andergrazh

334 posts

andergrazh banner
andergrazh

andergrazh

@graaazh

Katılım Haziran 2023
394 Takip Edilen47 Takipçiler
andergrazh retweetledi
Engineering at Meta
Engineering at Meta@Meta_Engineers·
Today we're announcing LevelUp: a free, four-week training program that takes people with no prior experience and prepares them to work as fiber technicians on data center construction sites across the US. We built this program with CBRE because the fiber technician field, and the broader construction industry, is facing a nationwide shortage at a time when data center demand is higher than ever. How it works: 🔧 Classroom instruction, hands-on labs + team activities covering transferable technical skills 🎓 Graduates have the opportunity to work at Meta's US construction sites through our contractor network 🤝 Open to everyone from recent high school grads to mid-career professionals Since 2010, Meta's data center projects have supported 30,000+ skilled trade jobs during construction + 5,000+ permanent operational roles. LevelUp is about building the pipeline to keep that going. Learn more: go.meta.me/0eb3f6
Engineering at Meta tweet media
English
364
1.5K
14.6K
8M
andergrazh retweetledi
Evan Luthra
Evan Luthra@EvanLuthra·
The Head of Claude Code at Anthropic said he hasn’t written code by hand in months. In 2 days he shipped 49 full features. All written 100% by AI. He just dropped a 30 min talk on exactly how he does it. Worth more than any $500 vibe coding course. Bookmark it:
English
171
759
9.1K
1.5M
andergrazh retweetledi
Milk Road AI
Milk Road AI@MilkRoadAI·
The man who turned 225 million dollars into 5.5 billion dollars explained on camera exactly why he made his biggest bet. This is Leopold Aschenbrenner, the same person whose Bloom Energy position is now worth close to 2 billion dollars after Oracle's 2.8 gigawatt fuel cell deal laying out the power math that drove every investment decision his fund has made. In 2022, the GPT-4 training cluster consumed roughly 10 megawatts of power and cost about 500 million dollars. AI compute has been scaling at roughly half an order of magnitude per year meaning the largest training cluster doubles in power requirement every 12 to 18 months without stopping. By 2024, the largest cluster was approximately 100 megawatts, the equivalent of 100,000 high-end GPUs and costs in the billions. By 2026, right now, the leading training cluster requires a full gigawatt of continuous power and that is the output of a large nuclear reactor. By 2028, the projection reaches 10 gigawatts, more electricity than most US states generate in total. By 2030, the trillion-dollar cluster, 100 gigawatts, over 20 percent of everything the United States currently produces in electricity, consumed by a single AI training installation. And that is just the training cluster. Inference, the continuous compute required to actually run AI products for hundreds of millions of users requires multiples of that on top. Meanwhile, total US electricity production has barely grown five percent over the last decade and the grid was not built for this. And the transformer shortage, the switchgear backorders, and the canceled data center projects that are making headlines right now are the first visible symptoms of a power system hitting a wall that Aschenbrenner saw coming years before the rest of the market. This is exactly why he built a 875 million dollar position in Bloom Energy, a company that generates electricity directly at the data center site using fuel cells, completely bypassing the grid bottleneck that is already stopping half of all planned US data centers from opening on schedule. The thesis was never complicated. The bottleneck in AI is not the models, not the chips, and not the software. The bottleneck is whether civilization can generate enough electricity to run the machines fast enough to matter.
Milk Road AI@MilkRoadAI

This is WILD A 24 year old who got fired from OpenAI just turned 225 million dollars into 5.5 billion dollars in under twelve months by betting on something the entire Wall Street AI trade completely missed. Leopold Aschenbrenner ran safety research at OpenAI until the company let him go. He then wrote a 165 page essay arguing that AGI was arriving faster than any investor understood, and that the people who would win were not the ones who owned the best AI model. They were the ones who owned the electricity. That thesis became a hedge fund called Situational Awareness LP, and his Q4 2025 SEC filing reveals one of the most concentrated bets in modern financial history. His largest single position is an 875 million dollar stake in Bloom Energy, a fuel cell company that generates power directly at the data center site, bypassing the power grid entirely. One of our analysts at Milk Road called this exact play two months ago, took a massive position in Bloom Energy, and it is already up over 40 percent. After Bloom announced a 2.8 gigawatt fuel cell deal with Oracle this week, the stock surged 15 percent in a single after-hours session and that 875 million dollar position is now worth close to 2 billion dollars. His other major positions follow the same electricity-first logic, 700 million dollars in CoreWeave, a massive short on Infosys betting that AI coding agents destroy the outsourced IT business, Intel call options printing multiples on a 53 percent run And a 10 percent stake in Core Scientific, a Bitcoin miner converting its power infrastructure into AI data center hosting. The entire Wall Street AI trade was piled into model companies and chip companies. Aschenbrenner looked at the same thesis and concluded the real bottleneck was whether the power grid could deliver enough electricity to run the models. He was right, and the returns are public record. Our analysts are finding the next plays before they make headlines. If you want access to the full thesis and what we are watching right now, go PRO. Link below!

