2026-08-09 Domingo

Notícias de cripto - Página 372

Mantenha-se a par do mercado de cripto. Notícias em tempo real, análises, preços, histórias em alta e análise de especialistas — tudo num só lugar.

Investors Are Now Hunting for AI Projects on Bilibili and Xiaohongshu

Investors Turn to Bilibili and Xiaohongshu to Source AI Projects The AI hardware boom is in full swing in 2025, with a surge in smart wearables like AI glasses, rings, toys, and companion robots. This frenzy has investors scrambling, not just sifting through business plans, but actively hunting for promising "under-the-radar" projects on youth and tech-enthusiast content platforms like Bilibili and Xiaohongshu. The logic is straightforward: for consumer-facing AI hardware, genuine user demand and potential pitfalls are often revealed earlier in public discussions, comments, and critiques on these communities than in formal pitches. As one industry insider notes, these products must ultimately be tested and understood by real people. This shift highlights a crucial challenge in the sector: user education. The success of AI hardware depends on moving beyond mere efficiency gains to fulfilling higher-order needs like "unleashing personal creativity." Products must convince users they are natural, unobtrusive additions to daily life. Early hype, as seen with devices like the Rabbit R1, often fades if the product fails to clearly solve real-world problems, leading to high return rates and market rejection. The market is now entering a shakeout phase. 2026 is seen as a year of commercial validation. Some projects have already stalled or been canceled due to market resistance, lack of differentiation, or financial woes. However, the long-term opportunity remains vast, with forecasts predicting a multi-trillion dollar global AI hardware market by 2030. The competition is intensifying. With giants like OpenAI and Meta preparing their own hardware, and Chinese companies launching diverse AI-powered products, the battle for user attention, product excellence, and market understanding is just beginning. The core principle endures: in the AI era, it remains a user-sovereign market.

marsbit06/12 05:07

Investors Are Now Hunting for AI Projects on Bilibili and Xiaohongshu

marsbit06/12 05:07

"Agents' Last Exam", Claude Fable 5 Actually Loses to GPT 5.5

Surprisingly, in the newly released "Agents' Last Exam" (ALE) benchmark from UC Berkeley, GPT-5.5 has outperformed the recently launched and highly-regarded Claude Fable 5. ALE tests AI agents on their ability to perform real-world tasks across 55 professional domains—such as 3D modeling in Siemens NX, creating game scenes in Unreal Engine, and visual effects work in Adobe After Effects—by granting them full GUI and command-line access. In the core task completion rate ranking, GPT-5.5 configurations secured the top two spots (24.0% and 23.0%), while Claude Fable 5 with Claude Code came in third (22.0%). Notably, the highest pass rate was only 24%, and the most difficult "Last-Exam" tier saw most top models, including GPT-5.5 and Fable 5, scoring zero. The benchmark also revealed significant cost and efficiency gaps: Fable 5 spent over four times more money than GPT-5.5's most expensive configuration for a slightly lower score, and was much slower. ALE differs from previous knowledge-based benchmarks by evaluating practical "ability to do" rather than static knowledge retrieval. Its tasks are derived from real expert projects, automatically scored, and designed to prevent cheating through a rotating pool of private challenges. The results suggest that high performance on traditional benchmarks does not necessarily translate to proficiency in complex, open-ended real-world work. The study also notes that agents often fail by prematurely declaring tasks complete without proper verification, and that no single model excels uniformly across all diverse domains.

marsbit06/12 05:01

"Agents' Last Exam", Claude Fable 5 Actually Loses to GPT 5.5

marsbit06/12 05:01

Retail Ecology Dwindles, ZKsync Bets on Bank Pilots for a Breakthrough

Amidst declining retail activity, ZKsync is pivoting to target institutional banking as its primary growth strategy. The article explores this shift, contrasting it with the competitive "survival of the fittest" narrative by highlighting a cooperative model inspired by naturalist Peter Kropotkin. ZKsync is developing infrastructure like its private, permissioned Prividium suite for banks (e.g., Deutsche Bank's use case via Memento), enabling private transactions with public verifiability via zero-knowledge proofs. This appeals to institutions needing privacy, compliance, and Ethereum-based settlement security, unlike fully private chains (e.g., JPMorgan's Kinaxis) or consortium models (e.g., R3 Corda). However, this strategic focus has coincided with a steep decline in its public DeFi ecosystem, evidenced by plunging TVL and the departure of major protocols like Aave due to low fees. The network's future now hinges on banking adoption, with upcoming pilots like the Cari Network involving regional banks holding over $600 billion in deposits. A significant challenge is balancing this institutional focus with ZKsync's decentralized governance. Banks must operate on a network where rules and fees (denominated in the volatile ZK token) can be changed via community vote, and where a Security Council holds emergency control—a stark contrast to the predictable, contract-bound environments of traditional finance. The coming 18 months will test whether ZKsync can successfully onboard traditional banks onto a dynamically governed public chain or if institutions will ultimately revert to proprietary solutions.

