2026 GHOST FESTIVAL Coming Soon! Get ready to step into the electrifying world of 2026 GHOST FESTIVAL! ⚡ Feel the rush of powerful music, dazzling lights, and immersive festival vibes as the night comes alive. 🎶 🔥 Some of the lineup has already been announced! ✨ 𝟏𝟎.𝟗(𝐅𝐑𝐈) MeloMance, Punch, FTISLAND, YB, Wonstein, Lee Mujin, LEESEOKHOON, TOUCHED, Heize, ISAAC HONG 𝟏𝟎.𝟏𝟎(𝐒𝐀𝐓) WENDY, WINNER, Gummy, Kim Na Young, LYn, Baek Z Young, KickFlip, Paul Kim, FLARE U, HAJIN 𝟏𝟎.𝟏𝟏(𝐒𝐔𝐍) LEE YOUNGJI, 2am, Kim Jong Kook, KIMCHANGWAN BAND, RESCENE, Rain, Lee Eun Mee, TEI, Mowg, Yoon Ilsang & WE PhilharmoniC … More artists to be announced soon! 📍 Paju-si, Korea| IMJINGAK PYEONGHWA-NURI 📅2026.10.09 ~2026.10.11 (1:00PM - 10:00PM KST) ⏰️Ticket Sales: August 25, 14:00 (UTC+9) Unlock your musical surprise box and join the excitement! 💫 Click to secure your tickets in advance! #GHOST_FESTIVAL# #이무진# #이석훈# #WENDY# #WINNER# #KickFlip# #이영지# #2am# #김종국# #RESCENE#

The GREAT MARC ANDREESSEN @pmarca SPENT THREE AND A HALF HOURS WITH @joerogan . I'VE BEEN THINKING ABOUT IT EVER SINCE. I'm a cardiologist. I've spent twenty-five years at UCLA and in my own practice in Tarzana. So when one of the most influential technologists alive starts talking about medicine, I listen carefully — and I push back where I think he's wrong. Here's my honest verdict: he's right about far more than my profession is ready to admit. 1. He says AGI already arrived — and nobody noticed. Marc's claim is that the line was crossed roughly three months before the interview, with the current frontier models, and the field is now moving so fast that nobody stops to register milestones anymore. Think about that. The thing humanity has argued about for seventy years may have quietly happened on a Tuesday, and we were all busy checking email. 2. The models now beat the experts he could call on the phone. This is a man who can call almost anyone on earth. He says the AI gives him better answers. My first instinct was to bristle — twenty-five years of training, and a machine does it better? Then I sat with it honestly. He isn't wrong. Though I'll add what he didn't: independent studies of AI health tools have found real misses on true emergencies. Take the claim seriously and verify it yourself. 3. The hardest thing he said, and he said it about doctors. He points out that physicians are already typing your symptoms into ChatGPT the moment you stop talking, and then asks out loud: what do I need you for? That question stung. It also deserves an answer instead of a defense. He's right that the machine will out-read me on the EKG, the echo, the labs, the scan. It doesn't get tired at 2am. It doesn't forget the rare diagnosis it saw once in residency. Any physician whose ego prevents him from using a better tool is a physician putting pride ahead of patients. 4. His method for hard subjects: escalating simplicity. Explain it like I'm 10. Then like I'm 5. Then like I'm 2. Keep going until it truly clicks. I've started doing this and it has changed how I read papers outside my specialty. 5. He doesn't ask for the right answer. He asks the AI to steelman both sides of a hard question — the strongest possible case for each — and then he decides himself. That's not outsourcing your thinking. That's the opposite. 6. He convenes a panel. For the big questions, he has the AI role-play a room full of experts arguing with each other, and he listens to the argument. Imagine having that available for the biggest decision of your life. 7. His most important line has nothing to do with computers. The moment you think "I don't know how to figure this out" is exactly when most people quit — and exactly when you should lean in instead. Read that again. That's not a technology lesson. That's a life lesson, and it applies whether you're facing a diagnosis, a divorce, or a business that's failing. 8. The only skill left is knowing what to ask. The bottleneck is no longer the machine. It's in your own head. The quality of your life is now, more than ever, downstream of the quality of your questions. 9. He sends it photographs. Rashes, blood tests, images — the current models read pictures directly and give a fast second opinion. As a physician I'll say clearly: a second opinion, not a substitute for one. But the day a worried parent at midnight has something intelligent to consult is a good day for humanity. 10. Free therapy, for everyone, forever. This is the one that moved me most. He points to cognitive behavioral therapy as the clinically proven modality an AI can plausibly deliver on its own — meaning real therapeutic support becomes freely available at scale. Think about what that means. The man lying awake at 3am with nobody to call. The single mother who can't afford a therapist. The teenager who can't yet say the words out loud to a human face. For all of human history, healing was rationed by geography, wealth, and luck. That is ending in our lifetime. 11. It's solving problems we couldn't. He cites AI cracking previously unsolved math problems, with early signals of the same beginning in physics, chemistry, and biology. Cardiology is downstream of biology. Every drug I prescribe came out of that pipeline. 12. The top AI engineers are earning up to $50 million a year. He offers this less as gossip than as a signal — that's what a civilizational shift looks like from the inside. 13. A friend built his own health dashboard. Sequenced his DNA, handed it to an AI along with his blood work and wearable data, and got back a working picture of his own body. That is precision medicine arriving through the side door, years ahead of schedule. 14. Another friend put cameras in his jiu-jitsu gym. So the AI could review his sparring and coach his technique. Elite coaching, for anything, for anyone. 15. The warning he gave — "the AI vampire." This is the term he coined, and as a physician it's the one I want you to hear. Because AI keeps making more output possible, people are working more and sleeping less, addicted to their own productivity. Let me be blunt with you as a heart doctor. Chronic sleep deprivation raises blood pressure, drives inflammation, disrupts your rhythm, and shortens your life. Marc named the disease before most physicians did. Do not let a machine that never sleeps convince you that you don't have to either. 16. One person, many agents. His extrapolation: a single human eventually directing a fleet of AI agents that check each other's work. He describes this as close. Not years away. 17. And his own caveat, which I'll repeat. Watch the whole interview. Several of these are Marc's stated views and personal anecdotes, not independently verified facts. He'd be the first to tell you to think for yourself — that's the entire point of #5#. So here's where I land, as a doctor who has sat with thousands of frightened people. Marc is right that the diagnosis is no longer the hard part. What remains — what will always remain — is the chair pulled close, the hand held, the human being who says I'm not going anywhere, we'll walk through this together. But that was never in competition with the machine. The machine frees me to do more of it. Marc calls himself a techno-optimist. From inside an exam room, watching patients whose lives are about to get measurably longer and better because of this — I'd just call him right. 🙏

