasteful on any day when nothing attacks. This is paranoia, optimized by natural selection. And it is correct. That is the part that stays with you. The octopus is not being irrational by dragging a pile of hardware across the sand on a quiet afternoon. It is being exactly as rational as it needs to be. Thousands of generations of ancestors who did not carry armor are not here to argue the point. There is a lesson buried in that, and it applies far beyond octopuses. The people and companies and systems that survive long enough to matter are almost never the ones that optimize for the average day. They are the ones that stay ready for the worst day, and pay the small, boring, invisible cost of readiness every single time. Everyone loves the dramatic escape. Almost nobody notices the thousand quiet days of preparation that made the escape possible. The clip you are about to watch is the escape. Remember it is also the receipt for all the preparation that came before. PART 9. WHY ENGINEERS CANNOT STOP STARING AT THIS There is a reason robotics labs obsess over octopuses, and it is not cute factor. Almost every robot built in the last century is rigid. Metal frames, joints, motors, gears. That design is predictable and easy to model, but it fails in messy environments. Rigid machines break on uneven ground, get stuck in tight spaces, and cannot adapt when the shape of a task changes. The octopus has none of those problems. It can squeeze through any opening larger than its beak, the only hard part of its body. It can wrap around an object of any shape. It can change its effective length, thickness and stiffness in real time. It operates in three dimensions without a single bone. This has spawned an entire field called soft robotics. In 2016, a team at Harvard unveiled what they called the octobot, a small autonomous robot made entirely of soft materials, with no batteries and no electronics, powered by a chemical reaction that moves fuel through tiny channels. It was a toy compared with the real animal. But it proved something important: you can build a machine without a skeleton. Researchers are now working on octopus-inspired grippers for handling delicate objects, surgical tools that can navigate tight anatomy without damaging tissue, and search and rescue machines that can wriggle through rubble. Others are studying the way the skin changes texture and color, hoping to build adaptive materials that could shift appearance on demand. And all of that is chasing the hardware. Nobody has come close to copying the software. How does a distributed nervous system coordinate eight arms and a brain without a single central controller? How does it decide, in less time than it takes to blink, which of a million possible responses is the right one? How does it learn from a single experience and generalize to a situation it has never seen? These are the questions that keep the best researchers in the field awake. Not because they are hard to ask, but because the octopus answers them every day, in the open water, without apparent effort, and refuses to explain how. What you are about to see is not a nature clip. It is a field demonstration of a computing architecture that humanity does not yet know how to build. PART 10. THE QUESTION THAT SHOULD KEEP YOU UP Now the big one. Humans and octopuses last shared a common ancestor more than half a billion years ago. That ancestor was a tiny, simple, wormlike creature with little more than a nerve net. It had no complex brain, no eyes worth the name, and no culture. From that same starting point, two lineages went on to build minds. One of them gave rise to mammals, then primates, then us. We got a centralized brain, a cortex, language, tools, cities. The other gave rise to cephalopods. They got a distributed nervous system, eyes that look uncannily like ours and evolved completely independently, skin that thinks, and a capacity for problem solving that rivals animals with far more familiar brains. Intelligence did not happen once. It happened at least twice, by completely different routes. Sit with that. For most of human history, we assumed that our kind of mind was the destination. That evolution was a ladder and we were at the top. The octopus is evidence that it was never a ladder. It is a bush. And on a very different branch of the bush, something else built a brain, and used it to invent armor out of seafloor debris. The philosopher Peter Godfrey-Smith has described octopuses as the closest thing we will ever have to meeting an intelligent alien. Not a metaphor. A literal observation. A mind that arose with no shared architecture, no shared history and no shared senses, and yet recognizes problems, weighs risks, makes plans, and appears to enjoy playing. Their genome is stranger still. The first octopus genome sequenced revealed roughly 33,000 protein coding genes, more than we have. It showed expanded gene families linked to neural development and a surprising amount of editing at the RNA level, a trick that lets the animal rewrite the instructions of its own nervous system on the fly, in response to temperature and environment. This animal does not just run software. It patches it live. And it does all of this without a childhood, without a culture, without a language, and without a single other octopus to learn from. If a mind like that can exist, and invent a defense system from scrap, what else is out there that we have not recognized as thinking, because it does not look the way we expect thinking to look? One more thing before you press play, because it changes how the ending lands. The animal in this clip has no way of knowing whether the plan will work. There is no feedback loop in the moment. No second attempt if the first one fails. It commits fully, with its whole body, to a story the predator may or may not believe. Every defensive success you have ever read about in nature is really a record of a gamble that paid off, written by the survivors. PART 11. HOW TO WATCH THIS You have the context now. Here is the operating manual. First pass: just watch. Let the strangeness of it wash over you. Do not try to analyze. Second pass: forget the octopus. Watch the fish. Its attention. Its speed. The moment the interest drains out of its body. That is the moment the entire trick either works or fails, and it happens faster than you will believe. Third pass: watch the timing. Count how long the whole sequence takes from first move to finished shape. Then remember what I told you about weeks to build a castle and months to forge armor. Fourth pass: ask yourself what you would have done. Two seconds. No tools. No time to think. No one to ask. Just you, a threat approaching, and whatever you happen to be holding. Most people, honestly, freeze. The animal in this clip does not. It does not hesitate, does not panic, does not look around for help. It executes a plan it was never taught, using materials it carried in advance, at the exact moment the plan becomes necessary. If that does not make you feel something, I do not know what will. I have posted thousands of clips. Almost none of them have made me feel quietly humbled. This one did. Not because it is beautiful, although it is. Not because it is surprising, although it is. Because it is a reminder that intelligence is far more common, far more inventive and far more alien than we ever gave it credit for. And that somewhere right now, in a stretch of ocean you will probably never visit, a creature the size of your hand is solving a problem you would not have been able to solve, with a brain you would not be able to recognize, and it has never once needed our permission to be brilliant. Press play. Then press it again. And when it ends, tell me honestly: who was really the predator? If this changed how you think about minds, reposts are how the idea spreads. Follow for more clips that make you question what you thought you knew. P.S. Save this post. In a week you will remember the fish, not the octopus, and you will want to watch the whole thing again with fresh eyes. Send it to the one friend who still says animals are simple.show more

