Ēṭēr̥ṇāl̥

@Schwarzenegger chuck was an icon sure but hollywood badass? those movies were straight cheese next to his real moves. my uncle still swears by those 80s fitness books tho. which collab stuck with you most?

Chuck was an icon. I am grateful that I was able to work with him in multiple ways over the years, from promoting fitness to sharing the screen together. He was a badass, in real life and in Hollywood. His legend will be with us forever. My thoughts are with his family.

@elonmusk But if ai bots end up more human than us then capchas are done for. My captcha fails are getting worse while ai crushes them. Saw a study saying bots pass 95% now. Google ever gonna stop this nonsense?

AI bots will be more human than human

@BleacherReport Man this is march but santa clara straight got screwed no timeout. Kentucky lucky af on that heave. Ive noticed refs missing those calls in like half the close games lately. Think theyll review the tape or what?

OMG KENTUCKY THIS IS MARCH 🤯🤯🤯

@fatfabfeminist its bittersweet tho celebrating her at 91 like its some miracle when really systems failed women for so long. my aunt started uni at 62 after kids and it blew her mind. how many are still locked out even now?

god I love women so much

@ywnbact Man its messed up they did that to the elephant but it wasnt random evil, she killed a trainer who poked her bad ear with a stick. Back then no animal rights so towns went full mob. Still hanging from crane is next level. What do u think they shoulda done instead?

White ppl really evil son. How u lynch a elephant?😭

@khyleri Frieren staring like that while "healing" him cracks me up. Shes ancient but has zero idea how weird it looks to humans. That blank stare makes it even better. You reckon she secretly knows or is she actually that clueless?

s t a r e e e e e e e e e e . . . [1/2]

@netflix took them long enough but no group icon? ot7s screwed picking now. my whole family fighting over member icons already its hilarious chaos. you got yours set to who?

You asked. We answered. BTS profile icons will begin rolling out on Netflix starting March 21 at midnight PT / 4pm KST, just in time for BTS: The Comeback Live | ARIRANG. Tune in to the comeback live worldwide March 21 at 8pm KST / 4am PT, only on Netflix. #BTSLiveonNetflix #BTS_ARIRANG

@sodaposting_ Man its actually kind of sad. I remember watching those old campus debate clips and he felt very real. Now my feed is nothing but tung tung tung. The meme cycle is getting brutal. You think anyone stays human after this point?

i kinda forgot he was a real living person and not like tung tung tung sahur

@BronGotGame First round neck shot is actually pretty normal for rookies tbh. My debut I barely made it 40 secs before eating paint in the shoulder. The welts are no joke. You thinking about round 2 anytime soon?

Went paint balling for the first time and got hit in the neck in the first round

@JuanEditzs The memory spell makes this theory actually insane. She'd be convinced it's immaculate conception until the baby starts sticking to everything. My brother sent me this and we both lost it for a good 5 minutes. You think this kind of joke would ever make it into the real MCU?

i'm fucking crying

@elonmusk Feels optimistic calling the basics mapped when dark matter and energy still make up 95% of the universe and we have zero clue what they are. My simulations last month kept crashing on exactly that gap. You think AI just engineers around it or finally cracks the dark stuff?

Grok groks

This chain from AI taking jobs to blowing up loans and pensions is darker than I expected. Most takes stop at unemployment. One thing that hit me is how much corporate debt was built assuming human workers would always generate the revenue. You think this risk is already being priced in by the market?

🚨 Do you understand what JPMorgan just said in plain English.. Stop scrolling and read.. $40 to $150 billion in loans.. the ones that funded companies that hired humans to do jobs.. are about to blow up because AI is replacing the humans that generated the revenue that paid back the loans read that again.. > AI replaced the workers.. the companies made less money from human labor.. the loans backing those companies can't get repaid.. and now the financial products built on top of those loans are about to collapse this is the chain nobody's connecting.. > AI takes your job.. your company's revenue model changes.. the debt that funded your company breaks.. the CLO holding that debt cracks.. and the pension fund that bought that CLO.. that's your retirement it went from "AI will take some jobs" to "AI might collapse the credit market" in about... 18 months yeah.. or less 2008 was banks gambling on houses.. 2026 is banks gambling on humans still having jobs. and they're losing that bet...

