Issue #429

Tuesday 11th July’s issue is presented by Firebolt

Are your queries not nearly as fast as you want them to be? Join Firebolt’s Live Product Showdown where you’ll learn how to deliver sub-second analytics over TB-scale datasets with the cloud data warehouse built for engineers. 

Get Started

Bottlenecks vs Bandpass

— Andrew Bosworth

tl;dr: To avoid bottlenecks in product development, horizontal teams should establish clear guidelines and standards, allowing vertical teams to work efficiently. This frees up time for horizontal experts to focus on complex issues and enables faster progress in the future.

Leadership Management Process

How To Do Great Work

— Paul Graham

tl;dr: Key takeaways include: (1) Recognizing the right kind of crazy: Good ideas that are innovative and groundbreaking often seem crazy or bad to most people. (2) Breaking rules: Being independent-minded, whether aggressively or passively, allows for rule-breaking. (3) Choosing the right problems: People tend to be more conservative when selecting problems to solve, favoring fashionable problems. And more.

CareerAdvice

People Love Snowflake, But…

tl;dr: Delivering sub-second analytics over large datasets can get pricey. When it comes to optimizing performance and cost, there are many great players. See how Snowflake compares to Firebolt, Clickhouse, Databricks, and more in the 2023 Cloud Data Warehouse Comparison Guide.

Promoted by Firebolt

Management Cloud

Executive Compensation

tl;dr: In this post, we cover some of the most important topics for late-stage companies putting together an executive compensation package, including: (1) How to think about the mix of base, bonus, and equity. (2) Best practices for benchmarking. (3) Selling a candidate on the value of equity. (4) What you’ll need to disclose about executive compensation. (5) How to make the offer.

Leadership Management

“If you’re afraid to change something it is clearly poorly designed.”

— Martin Fowler

Generating Code Without Generating Technical Debt?

— Reka Horvath

tl;dr: GPT and other large language models can produce huge volumes of code quickly. This allows for faster prototyping and iterative development, trying out multiple solutions. But it can also leave us with a bigger amount of mess code to maintain… This article explores several ways how to improve the code generated by these powerful tools and how to fit it into your project.

TechDebt ThoughtPiece AI

What Are Deployment Patterns?

— Dr Milan Milanović

tl;dr: The top three strategies for continuous deployment are: (1) Feature flags: toggle features on/off without deploying new code. (2) Blue / green deployments: run two environments simultaneously to test and switch traffic between them. (3) Use permission systems: grant access to a select group of users to test new features before releasing to all.

Management

Building Boba AI

— Farooq Ali

tl;dr: “We are building an experimental AI co-pilot for product strategy and generative ideation called “Boba”. Along the way, we’ve learned some useful lessons on how to build these kinds of applications, which we’ve formulated in terms of patterns. These patterns allow an application to help the user interact more effectively with a LLM, orchestrating prompts to gain better results, helping the user navigate a path of an intricate conversational flow, and integrating knowledge that the LLM doesn't have available.”

AI LLM

Joins 13 Ways

— Justin Jaffray

tl;dr: “Relational inner joins are really common in the world of databases, and one weird thing about them is that it seems like everyone has a different idea of what they are. In this post I’ve aggregated a bunch of different definitions, ways of thinking about them, and ways of implementing them that will hopefully be interesting. They’re not without redundancy, some of them are arguably the same, but I think they’re all interesting perspectives nonetheless.”

Database

Notable GitHub Repos

IGL: Cross-platform library that commands the GPU.

OpenChat: Less is more for open-source models.

QuestDB: OS time-series database for fast ingest and SQL queries.

Tinygrad: Deep learning framework.

How did you like this issue of Pointer?

1 = Didn't enjoy it all // 5 = Really enjoyed it

1 … 2 … 3 … 4 … 5