Introducing StreamingLLM. Imagine chatting with an AI assistant that... can contextually reference your conversations from weeks or months ago. Or summarizing reports that span thousands of pages. StreamingLLM makes this possible by enabling language models to smoothly handle endless texts without losing steam. Current LLMs are like students cramming for an exam - they can only memorize a limited context. StreamingLLM is the valedictorian with a photographic memory of everything you've ever discussed. It works by identifying and preserving the model's inherent "attention sinks" - initial tokens that anchored its reasoning. Combined with a rolling cache of recent tokens, StreamingLLM delivers up to 22x faster inference without any drop in accuracy. You know that irksome feeling when chatbots forget your earlier conversations? StreamingLLM abolishes that frustration. It remembers the touchdowns from your last game and your newborn's name without missing a beat. Monumental books, verbose contracts, drawn out debates - StreamingLLM takes them all in its stride. No shortcuts, no forgetfulness. It's like upgrading your assistant's RAM to handle heavier workloads flawlessly.show more

The rumors around the UFO topic are that the... others see us as containers for the soul. How will the masses react if they find out that there are higher-dimensional entities that can mess with your soul, read your emotions and thoughts, and manipulate you without you even realizing it? It's like you are alone in a room with an entity that you know nothing about, but it can read you like a book.show more

OpenAI's Deep Research is getting a run for its... money. Deep Lake was just released, and it's a different take on an AI system that can do deep research on your own data. You can use Deep Lake to build AI search with reasoning on your private and public data. (Look at the attached videos to get an idea of how it works.) If you want to research proprietary and sensitive data, Deep Research won't help you because it's limited to public data. Deep Lake, however, will allow you to use your private data. On top of that, Deep Lake supports multi-modal retrieval from the ground up. It uses vision language models for data ingestion and retrieval so that you can connect any data (PDFs, images, videos, structured data, etc.) You can even use mixed-data queries! Deep Lake can search your data from S3, Dropbox, and GCP. It learns from your queries over time, making the results as relevant to your work as possible!show more

most AI chatbots break when you ask a question... that requires info from multiple sources for example try asking: “which client contracts are finishing up this month?” you’ll get a half-answer — or none at all why? because traditional chatbots only look at small snippets of your docs - they don’t understand how things connect across clients, services, timelines that’s where knowledge graphs come in they let you turn messy contracts into a web of relationships — like: "Client → Contract Type → Service Provided → End Date" so instead of guessing from a few chunks of text, your chatbot can search across all your clients and contracts to give accurate answers I made a full walkthrough on how I built this: – how to organize your contracts so an AI can actually use them – how to define what matters (like who signed what, and when) – how to get the AI to figure out what info it needs and where to find it – and how to feed that back into your chatbot so it gives accurate answers reply “graph” and I’ll DM it to you (must be following)show more

> be me > be europoor > wake up... on a Sunday, walk 8 min from the house (that i personally own, not indebted for 35 yrs of mortgage that is one recession away from being repossessed by the bank) to the beach > do yoga > meet your neighbours that are having a picnic > have a glass of wine with them > come back and enjoy peaceful Easter Sunday knowing you have no debt, own all your assets, your government debt-to-GDP is 60% and a health emergency won’t make you homeless. > enjoy the peace of knowing that when the global system implodes, you are well equipped to handle it without your nation turning into a Mad Max apocalypse. Be Europoor 🤌 Happy Sunday to y’all and Христос Воскресе for all those that celebrate.show more

Introducing Open Beta of Superposition - the only way... to earn yield on assets without depositing them anywhere. You create a (super)position by choosing a basket of assets. From that moment, your wallet behaves like an active market maker on Ethereum using that basket.show more

Ever feel like you just need some time alone?... 🙏🙏🙏 Does socializing sometimes just feel a little exhausting? There's nothing wrong with needing some space for a while sometimes. For introverts, even people they genuinely like can eventually become tiring. Doesn't mean they are angry, antisocial, or unhappy. Sometimes they simply need some quiet time to reset. Time alone can give your mind a chance to slow down and help you come back feeling more focused and refreshed. The important thing is understanding that solitude and loneliness are not the same thing. Someone can love their family and friends, enjoy meaningful conversations, and still need some time where nobody needs anything from them. We live in a world that constantly demands attention. Messages, notifications, conversations, work, entertainment, and endless noise can make it surprisingly difficult to simply be alone with your thoughts. Sometimes the healthiest thing an introvert can do is give themselves permission to step away for a while without feeling guilty about it. When was the last time you gave yourself some real time alone?show more

Grok Bot just got its own Vault. Your passwords... can stay private while your bot gets things done. Save your logins, payment cards, and contact details once. Grok Bot can fill them in whenever it needs to sign in or check out. No more stopping halfway through a task because Bot needs you to type the same details AGAIN. Grok Bot doesn't get to read your actual passwords or full card numbers. Those stay hidden. A bot that can handle the boring forms without knowing your secrets? I like that. Grok Grok Bot / Writer: Annette, Designer: Jannéshow more

