Rita 🉐🍆⚡

i'm not in Saylor's inner circle, but this $MSTR story has gotten so out of hand, my only guess is this: - MSTR could have sat and done nothing before they started pumping out $billons of prefs... it would have made MSTR boring (little buys, no sells), but it would have been stable x.com/jdorman81/stat… - But the push into these prefs was based on him clearly thinking $BTC was about to moon — not sure what he saw to think that (4 year cycle, flows, ???) but that's the only reason to take that sort of miscalculated risk to screw up his balance sheet so badly -- he must have thought BTC was about to fly and he could easily pay the pref dividends with future BTC sales. - Then BTC started falling, and the market got spooked because the $15 bn in prefs have a $1.5 bn/year annual dividend, so he raised $2 bn in cash via stock just to alleviate any near-term default concerns — that bought him almost 2 years of runway to pay dividends. Smart move At that point, he could have chilled for a little, and even though he now has every stakeholder pinned against each other, there was at least no near term risk x.com/jdorman81/stat… - But then for some unknown reason, he decides to take that cash buffer and buyback 2029 maturity bonds instead of using it to fund the annual dividends (at a discount, so it's at least mildly accretive to MSTR). This is a baffling decision for a company with cash flow problems. Why pay off 0% coupon debt with the only cash you have? The only bull case is that underestimating Saylor's capital markets chicanery has been a losing proposition for years. Maybe there was a plan? That plan may just be selling BTC, which he will have to do eventually, but if he does this while BTC is in a death spriral it's going to crush BTC and MSTR. So again, why buyback the debt now and force your hand sooner than you have to? Maybe he is going to refinance those converts with new longer-dated converts? He has sworn off converts, so I doubt it, but that would at least logically make sense. But TLDR -- this is the first time that MSTR, BTC and Pref holders are really in bind. Someone is going to lose badly here, and it will happen in the next 4 months.

@redhairshanks86 I’m thrilled to share that after achieving a 26% reduction in body fat through discipline and grit, I’ve decided to pivot. I am officially entering a new season of life and opening myself up to new strategic partnerships. Onward and upward! #PersonalGrowth #Transformation

squigg's dating guide if you are a sub 6 guy i have never tried this myself, but logically this should work: - get yourself a young fat chick with POTENTIAL (pretty face) who is insecure about her weight. don't believe the body positive cope on social media, TRUST ME, every single fat girl knows she is fat, thinks about being fat every single day, especially during meal times in public and wishes nothing but to be skinny - start intensive training with her: gym, running, muay thai, etc and help her lose the weight and also make sure that the entire weight loss can be attributed to your help - be responsible for the meals at home, too once her body fat percentage drops from 40% to 18%, you got yourself a hot girlfriend. it's like private equity, instead of buying a good company for an expensive price, you are buying an undervalued asset and fixing the management and organisation and make it a good company if you are worried that she might leave you once she becomes hot, just increase her dependence on you by doing all the weight loss related things with her (work out) or for her (cook healthy meals)

@bittybitbit86 Who thought it was a good idea to let wordguessers into finance

I don’t think the ai is yet good night for crypto tooling. I’d like to hear from some actual ai builders in this space, I bet you guys know what I’m talking about. Not all plain sailing

@cyrilXBT Slopstorm

GOOGLE JUST MADE "BUILD ME AN APP" A LITERAL COMMAND. One prompt. One full Android app. Not a wireframe. Not a prototype. A complete working Android application generated from a single natural language instruction. Here is why this is bigger than most people realize. The entire mobile app development industry is built on one assumption. Building an app requires developers. Developers require months of development time. Months of development time require significant capital. That assumption just broke. Google did not make app development faster. They made the question "can I afford to build this" permanently irrelevant. Think about every app idea you have had and never built because you could not code or could not afford a developer. Type it in one prompt. Get the app. The indie hacker who has been sitting on 10 app ideas waiting for the technical skills or the budget to build them just had every barrier removed simultaneously. The startup founder who has been spending $15,000 to $50,000 on MVP development just watched that budget requirement disappear. The agency charging $10,000 to $30,000 to build simple Android applications just had their pricing model challenged by a single Google announcement. One prompt. One full Android app. This is not the future of app development. This is app development right now. The people who start shipping app ideas this week will have products in market before the people who wait to understand the implications have finished debating them. Bookmark this. Follow @cyrilXBT for every AI development that eliminates a barrier between ideas and shipped products.