Foresight News06/12 04:19

Retail Ecology Dwindles, ZKsync Bets on Bank Pilots for a Breakthrough

Foresight News06/12 04:19

The Recursive AI Anthropic Warned About: Tian Yuandong's New Company Has Just Taken the "First Step"

Anthropic recently highlighted the rapid progress toward "recursive self-improvement," where AI systems autonomously design and train their successors. In response, Recursive Superintelligence, a new company co-founded by former Meta researcher Tian Yuan Dong, has publicly demonstrated its first step toward automating AI research. The company released a system designed to autonomously execute the full AI research cycle: generating ideas, implementing code, running experiments, and learning from results. It validated this approach by achieving state-of-the-art results on three diverse benchmarks: 1. **NanoChat Autoresearch:** Optimizing a small language model's validation loss under a fixed 5-minute GPU budget, improving upon the community's best result. 2. **NanoGPT Speedrun:** Reducing the time to train a GPT model to a specific loss on 8 H100 GPUs from 79.7 seconds to 77.5 seconds, beating a highly optimized, human-driven community effort. 3. **SOL-ExecBench:** Improving the overall score on NVIDIA's suite of 235 GPU kernel optimization tasks by 18%, closing the gap to the hardware limit. The system discovered novel optimizations in this highly specialized domain without direct human expertise. Recursive's system operates as a general framework, capable of parallel exploration and cross-task knowledge transfer while incorporating safeguards against reward hacking. The company, backed by $650M in funding and a star-studded team including Richard Socher and Alexey Dosovitskiy, aims to create AI that recursively enhances its own research capabilities. This development represents an early but concrete move toward a new paradigm where AI accelerates its own advancement. It occurs alongside Anthropic's warnings about the need for industry coordination and potential pauses when recursive self-improvement thresholds are reached, highlighting the dual trajectory of rapid technical progress and growing calls for careful stewardship.

marsbit06/12 04:12

The Recursive AI Anthropic Warned About: Tian Yuandong's New Company Has Just Taken the "First Step"

marsbit06/12 04:12

The Gold Buy-on-the-Dip Guide: Watch Interest Rates, Not Just War

"Gold Buying Guide: Focus on Interest Rates, Not Just War" Four months ago, gold buyers likely didn't anticipate buying at a peak that even a war couldn't sustain. After hitting a record high of $5,596 on January 29, gold entered a bear market just 91 days later, its fastest decline since 2008. A key trigger was the Fed's hawkish shift, highlighting that monetary policy, not geopolitics, is the primary driver. The article argues that the traditional "buy gold in turmoil" script has changed. While the US-Iran conflict initially boosted prices, the sustained rally in oil prices heightened inflation fears, forcing central banks to maintain or consider tighter policy. Since gold yields no interest, higher rates increase its opportunity cost, eroding its appeal. This dynamic was evident when gold fell sharply on May 18 despite positive peace talks, as lower oil prices eased inflation and thus rate hike pressures. The recent sell-off is also part of a broader market deleveraging. Correlations between gold, Nasdaq, and Bitcoin spiked as leveraged investors sold liquid assets to cover losses, creating a synchronized downturn. Historically, gold bottoms align with policy shifts, not conflict resolutions. The 2008 and 2022 bear markets ended with shifts to extreme easing and peak inflation expectations, respectively. For potential buyers, the author suggests monitoring three signals: 1) Peak interest rate hike expectations, 2) Reopening of the Strait of Hormuz (to ease oil/inflation pressure), and 3) A return to net inflows for Gold ETFs, indicating the end of forced selling. While predicting the exact bottom is impossible, the author's personal strategy involves scaling into a position across price levels like $4000, $3700, and $3500, committing no more than 30% of the intended total allocation initially, and adding the remainder only if key signals emerge. The core conclusion: In turbulent times, watching interest rates is more crucial than watching wars.

marsbit06/12 04:06

The Gold Buy-on-the-Dip Guide: Watch Interest Rates, Not Just War

marsbit06/12 04:06

Recent On-Chain Review: No Clear Narrative Under U.S. Stock Market Pressure, Just Hype

This article analyzes the current state of the Solana meme coin and community token ecosystem, highlighting a market caught between two dominant forces: attention-based PvP and a gradual return to community-centric projects. The first part explores the "Attention PvP" dynamic, where success is driven by celebrity endorsements, viral events, and speed. Examples include $JOTCHUA, which surged after its meme creator's social media activity, and $WORLDCUP, which outperformed a similar Base chain project ($PITCH) largely due to influencer support. The recent "pump.fun GO" feature, allowing bounty tasks for token promotion, is critiqued for fostering sensationalist and often negative stunts—like people getting token tickers tattooed on their bodies for rewards—reminiscent of old internet shock content. In contrast, the article points to a resurgence of organic, community-driven tokens that survive market volatility through strong holder bases and shared ideology, not just hype. Influencer Ansem is cited, arguing that durable meme coins rely on communities willing to endure losses and promote their core message daily. Examples given are older tokens like $neet (anti-work ethos), $troll, $buttcoin, and $triplet, which have maintained relative price stability. A prime example of this community-build model is the new project $KINS, the token for the browser-based MMORPG Kintara. Its success stems not from advanced graphics but from consistently delivering updates, fostering player trust, and creating genuine engagement (e.g., in-game economies, events, property auctions). It has attracted a growing player base and even notable KOLs as participants, demonstrating that sustainable growth can come from building trust rather than orchestrating pumps. The article concludes by questioning whether the market is ultimately a game of mutual trust or mutual deception, expressing hope that such reflection might lead to a healthier ecosystem.

marsbit06/12 04:05

Recent On-Chain Review: No Clear Narrative Under U.S. Stock Market Pressure, Just Hype

marsbit06/12 04:05

活动图片