Premarket movers: Nvidia is the biggest gainer among Mag 7 stocks. The chipmaker’s partner Hon Hai reported a better-than-expected increase in quarterly profit, signaling robust global demand for AI hardware. (Nvidia +1.2%, Alphabet +0.8%, Meta +0.7%, Tesla +0.5%, Amazon +0.4%, Apple little changed, Microsoft -0.8%) Cava (CAVA) jumps 13% after the restaurant chain operator reported store comp sales for the second quarter that beat the average analyst estimate. Analysts again note positive trends for its pomegranate glazed salmon. CoreWeave (CRWV) rallies 18% after the cloud-computing provider reported second-quarter results that beat expectations. Analysts are positive about the company’s margins and note that AI demand remains robust. ERock (EROC) is up 13% after the power systems firm reported revenue for the second quarter that beat the consensus estimate, and said Anthropic has agreed to buy 470 megawatts of onsite power equipment. H&R Block (HRB) is up 15% after the tax preparation company gave a full-year forecast that was stronger than expected. It also reported fourth-quarter results that beat expectations. Hyliion Holdings (HYLN) is up 23% after the company boosted its full-year revenue forecast from $10 million to $15 million. Lumentum (LITE) gains 8%. Analysts are positive on the maker of optical equipment after it reported fourth-quarter results that beat expectations and gave an outlook above analyst consensus. Super Micro Computer Inc. (SMCI) jumps 9% after giving a revenue forecast for the current quarter that topped analysts’ estimates, a sign the booming artificial intelligence market continues to bolster sales of the company’s servers. US Antimony Corp. (UAMY) falls 14% after the natural resource company cut its full-year outlook for gross revenue.

On FSD 14.3.7 it delayed a necessary lane change for a right turn until the final 200 ft, even though space was available much earlier. The late lane change behavior remains an ongoing issue. Earlier positioning would make the drives feel more natural and confident @Tesla_AI

Save this 14% return strategy for the 2030 World Cup Smart traders spotted an inefficiency on the World Cup Winner market and made a 14% return on any size with almost no risk I’m sure there’ll be a similar opportunity at the next WС. So you should learn this strategy now It's not arbitrage. Not a bond. It's just a mispricing The logic is simple: one of the 14 favourite teams will win. And none of the bottom 20 will pull off an upset At this WC, buying shares for all 14 favorites cost just 86%. Traders didn’t have to wait for the final because only favourites remained by semi-finals. That 86% naturally became 100%, and they took their 14% profit I love this because the crowd never changes. People will always chase an underdog story, completely breaking the real odds Bookmark this tweet rn. Return to it two months before the 2030 World Cup to print money on this exact strategy

