DevSecOps Maturity Model

The 6 Core Competencies of Mature DevSecOps Orgs Understand the core competencies that define mature DevSecOps organizations. This whitepaper offers a clear framework to assess your organization's current capabilities, define where you want to be, and outline practical steps to advance in your journey. Evaluate and strengthen your DevSecOps practices with Datadog's maturity model.

The C-to-Rust Translator Is Coming! - Canonical and the University of Bristol are working on a way to automate the difficult process of translating C code into Rust. The goal is to make large-scale migration safer, faster, and more practical. Read more.

Claude finally remembers everything - Anthropic is giving Claude Cowork persistent memory, so it can remember useful context across different tasks. Instead of starting from zero each time, Claude can build on previous work and preferences. Read more.

Will the AI Boom Keep Going?

The AI industry has spent the last few years moving at a speed that has surprised even many people inside the technology world. Companies such as OpenAI and Anthropic are reporting huge increases in revenue, while billions of dollars are being poured into chips, servers, power infrastructure, and data centers. But the obvious question is becoming harder to ignore: how long can this continue?

The Forecasting Research Institute recently asked AI experts, superforecasters, and members of the public to look several years ahead and estimate where the industry could be heading. Their answers suggest that the AI boom is likely to continue, but not necessarily at the same explosive pace seen recently.

Data Centers Are Still Expected to Grow

One of the clearest signals from the forecasts is continued spending on data center infrastructure. Experts expect private US investment in data center structures to rise 76% in real terms by 2028, compared with 2025 levels.

That would take annual investment from around $40.8 billion in 2025 to $71.8 billion in 2028. Superforecasters were similarly optimistic, expecting investment to grow by about 70% over the same period.

This comes despite growing concerns around electricity shortages, slow grid connections, public opposition, and restrictions on new data center construction. In other words, the AI infrastructure race may continue even as building it becomes increasingly difficult.

Chips Could Slow Down

The forecasts become more cautious when looking at semiconductor stocks. The VanEck Semiconductor ETF, which tracks major semiconductor companies, has more than tripled over the past 2.5 years as AI demand pushed chip companies higher.

Experts don't expect that pace to continue. Their median forecast suggests the ETF could reach around 1.5 times its mid-2026 value by the end of 2028, rather than repeating its recent surge.

That doesn't necessarily mean the AI chip market is weakening. Instead, it suggests that expectations and stock valuations may have already priced in a significant amount of future AI growth.

Software May Have More Room to Run

Software stocks tell a slightly different story. Experts expect North American software equities to grow around 20–30% by the end of 2028, somewhat faster than their recent trend.

The big question is whether AI will become a powerful new feature for software companies or make some of those companies unnecessary altogether. Companies that successfully turn AI into products and revenue could benefit, while businesses whose core software becomes replaceable may struggle.

OpenAI and Anthropic Are the Wild Card

The most dramatic forecasts concern OpenAI and Anthropic. Superforecasters estimate a 34% chance that their combined annualized revenue run rate exceeds $400 billion by 2030, while putting an 18% chance on it falling below $100 billion.

The numbers are extraordinary, but so is the uncertainty. The AI boom still faces huge infrastructure costs, competition, regulation, and questions about how much businesses will ultimately pay for AI.

For now, the forecasts point toward continued growth rather than an immediate bust but with the easy gains of the AI boom potentially becoming much harder to repeat.

End of Microsoft Outages ft. K8s - Microsoft is adding more control over how AKS handles node disruptions, helping teams avoid unexpected changes to running workloads. The update gives Kubernetes operators better control over node replacement and consolidation. It’s a small change aimed at making automated infrastructure a little less chaotic. Read more.

AWS buys Duck! - DuckLabs, the team behind DuckDB, is joining AWS in a move that could bring its fast analytics technology to a much bigger audience. DuckDB itself will remain free and open source under the DuckDB Foundation. Read more.

Buzz of the Week!

Cache Stampede

A cache stampede happens when a popular cached item expires and many requests try to regenerate it at the same time. Instead of one request rebuilding the data while others wait, hundreds or thousands of requests may hit the database simultaneously. This can create a sudden traffic spike exactly when the system is already under pressure. It is especially dangerous for popular API responses, product pages, configuration data, and authentication metadata. Common ways to prevent it include request coalescing, cache locking, stale-while-revalidate, randomized TTLs, and background refreshes. The key idea is simple: an expired cache entry should not cause every request to rush to the database at once.

Things that launched. Things that went viral. Things you'll pretend to try.

sourcery

sourcery automatically detects opportunities to simplify and improve Python code while you work.

just

just is a modern command runner that can replace complicated Makefiles with much simpler task definitions.

mise

mise is a development environment manager for languages, runtimes, environment variables, and project-specific tools.

Build Braincells, Not Just Features

This weekend’s read: How much Internet is written with AI.

This week’s watch: Simply Explaining Financial Questions EVERYONE Has.

Meanwhile…