LIT surges past $5.2 to hit a new all-time high, gaining over 13% in the past 24 hours.

According to Bitget's market data, Brent crude oil futures hit the $100 per barrel mark, the first time since the end of July.

Insight Beating AI News Flash: DeepSeek plans to officially release its V4.1 Flash model around September 10. According to the official, after internal and external testing, the new model has fully outperformed V4 Pro in metrics including performance, cost, speed, and total latency. Following the launch of V4.1 Flash and prior to the release of V4.1 Pro, all requests originally sent to V4 Pro will automatically be routed to V4.1 Flash and billed at the cheaper Flash rate. Users do not need to modify their existing V4 Pro calling methods.

Beating AI Express: Coding Agent runtime Herdr has secured a $6 million seed round, led by Bessemer Venture Partners, with participation from Y Combinator and e2vc. Individual investors including Shopify CEO Tobi Lütke, Cloudflare CTO Dane Knecht, and fal co-founder G?rkem Yurtseven also joined the round. Herdr remains a 1-person team developed by Can Celik. It is essentially a tmux built specifically for Coding Agents, hosting tools like Claude Code, Codex, and Cursor Agent CLI to enable unified background running, switching, and status monitoring. The Agents continue operating even when the terminal is closed or SSH disconnects. Launched roughly four months ago, Herdr already boasts over 36,000 GitHub stars and 780,000 downloads, with the community having built more than 1,000 plugins and related projects. The funding will primarily be used for hiring; Celik plans to form a small team to focus on operational stability, performance, and scalability. Herdr’s core runtime will remain open-source under the Apache 2.0 license, with no plans to switch to a closed-source model post-funding. Notably, Herdr launched multi-machine management just the day before, allowing users to simultaneously manage Agents on local machines, servers, and Mac minis from a single terminal.

Trader Eugene Ng Ah Sio posted on his personal channel that over the past six months, he has been systematically learning AI stocks—a brand-new asset class—and has witnessed the largest volatility event in stock market history, gaining numerous new trading experiences from the episode. He noted that this experience is quite similar to when he entered the crypto market in 2021. Although trading advantages in the stock market are significantly weaker than in crypto, he does not regret his decision to shift to AI stocks. Eugene added that some trading patterns still hold up in the stock market, but overall, the market is more challenging to navigate, involving more influencing factors and stakeholders. By contrast, the crypto market functions as a relatively independent "sandbox." However, AI remains in the early stage of rapid development, with abundant notable opportunities across the industry and a very steep learning curve. He believes that for speculators, growing personal net worth tenfold via tech stocks may be easier than investing in crypto right now—a trend that will persist over the next three years as the market progresses toward artificial superintelligence (ASI). Eugene further stated that ultimately, choosing the right sector matters more than relative trading advantages. He views the AI semiconductor sector as entering a new rally, and Anthropic’s IPO and subsequent model releases could serve as key catalysts, adding that he is looking forward to potential upcoming market volatility.

Beating AI News: Claude Code head Boris Cherny posted that GPT-6 Astra’s prompt injection protection is “roughly on par” with Gemini Flash and Claude Opus 4.8, adding that his own Claude models had “practically solved” the problem around two months ago. Prompt injection refers to hiding malicious instructions in web pages, documents, and other content to induce AI agents to leak data or perform dangerous actions. Boris also stated that publicly evaluating and naming other labs is an effective way to push them to develop safer models: “We will continue doing this until other labs prioritize safety more.” The post was quickly countered with a Community Note on X. Independent testing by Gray Swan found none of the evaluated models are fully immune to prompt injection, with Claude also having successful attack cases. Another security researcher used a custom web page to target Claude Code Opus 5 Auto Mode, achieving a 60%–80% success rate in small-scale tests. OpenAI’s core product lead Thibault Sottiaux then directly parodied Boris’s phrasing. He said he was pleased to see Claude Code’s new background computer operations finally catch up to Codex: “And it’s our version from May this year.” Boris later acknowledged the original post was “more sarcastic than he intended,” stressing he was genuinely recognizing Astra’s progress. He also clarified that the claim that prompt injection was “solved” refers to the combined performance of Claude models, injection detection mechanisms, and Auto Mode—not that the model itself is inherently immune.

According to an official announcement, Binance’s Cross Margin, Unified Account Mode, and Unified Account Pro will add two new bStocks tokens — Hims & Hers (HIMSB) and Salesforce (CRMB) — as eligible collateral assets at 20:00 (UTC+8) on September 9, 2026. Corresponding bStocks trading pairs will also open for leveraged trading. Eligible users will then be able to use these bStocks tokens as margin collateral, expanding the collateral options for margin trading.