DeepFlowTech

Argentine President Milei faces a political crisis over the LIBRA token scandal. Investigations reveal that Milei had seven phone calls with key project figures before the token promotion and discovered a draft promotion payment agreement worth $5 million. Ultimately, the token's market value plummeted, causing tens of thousands of investors to suffer significant losses. Congress has now resumed investigation procedures.

OpenClaw released version 2026.4.7 on April 8, including the addition of a headless reasoning core, Webhook-driven TaskFlow, and memory-wiki knowledge layer, supporting multimodal reasoning and knowledge management. It is also compatible with Obsidian format and enhances support for various visual models.

The U.S. Department of Justice rejected Tornado Cash developer Roman Storm's request, stating that the Supreme Court's copyright ruling is unrelated to the criminal charges he faces. Storm has been convicted of operating an illegal money transfer business, and two other charges are still pending review.

Author: Deep Tide TechFlow
On April 7th, Anthropic did something never before seen in the AI industry: they officially released a model and then told the world, "You can't use it."
This model is called Claude Mythos Preview. It’s not a chatbot, nor a coding assistant. According to Anthropic, it is a "vulnerability discovery machine" that has autonomously found thousands of zero-day vulnerabilities over the past few weeks, covering all major operating systems and all mainstream browsers. Some of these vulnerabilities have been present in human code reviews and automated testing for over twenty years.
The oldest bug, hidden in the security-focused OpenBSD, is already 27 years old.
Anthropic has not released Mythos publicly.