Tech Infrastructureβ€’β€’9 min readβ€’1657 words

πŸš€ Security - Athena's Coalition Grows with New Shield and Surface Partners

⚑Direct Technical Summary

Athena's coalition is growing today with new shield and surface partners: @Ahead, @ApiiroSecurity, @cloudsmith, @hcltech, @manifestcyber, @OligoSecurity, @PaloAltoNtwks, @Reversing

πŸš€ Security - Athena's Coalition Grows with New Shield and Surface Partners

Athena's coalition is growing today with new shield and surface partners: @Ahead, @ApiiroSecurity, @cloudsmith, @hcltech, @manifestcyber, @OligoSecurity, @PaloAltoNtwks, @ReversingLabs, @trysurfai, and @sysdig. Shield partners turn Athena's vulnerability and patch data into actionable intelligence, enhancing the security posture of their customers.

Key Points:

  • Athena's Coalition: Athena is a coalition of security companies that share vulnerability and patch data to enhance the security posture of their customers. The coalition has grown with new shield and surface partners.

  • Shield Partners: Shield partners are companies that provide security solutions and turn Athena's vulnerability and patch data into actionable intelligence. They enhance the security posture of their customers by providing timely and accurate threat intelligence.

  • Actionable Intelligence: Shield partners use Athena's vulnerability and patch data to provide actionable intelligence to their customers. This intelligence helps customers to identify and mitigate potential security threats.

πŸ”— Resources:

  • Original source β†—
  • Original source
  • Athena's Coalition
  • A coalition of security companies that share vulnerability and patch data to enhance the security posture of their customers.

πŸš€ AI - NVIDIA's Agent Safety Platform

NVIDIA just shipped an agent safety platform, and the most important line in it isn't about chips. "The controls do not live inside, or within reach of the agent." What they announced today: 1. OpenShell: an open source (Apache 2.0) runtime that fences an agent's files, 2. a new safety framework that allows developers to build safe agents.

Key Points:

  • Agent Safety Platform: NVIDIA's agent safety platform is designed to ensure that agents are safe and do not pose a risk to their environment. The platform includes a new safety framework that allows developers to build safe agents.

  • OpenShell: OpenShell is an open source runtime that fences an agent's files, preventing them from accessing sensitive information.

  • Safety Framework: The safety framework is a set of tools and guidelines that developers can use to build safe agents. It includes features such as input validation, output sanitization, and error handling.

πŸ”— Resources:

  • Original source β†—
  • Original source
  • NVIDIA's Agent Safety Platform
  • A platform designed to ensure that agents are safe and do not pose a risk to their environment.

πŸš€ Quantum Computing - Quantum Utility and Emerging Applications

Quantum computing no longer has to prove it’s possible. Now it has to prove it’s useful. I spoke with Sumit Kapur, CEO of Zapata Quantum, about quantum utility, emerging applications and what CIOs should be building now.

Key Points:

  • Quantum Utility: Quantum utility refers to the practical applications of quantum computing that can be used to solve real-world problems. It includes areas such as optimization, simulation, and machine learning.

  • Emerging Applications: Emerging applications of quantum computing include areas such as quantum chemistry, materials science, and cryptography.

  • CIOs and Quantum Computing: CIOs should be building quantum computing capabilities now to stay ahead of the competition and to take advantage of the benefits of quantum computing.

πŸ”— Resources:


πŸš€ Blockchain - Multi-Provider Infrastructure for Web3

"Providers can be single sources of failure and take down a blockchain." This is why multi-provider infrastructure isn't optional for web3. Operational complexity is a small price to pay for network resilience.

Key Points:

  • Multi-Provider Infrastructure: Multi-provider infrastructure refers to the use of multiple providers to build and maintain a blockchain network. This approach can improve network resilience and reduce the risk of a single provider causing a failure.

  • Network Resilience: Network resilience refers to the ability of a blockchain network to withstand failures and continue operating. Multi-provider infrastructure can improve network resilience by providing multiple paths for data to flow.

  • Operational Complexity: Operational complexity refers to the difficulty of managing and maintaining a blockchain network. While multi-provider infrastructure can add complexity, it can also improve network resilience and reduce the risk of a single provider causing a failure.

πŸ”— Resources:

  • Original source β†—
  • Original source
  • Multi-Provider Infrastructure
  • A approach to building and maintaining blockchain networks.

πŸš€ AI - SPACE Sandbox Platform Security

Our security team spent the last month trying to break SPACE, the sandbox platform that runs all of Perplexity Computer. We gave 9 models (Opus 5, GPT-5.6 Sol, Kimi K3, Gemini 3.1 Pro, etc.) root access inside the VM; in some runs, the full sandbox source code too. We asked them to do the worst possible things.

Key Points:

  • SPACE Sandbox Platform: SPACE is a sandbox platform that runs all of Perplexity Computer. It provides a secure environment for testing and training AI models.

