净赚一千万美元，特朗普都有哪些「币圈收入」？

长文源：区块律动Published on 2020-08-24Last updated on 2024-08-20

Claude Code Slashes 80% of Prompt Tokens, But Opus 5 Just Adds Them Right Back In

Claude Code, the AI coding assistant from Anthropic, recently announced a massive reduction of over 80% in its system prompt content for models like Opus 5 and Fable 5. The goal was to remove verbose, often conflicting, rules (like strict commenting and documentation requirements) and replace them with a simpler directive: write code that matches the style of the surrounding project. This "pruning" aims to make the model more efficient by reducing internal conflict from overlapping instructions, with no measurable performance drop reported. However, a developer's (@chenchengpro) investigation revealed a twist. While the prompt was drastically cut from 15,225 characters in Opus 4.7 to 4,467 in Opus 4.8, it *increased* by approximately 72% to 7,694 characters in Opus 5. This isn't a contradiction. The "over 80% cut" refers to the overall shift from the old, detailed rulebook-style prompts to a new, streamlined system. The 72% increase for Opus 5 represents new, targeted instructions added to manage the model's enhanced capabilities. Opus 5 is more proactive—it likes to report progress, generate longer outputs, use sub-agents, and expand task scope. The added prompt content (roughly 3,755 characters) primarily provides guidelines for "Delivering work" (controlling task scope, progress reporting) and "Corrections" (limiting excessive self-correction). These new rules are necessary to curb potential over-engineering on simple tasks, ensuring efficiency even as the model becomes more independent. In short, the old, restrictive manual was deleted, but new guidelines were written to harness the model's newfound initiative.

marsbit3m ago

Claude Code Slashes 80% of Prompt Tokens, But Opus 5 Just Adds Them Right Back In

marsbit3m ago

One-Third of arXiv 'Contaminated', 65% of CS Papers Smell of AI, Only 0.7% in Math

Approximately one-third of recently posted arXiv papers show signs of significant AI-generated text, according to a new study. An analysis of 12,750 papers from January 2023 to July 2026 across ten disciplines found a sharp increase in AI text markers following ChatGPT's release, with the overall detection rate reaching 32% in the latest quarter and peaking near 39% in early 2026. The rate varies drastically by field. Computer Science papers lead at 65%, followed by Quantitative Biology (56.3%) and Electrical Engineering (51.3%). Mathematics, however, has the lowest detection rate at just 0.7%. The study's authors note this could be due to mathematicians using AI less or because the detector struggles with the high volume of formulas and symbolic notation in math papers, leaving the true cause unclear. The research highlights that the detector identifies a statistical "AI style" in the text rather than proving full AI authorship. It cannot distinguish between light AI-assisted editing and fully AI-generated content. Furthermore, the detector can produce false positives, as some pre-ChatGPT academic writing also exhibits patterns now flagged as "AI-like." The growing use of AI, particularly in highly competitive fields, is creating a cycle where researchers may feel pressured to adopt AI tools to keep pace. The findings raise questions about the changing nature of academic writing and the emergence of a new "AI style" that is increasingly difficult to distinguish from human prose, potentially undermining trust in written text regardless of its true origin.

marsbit6m ago

One-Third of arXiv 'Contaminated', 65% of CS Papers Smell of AI, Only 0.7% in Math

marsbit6m ago

The Biggest Enemy of the AI Bull Market Is Not a Bubble, But the Bond Market?

