VarBear

Jack Dorsey just cut nearly 1/2 of Blocks, from 10,000 to 6,000 people.

His reason is not financial trouble. Gross profit is growing and the customer base too.

He said it plainly: AI-powered tools paired with smaller, flatter teams are "fundamentally changing what it means to build and run a company."

A few things stood out to me:

👉He chose one hard cut over gradual reductions. His reasoning: repeated rounds of cuts destroy morale, focus, and trust more than a single honest action.

👉He's not hiding behind performance. Most layoff announcements are wrapped in corporate language. This one is not.

For developers and engineers, this is a signal! Many CEOs are thinking about the same thing, but not saying it out loud yet.

For entry-level and junior professionals, things are getting even tougher. Anthropic's CEO warned that AI could wipe out 50% of entry-level lawyers, consultants, and finance professionals within the next few months.

This is the new normal for tech companies.

Have a great week,
Aymen

AI is quietly moving from autocomplete to ops: agents that trigger off real production signals, run in isolated sandboxes, and leave you with reviews and PRs instead of TODOs. On the flip side, there is a renewed respect for fundamentals, like logging discipline and boring, sharp HTTP primitives in Go.

🤖 Build agents that run automatically
📱 I deleted my laptop from my dev workflow. My iPhone does the job now
📉 Logging at the wrong level
🧭 The Great Developer Divide: How AI Is Reshaping the Software Job Market Into Three Tiers
🧑‍💻 We Might All Be AI Engineers Now
🧰 Zen: A Minimalist HTTP Library for Go

Steal the patterns, skip the hype, and ship something sturdier this week.

Have a great week!
FAUN.dev() Team

At scale, IaC state management isn't a storage problem. It's a people and process problem. Join us live to learn how platform teams map dependencies, govern multi-team workflows, & stop mystery outages before they start. You’ll walk away with a practical checklist for scaling IaC.

Register now for this free virtual event on March 12 @ 12PM ET.

If you’ve managed Infrastructure as Code in production, scaled platforms under pressure, or built guardrails that held up at speed, we want to hear from you. IaCConf 2026 is seeking practitioners to present 40-min sessions on May 14 (virtual).

Submit your proposal by April 7.

If you're working with Kubernetes and exploring AI/ML in real-world environments, this 5-hour live workshop focuses on the practical side of running AI workloads in production.

🎟 Early Bird Offer – 50% Off (No Code Needed) : Build & Scale AI Workloads on Kubernetes Tickets, Sat, Mar 28, 2026 at 7:00 PM | Eventbrite

Unkey built Zen - a thin HTTP framework on Go's net/http. It restores precise middleware ordering and lets middleware run after errors to capture the final response.

Zen pools Session objects to cut allocations. It emits RFC7807 problem+json for tagged domain errors. It runs OpenAPI validation before handlers.

The team prefers a schema-first OpenAPI flow with generated types. They keep external deps minimal. Code lives at github.com/unkeyed/unkey.

A developer ditches the laptop and SSHs from an iPhone into an always-on Mac Mini.
The phone becomes a terminal and browser. The remote runs the dev server, the Claude Code/Codex CLI, hot reload, file watching, and pushes via Tailscale.
Persistent sessions (tmux) keep AI agents and services alive across dropped connections. Neovim, lazygit, bat, and fzf make phone terminals usable.

AI hiring has split dev work into three camps: Apex Tier, Hybrid Middle, and a shrinking Automatable Tail. Demand now favors AI orchestration, prompt engineering, fast code reading, and platform roles like platform engineer, fleet supervisor, and AI QA.

System shift: Organizations must rework career ladders, tighten platform stability, and refactor tooling to enable orchestration-centric hybrid roles. Ignore the change and hiring will outpace structure.

The author supervises AI agents that orchestrate concurrent graph traversal, multi-layer hashing, AST parsing, and file system watchers. The agents run traversal, hashing, and watcher loops. The engineer architects system behavior, verifies outputs, and probes agents in parallel to debug.

Agents trigger from schedules, Slack, Linear, GitHub, PagerDuty events, or custom webhooks.

They spin up cloud sandboxes. They run configured MCPs and models. They verify outputs. They use a memory tool.

Cursor automates security audits on pushes. Scores PR risk and auto-approves low-risk changes. Runs PagerDuty and Datadog triage that proposes PR fixes. Creates tests. Sends weekly digests.

System shift: Autonomous agents run builds, tests, audits, and triage in isolated cloud sandboxes. CI shifts toward continuous, agent-driven maintenance.

Open-source orchestration for zero-human companies

Animated sprite editor & pixel art tool

Secure environments for developers and their agents

Practice made claude perfect

Browser automation CLI for AI agents

Did you know that SQLite's test infrastructure runs hundreds of millions of test cases before each release, using multiple harnesses that systematically vary page sizes, journaling modes, and character encodings - not just random inputs? The TH3 harness achieves 100% branch coverage, test builds simulate malloc failures and I/O errors at the OS layer, and a dedicated fuzzer runs around 500 million cases every day. This level of rigor is why SQLite ships inside every Firefox and Chrome browser, every iPhone, and most Android devices.

Logging at the wrong level is one of the most quietly expensive habits in a codebase - DEBUG noise that gets promoted to INFO in prod, or errors swallowed into WARN because someone didn't want alerts to fire. Six months later, you're debugging an outage with logs that either tell you nothing or tell you everything except what actually failed. The log level contract is a design decision, not a cosmetic one - break it early and you'll spend the rest of the system's life compensating.