Corma raised $60M from Sequoia to build the defensive AI that cybersecurity is missing
AI models successfully planted persistent threats in 88% of simulations testing Fortune 500 company defenses. Corma raised $60 million from Sequoia to develop defensive AI that identifies security vulnerabilities by simulating attacks across dozens of security tools.
Brazil Signals Surging Confidence in Plan to Export Pix Globally
Brazil's central bank expressed increased confidence in linking its Pix instant payment system to similar platforms in other countries. The central bank stated that connecting Pix to other instant payment systems could lower costs and accelerate transactions.
Point2 raised $136M to solve AI’s wiring problem. Its RF cables beat copper on reach and optics on latency.
RF cables that outperform copper on range and optics on latency are being commercialized after Point2 Technology raised $136 million in Series B funding. The cables are designed to replace both copper and traditional optical connections inside AI data centres.
AWS Continuum integrates with OpenAI Codex and Anthropic Claude Code in major AI security push
AWS Continuum will integrate directly into OpenAI Codex and Anthropic Claude Code to provide vulnerability detection at the point where developers write code. AWS simultaneously expanded Security Hub Extended with a tenth security category for supply chain protection, adding Chainguard and Socket as partners.
When Zoo Elephants Are Rewilded in Africa, Here’s How Good Their Coping Skills Are
Elephants released from captivity into African wildlife systems successfully develop coping mechanisms and readapt to wild conditions. Researchers from the Elephant Reintegration Trust of South Africa examined behavioral patterns in rewilded elephants, finding they can function well within wildlife systems even after long-term captivity.
Robinhood's Head of Product Reveals How Robinhood Chain Hit 200M Transactions in 30 Days
Robinhood Chain processed 200 million transactions within 30 days of launching its blockchain. The platform enables permissionless stock tokens that users can redeem independently of Robinhood's systems.
Tokenized Real-World Assets (RWAs) Near $40B Milestone as US Treasury Debt Dominates
Tokenized real-world assets have reached approximately $38 billion in on-chain value, approaching a $40 billion milestone. US Treasury debt represents the dominant category within this tokenized assets ecosystem.
Texas Governor Orders Audit of Data Centers Seeking Grid Access, Heightening Oversight of AI and Bitcoin Mining Facilities
Governor Greg Abbott instructed Texas regulators to audit all data center projects seeking ERCOT grid access. The August 3, 2026 directive aims to heighten oversight of AI and Bitcoin mining facilities connecting to the state's power network.
OpenAI is hiring a power trader, because electricity is now a position on its books
OpenAI is hiring a power trader to manage its electricity trading and commodity hedging strategy across data centers. The new role will initially have no direct reports and represents the company treating electricity as a significant business position.
Mark Zuckerberg lays out Meta’s AI vision in a 6,500-word essay: 6 things to know
Meta announced a $1 billion fund to invest in communities where its data centers operate and introduced its new open-weight AI model called Muse Glimmer. Zuckerberg argued in his 6,500-word essay that open-source AI models prevent harmful centralization and empower users more effectively than closed systems.
Brex assumes its AI agents could do anything — so it watches the network, not the code
Brex built a network-level security system called CrabTrap to monitor AI agents like OpenClaw in production because traditional code-based security models failed to protect against unpredictable agent behavior. The system treats AI agents as "virtual employees" with email addresses and Slack presence that can perform internal roles, requiring monitoring at the network layer rather than code inspection.
OpenAI Expands Daybreak With Two Tiers and a New Cybersecurity Model
OpenAI split its Daybreak cybersecurity program into two access tiers on August 10, 2026, with Daybreak Blue providing approved defenders access to GPT-5.6 Sol for routine security work. Daybreak Red restricts the new GPT-5.6-Cyber model to more heavily vetted users for vulnerability research, exploit validation, and security testing.
TD Cowen Gives Crypto Clarity Act a 25% Chance of Passing This Fall
TD Cowen estimates the Crypto Clarity Act has a 25% probability of passing this fall. The bank indicated the long-awaited crypto market structure bill remains alive despite slim passage prospects.
Vitalik Buterin's Ethereum Roadmap Prioritizes Quantum Readiness and Privacy
Ethereum's roadmap reprioritized quantum safety above EVM improvements and removed Verkle trees and state expiry from planned work. Privacy, post-quantum scaling, and native rollups, which were absent from the 2023 diagram, now receive focus in the updated priorities.