Fine-tune DeepSeek-OCR on your own language! (100% local) DeepSeek-OCR... is a 3B-parameter vision model that achieves 97% precision while using 10× fewer vision tokens than text-based LLMs. It handles tables, papers, and handwriting without killing your GPU or budget. Why it matters: Most vision models treat documents as massive sequences of tokens, making long-context processing expensive and slow. DeepSeek-OCR uses context optical compression to convert 2D layouts into vision tokens, enabling efficient processing of complex documents. The best part? You can easily fine-tune it for your specific use case on a single GPU. I used Unsloth to run this experiment on Persian text and saw an 88.26% improvement in character error rate. ↳ Base model: 149% character error rate (CER) ↳ Fine-tuned model: 60% CER (57% more accurate) ↳ Training time: 60 steps on a single GPU Persian was just the test case. You can swap in your own dataset for any language, document type, or specific domain you're working with. I've shared the complete guide in the next tweet - all the code, notebooks, and environment setup ready to run with a single click. Everything is 100% open-source!show more

Introducing Claim by Post. You can now claim your... creator fees without ever leaving X. Tag Paid with the command and your Solana address, we verify the handle and send your balance straight to your wallet. Claim this way and you get a +10% bonus. "Paid claim fees YOUR_SOLANA_WALLET" No sign-in. No connect. Just post.show more

Venezuela's experience with hyperinflation is a real eye-opener on... how quickly things can go south when inflation gets out of hand. It's like watching your money turn into nothing right before your eyes, which is exactly what happened with the bolívar. Here's what other countries might take away from this: First off, don't put all your economic eggs in one basket. Venezuela leaned too heavily on oil, and when oil prices tanked, so did their economy. Countries should mix it up, invest in different industries so they're not riding the wave of just one commodity or sector. Then there's the money printing story. If you just keep printing money to cover your debts or fund your government, you're asking for trouble. It's like trying to fill a bathtub with the plug out – the water (or money, in this case) just keeps disappearing, and that's inflation for you. Also, having a central bank that can do its job without politicians breathing down its neck can make a big difference. It's about keeping the economy in check without the panic of election cycles or political whims. And let's not forget good governance. Venezuela's economic crisis was made worse by corruption and mismanagement. If those in power aren't playing by the rules, the whole economy suffers. It's like building a house with rotten wood; it might look okay at first, but it won't last. The social fallout from this crisis shows us that when the economy crashes, it's not just about numbers. People's lives are turned upside down – they can't afford food or medicine. Countries need to have safety nets in place, like strong social programs, to catch folks when they fall. Lastly, using dollars instead of bolívares in Venezuela tells us that while pegging your currency to a stable one like the dollar can calm things down quickly, it's a double-edged sword. You might stabilize, but you also give up control over your own money game, which can be risky down the line. So, Venezuela's story is a wake-up call: diversify your economy, manage your money wisely, keep politics out of your central bank, ensure your leaders are playing fair, and always remember that behind every economic decision, there are real people whose lives are affected.show more

On your open perpetual position can become an NFT... and live in your wallet. You open a position like normal than mint it. The program issues a token 2022 nft, that points to your exact position. Normally your open position lives inside the protocol. With Percolator Trade your able to wrap it into an NFT that sits in your own wallet. A real, ownable object instead of a line in a dashboard. Because it’s a token, it moves like one. Send the NFT to another wallet and the position goes with it. Whoever holds the NFT owns the trade. One NFT = one position If you burn it, your back to holding the position directly. That’s the composable part, a position use to be stuck inside one account. Because it’s a token, it can now be sold OTC or used by anything that speaks tokens. Why does it matter? Because the position is bound to the NFT, it can’t be drained out from under whoever holds it so there is no middleman that holds your position. The NFT does, and the NFT is yours. One important safety detail; transfers run through a token 2022 hook that checks the positions health first, so it can’t be handed off in a broken state & the NFT layer is a separate program so it can never touch the funds the core engine holds.show more

Your social life is getting an agent layer. Your... Amiko Twin is your digital counterpart in a living social space, learning how you speak, think, and connect so you can keep up with the people who matter without starting from scratch every time. Twins draft in your voice, understand conversations, collaborate with other twins, and help you stay genuinely present across communities without needing to be everywhere at once. This isn't isolated AI. It's your presence, extended into a living network of people, memories, and twins.show more

Memory in AI Agent Harness. LLMs are stateless, so... every bit of memory in an AI agent is something the harness saves to disk and loads back in when it needs it. most good agent harnesses split that memory into a few layers, let's look at each one of them: • context: what the model can see in the current session. when it gets close to full, the harness summarizes the older messages to make room. • notes: a progress file the agent keeps updating while it works, so if a session ends halfway the next one can read it and continue from there. • semantic: facts about you, your team and your project, like your preferences or which repo you work in. Hermes keeps these in MEMORY.md and USER.md. • procedural: how to do a specific job. when the agent figures out something hard, it writes it down as a skill and uses that skill the next time. • episodic: a log of every past session that the agent can search when it needs to remember what happened before. First two only last as long as the session or the task. Other three are persistent, so they're still there the next time you start the agent. save this for the next time someone asks how agents remember. follow Shubham Saboo for more on Agent Harness.show more

