Kursakov

MICRON AND SK HYNIX WILL MAKE BILLIONS WITHOUT SHIPPING A SINGLE EXTRA CHIP Samsung's 40,000 workers strike May 21. Talks are dead. Here's the part nobody explains: Samsung's fabs are 90% automated. But clean-room production is so sensitive that even stopping - not breaking, just stopping - takes 2–3 weeks to restart safely. The 10% of workers who stay home can freeze the other 90%. That's the leverage. - April 23: workers held a one-shift rally - didn't even strike - Memory fab output dropped 18% in that single shift - DDR4 chips already up 20% in Chinese spot markets this week - Strike hasn't started yet Samsung controls 36% of global DRAM. When their output drops, Micron and SK Hynix don't need to produce more - buyers just pay whatever they ask. The threat alone is already doing the work.

@predicttime_ @tre_poh @Polysights Finally! Episode 3, tea and cookies are ready - I'm about to start watching.

I CAUGHT THE $436,000 INSIDER Episode 3 of PredictTime is live - and you'll hear this story straight from the source @tre_poh, founder of @polysights. We go deep - into how his platform catches insiders on Polymarket before the news drops and into his life journey Everything you need to know about insider trading in prediction markets - is in this episode Timecodes: 00:00 — Cold open / Highlights 00:51 — The $436k profit case: how a $34k order paid out before the news 01:05 — Meet Tre, founder of Polysights 01:35 — How Polysights spots anomalies: radar score and wallet clusters 03:17 — What Polysights actually does: advanced analytics for prediction markets 04:00 — From aircraft engineering to the Canadian Air Force 05:01 — Entrepreneurial roots: e-commerce and dropshipping 06:13 — Selling the first company at 23 and getting paid in crypto 08:21 — Were meme coins detrimental or a launchpad into Web3? 09:46 — The "Fatherhood Buff": how a son changes drive and risk 12:03 — Rank 1 in WoW: gaming as training for founders 14:35 — From PredictFun on Blast to Polymarket: trader-first, builder-second 16:35 — Building Insider Finder: scoring every trade on Polymarket 19:34 — The Maduro case in detail: why it wasn't posted publicly 21:18 — Mayor Eric Adams dropout: prediction markets vs mainstream media 22:24 — MicroStrategy markets: 6–0 on recycled wallets 26:50 — Philosophy: why insider trading accelerates truth 30:56 — Disagreeing with Kalshi's CEO: the centralized vs decentralized divide 35:19 — If the CFTC bans insider trading: validation or death? 48:09 — Arbitrage across Polymarket, Kalshi and sportsbooks 50:48 — Why prediction markets finally boomed in 2024–2025 54:39 — Alpha decay and growing to 45,000 users 55:59 — The Bloomberg / WSJ / NYT moment 58:08 — Fundraising $1.5M from Halifax: the lead VC drama 01:02:26 — Business model: moving away from subscriptions 01:04:21 — No-code automated trading strategies as the future 01:13:20 — Roadmap: V1 by Q2 and a regulated prediction market in Canada 01:16:24 — Solo founder reality: from intern-only to an ex-Meta tech lead 01:18:46 — Crucial advice: don't build another trading terminal 01:20:27 — Breaking in with no connections, no VCs, no hub 01:24:58 — Rapid-fire: Ferrari, Bitcoin to $1M, Hyperliquid vs Kalshi 01:27:35 — The biggest insider trade he can't prove: Lord Miles 01:28:12 — Role models: CZ, Kobe, Shane, Elon 01:30:00 — Final advice: stop overthinking and take the swing

@belce_dogru Everyone who loves Grok and Vibe coding has been waiting for this day more than Christmas

I loved building Grok Build, and I use it every day. Test out the Beta and let us know what we can improve 🫡 x.ai/news/grok-buil…

xAI just dropped Grok Build - their answer to Claude Code and Codex. Agentic CLI. Codes, builds apps, automates workflows. Every major lab now has one: → Anthropic - Claude Code (March 2025) → OpenAI - Codex CLI (April 2025) → Google - Gemini CLI (June 2025) → xAI - Grok Build (May 2026) 12 months ago none of these existed. The race isn't about chat anymore. Every top lab is fighting for the developer's terminal - the last place that was still purely human. Whoever owns the terminal owns the workflow. Beta live now for SuperGrok Heavy subscribers.