LangGraph. CrewAI. Agno. Which one to pick? The good... news is that this will not matter soon! Finally, we have a full picture of how the industry is solving this with just three open protocols that work across ALL frameworks. It's not about picking the best framework. Instead, it's about understanding how protocols create interoperability. The Agent Protocol Landscape shows how three complementary protocols are creating a universal language for Agents: > AG-UI (Agent-User Interaction): - The bi-directional connection between agentic backends and frontends. - This is how agents become truly interactive inside your apps, not just as chatbots, but collaborative co-workers. > MCP (Model Context Protocol): - The standard for how agents connect to tools, data, and workflows. > A2A (Agent-to-Agent): - The protocol for multi-agent coordination. - How agents delegate tasks and share intent across systems. These aren't competing standards. They're layers of the same stack and have handshakes with each other. So instead of building point-to-point integrations, you build to protocols. Moreover, you can integrate LangGraph, CrewAI, or Agno into the same frontend, without rewriting your UI logic. These protocols let everything work together. For instance: - Your LangGraph agent pulls data via MCP. - It delegates analysis to a CrewAI agent via A2A. - Results stream to your React app via AG-UI. - Users see real-time collaboration in your interface. This way, you can focus on building agent capabilities instead of integration mechanics. The protocols handle interoperability automatically. CopilotKit unifies this entire stack into one framework so you can build "Cursor for X" style apps without implementing each protocol from scratch. It gives you all three protocols, generative UI support, and production-ready infrastructure in one framework. I have shared this playbook in the replies! It breaks down handshakes, misconceptions, and real examples and shows exactly how to start building.show more

The new Product Page Header image in the App... Store will not matter. First you need users to actually want to enter your App Store listing. And the reality is that a lot of users NEVER enter the listing. They search for something, see the app in the search results and download it directly from there. So the important update here is the new Search Results image. It's an app banner inside the App Store search results. And you should think about it as a persuasive ad. COPYWRITING will be what matters. Not the fancy "branding". Users are searching because they want an answer to something. Your creative needs to give them that answer immediately. And this becomes even more important with Apple Ads. You can pay to appear for the keyword, but your creative still needs to convince the user that your app is the answer they were searching for.show more