English
72
215
3K
1.2M
andergrazh retweetledi
The Kobeissi Letter
The Kobeissi Letter@KobeissiLetter·
The IPO market is about to make history: Anthropic's pre-IPO stock, one of the most anticipated IPOs of all time, now has an implied market cap of $851 billion on Jupiter. In October 2025, the value of a share of Anthropic's pre-IPO stock on Jupiter was as low as ~$122. Just 7 months later, and these shares are up nearly +640%, to ~$900. These pre-IPO shares are structured instruments backed 1:1 by SPV exposure, giving investors direct price exposure to Anthropic before the IPO launches, allowing the market to dictate potential valuation. Between SpaceX's anticipated $1.7+ trillion IPO and both Anthropic and OpenAI targeting $1+ trillion, the IPO market is about to witness history. The AI Revolution is accelerating.
The Kobeissi Letter tweet media
English
231
421
4.2K
666.6K
andergrazh retweetledi
Cointelegraph
Cointelegraph@Cointelegraph·
🔥 LATEST: Hyperliquid, built by Jeffrey Yan, reportedly generated over $900M in profit with just 11 employees, making it one of the most profitable startups per employee globally.
Cointelegraph tweet mediaCointelegraph tweet media
English
245
683
6.9K
496.4K
andergrazh retweetledi
Elon Musk
Elon Musk@elonmusk·
Engineering is real magic
English
12.3K
44.5K
526.7K
100.4M
andergrazh retweetledi
TheNewPhysics
TheNewPhysics@CharlesMullins2·
🚨 BREAKING: Scientists just learned how to control magnetism at the atomic level. Not materials. Not circuits. Individual spin patterns. Read that again. Instead of using electric charge… they’re using the spin of electrons to store and process data. And it gets crazier: They can create tiny magnetic whirlpools called skyrmions… that move with almost no energy and can store massive amounts of data This means: Faster computers Lower power usage Ultra-dense memory But the real shift is this: We’re not just building electronics anymore… we’re engineering structure at the smallest possible scale. So the real question is: If information can be stored in spin itself… what limits computation? Follow me I’m tracking where physics becomes technology.
English
529
3.1K
14.5K
666.5K
andergrazh retweetledi
Muhammad Ayan
Muhammad Ayan@socialwithaayan·
SOMEONE JUST KILLED EVERY SKETCHY VIDEO DOWNLOADER SITE ON THE INTERNET. It's called ReClip. Self-hosted. Open source. Free. Paste a link from YouTube, TikTok, Instagram, Twitter/X, or 1000+ other sites. Download as MP4 or MP3. That's it. No ads. No popups. No trackers. No rate limits. No account. No sketchy installer. Your machine. Your downloads. Your data. Here is the full feature set: -> 1000+ supported sites including YouTube, TikTok, Instagram, and Twitter/X -> MP4 or MP3 output, your choice -> Resolution selector before you download -> Batch download. Paste multiple links at once. -> Clean web UI that runs in any browser -> Lightweight self-hosted setup It went from 0 to 1.4K GitHub stars in 9 days. 239 forks already. Here is why that number matters: Every person who found a janky ad-covered downloader site in the last 10 years was waiting for this. A clean, private, self-hosted alternative that just works. No business model built on your data. No premium tier. No upsells. Built in HTML. MIT License. Two contributors. Ships in days. 100% Open Source. Free forever.
Muhammad Ayan tweet media
English
249
1.4K
13.3K
777.2K
andergrazh retweetledi
Anish Moonka
Anish Moonka@anishmoonka·
The battery in your phone exists because a physicist in 1799 tried to copy a fish. That fish was the electric eel, and 80% of its body is a living power source. An electric eel is not an eel. It is a knifefish, closer to catfish than to any real eel. And its body is built backward. All the normal organs (heart, stomach, brain) are crammed into the front 20%, right behind its head. The remaining 80% is the electric organ. It has three separate electric organs, each doing a different job. Two of them produce high-voltage shocks for hunting and scaring off predators. The third one puts out weak 10-volt pulses that work like built-in sonar. The eel has awful eyesight and lives in dark, muddy Amazon rivers, so it uses those pulses to "see" by sensing how the electrical field bends around nearby objects. The cells that make all of this happen are called electrocytes, tiny disc-shaped muscle cells that gave up the ability to flex and instead learned to produce a small electrical charge. Each one makes about 0.15 volts on its own. But a full-grown eel stacks around 6,000 of these cells end to end in a single column, with roughly 35 columns running side by side on each half of its body. The voltages add up. Same principle as stacking batteries in a flashlight. Until 2019, scientists thought there was only one species of electric eel. Then a team, including researchers from the Smithsonian, found there were actually three. The strongest one, Electrophorus voltai, was measured at 860 volts. Roughly seven times what comes out of a US wall socket. Kenneth Catania, a biologist at Vanderbilt, published a paper in Science in 2014 after years of studying how eels use all that voltage. He found that the eel's high-voltage attack works almost identically to a Taser (the stun device law enforcement uses). A Taser fires 19 high-voltage pulses per second to override the nerves controlling your muscles, forcing your whole body to seize up involuntarily. The eel fires 400 pulses per second. Twenty-one times faster. It can freeze a fish solid in 3 milliseconds without even touching it. Catania also discovered the eel has a second trick: it sends paired electrical pulses that force hidden fish to twitch against their will, creating a tiny ripple that gives away their hiding spot. The connection to the battery in your phone is not a metaphor. In the late 1790s, an Italian physicist named Alessandro Volta noticed the stacked-cell structure in electric fish and tried to replicate it with alternating discs of zinc and copper separated by cardboard soaked in salt water. It worked. He built the first battery ever made in 1799 and called it an "artificial electric organ." The unit we call the "volt" is named after him. The strongest eel species, found 220 years later, was named Electrophorus voltai in his honor, closing a loop that started when a fish taught a physicist how to store electricity.
Nature is Amazing ☘️@AMAZlNGNATURE

