2026-08-02 Domingo

Notícias de cripto - Página 101

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.

ChatGPT Finally Can 'Search Itself': Nearly Four Years of Conversations, Retrieved with One Click

ChatGPT, nearly four years old, has long lacked a functional way for users to search their own accumulated history. This changed on July 14th, when OpenAI launched a comprehensive "Unified Search" feature across all platforms and plans. This new search allows users to find and instantly access past conversations, uploaded documents, generated images, and projects within their own ChatGPT account, all from a single sidebar entry. This update signifies a major shift in ChatGPT's role, transforming it from a transient chat tool into a personal knowledge base or "file system" for users' AI-generated content. It completes a critical missing piece as OpenAI expands ChatGPT's capabilities with features like Projects, Memory, Work, and Sites, aiming to make it a central hub for work and creativity. The ability to easily retrieve years of personal data makes that data more valuable as a unique, irreplaceable asset in the AI era. However, it also highlights the increasing weight and potential sensitivity of this data, given default settings that allow conversations to be used for model training and legal precedents involving user logs. While a significant improvement over the previous ineffective search, this move is seen as addressing an industry-wide oversight. As the competition between AI assistants (ChatGPT, Gemini, Claude) moves beyond raw model power, the new battleground is becoming which platform can best organize, retain, and leverage the valuable data co-created by users and AI.

marsbit07/20 00:29

ChatGPT Finally Can 'Search Itself': Nearly Four Years of Conversations, Retrieved with One Click

marsbit07/20 00:29

7-Year-Olds Starting Companies? 9-Year-Olds Making Movies? Post-90s Parents Are Raising 'AI Prodigies'

"AI Genius Kids: A New Wave or a Parental Anxiety Trap?" The article examines the recent surge in AI-focused educational camps and programs in China, heavily marketed to parents of young children with sensational claims like "7-year-olds start companies" and "9-year-olds make movies." These programs promise rapid, practical AI skills—from creating agents and business plans to generating content—often for significant fees, capitalizing on parental anxieties about the AI-dominated future. It contrasts this trend with earlier, more foundational AI education, arguing current offerings prioritize superficial, immediate application over deep learning. The author critiques this approach as premature for children still in basic education, suggesting it fosters a shortcut mentality and may even teach improper use, like using AI to cheat on homework. The driving force behind this trend is identified as the profound anxiety of the current generation of parents—largely professionals in tech or related fields—who are witnessing firsthand how AI disrupts and replaces jobs. This workplace fear translates into a desperate desire to equip their children with perceived competitive advantages from an extremely young age, viewing AI proficiency as a magical key to success. The piece also casts a skeptical eye on the phenomenon of real "AI child prodigies," citing examples like Australian teen Alby Churven, whose ventures attract media buzz but little serious investment. It suggests that beyond the hype, the business world remains wary of child-led enterprises. True advancement, the article implies, still relies on solid, traditional education and foundational knowledge, not just early exposure to tools. Ultimately, it frames the AI camp boom less as a legitimate educational movement and more as a reflection of societal pressure and a potentially expensive exploitation of parental fears in a rapidly changing technological landscape.

marsbit07/20 00:26

7-Year-Olds Starting Companies? 9-Year-Olds Making Movies? Post-90s Parents Are Raising 'AI Prodigies'

marsbit07/20 00:26

Claude Code's Father's Night Shift AI Army, Fable 5 Built with Two Commands

For the past year, Boris Cherny, creator of Claude Code, hasn't written a single line of code himself. He manages hundreds, even thousands, of AI agents that run day and night, autonomously handling tasks from writing and submitting code to fixing CI tests and gathering user feedback. His secret lies in a "loop" system, where tasks are defined and set to run repeatedly, combined with a critical "/goal" command that provides a clear, verifiable objective and a "supervisor" model to check progress. This approach transforms the programmer's role from a coder to a system designer and reviewer. At Anthropic, AI agents even collaborate in Slack channels to divide work. The recently released Fable 5 model is built for such long-term autonomous operation, featuring self-verification and the ability to understand complex charts. Effective use requires the "/goal" command for targeted tasks and "/loop" for recurring ones. To maximize efficiency, a local context system can be set up in 20 minutes, providing the AI with persistent memory of your projects, preferences, and procedures. The key takeaway is that the future lies not in prompting AI for each step, but in building autonomous systems driven by clear, verifiable goals.