Two posts in one day calling for the end... of John Cleese. One says he has allied himself with the darkest forces in Britain and has become the caricature he once mocked. The other, from a Muslim outlet, argues plainly that he should be de-platformed by respectable institutions. De-platformed. Respectable institutions. Say those words slowly and try to hear how they would have landed in 1972. The man co-wrote a film so blasphemous the BBC would not touch it, that was banned outright in Norway and by councils across England, and that two bishops went on television to denounce while he sat there and let them. Life of Brian survived the actual Church of England. It is now being reviewed by a newsletter. And the accusation is that he has become the caricature he once mocked, which is a remarkable sentence to write about a man whose entire body of work was aimed at exactly this. The Ministry of Silly Walks was a joke about a government department funding nonsense with a straight face. Now we have men in offices deciding which comedians remain permissible, and doing it with no jokes at all. At least the Ministry had the decency to be funny. Let me be fair, because I am always fair, and because being fair is what makes the rest of this land. Criticism is not censorship. Both of those men have every right to say Cleese is wrong, tasteless, or past it. That is speech. Nobody owes an 86 year old comedian applause. But there is a hard line between you are wrong and you should not be permitted to speak, and one of those posts crossed it in the headline. The first is an argument. The second is an eviction notice. And notice what the eviction is actually for. He said things about a religion. Not about a person. A religion, which is a set of ideas, and ideas do not have rights. People do. In England you can still say anything you like about the Church of England, and Cleese made a career of it, and nobody drafted a newsletter about respectable institutions. Here is what is wrong with Britain, and I say it with real affection because I am going there next month. A country that gave the world Magna Carta, the common law, and the funniest people who have ever lived has developed the habit of asking whether a joke is allowed. That question has an answer over here. John, if it gets any worse, come to America. Say every single thing you think. Say the wrong ones. Say the ones that would end you at home. Nobody will de-platform you and nobody will send a constable. Bring the silly walk. We have room. 🦋show more

CLAUDE CAN NOW BUILD THE KIND OF WEBSITE THAT... USUALLY COSTS $3,000+, FROM JUST A DESCRIPTION, NOT A TEMPLATE no drag-and-drop builder, no theme marketplace, no dev handoff, just a conversation tell it what the business does, who it's for, what tone it should hit, and Claude writes the copy, lays out the sections, and ships working code in the same thread the part people miss: the first draft is never the final draft, the real leverage is that revisions take one sentence instead of a support ticket to a freelancer "make the hero section darker," "add a pricing table," "this feels too corporate, loosen it up" each one lands in seconds, not days what used to be a week of back-and-forth with a designer is now an afternoon of talking to Claude until it feels right the tool isn't the moat anymore, knowing what to ask for isshow more

Running cold email campaigns just became a whole lot... easier Smartlead now runs an MCP server, which in plain terms means Claude can read and act on your live campaign data directly instead of working off a spreadsheet that went stale the moment you exported it. The workflow is worth walking through properly, because it is shorter than people expect. You generate an API key inside your account, point Claude at the server once, and from then on you ask for what you want in a sentence. Here is a prompt worth stealing in full: "Fetch all Smartlead clients, then get today's performance for each: emails sent, replied, positive replies, unique lead count. Compute reply rate per client, run a top and bottom performer analysis, format it as a daily client performance report, and post it to Slack." One paste, and it pulls live figures for every account, does the arithmetic, ranks the strongest and the weakest, and delivers the finished thing into the channel your team already sits in, before anyone has logged on for the day. Be clear about the division of labour, because it is what makes this useful rather than a novelty. Smartlead is the engine holding the campaigns, the mailboxes, the warmup and the reply data, and Claude is simply the interface you operate all of it through, so nothing about your sending changes and everything about how you interrogate it does. The effect people underestimate is on the questions you start asking. Once a report costs you a sentence rather than an afternoon, you stop rationing the ones that used to feel like too much trouble, and problems that used to surface on a Friday start surfacing on a Tuesday. Connect it with Claude through MCP and run one prompt against your own account today.show more