Electric Eel power demonstration using LED's

English
33
1.3K
9.6K
716.1K
andergrazh retweetledi
Dom Lucre | Stealer of Narratives
🔥🚨BREAKING: Nectome The Portland Oregon tech company that Sam Altman is on the waitlist for since 2016 to kill him so they can upload his Brian to the internet which would allow him to live forever just confirmed that they have successful reanimated a dead pig’s brain meaning they are about to be ready to take human brains like Sam Altman’s and upload their consciousness online. An entire mammalian brain has been successfully preserved using a technique that will now be offered to people who are terminally ill. The intention is to preserve all the neural information thought necessary to one day reconstruct the mindof the person it once belonged to. Borys Wróbel at Nectome: “They would need to donate their brain and body for scientific research,” in Portland, Oregon, a research company focused on memory preservation. “But what we are offering, as a company, is for their body and brain to be kept, essentially indefinitely, in the hope that sometime, in the future, it would be possible to read out the information from the brain and reconstruct the person… to allow them to continue, in effect, with their life.”
English
480
333
1.9K
493.5K
andergrazh retweetledi
Adam Back
Adam Back@adam3us·
i'm not satoshi, but I was early in laser focus on the positive societal implications of cryptography, online privacy and electronic cash, hence my ~1992 onwards active interest in applied research on ecash, privacy tech on cypherpunks list which led to hashcash and other ideas.
English
2K
3.4K
28.5K
3M
andergrazh retweetledi
Josh Kale
Josh Kale@JoshKale·
This is big... Anthropic just announced a model so powerful they won't release it to the public out of fear over the damage it will cause 😨 Claude Mythos Preview found thousands of zero-day exploits in every major operating system and web browser... The numbers are hard to believe: > $50 to find a 27-year-old bug in OpenBSD, one of the most security-hardened operating systems ever built > Under $1,000 to find AND build a fully working remote code execution exploit on FreeBSD that grants unauthenticated root access from anywhere on the internet > Under $2,000 to chain together multiple Linux kernel vulnerabilities into a complete privilege escalation exploit For context: these are the kinds of findings that previously required elite security researchers working for weeks. Anthropic engineers with no formal security training asked Mythos to find exploits overnight. They woke up to working code the next morning. The results were so impressive Anthropic assembled Apple, Google, Microsoft, Amazon, NVIDIA, and seven other organizations into Project Glasswing: A $100M defensive coalition. They're not releasing this model publicly. Instead, they're racing to patch the world's infrastructure before models like this proliferate.
Anthropic@AnthropicAI

Introducing Project Glasswing: an urgent initiative to help secure the world’s most critical software. It’s powered by our newest frontier model, Claude Mythos Preview, which can find software vulnerabilities better than all but the most skilled humans. anthropic.com/glasswing