🗺️ Even frontier GPT-5 succeeds on just 14.4% of real-world spatial tasks. A new benchmark goes beyond staring at a static image and exposes how weak AI agents still are at active spatial reasoning. Title: SpatialWorld: Benchmarking Interactive Spatial Reasoning of Multimodal Agents in Real-World Tasks URL: 📝 Overview SpatialWorld measures whether multimodal LLMs can solve tasks by actively exploring 3D environments from a vision-only, egocentric viewpoint. It unifies eight different simulators across indoor, outdoor, and digital-game settings under a shared protocol, and evaluates 15 frontier models on 760 human-annotated tasks. The agent gets no prior map and no reference solution; it has to look, move, and decide on its own. ❓ Challenges Solved Prior spatial-reasoning benchmarks relied on passive evaluation via static VQA or pre-recorded video. That can't capture the interactive spatial understanding the real world demands, where an agent must move its own viewpoint to gather visual evidence and replan on the fly under partial observability. There was a large gap between recognizing a static scene and actually moving through an unfamiliar space to get a task done. 💡 Methodology & Proposed Approach ・The task is framed as a vision-only POMDP (Partially Observable Markov Decision Process) ・The agent receives only a natural-language goal and a single native-resolution egocentric RGB image, with no depth, maps, or semantic metadata ・Actions are issued through a high-level text interface covering navigation, viewpoint control, object interaction, and task completion ・It integrates eight backends: indoor (AI2-THOR, ProcTHOR, VirtualHome), outdoor (CARLA, EmbodiedCity), and digital games (Block3D, Snake3D, Rubik's Cube) ・Success is judged by whether the final terminal state satisfies the goal, not by matching the trajectory, and is validated by human annotators ・Beyond success rate, it measures step efficiency against human reference trajectories to surface inefficient behavior 🎯 Use Cases It offers a unified, fair way to evaluate the spatial abilities of home robots and autonomous agents before real-world deployment. It can systematically diagnose where long-horizon tasks that combine navigation and manipulation break down, serving as a rigorous testbed for improving spatial-reasoning models. 📊 Experimental Results ・Across 15 frontier models, physical-task success was 14.4% for GPT-5, 12.2% for Qwen-3.5-397B, 9.2% for Gemini-3.1-Pro, and 9.2% for Kimi-K2.5 ・On digital games, Gemini-3.1-Pro led at 39.0%, followed by GPT-5 at 36.4% ・By complexity, interaction-only tasks averaged 50.2%, navigation-only dropped to 8.6%, and combined navigation-and-interaction collapsed to just 4.2% ・Models with similar success rates showed very different efficiency scores, revealing heavy reliance on trial-and-error exploration ・Model rankings shifted dramatically across environments, with no single model dominating every category #AIAgents# #SpatialReasoning#

🚀BITCOIN BULL RUN IMMINENT🚀 WE LIVE IN A SIMULATION. Bitcoin bear markets priced in GOLD have lasted: 406 days. 416 days. 440 days. Historical average: 421 days. So naturally the latest Bitcoin/gold bear market bottoms after... 433 DAYS. Then Bitcoin proceeds to rip +41% against gold. Apparently the monetary system has a RESPAWN TIMER. We spent 14 months watching everyone declare Bitcoin dead for the 900th time while the hardest-money ratio was quietly approaching the exact same historical expiration date like a frozen pizza in God’s microwave. And the funniest part? Bitcoin even made a LOWER LOW in dollars afterward... ...while making a HIGHER LOW against gold. Now get this. We are 6 months and 5 days passed the bottom now. Walk forward that length of time from each BTC/XAU bottom in history. Forward returns in USD: Six months forward +36.8% -9.1% +55.1% Median: +36.8% Two of the three periods were strongly positive, while 2019 ran directly into the COVID-era disruption shortly afterward. Twelve months forward The results become much more interesting: +144.5% +1.6% +87.2% Median: +87.2% And crucially: 3 out of 3 were positive 12 months later. If today's cycle merely matched those historical medians mechanically: 6-month median analogue $80,279 × 1.368 ≈ $109,800 BTC 12-month median analogue $80,279 × 1.872 ≈ $150,300 BTC Fiat chart: “OH GOD IT’S OVER.” Hard-money chart: “sir, your bear market subscription has expired.” At some point you have to consider the possibility that Bitcoin is not trading randomly. It is simply running a recurring program designed to transfer wealth from people who panic after 400 days to people autistic enough to count them.

Felix Vanilla market utilization has held at an average of 79.4% over the past 30 days. Borrowing demand from Hyperliquid traders remains strong relative to supply, creating an opportunity for USDC lenders to earn higher natural yield over longer durations than other money markets on other chains. Earn 14.33% APY on USDC now:

AI filmmaking is steering industry conversation at Toronto International Film Festival, and CapCut is setting the pace.

TIFF: The Market is a new extension of @TIFF_NET where entertainment and innovation professionals shop what’s next in content. Check out CapCut’s #TIFF26# programming Sept. 13-14: 🟠 Selected screenings from the CapCut CRE[AI]TE AI Festival 🟠 Q&A with creators, moderated by Justin Hackney @WonderStudiosX 🟠 An industry panel with @ai_massive 🟠 A major upcoming project announcement Social and AI-driven content are now key players in an industry looking to capture attention; creators in these fields are naturally at the forefront. Meet us at The Market, or follow along here for updates throughout TIFF weekend.