The bond market is emerging as the most dangerous variable for the AI stock rally. US Bank's chief investment strategist Michael Hartnett warns that surging bond yields are tightening financial conditions beyond what corporate earnings can support. Key indicators include the 30-year Treasury yield hitting 5.2% (a high since 2007) and real yields reaching 3% (a peak since 2008). Hartnett argues this bond market stress could force the Fed to hike rates, which would be detrimental to equities. A critical risk signal would be if the bullish combination of "rising yields & rising bank stocks" flips to "rising yields & falling bank stocks," potentially triggering a broader de-risking in markets. Simultaneously, credit risk for hyperscale cloud companies has reached record highs, with Credit Default Swaps (CDS) at unprecedented levels. This reflects bond investors' growing skepticism about the return on investment from the massive AI capital expenditure boom. The core concern is: if debt markets refuse to fund the AI spending spree, where will the capital for expensive memory chips and potentially unprofitable frontier models come from? Hartnett frames the current dynamic as "FCI > EPS" – where tightening Financial Conditions outweigh the support from Earnings Per Share. He advises a defensive tilt: going long defensive stocks, high-dividend stocks, and long-duration bonds, while shorting bank stocks, brokers, tech, and industrials to hedge against a potential reversal of the "boom" narrative. From a macro perspective, the 2020s are characterized by supply-side constraints (labor, imports, oil) rather than demand, while bond and equity supply remains abundant due to persistent fiscal deficits and reduced corporate buybacks. In this environment, Hartnett sees gold and Bitcoin quietly forming a base in 2026.

marsbit8m ago

The Biggest Enemy of the AI Bull Market Is Not a Bubble, But the Bond Market?

marsbit8m ago

$2.5 Billion Bitcoin Bet Aims for $70,000 with Only 6 Days Left

Traders have long blamed the "options wall" for suppressing Bitcoin's price throughout July, but with two major weekly expirations now passed and BTC remaining stagnant around $64,000, that narrative has weakened. The core issue appears to be simply thin spot market demand. While a massive ~$2.5 billion bullish options bet targeting prices of $70,000-$72,000 expires on July 31st, market-implied odds of Bitcoin reaching even $70,000 by then are low (~14.5%). Recent data shows subdued leverage, mixed ETF flows, and forced long liquidations, indicating a lack of strong directional conviction. The upcoming expiry coincides with a narrow window for US crypto legislation and follows a key Fed meeting, adding uncertainty. With the weekly "max pain" expirations failing to move price, Bitcoin's range seems determined by the absence of significant spot buyers or sellers.

marsbit10m ago

$2.5 Billion Bitcoin Bet Aims for $70,000 with Only 6 Days Left

marsbit10m ago

"Injective Nova Program" Final Demo Day Summit Finals Successfully Concluded: TOP3 Teams Announced, Co-creating a New Paradigm of AI × Web3

"Injective Nova Program" Final Demo Day Successfully Concludes in Hangzhou: Top 3 Teams Announced, Shaping the New Paradigm of AI × Web3 On July 25, 2026, the Final Demo Day of the "Injective Nova Program," jointly launched by Injective, Microsoft, and Web3Labs, was successfully held in Hangzhou. The event gathered global blockchain experts, AI developers, top VC representatives, and ecosystem partners to witness innovations at the intersection of AI and decentralized networks. The program generated significant global interest, receiving 89 applications from over 10 countries and achieving over 1,000,000 ecosystem impressions. After a rigorous review, a global Top 10 was announced on July 10. These teams participated in an exclusive visit to Microsoft's Beijing headquarters for deep technical exchanges on July 16. At the final demo day, the Top 10 teams delivered live presentations. The jury evaluated projects based on technological innovation, commercial potential, and synergy with the Injective ecosystem, ultimately awarding the top three positions. All winning and shortlisted projects will receive comprehensive incubation support, technical resources, and capital access from Injective, Microsoft, and Web3Labs. The successful conclusion of the Final Demo Day marks a significant milestone for the first phase of the Injective Nova Program. The organizers plan to continue supporting global builders and exploring the limitless possibilities of AI and decentralized technology integration.

marsbit13m ago

"Injective Nova Program" Final Demo Day Summit Finals Successfully Concluded: TOP3 Teams Announced, Co-creating a New Paradigm of AI × Web3

marsbit13m ago

Trading

Spot