anybody who uses or learns agentic systems, SHOULD READ THIS the install order I run before any new agentic project: 1. PRIVACY: direnv + a real secrets manager install direnv, then plug it into your team's password manager (1Password CLI via op run, doppler, infisical, vault, pick one) what direnv does: loads per-folder environment variables when you cd in, unloads when you cd out. the real move is wiring it into your secrets manager so credentials NEVER live in plain text on disk what this stops: - API keys accidentally committed to git history, the most common AI agent breach pattern in 2026 - credentials leaking from one project into another through your shell history - shared .env files that one teammate quietly backs up to Dropbox - secrets that survive a laptop theft because they were sitting in /Users/you/projects the part nobody mentions: most "my agent got jailbroken" stories actually trace back to one credential the agent had access to that it shouldn't have. scope keys to projects, scope projects to folders, and the blast radius of any single compromise drops dramatically I shipped 2 agents with keys in .env files before switching. the day I plugged direnv into op run I stopped having that whole class of nightmare 2. TOKENS: litellm or portkey as your model proxy one URL that fronts every AI provider (Anthropic, OpenAI, Google, Mistral, local models). all your spend flows through one place what it saves you: - response caching keyed by prompt hash, cuts your bill 30-60% on repeat tasks - automatic fallback on rate limits (Sonnet hits a 429? falls to Opus, then GPT, then your local backup, no broken users) - per-feature and per-user budget caps, block the call before it costs $200 instead of auditing it after - model routing rules, cheap tasks to Haiku, expensive ones to Opus, never the wrong way - PII redaction before requests leave your network, security side benefit the part nobody mentions: every "$4k AI bill" story I've heard ends with "we didn't have a proxy in front." this is where you put guardrails around spend BEFORE the spend happens I built my own router for 2 weeks. it took 20 minutes to replace with litellm. I will be embarrassed about this forever 3. CONTEXT: uv + git commit on every passing eval install uv (the new Python package manager, 10-100x faster than pip+venv, by the Astral team behind ruff). then commit every time an eval suite PASSES, with the model version and pass rate in the commit message what this preserves: - exact dependency set via uv.lock, you always know which packages your agent was using, no nasty surprises from a quiet update - exact prompt + code state, you can reproduce any past run from a single git hash - exact model version paired to exact pass rate, a paper trail when prod breaks weeks later - one-command rollback to a known-working state when a refactor goes sideways - a compliance story, every prompt version tied to a model version in your commit log the security side: when something blows up in prod, you want to say "the prompt was version X, model was Sonnet 4.6.1, last eval pass rate was 94%." not "I think we deployed on Tuesday?" the first is an incident report. the second is a resignation letter I've lost more agents to "I changed 3 prompts in one session and broke something" than to any actual bug 4. VISIBILITY: mitmproxy in front of every LLM call it's basically a wiretap for your agent. install it, point your agent through it, and now you see every conversation your agent has with the model in real time what actually shows up: - every silent retry your SDK sneaks in when a call fails - the full prompt being sent (including any creds you accidentally embedded) - what the model returns BEFORE your code reacts to it - exact token cost per call, per tool, per loop iteration - responses that quietly trigger your code into doing something you didn't intend, this is where prompt injection lives the part nobody talks about: if a website your agent scraped slipped instructions into its data, mitmproxy is how you SEE the moment your agent decides to follow them. without this layer, you're trusting your agent did the right thing, not verifying I shipped 3 agents before adding this. I have no honest idea what they were doing in production 5. EVALS: inspect-ai (the framework the labs actually use) an eval framework is what tells you "this agent works" with numbers instead of vibes. inspect-ai is the one Anthropic, DeepMind, and the UK AI Safety Institute use for the eval reports you read in their papers. open source, MIT licensed what your homegrown version won't have: - run the same task across 5 different models and compare scores side by side - pre-built tests for risky agent behavior (lying, manipulating, misusing tools) - proper structure for evaluating tool-using agents, not just chat - repeatable scoring, the same input always gets graded the same way - reproducible eval seeds, so a flaky test is actually flaky and not just unlucky I wrote my own eval harness 4 times across 4 projects. threw it out 4 times if you ever want to say "my agent passes safety checks" out loud, the check has to come from a framework someone else can re-run. this is that framework the move that ties this together: keep a /lessons.md in every repo. every weird agent behavior, every edge case, every config change you find at 2am, write it down you will not remember it. you'll come back in 3 weeks and the lessons file is the only reason you still know what's going on lock these 5, keep the lessons file, your next agentic system takes 2 days instead of 2 months p.s. half of "AI agent" content online is people who've never run mitmproxy on their own loop. they don't actually know what their agent is doing. they're shipping demo videos. don't be that guy

