In partnership with
Hey {{first name | there}}. This week, the people building AI are rethinking how software gets reviewed, security researchers are uncovering new ways attackers exploit the AI boom, and tech leaders are making some questionable suggestions about how we should use these tools.
There’s plenty to unpack.
In today’s roundup:
Entire launches a GitHub rival that reinvents code review
Datadog builds a specialist AI model for SRE investigations
Fake ChatGPT and Claude apps distribute malware and crypto miners
GitHub plans a major infrastructure expansion to handle AI-driven traffic
AI executives face criticism over their vision for AI in everyday life
TOGETHER WITH EVERYTHINGDEVOPS
Your AI prompts are leaving out 80% of what you're thinking.
When you type a prompt, you summarize. When you speak one, you explain. Wispr Flow captures your full reasoning (constraints, edge cases, examples, tone) and turns it into clean, structured text you paste into ChatGPT, Claude, or any AI tool. The difference shows up immediately. More context in, fewer follow-ups out.
Used by teams at Microsoft, Amazon, and Notion. Try Wispr Flow free on Mac, Windows, iPhone, and Android.
LATEST DEVELOPMENT
Image credit: TheRegister
The Rollout: GitHub is planning to add four million CPU cores and 160 petabytes of high-speed storage as AI-generated code drives record traffic and recurring outages. The company is also redesigning its architecture to handle up to 35 times more throughput.
The details:
GitHub recorded 7.4 billion code commits in September 2026, five times the volume from a year earlier.
The company plans to move to a multi-region Azure deployment and separate compute from storage, with migration expected to finish early next year.
GitHub also launched Copilot Fusion, a model-routing service, and introduced AI security scanning, package firewall capabilities, and additional Copilot features.
Why this matters: AI coding agents are increasing demand on the infrastructure developers rely on every day. GitHub's expansion highlights the operational challenges of supporting that growth while maintaining reliability, scaling CI/CD workloads, and securing increasingly large volumes of code.
Image credit: Datadog
The Rollout: Datadog has introduced Scout, a specialised AI model designed to investigate production incidents inside its Bits Investigation product. In a closed preview, Scout analyses telemetry and suggests root causes, aiming to help engineers diagnose problems faster and at a lower cost.
The details:
Datadog says Scout's inference costs are three to eight times lower than comparable general-purpose models on its AlertEval benchmark.
On SREGym's 90-fault evaluation, Scout resolved 80 cases at its highest effort setting, compared with 82 for GPT-6.1 Sol.
Scout can investigate incidents and recommend recovery actions, but engineers remain responsible for making changes. General availability has not been announced.
Why this matters: Investigating every alert with a general-purpose AI model can be expensive. Datadog is betting that a specialised model can make automated incident investigation affordable enough to run more often. However, the results come from company-run evaluations, so independent testing will help establish how well Scout performs in real production environments.
Image credit: TheNewStack
The Rollout: Entire, founded by former GitHub CEO Thomas Dohmke, launched a GitHub alternative featuring Trails, an agent-driven review system that evaluates code changes and tracks the intent behind them.
The details:
Trails runs configurable AI evaluations on every push, flagging risks and intent drift so coding agents can address issues before developers approve changes.
Entire combines Git-compatible hosting, agent memory, and distributed repositories, with private hosting available in the US, Europe, India, and Australia.
The platform is free, with a $100 starter credit. After that, users pay for the tokens consumed by review agents.
Why this matters: As AI coding agents produce more code, reviewing it all becomes a growing bottleneck. Entire is betting that tracking an agent’s intent and work history can make reviews more scalable, though relying on automated evaluations to approve merges still carries risks.
🧠PRESENTED BY EVERYTHINGDEVOPS
AI agents are spending more time in the terminal, running commands, searching code, and working with JSON.
If you can’t follow what they’re doing, it’s difficult to know when something has gone wrong.
This video covers five CLI tools that make everyday terminal work faster: zoxide, ripgrep, jq, fzf, and tmux.
You’ll see what each one does, how they fit into your workflow, and why you should still know the classic Unix commands underneath them.
Subscribe: @EverythingDevOpsHQ for deeper dives on agents, Kubernetes, MLOps, and AI infrastructure.
Image credit: The Hacker News
The Rollout: A campaign dubbed Yoba uses fake GitHub projects advertising AI desktop apps to distribute malware that eventually installs a cryptocurrency miner, hiding malicious components alongside legitimate, digitally signed executables.
The details:
Attackers promote fake ChatGPT and Claude apps through GitHub repositories and websites, using search engine manipulation to make downloads appear credible.
Malicious DLLs use sideloading, anti-analysis checks, and scheduled tasks to evade detection and maintain persistence on infected systems.
YobaLoader uses Telegram, generated domains, and Google’s DNS-over-HTTPS service to retrieve commands and download its next-stage payload.
Why this matters: The growing interest in AI tools gives attackers a convincing disguise for malware. Developers should download AI applications only from verified official sources, since a GitHub listing or signed executable alone does not guarantee that a download is safe.
Image credit: Nhat V. Meyer/Bay Area News Group/ Getty Images
The Rollout: TechCrunch argues that Silicon Valley leaders are increasingly pitching AI as a way to manage parenting, friendships, and emotional needs, raising questions about whether making life frictionless could come at the expense of meaningful human connection.
The details:
Marissa Mayer showcased how her AI app Dazzle helped surface details about her daughter, sparking criticism that parents should know these things themselves.
Sam Altman has suggested using AI to track children’s interests and milestones, while Mark Zuckerberg predicts greater acceptance of AI companions and therapists.
A Pew Research Center survey found that 50% of US adults expected AI to reduce people’s ability to form meaningful relationships, compared with 5% who expected it to help.
Why this matters: AI can help with planning and routine tasks, but applying efficient thinking to relationships raises a different question: what happens when we delegate the attention, effort, and everyday interactions through which human connections grow?
QUICK LINKS
📅EVENTS
🇪🇺 🇬🇧Europe
Title | Date | Location |
October 31 | Newcastle upon Tyne, United Kingdom | |
November 1- 3 | Italy | |
November 6 | Brighton (UK) |
🇺🇸🇨🇦North America
Title | Date | Location |
November 2 | Join online or in San Francisco | |
November 2 - 4 | Brooklyn | |
November 6 | Montreal, Canada | |
November 8 | Salt Lake City, UT |
🇳🇬Africa
Title | Date | Location |
October 31 | Kansanga Ggaba Road, Kampala, Uganda | |
October 31 | Kaduna, Nigeria | |
November 5 - 6 | Nairobi, Kenya |