Porsche’s ‘long haul’ Web3 project ends in less than four years
Related
Michael Saylor: If Bitcoin's long-term price increase exceeds 3.3%, BTC capital gains could fund STRC dividends indefinitely.
According to ChainCatcher, Bitcoin Treasuries.NET posted on the X platform that Strategy's Michael Saylor stated that if Bitcoin's long-term price increase exceeds 3.3%, BTC capital gains can fund STRC dividends indefinitely. Even if BTC's annual price increase is 0%, Strategy will still have dividend funds for 31 years.
Four wallets long in DEXE on Aster have generated a floating profit of approximately $1.32 million.
PANews reported on July 6th that, according to Lookonchain, the price of the token DEXE continued to rise today. Four wallets held 2x leveraged long positions on the decentralized exchange Aster_DEX, with a combined unrealized profit of approximately $1.32 million. The highest return for a single wallet reached approximately 104.57%.
SEC sends crypto custody rule overhaul to White House for review
The securities regulator’s crypto custody overhaul could clarify how investment advisers and funds hold digital assets for clients as the proposal moves through review.
Vitalik Buterin outlines Ethereum's long-term roadmap, predicting Lean Ethereum will be the third major iteration.
According to ChainCatcher, Vitalik Buterin recently reported that Ethereum researchers met in Berlin and updated the protocol's long-term roadmap (strawmap.org). Vitalik pointed out that "Lean Ethereum" is not a single upgrade, but a series of improvements to be implemented in phases over the next three to four years. Its importance is comparable to a "merge," encompassing the restructuring of almost every core module of the protocol. Key features include: Verification Mechanism: Introducing recursive STARKs to replace the existing direct re-execution method, becoming a core component of the protocol at the primary level; Quantum Security: Significantly elevated priority, all quantum-fragile components will be replaced, and the quantum-safe Blob design is underway; Consensus Layer: Decoupling the available chain from finality, achieving one to two rounds of finality, resulting in better security and lower latency; State Layer: Existing dynamic states remain unchanged, but new, more scalable states (such as UTXO storage, circular buffers, etc.) will be added. It is projected that by 2030, Ethereum will have 2 TB of dynamic states + 100 TB of new states, and the migration of applications such as ERC20 and NFTs can achieve a gas fee reduction of over 10 times; Privacy: Upgraded from an additional feature to a primary goal, permeating the design of Mempool, state trees, etc.; VM: In addition to the EVM, leanISA or RISC-V will be introduced, with the long-term goal being direct at the protocol layer.
"Whale Tracking"long that a whale holding 20 million shares of a long position in a basket of semiconductor stocks is now deeply trapped, incurring losses exceeding $5.5 million.
According to BlockBeats, on July 8th, Hyperinsight monitoring showed that a whale with the account starting with 0x519c currently holds 7 AI semiconductor-related positions, all of which are 5x isolated margin long positions, totaling approximately $19.541 million. The current unrealized loss is approximately $5.502 million, with all 7 positions showing losses. The losses are mainly concentrated in: MRVL: Long positions amount to approximately $7.456 million, with a floating loss of approximately $2.609 million; SNDK: Long positions amount to approximately $3.836 million, with a floating loss of approximately $1.162 million; SKHX: Long positions amount to approximately $3.883 million, with a floating loss of approximately $676,000; MU / NBIS: Unrealized losses of approximately US$281,000 and US$284,000 respectively. The other two long positions are also in the red, and the total unrealized loss for this address is currently $5.56 million. All four positions have lost more than 100% of their initial investment, making this the address with the largest losses in the market conditions of last night and this morning. Previous report from a whale: South Korea's largest chip production expansion in history failed to stem a massive sell-off, with one major whale who chased SKHX at a high price incurring a loss of over $2.2 million in a single day.
MIRA, an open-source 5B multiplayer world model, uses DINOv3 representation to mitigate long-term drift and can simulate 2v2 battles in Rocket League in real time.
According to Beating, AI research firm General Intuition, in collaboration with French AI lab Kyutai and Epic Games, has launched MIRA, a multiplayer interactive world model. As a generative game simulator supporting real-time multiplayer interaction, MIRA can simulate 2v2 battles in Rocket League in real time, based solely on historical footage and player button presses, without requiring a physics engine, rendering engine, or explicit 3D representation for inference. Unlike the "decoupling of logic computation and image rendering" approach adopted by companies like Odyssey, MIRA takes a generative simulation approach based on video latent space. MIRA boasts 5 billion parameters, and its core design builds the latent prediction space on a frozen general-purpose visual encoder, DINOv3-L. Leveraging pre-trained visual features, the generated latent states can more stably fall within the effective representation space, significantly mitigating image drift and divergence during long-term prediction. For multi-screen alignment, MIRA stitches the latent images from four player perspectives into a unified grid, enabling spatial attention mechanisms to operate naturally across viewpoints and improving the spatial consistency of vehicles, the soccer ball, and key events across multiple perspectives. The Action Dropout introduced during training also helps the system complete the game behavior of vehicles not controlled by commands when parts of the motion flow are missing. Currently, MIRA can run in real time at 20 frames per second on a single NVIDIA B200 graphics card. The team has open-sourced the training and inference code and released the Rocket Science dataset, which contains 1,000 hours of matches, approximately 4,000 hours of video, motion flow, and physics data from four perspectives; the complete training of the model used approximately 10,000 hours of clean match data.