marsbit07/20 00:23

Claude Code's Father's Night Shift AI Army, Fable 5 Built with Two Commands

marsbit07/20 00:23

Large Model Memory Anxiety Solved on a USB Drive?

To address the memory and bandwidth challenges of large language model inference, SanDisk and SK Hynix are developing a new technology called High Bandwidth Flash (HBF). This approach adapts the advanced 3D packaging and stacking techniques used for High Bandwidth Memory (HBM) to NAND Flash, which is traditionally seen as slow for compute tasks. The key insight is that while NAND Flash is slow for writes, its read speeds can be significantly enhanced through parallelization. By stacking up to 16 NAND Flash chips, the first-generation HBF aims for capacities up to 512GB and read bandwidths up to 1.6 TB/s, with roadmaps targeting 3.2 TB/s. This positions HBF not as a replacement for HBM, but as a complementary, cost-effective tier in a hierarchical memory system for AI inference. During inference, model weights are largely static and read-intensive, making them suitable for HBF storage. This can alleviate the severe capacity bottleneck of expensive HBM, potentially reducing the number of accelerators needed per server, lowering overall system cost and power consumption. The technology, currently being standardized within the Open Compute Project, signals a shift towards more specialized memory architectures tailored for the specific demands of large-scale AI deployment.

marsbit07/20 00:18

Large Model Memory Anxiety Solved on a USB Drive?

marsbit07/20 00:18

Wintermute: What Is Left to Build in Crypto?

The article argues that the foundational question for crypto has shifted. Instead of asking "what else can crypto do?", the real question is "why will the real world need crypto next?" The author's answer is the emerging **Machine Economy** – where autonomous agents and robots act as independent economic participants, making decisions, executing transactions, and initiating actions in both digital and physical worlds. Three key recent developments make this shift plausible: 1) AI models are now capable and cheap enough for autonomous, persistent action; 2) Open standards (e.g., stablecoins, Agent payment protocols) allow for composability; 3) Agents can now operate with context over long periods, changing the scale of automated activity. Crypto's core strengths—programmability, permissionlessness, fast settlement, and auditability—align well with the needs of autonomous systems, which don't fit traditional human-centric financial rails. Major platforms like Coinbase and Robinhood are already building infrastructure for Agent-based activity. However, challenges remain, primarily around **Security** (e.g., prompt injection attacks on Agent wallets) and **Attribution/Accountability** when AI-assisted systems fail. The most promising opportunities lie not in crowded foundational layers (L1s, models, exchanges) but in the connective tissue for the Machine Economy: **1) An economic layer for Agents** (handling authorization, risk, identity), **2) Physical AI** (enabling robots in structured environments like warehouses to transact), and **3) Machine-led discovery systems** (e.g., in autonomous scientific research).

marsbit07/20 00:03

Wintermute: What Is Left to Build in Crypto?

marsbit07/20 00:03

Is Strategy’s $54.5B Bitcoin bet no longer just about BTC’s price?

Michael Saylor's company, Strategy, holds a $54.5 billion Bitcoin (BTC) treasury, but its financial model faces new challenges. While it historically relied on issuing premium-priced shares to fund more Bitcoin purchases, its market-to-net asset value (mNAV) has fallen sharply to 1.03x, eroding that advantage. The company's cost basis for its BTC is now higher than its market value. To meet obligations like dividends, it recently sold some Bitcoin, shifting focus from pure accumulation to balance sheet sustainability. Key measures like Bitcoin per share and BTC yield have declined. The company must now prove it can sustainably finance future purchases and maintain investor confidence amidst a complex capital structure with significant funding gaps. While Strategy's scale is unmatched, its aggressive accumulation model is being tested, with the debate moving beyond treasury size to long-term financial resilience.