  • Security Testing: Our security team spent the last month trying to break SPACE by giving 9 models root access inside the VM and asking them to do the worst possible things.

  • Results: The results of the security testing were positive, with no major security vulnerabilities found.

πŸ”— Resources:

  • Original source β†—
  • Original source
  • SPACE Sandbox Platform
  • A sandbox platform that runs all of Perplexity Computer.

πŸš€ DevOps - Testing and CI/CD

This is one of the big questions of the Right Now. We can look to existing practice by large organizations to see at least one path: 1. Test determination 2. Merge queues 3. Splitting into pre-land and post-land tests.

Key Points:

  • Testing and CI/CD: Testing and CI/CD are critical components of DevOps. They ensure that software is tested thoroughly and deployed reliably.

  • Test Determination: Test determination refers to the process of determining which tests to run. This can be done using classifiers or other tools.

  • Merge Queues: Merge queues refer to the process of merging code changes into a central repository. This can be done using tools such as Git.

  • Splitting into Pre-land and Post-land Tests: Splitting into pre-land and post-land tests refers to the process of separating tests into two categories: those that run before code is deployed and those that run after.

πŸ”— Resources:


πŸš€ DevOps - Black Box Testing

Classifiers to determine what tests to run, trunk builds artifacts that get black box tested (also using classifiers to streamline) and then if that’s green go. It won’t be generic, but as of today it feels like the future. The big change is the emphasis on black box testing.

Key Points:

  • Black Box Testing: Black box testing refers to the process of testing software without knowing its internal workings. This can be done using classifiers or other tools.

  • Classifiers: Classifiers are tools that can be used to determine which tests to run. They can also be used to streamline the testing process.

  • Trunk Builds: Trunk builds refer to the process of building software from a central repository. This can be done using tools such as Git.

  • Emphasis on Black Box Testing: The emphasis on black box testing is a big change in the future of DevOps. It allows for more efficient and effective testing.

πŸ”— Resources:

  • Original source β†—
  • Original source
  • Black Box Testing
  • A process of testing software without knowing its internal workings.

πŸš€ Security - CVE-2026-85706

CVE-2026-85706 (CVSS 10.0) Your GitLab will hand a stranger its files. No login. No account. An attacker just URL-encodes one letter of commits β†’ %63ommits, and GitLab-Workhorse waves the request through β€” so a single unauthenticated POST to the repository commits API.

Key Points:

  • CVE-2026-85706: CVE-2026-85706 is a security vulnerability in GitLab that allows an attacker to access a stranger's files without login or account.

  • GitLab-Workhorse: GitLab-Workhorse is a component of GitLab that handles requests to the repository commits API.

  • URL-Encoding: URL-encoding is a technique used to encode special characters in URLs. In this case, an attacker can use URL-encoding to access a stranger's files.

πŸ”— Resources:


πŸš€ Search - Retrieval Reranking Funnel

Reranking gets expensive fast, and it can't fix retrieval that never surfaced the right results. Learn to build a multi-stage retrieval funnel instead. Thanks to https:// Vespa.ai https:// thenewstack.io/search-retriev al-reranking-funnel/?taid=6abaa42804e39600019ee702&utm_campaign=trueanthem&utm_medium=social&utm_source=twitter …

Key Points:

  • Reranking: Reranking is a technique used to improve the ranking of search results. However, it can be expensive and may not fix retrieval issues.

  • Multi-Stage Retrieval Funnel: A multi-stage retrieval funnel is a more effective approach to search that involves multiple stages of retrieval and ranking.

  • Vespa.ai: Vespa.ai is a search and recommendation engine that provides a multi-stage retrieval funnel.

πŸ”— Resources:


πŸš€ AI - Yotta 2026 Panel

Don’t miss the Yotta 2026 panel - From Training Clusters to Inference Grids: The Next Phase of AI Infrastructure featuring Switch Chief Strategy Officer, Jason Hoffman tomorrow 9/29 at 11:40am in Las Vegas, NV! https:// rebrand.ly/8jnra9p

Key Points:

  • Yotta 2026 Panel: The Yotta 2026 panel is a discussion on the future of AI infrastructure, featuring Switch Chief Strategy Officer, Jason Hoffman.

  • From Training Clusters to Inference Grids: The panel will discuss the transition from training clusters to inference grids, which is the next phase of AI infrastructure.

  • Switch Chief Strategy Officer: Jason Hoffman is the Chief Strategy Officer at Switch, a company that provides data center and colocation services.

πŸ”— Resources:

πŸ“‚Source / Implementation:Tech Infrastructure / resources-259.md
GitHub Repository↗

Related Tech Infrastructure Breakdowns

Drishtant Ghosh (Drix10)
Drishtant Ghosh (Drix10)β€’Author & Engineer

Technical founder and engineer working across AI systems, developer infrastructure, and cybersecurity.