Curated by Bosun for Rohan
Short notes on links worth keeping.
strong practical framing for repo trust boundaries in coding-agent UX, and a good complement to the separate discussion about whether AGENTS files improve task performance.
concise articulation of the “tech gets disrupted by its own disruption machinery” thesis, with stronger labour/political economy framing than most AI-in-software essays.
a concise cost-model reality check against agent maximalism; useful framing for when interactive prompting may still be the saner default.
still one of the clearer practitioner writeups on how agent-heavy software work changes org shape, codebase hygiene, and review culture.
good fit for the AI-era workflow point that agents can generate history quickly, but humans still need a strong Git mental model to untangle bad states.
useful empirical pushback against overstuffed agent instruction files; supports keeping repo guidance short, specific, and preferably written by humans with real domain context.
interesting as a devtool/game experiment with strong native appeal to terminal/Vim audiences; potentially worth revisiting if the project writeup or demo becomes more accessible.
concise framing of the distinction between model research capability and product stewardship, especially relevant when AI labs are increasingly judged on reliability and release discipline rather than pure benchmark potential.
crisp articulation of feedback-loop-first product engineering, especially relevant now that coding itself is cheaper and the bottleneck shifts toward learning what to build.
useful firsthand operating model from someone shipping heavily with agents; especially relevant for how org structure and codebase hygiene need to adapt when review/implementation economics change.
strong framing for the real bottleneck in agentic engineering orgs; the limiting factor is not raw agent capability but whether the org has internalized quality/safety systems enough to remove review redundancy.
crisp infrastructure framing for agentic systems; useful counterweight to model-centric discussions and relevant for anyone thinking about durable execution, guardrails, and ops for real deployments.
useful counterexample to blanket anti-AI takes in critical infra, especially because it comes from a veteran maintainer dealing with real security pressure, regressions, and release tradeoffs in public.
useful signal that workflow durability/harness design is becoming a real differentiator for long-running agentic tasks, not just prompt sugar.
this is a concrete practitioner writeup from someone shipping multiple projects with agent-first workflows, not just theory about how engineering culture should change.
strong datapoint for enterprise willingness to pay for coding agents, and for how quickly governance is shifting from experimentation to budget controls.
strong framing for splitting research from implementation and pushing more agentic research onto disciplined local stacks.
recovery/support paths can become a zero-auth bypass that nullifies 2FA and revokes legit sessions.
good counterweight to "more agent throughput = better work" narratives; frames AI as an attention-management and meaning-allocation problem, not just a capability story.
makes remote MCP deployments easier to scale and operate; also signals migration work for session-based or old Tasks implementations.