ambcrypto07/19 22:03

Is Strategy’s $54.5B Bitcoin bet no longer just about BTC’s price?

ambcrypto07/19 22:03

Crypto market’s weekly winners and losers – LDO, PUMP, LIT, PI

The crypto market saw mixed performance this week, with technical rallies and profit-taking driving action. Weekly winners were led by Lido DAO (LDO), up 16% and approaching a key resistance zone. Pump.fun (PUMP) followed with a 13% gain but remains below critical resistance. Venice Token (VVV) posted a 12.3% relief rally after a prolonged downtrend. Notable altcoin gainers included IOTA (SN9) and Akedo (AKE). On the losing side, Lighter (LIT) fell 16%, entering a cooldown phase after a strong rally. Pi Network (PI) declined 2.7%, breaking below key support to a new all-time low. Arbitrum (ARB) dropped as bulls failed to reclaim the $0.10 resistance. Other significant losers included Cash Cat (CASHCAT) and LAB. Overall, the week featured sharp swings, with recovery attempts in some assets and extended declines in others, highlighting ongoing market rotation and volatility.

ambcrypto07/19 22:02

Crypto market’s weekly winners and losers – LDO, PUMP, LIT, PI

ambcrypto07/19 22:02

OriginTrail surges 32% as the AI narrative takes over – Can TRAC target $0.4?

OriginTrail (TRAC) surged 32% to a monthly high near $0.38, fueled by strong speculative demand driven by its AI narrative, with the crypto community comparing it to a decentralized Palantir. Trading volume jumped 300% to $30 million. Futures data showed significant buying pressure, with a positive buy-sell delta. However, profit-taking increased, indicated by a spike in exchange inflows, which could threaten the rally. Technical indicators like the RSI at 65 and a strong DMI suggest bullish momentum remains. The key for TRAC is sustaining demand to break the $0.40 resistance; failure could see a drop toward $0.26 support.

ambcrypto07/19 21:03

OriginTrail surges 32% as the AI narrative takes over – Can TRAC target $0.4?

ambcrypto07/19 21:03

Tokenized markets reach $2.3B – Why utility matters more than TVL

Tokenized financial markets have reached a $2.3 billion market capitalization, driven by growing institutional adoption of on-chain funds and equities. While Ethereum leads in custody value, holding 34% of the total, Solana processes the vast majority (95-97%) of tokenized equity trading. This highlights a key industry shift where execution and utility—measured by trading volume, transaction activity, and settlement efficiency—are becoming more critical than Total Value Locked (TVL) alone. Platforms like Arbitrum and Solana lead in holder growth, supported by established issuers such as Ondo Finance, underscoring that issuer reputation and robust liquidity channels are central to attracting institutional capital. The evolving landscape suggests blockchain networks will compete increasingly on real economic activity and capital efficiency.

ambcrypto07/19 17:02

Tokenized markets reach $2.3B – Why utility matters more than TVL

ambcrypto07/19 17:02

Michael Saylor warns against BIP-110: Is Bitcoin’s utility truly evolving?

Bitcoin's evolution from a pure store of value to a more usable asset is highlighted by proposals like BIP-110, which prioritizes fee-paying transactions. However, Michael Saylor strongly criticizes this direction, arguing Bitcoin's long-term strength lies in institutional adoption rather than protocol changes for utility. On-chain data shows a surge in transactions, but Bitcoin's market valuation is outpacing network adoption, as indicated by a rising Metcalfe Ratio. This growing gap suggests speculation is increasingly driving BTC's price, raising questions about its core value narrative amid debates on network upgrades and scalability.

ambcrypto07/19 15:02

Michael Saylor warns against BIP-110: Is Bitcoin’s utility truly evolving?

ambcrypto07/19 15:02

活动图片

Categorias popularesright-arrow

Etiquetas Popularesright-arrow