Insights
Insights
Practical writing on agentic development, engineering leadership and the decisions that matter when adopting this work well.
- Engineering practice4 min read
Standardising on one AI model is a risk, not a simplification
Sakana's new system, Fugu, does something quietly interesting. Rather than trying to be the single best model, it farms each request out to a pool of models through one API — an orchestration layer that, in Sakana's description, "chooses helpers, assigns work, checks results, and merges answers." The company pitched it partly as a hedge: frontier capability without depending on any one provider, after export controls disrupted access to some top models.
- Productivity4 min read
The 39-point gap: why developers can't tell if AI is making them faster
There is a study everyone deploying AI in engineering should read: in July 2025 the research nonprofit METR ran a randomised controlled trial on whether AI tools actually make developers faster.
- Engineering practice4 min read
AI writes 80% of the code. The expensive 20% is still yours.
Describe what you want to an AI coding tool and there is something working on screen within minutes. It looks close to done. But making it actually done is where the time goes, and that stretch takes a lot longer than the fast bit made you expect.
- Engineering leadership5 min read
What agentic development does to the shape of an engineering org
Search "AI-native engineering team" and you will get a dozen near-identical diagrams. A wheel with a data scientist, an ML engineer, an MLOps engineer, a prompt engineer, and recently also some agent-shaped box with a freshly invented title in capital letters. AI Reliability Engineer, or Agent Supervisor. The diagrams are clear, tidy and confident, but one engineer who actually does this work, puts it well: they do not look much at all like the teams that are actually shipping.
More writing is on the way. If a topic would be useful for your team, tell us in your briefing.