Explore all Think content

Learn how to build a streamhouse with our technical guide covering streamhouse architecture patterns, streaming data pipelines, and implementation best practices.

DoorDash and Uber Eats logged just hundreds of AI agent orders out of millions, even as AI and robotics reshape food delivery and restaurant operations.

Explore the hidden costs of enterprise AI, from inference and deployment to cost per useful outcome, and how IT leaders can gain greater control.

IBM Consulting and AWS on security in the AI era: securing agentic systems, building compliant landing zones, and preparing for post-quantum cryptography.

A Knicks phishing test dangles free rings, fake IT desks flood Teams chats and PHP flaws get exploited. Who can you trust in your inbox?

The Cyber Resilience Act (CRA) is a European Union (EU) regulation that establishes horizontal cybersecurity requirements for products with digital elements (PDEs) made available on the Union market.

Learn how organizations can improve analytics performance and infrastructure utilization by optimizing workloads before adding more compute.

Learn how organizations can improve analytics performance and infrastructure utilization by optimizing workloads before adding more compute.

Scientists have spent years trying to figure out what happens inside AI systems, but the answer may be simpler than they thought. Nick Cammarata of AI research company Goodfire has reopened the debate with a provocative essay.

Attackers hijacked 333 Tor exit nodes to intercept SSH connections and target unencrypted email, exposing where anonymity alone falls short.

Credential lifecycle management is the practice of monitoring and protecting digital credentials across each stage of the credential lifecycle, from provisioning to issuance, renewal and revocation.

Cohere Labs tested 28 AI models on 6,000 cultural riddles, revealing gaps in AI’s cultural understanding and how models handle uncertainty.

You can't write all your own code, but you can know who answers for every piece. Here's how sovereign zero trust secures modern software supply chains.

AI uncovers PHP2Shell flaws, Block's Dessa takes a new approach to voice AI, and cultural riddles expose gaps in models' understanding. Plus, highlights from IBM TechXchange 2026. All that and more on Mixture of Experts.

Can AI get scared? Not really. But researchers are using frightened AI and creepy robots to reveal how humans respond to machines.

A PHP bug threatens 70% of the web, Anthropic embeds outside evaluators and hijacked Tor relays steal passwords. Who’s checking the systems we trust?

Learn how IBM Concert Protect helps teams respond to PHP2SHELL with proof of exploitability, application-centric risk prioritization and automated remediation.

GxP compliance (an abbreviation of “good (x) practice”) is the practice of meeting and demonstrating patient safety standards and high product quality standards in the life sciences industry.

Quantum machine learning is one of the most talked about areas of quantum computing—but what is it, really? Discover the promise, challenges, and future of QML in this week’s episode of The Coherence Times.

Follow live news from IBM TechXchange 2026 as developers, engineers, architects and researchers take on AI, cybersecurity, data, infrastructure and emerging tech for the real-world enterprise.

Learn how IBM Concert Protect helps organizations respond to China-linked cyber threats with application context, proof of exploitability and automated remediation.

AI can tackle mathematical problems that have stumped scientists, but physicist Jim Al-Khalili says understanding time may require something that today’s machines lack: imagination.

Anthropic shows small AI models nearing self-replication, a JAMA study tracks how doctors use OpenEvidence and AI agents network at a founders’ dinner. All that and more on Mixture of Experts.

Secure by default is a development principle that encourages teams to deploy products in a hardened state out of the box, rather than requiring users to configure initial basic security settings themselves.

Client AI agents now reach banks directly. Here's why agent-to-agent interaction makes governance, permissions and audit trails the new edge.

IBM Trusteer examines three ChatGPT-themed Chrome extensions and the data access behind their features—from collecting AI conversations and routing prompts through third-party servers to reading session tokens and accepting remote configuration.

Wiz researchers used AI tools to uncover critical PHP security vulnerabilities affecting WordPress, Drupal and Joomla, prompting urgent calls for organizations to patch vulnerable systems.

A study of 39 million physician AI queries found doctors are consulting with AI more on treatment decisions than diagnosis, raising questions about expertise, judgment and when to consult specialists.

A streamhouse is a data architecture that continuously captures, transports, transforms, governs and serves the changing state of a business.

AI is flooding the music world with everything from streaming scams and fake bands to experimental orchestras and endless death metal. Here’s what happens when algorithms start making the music.