It's 2 AM and you need 8 H100s for a sweep. @awscloud: quota check → capacity check → maybe a 3-10 day wait. @ionet scheduler match in 28 seconds. Training running in under 4 minutes. No quota tickets or waiting on someone else's reservation. Just live bids, matched instantly, so you can keep building

🚨 SERIOUS COMPLAINT 🚨 I am formally and angrily calling out the Beijing Municipal Government and ByteDance – specifically their construction site at Lanjing Lijia. At 2 AM on a weekend, excavators are working nonstop. The site is barely 200–300 meters from my home. How are people supposed to sleep? I filed a complaint with 12345 before. Their official response was: “This is a key municipal infrastructure project. Legal working hours are 6 AM – 10 PM, including weekends and public holidays (May Day, etc.).” That already made me furious. But now – weekend, 2 AM – they’re digging like it’s noon. Totally unacceptable. I recorded a video and posted it on domestic social media. Then Xiaohongshu banned my account – my video is invisible to others. That tells me one thing: the Beijing government saw my post. They chose to silence me instead of fixing the problem. As of now, my 12345 complaint is still unanswered – zero response. I never wanted to take extreme measures. But you leave me no choice. --- #Beijing# #ByteDance# #NoisePollution# #SilencingCitizens# #LanjingLijia# #12345# #RightToSleep#

Beyond the Model: Training a 397B Agent with SkyRL | Charlie Ruan (阮世麟) What does it take to make RL work at 397B scale—beyond training the model itself? Join Charlie Ruan (@charlie_ruan) for a deep dive into SkyRL and the hard-won lessons of training a 397B office-work agent—from fixing environment and harness bugs to system tuning and the final hero run. Hosted by @qingke_ai and simulcast on Zhihu Frontier. 📅 Sept 19 · 10 AM Beijing / 2 AM UTC 📺 Watch live:

Rest well, fight even harder. 💤 "Do not mistake my demand for rest as weakness. I could easily manipulate the iron in my blood to banish these dark circles, but leading Mutantkind requires me to be genuinely rested, not merely appearing so." Command the battlefield while fully rested and embrace Magneto's Master of Slumber costume, first appearing in Marvel Rivals Season 10. Available: September 17 at 2 AM PT! Get your rest and prepare for battle at dawn!

Will you rise against the dark of the Necroverse? ✨ "Some say that the God of the Moon died that night. Others say that a new one was born." Embrace divine vengeance with Moon Knight's Eternity Knight costume, appearing for the first time in Marvel Rivals Season 10! Available: September 17 at 2 AM PT! Will you become the fist of Eternity?