ibm

IBM announced a new modular cryogenic architecture to connect quantum processors, enabling scalability and future fault-tolerant quantum systems as part of their quantum roadmap.

IBM and OpenAI announced a partnership to bring OpenAI's models and tools to enterprise customers through IBM Consulting, including training tens of thousands of consultants and integrating GPT-5.6, Codex, and ChatGPT Work into IBM's AI platform.

This paper introduces BenchDrift, a method for quantifying how LLM benchmark performance changes when problems are rephrased without changing meaning or answer. It shows that rephrasing causes bidirectional correctness flips across models and benchmarks, with stronger models becoming more sensitive to phrasing.

IBM and collaborators demonstrated three techniques for verifiable quantum advantage on noisy hardware, tackling the challenge of validating results that cannot be classically simulated.

The author conducts an independent spin audit of IBM's SQD/QSCI iron-sulfur cluster benchmarks, finding that the quantum results converge to high-spin states rather than the target singlet. Using IBM's own archived hardware data, they show the shipped mitigation fails to fix the spin problem, and a Claude-assisted review is provided.

IBM and collaborators report demonstrations of quantum advantage with built-in validation using doped Clifford sampling and error-mitigation techniques, enabling trustworthy quantum computations beyond classical verification.

Quantego is a project offering free LEGO build instructions and interactive 3D models of IBM Quantum System One and System Two, with educational quantum circuit simulations.

IBM announces three new entries in its quantum advantage tracker, each using different approaches to overcome errors and validate quantum results, demonstrating quantum advantage in ways that can be trusted even when classical verification is infeasible.

IBM has launched a free 1-hour course on building agentic knowledge graphs, covering topics from introduction to multi-agent orchestration.

Major tech companies including Microsoft, NVIDIA, Meta, IBM, and Palantir issued a joint letter urging Washington regulators to avoid policies that would cripple open-weight AI models, warning of negative consequences for innovation.

IBM reported a 42% drop in mainframe sales, leading to a 25% stock crash, but CEO Arvind Krishna insists the decline is temporary due to customers prioritizing AI hardware spending, not a sign of mainframe demise.

IBM's stock plunged 24% after CEO Arvind Krishna warned of worse-than-expected quarterly performance due to customers shifting budgets to AI datacenter components, causing delayed mainframe purchases and disrupted deals.

The world's largest companies are starting to build their own custom software to replace off-the-shelf solutions. Starbucks dropping IBM and Microsoft in favor of internal systems is seen as a major signal for the AI consulting business.

Starbucks announced it will move away from IBM and Microsoft to build its own custom systems internally, leading to declines in IBM and Salesforce stock prices.

The author built a daily digest POC using IBM Bob, IBM's AI coding agent, and shares their experience with its new v2 features including parallel tool calls and simplified modes, while noting limitations like lack of reusable skills. They ask the community for their own experiences with Bob.

A detailed examination of two circuit boards from the Space Shuttle's I/O Processor, explaining its multi-threaded architecture, network interfaces, and microcode storage using fuse-based PROMs.

IBM announces a breakthrough chip architecture called NanoStack, achieving a sub-1nm node (0.7nm) that packs nearly 100 billion transistors on a fingernail-sized chip, with 50% better performance and 70% energy efficiency than its 2nm chip.

IBM signals a renewed focus on efficiency in AI, marking its comeback in the industry.

IBM announces a new nanostack transistor architecture that enables sub-1 nanometer node chips, offering significant performance and energy efficiency gains for AI data centers.

IBM has unveiled a new chip architecture called nanostack that vertically stacks transistors in two layers, enabling twice the transistor density of its previous technology. This breakthrough could extend Moore's Law for another decade and improve chip performance by 50% and energy efficiency by 70%.