AI Industry Development & Policy Dynamics

Cover foundation model launches, product releases, major company developments, funding activity, regulation, compliance, and compute infrastructure trends shaping the global AI landscape.
All
AI Agent
AI Industry News
AI Tokens
AI Tools & Apps

Trump, China Both Dismiss AI Slowdown Proposal This Week

According to AP News and official statements, President Trump and China's Foreign Ministry recently rejected Anthropic CEO Dario Amodei's call to slow frontier AI development. On Truth Social, Trump dismissed the proposal as part of a hoax, stating "AI taking over the World, destroying Humanity, and
GateNews·09-20 02:32
Trump, China Both Dismiss AI Slowdown Proposal This Week

Tencent: Xi Dan said support for basic scientific research remains unchanged, and the 77-billion-parameter preview version of Hy4 has been open-sourced.

On Sept. 19, Tencent made two announcements simultaneously. At the New Cornerstone 50² Forum, Tencent Senior Vice President and Chief Talent Officer Xi Dan said, “Tencent’s determination to support basic research will not change. This is also our long-term commitment to going beyond the boundaries of business,” emphasizing Tencent’s commitment to the Science Exploration Award and the New Cornerstone Researchers Program. On the same day, Tencent released the Hy4 preview large language model as op
Bisc·09-20 02:29
騰訊首席人才官奚丹聲明基礎科研支持不變

Claude Code Now Natively Supports AGENTS.md, Finally Answering the Shopify CEO’s Request

Anthropic released Claude Code version 2.1.277 on September 18, with native AGENTS.md support listed as the first item in the changelog. At the end of August, Shopify CEO Tobi Lütke publicly complained that Claude Code recognized only CLAUDE.md and said he was considering banning it, making this one of the industry pressures that directly drove the update. AGENTS.md Support in Claude Code 2.1.277: Configuration Methods and Applicable Conditions for Four Modes According to Anthropic’s changelog a
Bisc·09-20 02:23
Claude Code原生支援AGENTS.md

Bart Smith: AI agents flooding into finance will exhaust L1 blockspace, while traditional markets move toward 24/7 trading.

Avalanche Treasury CEO Bart Smith warned at Avalanche Summit that as AI agents enter financial markets on a massive scale, blockchain blockspace will come under unprecedented pressure: “Blockspace is no longer unlimited.” He also predicted that traditional financial markets are highly likely to move toward 24/5 trading by mid-2027. Smith’s Warning on L1 Blockspace Scarcity: The Scale of AI-Agent Financial Trading and Differences in L1 Technology In an exclusive interview with The Block, Smith sa
AVAX-0.10%
SOL-0.77%
ETH-1.26%
RWA-0.39%
Bisc·09-20 01:56
Bart Smith警告AI代理湧入金融將耗盡L1區塊空間

Two Perplexity engineers built CobbleDB themselves, cutting latency fivefold and replacing Amazon’s system

The Perplexity engineering team published a technical article on Sept. 14 explaining how it replaced Amazon DynamoDB with CobbleDB, an in-house key-value Storage system comprising approximately 40k lines of Rust code. After the migration, median batch-read latency dropped from 31.4 milliseconds to 5.60 milliseconds, while p99 latency fell from 123 milliseconds to 24.2 milliseconds. CobbleDB Performance Data: Measurement Conditions for Five- to Sixfold Latency Reduction and 20% Cost Savings Accor
AMZN+2.23%
Bisc·09-20 01:52
Perplexity兩名工程師自建CobbleDB延遲降五倍

OpenAI’s Five-Year Hashrate Spending: 856 Billion; 2030 Revenue Target: 350 Billion

The UK's Financial Times (FT) reported on September 18 that an internal OpenAI presentation from July showed that the company estimates it will incur a cumulative $278 billion in negative free cash flow over the five years from 2026 through 2030. The presentation estimates approximately $856 billion in Hashrate and infrastructure spending, exceeding cumulative revenue of $840 billion over the same period. This means that all the revenue generated over those five years would still be insufficient
ORCL-8.84%
NVDA-1.27%
Bisc·09-20 01:47
OpenAI五年算力支出8560億美元

Elon Musk: AI risks are rising exponentially; proposes that major AI companies test each other’s models

Elon Musk said during a conversation at the All-In Summit that, at this stage, the dangers of AI have become “very significant,” and that as model capabilities continue to improve, AI risks are increasing exponentially. Musk’s proposed solution is not to establish a new global AI regulatory agency, but to require major AI companies to conduct safety testing on one another before officially releasing their models, including companies such as OpenAI and Anthropic. Musk’s AI Danger Warning at the A
Bisc·09-20 01:43
馬斯克警告AI風險指數式上升