/Murat Demirbas

Looming Liability Machines (LLMs) tl;dr: “We discussed a paper that uses LLMs for automatic root cause analysis (RCA) for cloud incidents. This was a pretty straightforward application of LLMs. The proposed system employs an LLM to match incoming incidents to incident handlers based on their alert types, predicts the incident's root cause category, and provides an explanatory narrative... The use of LLMs for RCAs spooked me viscerally.”

featured in #544


Advice To The Young tl;dr: (1) It is called foundations, not theory. (2) Keep your hands dirty, your mind clean. (3) Ship something, anything, weekly. (4) Cram for the deadlines, put your heart into it. (5) Ask and you shall receive. (6) People skills are very important. (7) Manage the stories you tell yourself. (8) Cultivate deep focus through deliberate practice.

featured in #537


Know Yourself tl;dr: “MongoDB has a nice leadership development program internally. They suggested that filling / sharing this questionnaire would be useful to get you acquainted with the people you work with daily." Questions include: (1) What are you amazing at? Where do you want to improve? (2) What makes you most excited about your work / role? (3) Describe an ideal workday. (4) What is your meeting participation style? (5) What is something that people incorrectly assume about you?

featured in #522


The Demise Of Coding Is Greatly Exaggerated tl;dr: “Natural language is ambiguous and not suitable for programming. LLMs still need to generate code to get things done. If not inspected carefully, this incurs tech debt at monumental speed of the computers. The natural language prompts are not repeatable/deterministic, they are subject to breaking any time. This makes "natural language programming" unsuitable for even small sized projects, let alone medium to large projects.” Murat also believes that certain tasks require too much expertise to be completed by an LLMs as they stand. 

featured in #498


Looking Back At Postgres tl;dr: The Postgres origin story: “Riding on the success of Ingres project at Berkeley, Stonebraker began working on database support for data types beyond the traditional rows and columns in the early 1980s. A motivating example was to provide database support for CAD tools for the microelectronics industry, including "new data types such as polygons, rectangles, text strings, etc...

featured in #483


The Checklist Manifesto tl;dr: This book advocates for integrating checklists as potent safety and fault-tolerance tools across diverse domains. While the author, a prominent surgeon, enriches the narrative with numerous surgery cases, he also discusses their use in the construction and aviation industries. Checklists significantly reduce cognitive load, enabling complex tasks and effective team collaboration. Murat questions why we don’t use checklists more frequently in software development.

featured in #480


Hints for Distributed Systems Design tl;dr: “I have seen these hints successfully applied in distributed systems design throughout my 25 years in the field, starting from the theory of distributed systems, immersing into the practice of wireless sensor networks, and working on cloud computing systems both in the academia and industry ever since. These heuristic principles have been applied knowingly or unknowingly and has proven useful. I didn't invent any of these hints. These are collective products of distributed systems researchers and practitioners over many decades.”

featured in #455


SQLite: Past, Present, And Future tl;dr: "SQLite is an awesome little engine; it reaps the benefits of its size and deployment constraints - single node and mostly single threaded - to keep things simple, agile, and reliable. It is a great motorcycle, but the world also needs fleets of vans and 16-wheelers for high performance scale-out workloads." Murat guides us through SQLite's architecture. 

featured in #353


Learning A Technical Subject tl;dr: Murat believes cultivating the following attitudes helps navigate learning more easily: (1) curiosity, (2) relentlessness, (3) being a hands-on maker, (4) being social & learning to communicate well, (5) analyzing and drawing lessons, (6) leveraging previous experience. 

featured in #329


Master Your Questioning Skills tl;dr: Murat committed to asking "crazier" left-field questions in each post he wrote. After 70 posts, he learned that these: (1) Are useful as they took him out of "default mode." (2) Keep the brain more engaged. We tend to ask easier questions. Harder ones feel uncomfortable or seem impolite. (3) By calling them "MAD" questions, they give him the license to be more uninhibited. 

featured in #322