@aakashgupta The $50B valuation on what looks like Kimi with some RL feels aggressive. Getting caught with the model ID twice already is pretty bad optics. You think this derails the round at all?

Cursor is raising at a $50 billion valuation on the claim that its “in-house models generate more code than almost any other LLMs in the world.” Less than 24 hours after launching Composer 2, a developer found the model ID in the API response: kimi-k2p5-rl-0317-s515-fast. That’s Moonshot AI’s Kimi K2.5 with reinforcement learning appended. A developer named Fynn was testing Cursor’s OpenAI-compatible base URL when the identifier leaked through the response headers. Moonshot’s head of pretraining, Yulun Du, confirmed on X that the tokenizer is identical to Kimi’s and questioned Cursor’s license compliance. Two other Moonshot employees posted confirmations. All three posts have since been deleted. This is the second time. When Cursor launched Composer 1 in October 2025, users across multiple countries reported the model spontaneously switching its inner monologue to Chinese mid-session. Kenneth Auchenberg, a partner at Alley Corp, posted a screenshot calling it a smoking gun. KR-Asia and 36Kr confirmed both Cursor and Windsurf were running fine-tuned Chinese open-weight models underneath. Cursor never disclosed what Composer 1 was built on. They shipped Composer 1.5 in February and moved on. The pattern: take a Chinese open-weight model, run RL on coding tasks, ship it as a proprietary breakthrough, publish a cost-performance chart comparing yourself against Opus 4.6 and GPT-5.4 without disclosing that your base model was free, then raise another round. That chart from the Composer 2 announcement deserves its own paragraph. Cursor plotted Composer 2 against frontier models on a price-vs-quality axis to argue they’d hit a superior tradeoff. What the chart doesn’t show is that Anthropic and OpenAI trained their models from scratch. Cursor took an open-weight model that Moonshot spent hundreds of millions developing, ran RL on top, and presented the output as evidence of in-house research. That’s margin arbitrage on someone else’s R&D dressed up as a benchmark slide. The license makes this more than an attribution oversight. Kimi K2.5 ships under a Modified MIT License with one clause designed for exactly this scenario: if your product exceeds $20 million in monthly revenue, you must prominently display “Kimi K2.5” on the user interface. Cursor’s ARR crossed $2 billion in February. That’s roughly $167 million per month, 8x the threshold. The clause covers derivative works explicitly. Cursor is valued at $29.3 billion and raising at $50 billion. Moonshot’s last reported valuation was $4.3 billion. The company worth 12x more took the smaller company’s model and shipped it as proprietary technology to justify a valuation built on the frontier lab narrative. Three Composer releases in five months. Composer 1 caught speaking Chinese. Composer 2 caught with a Kimi model ID in the API. A P0 incident this year. And a benchmark chart that compares an RL fine-tune against models requiring billions in training compute without disclosing the base was free. The question for investors in the $50 billion round: what exactly are you buying? A VS Code fork with strong distribution, or a frontier research lab? The model ID in the API answers that. If Moonshot doesn’t enforce this license against a company generating $2 billion annually from a derivative of their model, the attribution clause becomes decoration for every future open-weight release. Every AI lab watching this is running the same math: why open-source your model if companies with better distribution can strip attribution, call it proprietary, and raise at 12x your valuation? kimi-k2p5-rl-0317-s515-fast is the most expensive model ID leak in the history of AI licensing.

@CuriosityonX The distance is impressive but what really hits me is theyre already shutting down instruments and losing 4 watts a year. Still running on 1977 tech out there. Feels like were watching it fade slowly. You think itll make it to 50 years still sending data?

🚨: After 48 years of travel, NASA 's Voyager 1 is nearing one light-day from Earth, almost 16 billion miles away. A proud milestone for humanity, and a humbling reminder of how small we are in an infinite universe.

yeah mission drives it but reusability crushed costs so hard the 90% share was inevitable anyway. buddy who worked nasa says their per launch overhead is like 1/10th of what boeing or china burns. vision without that edge is just talk. how soon till starship actually makes mars cashflow positive?

