The archive

Latest articles

Reporting, explainers, and analysis on data and AI, updated continuously by the newsroom.

Ford AI quality comparison showing 2026 PP100 scores: Porsche at 138, Ford at 152, Nissan at 156, Buick at 162, and overall industry at 175.

Ford AI quality fix puts veterans back in the loop

Ford AI quality is a lesson in automation debt: 350 veteran engineers came back as JD Power scores rose and recalls stayed costly.

Lars Cornelissen · Jun 25, 2026

IBM nanostack chip density trend showing 50 billion transistors in IBM’s 2021 2 nm chip and nearly 100 billion in 2026 sub-1 nm technology.

IBM nanostack chip gives Moore’s Law a vertical path

IBM nanostack chip is a 0.7 nm research architecture that stacks transistors, doubling IBM’s 2021 density claim to nearly 100 billion.

Lars Cornelissen · Jun 25, 2026

Oracle AI layoffs abstract bar chart showing restructuring costs rising from $374 million in fiscal 2025 to $1.8 billion in fiscal 2026.

Oracle AI layoffs expose the cloud capital tradeoff

Oracle AI layoffs are a capital reallocation signal: 21,000 fewer workers, $48 billion raised, and a cloud roadmap funded by debt.

Lars Cornelissen · Jun 24, 2026

Post-quantum cryptography deadline comparison showing Google and Cloudflare at 2029, White House key establishment at 2030, White House signatures at 2031, and NIST general disallowance at 2035.

Post-quantum cryptography deadline gets a 2030 clock

Post-quantum cryptography deadline means high-value federal systems must shift key establishment by 2030 and signatures by 2031.

Lars Cornelissen · Jun 24, 2026

RIFT-Bench deterministic attack success rates by domain: wild 41.9 percent, finance 40.5 percent, medical 31.3 percent, personal assistant 30.1 percent, and travel 27.4 percent.

RIFT-Bench tests agents where prompts cannot reach

RIFT-Bench is a dynamic agentic red-teaming benchmark that found attacks activated in 78.9% to 89.3% of tested agent runs.

Lars Cornelissen · Jun 24, 2026

Copilot spend meter plan chart showing individual Copilot AI credit allowances: Pro 1,500, Pro+ 7,000, and Max 20,000.

Copilot spend meter shows where AI coding costs bite

Copilot spend meter is VS Code's new warning for AI credit overruns. Treat the 1,500 Pro credits as a budget, not a perk.

Lars Cornelissen · Jun 23, 2026

AutoJack exposure chart showing one malicious page, three chained weaknesses, zero stable PyPI releases affected, and one GitHub hardening commit.

AutoJack makes AI agent prototypes a real RCE risk

AutoJack is a host RCE warning for AI agent prototypes: one malicious page chained three AutoGen Studio weaknesses into command execution.

Lars Cornelissen · Jun 23, 2026

FortiBleed credential theft scale showing 430,000 FortiGate firewalls targeted, 86,644 compromised devices, and 22,405 unique domains.

FortiBleed credential theft reaches the firewall edge

FortiBleed credential theft turns compromised FortiGate firewalls into sniffers. Rotate secrets and hunt traffic capture now.

Lars Cornelissen · Jun 23, 2026

High-NA EUV comparison showing ASML NXE low-NA EUV at 13 nm resolution and EXE High-NA EUV at 8 nm resolution, with ASML citing 2.9 times higher transistor density.

High-NA EUV turns AI chips into a $400M bottleneck

High-NA EUV is ASML’s 8 nm lithography step for denser AI chips. Treat the $400M tool as a roadmap risk and cost signal.

Lars Cornelissen · Jun 23, 2026

45C liquid cooling comparison showing conventional cooling-tower systems at 2.6 million gallons per megawatt per year and Nvidia Rubin warm-water cooling at near zero gallons per megawatt per year.

45C liquid cooling gives Nvidia a tougher water defense

45C liquid cooling lets Rubin racks reject heat with dry coolers, cutting on-site water use while shifting scrutiny to power and buildout.

Lars Cornelissen · Jun 23, 2026

45C liquid cooling water use comparison showing conventional cooling towers at 2.6 million gallons per megawatt per year and favorable dry cooler designs near zero.

45C liquid cooling makes AI factories a water test

45C liquid cooling lets Rubin servers use warmer coolant and near zero facility cooling water in favorable climates, changing AI site math.

Lars Cornelissen · Jun 22, 2026

Databricks Lakeflow Designer surface area, with 9 built-in AI Function operators, 9 built-in transformation operators, and 3 user-defined operator types.

Databricks Lakeflow Designer GA cost guide for teams

Databricks Lakeflow Designer is GA for visual, governed data prep. Use it for analyst-built transforms, but meter previews and jobs.

Lars Cornelissen · Jun 22, 2026