A CARNIVAL GAG LIKE THIS IS THREE TIMED EVENTS,... NOT ONE Bank-a-ball booth, plush prizes hanging overhead. She throws from behind camera, the ball bounces off the board and misses. Right as it lands, a guy in a cap pops up from behind the stand trying to catch it. She laughs, fixes her skirt, leans down for the next ball. None of it happened - not the booth, not the throw, not the guy who popped up at exactly the wrong moment. - Three moving things have to land on the same beat The ball needs its own arc, bounce, and miss off the board. Her throw needs the weight shift and follow-through of an actual pitch, not a mimed one. And the guy behind the stand has to time his entire pop-up to the second the ball lands, or the joke reads as staged instead of caught. None of the three gets generated with the other two in mind by default - getting all three to agree on one beat is the actual trick. Landing three unrelated timings on the same beat, on command, is the same process the article below turns into a business - across five different characters, every time.show more

Build AI agents on a time-aware knowledge graph! Utopia... is an open-source knowledge system that turns documents, databases, and connected sources into a temporal graph your agents can reason over. Most RAG systems are optimized for one question: what is relevant right now? That works until the underlying knowledge changes. A customer contract gets updated. A project owner changes. A policy is revised. A previous fact may no longer be true, but simply overwriting it means the system loses the history behind that change. Utopia handles this with a bitemporal knowledge graph. Each fact can track both when it was true in the real world and when the system learned about it. When something changes, the old fact is preserved instead of silently disappearing. That means an agent can reason about questions like: • What is true now? • What was true three months ago? • When did this information change? • What evidence was the conclusion based on? The graph is also ontology-aware, so documents are represented as entities, facts, and relationships instead of only chunks and embeddings. That gives the system more structure for reasoning across relationships, resolving entities, detecting conflicting facts, and deriving new information through explicit rules. Key capabilities: • Bitemporal knowledge graph for tracking how facts change over time • Provenance on facts so agents can trace where information came from • Conflict detection instead of silently overwriting contradictory information • Ontology-based reasoning across entities, relationships, and derived facts • Hybrid retrieval across full-text search, vector search, and graph traversal • MCP and agentic RAG support for exposing the knowledge layer directly to agents The interesting part is that this turns the knowledge base into more than a retrieval system. Instead of only finding relevant information, an agent can reason over what changed, what is still valid, how facts are connected, and where each conclusion came from. 100% open source. I've shared the GitHub repo in the comments!show more

There’s been two papers released in the past couple... months, one by Google and one by NVIDIA, that argue that ordering the documents retrieved by RAG systems can enhance performance. However, they both give two different strategies on HOW these documents should be ordered 🤔 Both papers agree on two main points: 1️⃣ There’s a fundamental issue in RAG - as more documents are retrieved, more irrelevant context (e.g., hard negatives) are introduced, which leads to confusion for the LLM and eventually degrades the quality of the generated output. This is called an inverted-U performance curve. 2️⃣ Ordering the retrieved documents is a key lever for optimizing RAG performance. Google Cloud researchers proposed ordering results based on relevance scores: The authors in this paper argue for relevance-based reordering, or ordering the retrieved chunks based on their similarity scores, so the most relevant documents are at the beginning and the end of the inputs to counter the “lost in the middle” effect. NVIDIA researchers proposed ordering results based on the original sequence of document chunks: The authors of this paper argue for Order-Preserving Reordering, or Order-Preserve RAG (OP-RAG), to maintain the logically coherent content flow of the document. So they preserved the original order of retrieved document chunks in the source text, instead of ranking them by relevance scores. So which one is right? It probably depends on the specific use case and dataset - relevance-based reordering could perform better in tasks where you need fast access to the most critical information (e.g., fact retrieval, QA systems), while order-preserving RAG might be better where you need to understand the sequential structure of information (e.g., narrative or legal documents). There are still so many uncertainties in AI - we don’t actually know what we’re doing, and it takes awhile to figure out the best strategies for most things! Excited to see more research about this.show more