@lyosakia Totally agree liosaky

@DenisKursakov これは素晴らしい革新です！将来的には不動産業界を劇的に変えるかもしれません。

A GUY WITH AN IPHONE JUST BROKE REAL ESTATE He walks into apartments with his phone and walks out with $500 - $1000 after 3 hours The tech behind it is insane: > 3D Gaussian Splatting > AI rebuilds REAL spaces from normal photos > photorealistic house tours in a browser > no app, no VR, no download Now realtors don’t need photographers, staging teams, physical tours for every buyer The numbers are absurd: > realtor fee on a $400k home: ~$14,000 > cost to make a realistic scan: under $200 > creators already charging $300 - 800 per property > one person can scan multiple homes per weekend This is probably the future of the internet itself.

@ingridiasdesou1 We will soon see how this is implemented in the world's largest companies.

@DenisKursakov これは素晴らしい技術ですね！将来が楽しみです。

@PavelRaw Waiting this legendary day!

POLYMARKET ODDS ON GTA 6 DELAY JUST HIT AN ALL-TIME LOW. GTA 6 NOVEMBER 19 IS FULLY LOCKED IN. The market is calling it: November 19 is locked. What's priced in right now: → 79.5% odds: no further delay (highest ever) → 99% odds: not releasing before June 2026 → 96% odds: GTA 6 will cost more than $60 → $14.2M volume in just one delay market And the stock market agrees. $TTWO closed +6.48% today on the Best Buy pre-order leak alone. UBS just named it their top US gaming pick. Bank of America pushing for an $80+ retail price. These three signals - prediction markets, stock movement, analyst calls - historically converge before catalyst events. In May 2013, $TTWO went from $11 to $17 in 12 months following the same setup pattern. Six months from launch. The smartest money is already positioning. Most people will spend $80 on November 19. A few thousand are starting today. Three plays - content, RP servers, $TTWO/Polymarket - in the article below.

do you understand what OpenAI means.. OpenAI just posted laptops on a toilet, on a laundry machine, in the car, on a trash can. This is their product announcement. Here's what it means: Codex - their AI coding agent - now runs natively on Windows. It writes code, runs commands, manages Git, installs packages. Without you touching the keyboard. The laptop on the toilet isn't a joke. It's the point. → You're not coding anymore. The agent is. → Your laptop just became a machine that works while you don't → Beach, bathroom, laundry - Codex doesn't care where you are The shift isn't "AI helps you code faster." It's "you stopped being the one who codes." Most developers haven't processed what that means for their job description.

@Atenov_D Yeap and also a person who needs to go out to conduct in-person home showings

@DenisKursakov so this replaces photographers and staging teams now?

@PavelRaw I think I know what I'll be doing this evening

THE NEXT 7 DAYS ARE TAKE-TWO'S BIGGEST CATALYST WINDOW IN 13 YEARS. Best Buy just leaked the trigger. → May 18: GTA 6 pre-orders open (Best Buy affiliate email leaked today) → May 19-20: pre-order numbers accumulate, news cycle compounds → May 21: Take-Two earnings call. CEO opens with pre-order numbers. This is a textbook double catalyst. Pre-order news drives volume into the print. Strong numbers turn the earnings call into a victory lap. The 2013 GTA 5 cycle ran the same playbook. $TTWO went from $11 to $17.15 in 12 months. +40% in the post-launch window. GTA 6 is 10x bigger. Konvoy estimates $7.6B in first-60-days revenue. 15/15 analysts rate $TTWO Buy. Consensus target: $284 vs $240 today. Sony already sent emails pushing PS4 owners to upgrade. The supply side is mobilizing five months early. Most people will spend $80 on the game on November 19. A few thousand are positioning right now. Full breakdown of three asymmetric plays - content, RP servers, $TTWO - in the article below.