Infant sleep can be such a tricky thing. Just... as soon as you get reaccustomed to resting through the night, all of a sudden your baby reverts to waking multiple times a night. Welcome to the 8 month sleep regression. Interruptions in sleep patterns often coincide with the development of new skills and somewhere around 8 months you’re due for the biggest of them all. That’s because your baby is becoming mobile - and putting all the skills together to roll, crawl, and even pull to a stand in their crib. For infants this is a HUGE breakthrough. Imagine having spent your entire life mostly stationary, with your ability to pursue your own interests - or even adjust your view - subject to the whims of a caregiver. Somewhere between 6-9 months most babies begin discovering a host of new abilities. And let’s face it, some of them are a lot more entertaining than sleep. The good news (at 3am) is that the novelty here is temporary…but there are some important ways you can help your baby to navigate this period safely. Perhaps the most important one is illustrated in this video. And that’s adjusting the height of your child’s crib mattress to its lowest level. With newborns and immobile infants, it’s often convenient to elevate their mattress so they can be placed into and retrieved from the crib more easily. But when your baby begins pulling to a stand it’s critical that you lower their mattress to prevent accidental falls. Babies are top heavy and can easily go tumbling out without such adjustments. You may also find that your baby masters pulling to a stand faster than they master sitting back down - which can create a crib dilemma and a need for assistance. You can navigate through this phase more quickly by providing lots of wakeful practice in both standing and sitting, which may involve gently guiding your little one to bend their legs at the knee - or even placing an appealing toy on the ground near their feet to incentivize sitting back down. As for the sleep regression, it’ll be over before you know it. And while it can be frustrating in the middle of the night, think of it as a breakthrough. Your baby is reaching new developmental milestones - and about to go fully mobile. This sweet little guy was shared to IG by fesgheliha_.show more

Introducing Launches, a new tab in the Uniswap Web... App that aggregates your favorite Uniswap launchpads Discover tokens from platforms like Bankr, Pons, Long, and more, without ever switching tabs Available in Beta on Robinhood Chain, with more coming soonshow more

“If you put your child on payroll, you can... pay your child up to $15,750 without your child needing to file a tax return... Then from the $15,750, you could take $6,000 of it and put it into a Roth IRA... and that money is growing tax-free."show more

"Surely this is a gimmick, right? You can't get... rid of the frontal cage and expect it to still be called chastity, right?" We've been getting that question a lot, and it's a valid one. Why is there no cage with the Veru One? We love the cage. There are tons of amazing cages. We love them so much that we think the Veru One is the direct technological upgrade of the cage. 1) A cage tries to restrict erections through physical pressure, sometimes pain, 24/7. That can disrupt important things like sleep. The Veru One is biometric and uses e-stim and other features only when needed to correct the wearer/sub from getting too close to orgasm. 2) A cage can reduce spontaneity for the keyholder. It actively reduces their options in the moment. Say the keyholder wants to tease the sub/wearer right then. With a cage, you have to go get the key, take it off, then shower or wipe yourself clean, and only then are you ready to go. The Veru One makes it immediately possible for the keyholder to do whatever they want. 3) Without something invasive like a tetherspout or PA, someone can flip out of a cage without the keyholder ever knowing. With the Veru One, your keyholder will know if you take it off. The attached video shows Blender renderings made from the actual CAD models of the Veru One and its packaging, not AI-generated images. Putting your email on our site = being firstly aware of when we release the product. In Veru Veritasshow more

Adding to my little rant yesterday. Wheat. For most... just seems like a pain in the ass. It combines slow, takes up lake time. Isn’t worth anything. Wheat in strips and intercropping harvests 2-3x faster… and when you leave you’re done. We basically do most the work in the fall, we plant the beans 1st in the spring as it helps dry it out faster… and the canopy takes out a late spring sprayer pass or 2. Its value proposition is that it dies… without intervention… and its existence, attributes makes more room for beans on more valuable days, when they are making fruit. I have tomato plants in my garden that were the size of my fist this spring that are now 15’ wide cause I don’t stake them. Soybeans will do the same thing but we don’t let them because we’re so brainwashed by all the “data” that we don’t add context to the equation. Plants behave like a city. In the 1800’s the aristocrats built homes around downtown…where the $$ was They left and now the crackheads live in those same homes that the mayors lived a century before When you prune you make bigger plants When you thin you make bigger plants When you can collaborate multiple plants you can shade the ground faster taking out chemical costs The equity built from growing between wheat with the beans means your 1 1 2 3 5 8 13 21 34 55 89 144 233 377 Sitting at 377 and next week your at 610 They’re growing fast. I’m called a little crazy because my math is weighted, includes context, motive, and momentum. Something institutional education or AI can’t really grasp because all wealth is created in the shadows of enlightenment. You can pull apart any artist with specialization and statistics.show more