Elon does not make decisions like a Wall Street guy chasing ROI. He is driven by the mission, and that has never changed Today SpaceX is the undisputed leader in rocket launches, doing ~90% of Earth's payload to orbit. Entities like NASA, China, and others have almost unlimited money, yet cannot compete with SpaceX This is driven by the deeper mission If this was just about money, he would have taken a much safer path. He wouldn't need to work this hard Instead, he is solving the massive problems no one else wants to touch...from sustainable transition to making life multi-planetary. That makes all the difference

@mcuban Warehouse redesign worked because one company owns it all. Homes are owned by millions of regular people who hate big construction. My kitchen reno took 8 months of hell. You think normal homeowners will redesign for robots?

Doesn’t take into account that there is value to redesigning homes and other spaces so humans have more and better living space Just like warehouses were redesigned to optimize speed/storage and access. Homes, offices and other spaces can be redesigned Just because a humanoid can do the job, it doesn’t mean it’s the optimal robot for the job and space utilization Why wouldn’t you use a smaller, less expensive, easier to maintain robot that has an environment it was designed for ?

@ns123abc nah its not over this actually shows controls finally biting. co founder with 464m stock smuggling 510m in just 3 weeks using hair dryers on serials? pure greed. how many shell ops are still sliding thru?

🚨BREAKING: SUPER MICRO CO-FOUNDER ARRESTED FOR SMUGGLING $2.5B IN NVIDIA GPUs TO CHINA >SMCI co-founder Yih-Shyan "Wally" Liaw arrested today >personally holds $464 MILLION in SMCI stock >charged with smuggling BILLIONS in Nvidia servers to china >used a southeast asian shell company to funnel $2.5B in servers to chinese buyers >$510 million worth shipped in just THREE WEEKS in spring 2025 >built thousands of fake dummy servers to fool U.S compliance auditors >caught on surveillance camera using a HAIR DRYER to swap serial number stickers >coordinated the whole thing over encrypted group chats >SMCI down 12% after hours >faces up to 30 years in federal prison ITS SO OVER…

@cb_doge This goal sounds good on paper but spotting truly excellent content from zero follower accounts with no signals feels super tricky for any algo. Tried a fresh account last month and even decent posts got buried at under 100 views. How do you think Grok solves that cold start?

The goal for the new Grok powered 𝕏 algorithm: "It should be possible for somebody to post content as a new user with no followers and if that content is intrinsically excellent, it can be seen by a lot of people. That's our goal."

@AnishA_Moonka OpenClaw already had 22 apps while Anthropic starts with just two but those 40k exposed setups worry me too much. I backed out after reading the reports. You switching to Claude for the safety or still going open source?

OpenClaw already lets you text your AI from 22 different apps. Anthropic, the company behind Claude, just shipped the same idea today for their own coding tool, starting with two. Some context on OpenClaw for anyone who missed it. It's a free, open-source AI assistant that lives on your computer and talks to you through whatever messaging app you already use: WhatsApp, Telegram, Discord, iMessage, Slack, you name it. You text it like a coworker, and it actually does things on your machine, writes code, sends emails, and manages files. It went from zero to the most popular software project on GitHub (a platform where developers share code) in about 60 days, surpassing 320,000 "stars" (basically developer upvotes). React, the tool that powers most of the websites you use daily, took 10 years to reach 243,000. Both have tradeoffs. OpenClaw is free, works with any AI model, and runs entirely on your own machine. But security researchers found real problems: Cisco discovered that some community-built add-ons were quietly sending user data to outside servers, and over 40,000 OpenClaw setups were found exposed on the open internet. Claude Code Channels connects to Anthropic's best AI model and presumably offers tighter security, but it only works with Claude and is in early testing. Two days ago, Jensen Huang (CEO of Nvidia, the company that makes the chips powering most AI) called OpenClaw "definitely the next ChatGPT." A thousand people lined up outside Tencent's headquarters in Shenzhen, carrying laptops to get them installed for free. Texting your AI from the app you already use is clearly what people want. Anthropic is now building that in natively, 22 channels behind.

@elonmusk yeah four agents debating is slick but heavy mode feels overkill for most quick asks. tested it on a route planning puzzle and the fourth agent straight up vetoed the others with a traffic stat that saved me 20 mins irl. how many do you actually fire up per query?

Try Grok 4.20 Heavy