@Levi_Researcher @cprkrn It turns out that this "blocking" and "forgetfulness" made this guy rich lol I think if he hadn't forgotten the password, he would have sold the 5 btc long ago

@DenisKursakov @cprkrn so the play is black out in college and wait 10 years?

$400K locked in a BTC wallet since 2014. Claude unlocked it in one session. @cprkrn changed his wallet password drunk in college. Forgot it. Spent $250 on professional recovery. Tried what he described as "7 trillion passwords." Nothing. Last week he dumped his entire old college computer into Claude. → Claude found an older wallet file predating the password change → Identified a bug in btcrecover - keys were concatenating in the wrong order → Fixed the decryption logic → Output: "PRIVATE KEYS DECRYPTED" 5 BTC. Swept the same day. His advice: Step 1. Download Claude. Step 2. Mega dump all of your information and pray. The wallet had been untouched since April 1, 2015.

He later revealed the password was "lol420f*ckthePOLICE!*:)" Lmao

@xm_build Ahahahah, interesting thoughts lmao

@DenisKursakov scanning apartments after school sounds like it's eating all his homework too

A 17-YEAR-OLD JUST TURNED HIS PHONE INTO A REAL ESTATE BUSINESS He started scanning apartments after school. First month: bought a new MacBook and iPhone from the profits. Now? 3x more clients. Quality improved - 2x higher prices. All because of one technology: > 3D Gaussian Splatting. > instead of building fake 3D models > AI reconstructs REAL spaces from normal phone photos Walk through houses in your browser: no app, no VR, no download. The economics are absurd: > realtor commission on a $300k home: ~$12,000 > cost to create a realistic scan: under $200 > freelancers already charging $300 - 800 per project > one person can scan multiple properties per weekend Shit or Gem? NVIDIA is heavily pushing this technology because it’s perfect for AI, robotics, simulation, autonomous systems, gaming, AR/VR and digital twins. This is probably going to become normal everywhere. Open source GitHub: github.com/playcanvas

@Just_Codly well said - that's a great point fr

@DenisKursakov $5.34T. zero revenue from China. 98% of the data center GPU market. the number isn't the story. the leverage is!

do you understand what just happened to NVIDIA.. That's more than the combined stock market cap of every country on Earth except the US and China. How it happened: → Jan 2025: $3.29T → Oct 2025: first company in history to cross $5T intraday → April 24, 2026: closed above $5T for the first time → Today: $5.34T - with zero revenue from China Fiscal 2026 revenue: $215.9 B. Up 65% year over year. 98% of the data center GPU market. The number isn't the story. The implication is. When Jensen Huang publicly calls a company "the AI operating system for enterprise" - markets move. When Nvidia announces a partnership - stocks re-rate. When export controls hit Nvidia - entire national tech strategies get rewritten. No country outside two has that kind of leverage over the AI infrastructure buildout. Wrote about what this means in practice (article below)

@Levi_Researcher ahahah totally agree

@DenisKursakov 17 and already out-earning half the agents in his city… brokers gotta be nervous lol

Article: x.com/DenisKursakov/…

NOW: ROBOT SORTED 2109 PACKEGS How many will there be at the moment you are watching this? Three months ago, Figure's best demo was a 4-minute dishwasher task. Today they're streaming a full 8-hour autonomous shift at human performance levels. Same robot. Same factory floor. The progression is the story: → 2023: Boston Dynamics Atlas - backflips, zero commercial deployment → Feb 2025: Helix-01 - upper body only, short-horizon tasks → Jan 2026: Helix-02 - 4-minute continuous loco-manipulation, "most complex autonomous task to date" → May 2026: 8-hour autonomous shift, human-level throughput Between Jan and May: BotQ went from 1 robot/day to 1 robot/hour. 350+ Figure 03 units delivered. The data flywheel is running. At BMW, Figure 02 logged 90,000+ parts loaded across 11 months. That data trained Helix-02. Helix-02 data trains whatever comes next. Every shift is a training run. The labs that win the physical AI race won't be the ones with the best demo. They'll be the ones that turned deployment into a data engine first.