English
711
2.5K
18.2K
4M
andergrazh retweetledi
Ben Sigman
Ben Sigman@bensig·
My friend Milla Jovovich and I spent months creating an AI memory system with Claude. It just posted a perfect score on the standard benchmark - beating every product in the space, free or paid. It's called MemPalace, and it works nothing like anything else out there. Instead of sending your data to a background agent in the cloud, it mines your conversations locally and organizes them into a palace - a structured architecture with wings, halls, and rooms that mirrors how human memory actually works. Here is what that gets you: → Your AI knows who you are before you type a single word - family, projects, preferences, loaded in ~120 tokens → Palace architecture organizes memories by domain and type - not a flat list of facts, a navigable structure → Semantic search across months of conversations finds the answer in position 1 or 2 → AAAK compression fits your entire life context into 120 tokens - 30x lossless compression any LLM reads natively → Contradiction detection catches wrong names, wrong pronouns, wrong ages before you ever see them The benchmarks: 100% recall on LongMemEval — first perfect score ever recorded. 500/500 questions. Every question type at 100%. 92.9% on ConvoMem — more than 2x Mem0's score. 100% on LoCoMo — every multi-hop reasoning category, including temporal inference which stumps most systems. No API key. No cloud. No subscription. One dependency. Runs on your machine. Your memories never leave. MIT License. 100% Open Source. github.com/milla-jovovich…
Ben Sigman tweet media
English
449
796
7.9K
3M
andergrazh retweetledi
Jeremy Nguyen ✍🏼 🚢
Jeremy Nguyen ✍🏼 🚢@JeremyNguyenPhD·
Milla Jovovich (actress from The Fifth Element) created a world-beating Claude memory system with @bensig?! - 100% on LongMemEval — first perfect score ever recorded. Free and 100% open source. Github link in the quoted post from Ben. I'm keen to hear how it works for you.
Ben Sigman@bensig

My friend Milla Jovovich and I spent months creating an AI memory system with Claude. It just posted a perfect score on the standard benchmark - beating every product in the space, free or paid. It's called MemPalace, and it works nothing like anything else out there. Instead of sending your data to a background agent in the cloud, it mines your conversations locally and organizes them into a palace - a structured architecture with wings, halls, and rooms that mirrors how human memory actually works. Here is what that gets you: → Your AI knows who you are before you type a single word - family, projects, preferences, loaded in ~120 tokens → Palace architecture organizes memories by domain and type - not a flat list of facts, a navigable structure → Semantic search across months of conversations finds the answer in position 1 or 2 → AAAK compression fits your entire life context into 120 tokens - 30x lossless compression any LLM reads natively → Contradiction detection catches wrong names, wrong pronouns, wrong ages before you ever see them The benchmarks: 100% recall on LongMemEval — first perfect score ever recorded. 500/500 questions. Every question type at 100%. 92.9% on ConvoMem — more than 2x Mem0's score. 100% on LoCoMo — every multi-hop reasoning category, including temporal inference which stumps most systems. No API key. No cloud. No subscription. One dependency. Runs on your machine. Your memories never leave. MIT License. 100% Open Source. github.com/milla-jovovich…

English
236
777
6.9K
1.7M
andergrazh retweetledi
Mario Nawfal
Mario Nawfal@MarioNawfal·
🇫🇷🇺🇸 France just pulled all its remaining gold out of the New York Federal Reserve. 129 tonnes moved from U.S. custody to Paris vaults between July 2025 and January 2026. All 2,437 tonnes of French gold reserves are now entirely in France for the first time since the 1920s. France says it's "not politically motivated," just upgrading to modern standard bars. But the timing is notable given global tensions and the Iran war. They also made €13 billion profit on the swap due to rising gold prices. Source: @mining
Mario Nawfal tweet mediaMario Nawfal tweet media
English
645
895
4K
382.6K
andergrazh retweetledi
Science girl
Science girl@sciencegirl·
Amazing project from ETH Zürich! This robot combines wheels with tilting propellers that push it against a wall while also generating upward force, allowing it to climb vertically with impressive agility. 📹 mauricio.frizzarin
English
130
1.1K
8.2K
375.5K
andergrazh retweetledi
Michael Saylor
Michael Saylor@saylor·
Bitcoin has won. Global consensus is that $BTC is digital capital. The four-year cycle is dead. Price is now driven by capital flows. Bank and digital credit will determine Bitcoin’s growth trajectory. The biggest risk is bad ideas driving iatrogenic protocol changes.
English
2.5K
3.5K
26.1K
2.5M
andergrazh retweetledi
All day Astronomy
All day Astronomy@forallcurious·
Since its release 11 years ago, just one hour and 36 minutes have passed on Miller's planet in Interstellar.
English
23
213
1.9K
73.2K
andergrazh retweetledi
Interesting things
Interesting things@awkwardgoogle·
1:1 scale recreation of the Titanic leaving Belfast docks using 1000 drones 🤯
English
167
2.6K
19.7K
1.3M