Interesting signal is mentioned 3 times in #TiDBSCaiLE: 1.TiDB listed 50× faster table creation as one of its most important improvements. 2.@Atlassian wants a horizontally scalable DDL pipeline. 3.@ngaut1 said he doesn’t know who would need to create over 500k tables in a DB but

Try to steal from Alex Webb.

Only in SF I can see so many startups’ signboards

hotpot night in @mdm822

In #SupabaseHackathon

Joining #SupabaseSelect

Tonbo’s new release is on the way! Highlight: Query and analyze S3 data directly within Node/Deno, with zero local disk dependency. Means-> closer-to-edge computation and serverless analytics. Calling on developers to join our testing. You may also receive a little tonbo gift.

To get you ready for our @P99CONF 2025 event on October 22-23, here is a preview of some of our highly-anticipated #Rustlang talks featuring speakers like @glcst, @astuyve, @problame, @yochowgwo, and @GeeWengel > https://t.co/EYAq2yVKAe #ScyllaDB #P99CONF

A new potential interview step: no handwritten code allowed — all code must be generated directly by Claude Code.

I missed when it was on the Hacker News front page

Just released: https://t.co/s7sNzw8ag8. When working with arrow-rs, we noticed that schemas are declared at runtime. typed-arrow provides pure compile‑time Arrow schema declaration: ✅ No runtime errors ✅ No runtime overhead Heavily using Rust macros and abstract type system.

World: exists hundreds of vector databases. Claude Code: uses JSON to persist historical conversations.

After three weeks of intense vibe coding, I found that it still need to rely on human prior knowledge to do manual context management: Proper manual execution of /compact, /clear and /init is very important for the quality of code generation

Our neighbor alumni seems like a big fan of Teenager Engineering

Since Tokio is built around epoll it limits support for io_uring since these I/Os use incompatible trait bounds. @Tonboio's @yochowgwo will share the simple way to switch between different I/O runtimes at compile time at @P99CONF. https://t.co/PLYga8vHYL #ScyllaDB #P99CONF

I built a coding agent that serves as a general-purpose agent, dynamically accessing APIs like Linear and GitHub via GraphQL. Instead of relying on predefined scripts or tools, it performs schema introspection on-the-fly to understand API semantics and generates GraphQL queries

I built a Python GraphQL framework before, and I've always found it's API semantics frustrating after that, concluding that its performance are fundamentally flawed. Surprisingly, its composability and schema introspection make it ideal for AI agents executing ad-hoc queries.

Using Claude Code wrote another simple coding agent based on extended thinking mode Claude, as a slack bot to help us on task management across various tool: Linear, Slack, GitHub, Notion, etc.