New Treatment Shows Remarkable Motor Improvements in Boys With Muscular Dystrophy—in Just 12 Weeks
Vamorolone treatment produced motor improvements in young boys with Duchenne muscular dystrophy within 12 weeks. Researchers at Binghamton University are studying early intervention with the drug in very young boys before obvious symptoms develop.
Nvidia Taps Wall Street for $500 Billion Funding Commitment
Nvidia secured a $500 billion financing commitment from Apollo Global Management, Blackstone, BlackRock, and Brookfield Asset Management for artificial intelligence infrastructure. The funding will support the development and deployment of AI computing systems.
New bank targeting 'innovation economy' closes in on $1.5bn funding round
Erebor, a new bank backed by tech billionaires, is closing in on a $1.5 billion funding round. The bank aims to fill the market gap created by Silicon Valley Bank's collapse.
Dyna Robotics Trains DYNA-2 on a Million Hours of Human Video, No Robot Data
A robot model was trained on one million hours of human video instead of teleoperated robot data to develop physical intuition. The approach uses a World-Action Model to transfer learning from human movement to robot hardware without requiring expensive direct robot guidance.
Meta returns to open source with Muse Glimmer, an Apache 2.0 licensed 30B parameter AI model optimized for agents — available now
A 30-billion-parameter open-weight model called Muse Glimmer was released under the permissive Apache 2.0 license to enable autonomous AI agents to run on consumer hardware like high-end Macs and PCs. The model is available on Hugging Face with support rolling out this week through platforms including Ollama, LM Studio, vLLM, and SGLang.
Robinhood Launches Crypto Trading in the UK
Robinhood began offering cryptocurrency trading to eligible UK customers this week. The launch represents the company's expansion into the British market, following a different strategy than competitor Coinbase.
Newsom Orders AI Cyber Defense Program for California Critical Infrastructure
Governor Newsom directed California to establish an AI Cyber Defense Program within the California Cybersecurity Integration Center on August 10, 2026. The program applies AI to vulnerability detection, network hardening, and incident response for state systems, local governments, and critical infrastructure.
New GLP-1 pill delivers up to 12% weight loss in 36 weeks
An experimental GLP-1 pill called aleniglipron produced up to 12.1% body weight loss in 36 weeks. The once-daily oral medication differs from injectable alternatives by being a small molecule that could be manufactured more easily at scale.
Most Enterprise AI Isn’t Secure. Here’s What Businesses Can Do
Enterprise AI deployments rely on security frameworks designed for older technologies, leaving vulnerabilities when employees use browser-based AI platforms. Businesses must rethink their approach to protect generative AI adoption while maintaining competitiveness.
Token-maxxing is dead. Agentic memory is what comes next.
Token consumption became a vanity metric in early 2026 before being abandoned as it measured activity rather than outcomes. Organizations are now recognizing that the context window itself is the scarce resource, shifting focus from maximizing token volume to deciding which information belongs in each prompt.
The Understanding Gap: Why Smarter AI Still Struggles to Deliver Business Value
Global corporate AI investment reached $252.3 billion, yet enterprises struggle to convert advanced AI capabilities into measurable business value. The gap exists not because AI lacks intelligence, but because organizations fail to ensure AI systems genuinely understand their specific operational contexts and business problems.
Lion Nebula Roars to Life With NASA’s Webb
NASA's James Webb Space Telescope captured new images of NGC 2392, a planetary nebula nicknamed the Lion Nebula. The telescope revealed details of the structure previously observed by Hubble in 2000, including the lion-shaped face and comet-shaped objects forming the mane.
Why Agentic AI Will Fail Without Trusted Asset Data
Agentic AI systems can autonomously retrieve data from multiple systems, review maintenance history, check work orders, and compare crew schedules to generate recommendations without manual human intervention. This capability requires trusted asset data because agentic AI acts directly on information rather than simply summarizing it like generative AI does.
1 in 5 people may carry this hidden genetic heart risk
High levels of Lp(a), an inherited cholesterol particle, were linked to increased risk of stroke and cardiovascular death in a study of over 20,000 people. A blood test can detect elevated Lp(a) that standard cholesterol tests typically miss.
Scientists discover a hidden switch inside silver nanocatalysts
Silver nanocatalysts switch their reaction location between different parts of the catalyst depending on whether a solid oxide cell generates electricity or hydrogen. This switching behavior was discovered through analysis of how silver nanoparticles respond to the cell's operational mode.
The Download: AI agents for science, and the “censorship-industrial complex”
AI systems designed for scientific research require advanced reasoning capabilities beyond simple data processing. Schmidt Sciences and researchers emphasize that developing these agents demands computational methods that can work through complex problems methodically.