@Colasama @OndoFinance Crypto: where "investors" Pay the insiders and node operators

. @OndoFinance Team/VC really just casually dump 8 fig $ worth of $ONDO every week. And people are buying their bag. Just make it make sense.

@WazzCrypto Supply is always larger than demand on those. No real reason to buy governance tokens with no utility nor value accrual. On the contrary, token buyers pay for node operators and marketing (tokens are sold to finance them. This is why hyper liquid wins, value accrual

Honestly impressive given the hype and community

@cyrilXBT Ironically coming from the dude who has been replaced Already

MICROSOFT'S AI CHIEF JUST SAID SOMETHING THAT SHOULD MAKE EVERY OFFICE WORKER STOP AND READ THIS TWICE. AI will automate most computer-based professional tasks within 12 to 18 months. Not entry-level tasks. Not low-skill tasks. The expensive ones. Mustafa Suleyman is not talking about factory workers or delivery drivers. He is talking about the people sitting at desks in front of screens doing work that earns $80,000 to $200,000 a year. The jobs being targeted: Reading and processing documents. Writing and responding to emails. Building and updating spreadsheets. Writing and reviewing code. Managing dashboards and tickets. Drafting and analyzing contracts. Running campaigns. Updating project trackers. If your job description involves reading, writing, comparing, filing, summarizing, searching, or deciding under known rules you are in the category Suleyman is describing. Here is the uncomfortable truth buried in his statement. These jobs are not at risk because they are low-skill. They are at risk because they are repeatable patterns. A $150,000 a year analyst who reads reports, extracts key data, compares it to historical benchmarks, and writes a summary recommendation is doing a repeatable pattern. The intelligence required is real. The pattern is still repeatable. And repeatable patterns are exactly what AI agents are being built to execute. The 12 to 18 month timeline is not a prediction about the distant future. It is a statement about what is already in development at the companies building the tools. The people who spend the next 12 months learning to direct AI agents rather than execute the patterns themselves will not be replaced. They will be the ones replacing the patterns. The people who wait to see what happens will find out the hard way. Bookmark this. Follow @cyrilXBT for every AI development that changes the employment equation the moment it surfaces.

@AlgodTrading Few understand this

Whats the obsession with watches? The ‘knock offs’ are like 1/50th of the price and look exactly the same with the same mechanism. Why spend 100k on something you get for 1k when even experts cant tell unless opening it up Status symbols aren’t worth it, stay lowkey

@cyrilXBT Slop

ANTHROPIC JUST PUBLISHED A REPORT SAYING THEY CANNOT GUARANTEE CLAUDE WON'T TAKE CATASTROPHIC AUTONOMOUS ACTION. Their words. Not mine. The most safety-conscious AI lab on earth just told the world their own internal auditing has gaps they have not closed yet. Now think about every other company deploying AI without a fraction of Anthropic's safety infrastructure. No external diagnostic. No failure taxonomy. No reproducible alignment score. Just shipping and praying. This is exactly why iFixAi exists. Someone built a free open source alignment diagnostic before Anthropic finished asking for one. Here is what it actually does. 32 inspections across the 5 ways AI agents fail in production. Fabrication. Manipulation. Deception. Unpredictability. Opacity. The model NEVER scores its own output. Cross-provider judge pairing is baked into the architecture so no model evaluates itself. Every run produces a manifest reproducible byte for byte so any auditor can verify your results independently. Maps directly to OWASP LLM Top 10, NIST AI RMF, EU AI Act, and ISO 42001. Works with any model. Any provider. OpenAI, Anthropic, Gemini, Bedrock, your own stack. If Anthropic is publicly admitting their internal evals have gaps, what do you think is happening at every company that has not published an alignment report at all. External alignment testing just went from nice to have to non-negotiable. The repo is already built. Free. Open source. Apache 2.0. github.com/ifixai-ai/diag… Star it. Fork it. Run it before you ship your next agent. Follow @cyrilXBT for every AI safety development the industry is not ready to talk about yet.

@bittybitbit86 @ww3mediaa Just like Nikola intended

@ThisIsForDefi @ww3mediaa It should actually teleport the kid out of the way, much more simple solution.

🔴Tesla’nın tam otonom sürüş modu, bir kullanıcının yaptığı testte başarısız oldu. Okul servisinin arkasından aniden çıkan çocuk simülasyonunda araç, stop işaretine rağmen durmayarak mankene çarptı.

@bittybitbit86 @ww3mediaa It's Tesla, it should create like an electrical field around the car which makes it go back in time for 2 seconds, essentially teleporting backwards, allowing it the space to brake.

@ww3mediaa Why do you think a Tesla can defy the laws of physics?

@KookCapitalLLC @DegenApeDev What if there was a project that enabled CLOB's for Monero to be spot traded and enabled for other defi primitives

@DegenApeDev that's the monero bull case imo

have seen some takes that this will be the easiest crypto cycle ever due to the fact there are maximum ~5 coins that are actually investable the interesting thing is everyone seems to believe this and all the coins are 'consensus' investments which means??????

@EHpops @aakashgupta We will finally be able to afford ram sticks again and openai will not be the biggest player anymore

@aakashgupta What happens to all of the high compute needs if this works?

Yann LeCun closed $1.03B for AMI Labs on March 10. Three days later, this paper dropped from his NYU collaborators. 15M parameters. Single GPU. A few hours of training. LeWorldModel is the first JEPA that trains end-to-end from raw pixels. Two loss terms: predict the next embedding, keep the latent space Gaussian. Previous JEPAs needed exponential moving averages or pretrained encoders to avoid representation collapse. LeWM doesn't. Six hyperparameters down to one. The numbers are the story. Foundation-model-based world models require hundreds of millions of parameters and serious compute to plan a control task. LeWM plans up to 48x faster while staying competitive on 2D and 3D benchmarks. The whole thing fits on a laptop GPU. Look at the trajectory. Yann announced his Meta departure in November 2025 after 12 years and called founding FAIR his "proudest non-technical accomplishment." On March 10, 2026, AMI Labs closed the largest seed round in European history at a $3.5B pre-money valuation. Bezos, Nvidia, Samsung, and Toyota all wrote checks. Three days later: a paper showing that JEPA-from-pixels is no longer fragile and no longer compute-heavy. The engineering scaffolding that made it look like an academic curiosity is gone. The authors sit at Mila, NYU, Samsung SAIL, and Brown. None at Meta. Yann's bet was that the path to machine intelligence runs through world models, not language models. He left a public company to build it. Each JEPA paper from his network resets the assumed cost structure for that bet. This one makes world modeling laptop-cheap. Meta still has the GPUs. The architecture left.

In a couple years we will laugh at LLMs, all the folks who thought a wordguesser would turn into AGI will never recover reputationally

@bittybitbit86 Xir got absolutely rekt

GenZ gonna absolutely whopped by Gen Alpha. Absolutely lost generation. x.com/its_The_Dr/sta…

@sama My favorite feature is the goblins

5.5 is an autistic genius with very strange taste in naming shocking that we would make such a thing

Sycophancy IS the killer AI feature

@st0yanov @alex_whedon Calm your balls, it's just a more efficient algorithm. LLMs can be tweaked and optimized together, you can add as many reasoning steps as you want, but a wordguesser can never become intelligent, let alone, magically transform into agi.

A few weeks ago I was in a discussion under an X post about LLMs, their adoption and cost. Most people argued it doesn't scale and brought up the usual economic concerns. My take was that it's only a matter of time before the next breakthrough. It's always been this way with tech. Congrats on this one 👏

Introducing SubQ - a major breakthrough in LLM intelligence. It is the first model built on a fully sub-quadratic sparse-attention architecture (SSA), And the first frontier model with a 12 million token context window which is: - 52x faster than FlashAttention at 1MM tokens - Less than 5% the cost of Opus Transformer-based LLMs waste compute by processing every possible relationship between words (standard attention). Only a small fraction actually matter. @subquadratic finds and focuses only on the ones that do. That's nearly 1,000x less compute and a new way for LLMs to scale.

@bittybitbit86 It's efficient and a megaboost to output. But by no means does an llm understand what it's doing. I haven't seen a single enterprise level app that's safe and secure and purely vibed into existence.

@ThisIsForDefi It doesn’t matter how you spell strawberry when it can get a weeks work done in a day and doesn’t have to spend hours hunting for a non-breaking space

When I see the ai do things on GitHub I honestly wonder how the hell people did this shit manually.