captaintobs

i spent 8k and get close to a 2-3x slower sonnet. running 2x spark deep seek v4 flash

18xx, hardcore economic train games. Think war games with money.

They're using it for data transformation.They're long time dbt users, but are switching to SQLMesh because it's extremely efficient, provides a better development experience, and can help them become warehouse agnostic.

There are many teams using SQLMesh in production. Fivetran, Harness, Hopper, Pitchbook to name a few.

You can read some case studies here https://tobikodata.com/harness.html or join Slack to meet with folks to learn more about their experiences.

Thanks! Yes, it's a much requested feature but it's difficult to get right!

Why is this faster than the stdlib? What does it do to achieve better performance?

Very cool work!

It's mostly affecting the amount of AI content I have to wade through these news aggregators...

There's a lot of hype with ruff, but I've been doing fine with black and autoflake. I have a pretty sizeable project and have never thought to myself it's problematic because it's slow.

We used both svb and first republic!

This is a garbage/sensationalist/click bait article. They source blind conversations.

Monorepos at scale only really works if you have the tooling and infra to support it. Otherwise, you're going to be miserable with slow builds, pushes, pulls, impossible merge conflicts.

It's fun having the ritual of grinding coffee. Maybe I can't tell the difference but I enjoy it!

This and Astral (ruff) are similar. “hot” python frameworks, but unclear how they’ll monetize.

It's surprising how these rash decisions are made with seemingly little thought. You'd think someone so successful would be logical... I guess it's optional.

The real challenge is can a GPT deliver a product to a PM / Executive who doesn't know what they want??

Go and rust aren't going to help you as data engineer. Data Engineering is dominated by python, sql, java, and scala.

Writing a pipeline in those languages isn't particularly great unless you have very specific reasons for it because there's no ecosystem for them.

In general removing the gil will make things slower because you need to make sure things are safe. If most code out there is single threaded, removing the gil may not really have a benefit.

Why is this linking to a random markdown page?

Ahh the big company reorg. Executives need to somehow look busy, gaining alignment etc etc...

It's really depressing to see the world crumble like this. It's not just twitter, but look at what's happening in Florida.

Although I have a degree in mathematics, I was never very good at deriving proofs.

Google had every opportunity to capitalize on the architectural innovation and failed tremendously.

I've always preferred using notebooks (Jupyter Labs) for data. It preserves history and I can bounce around and execute arbitrarily complex scripts.

It's such a funny story that Blockbuster had a chance to buy Netflix for $50 million.

How does this compare to Flink and Spark? Is it performant for batch processing as well?

It's all a facade.

Nice and simple! I had a similar design on the iPhone back in the day where you could play head to head on the same phone.

Great work!

It's just not gonna happen unless the government does something.

i was looking to do yc but decided against it because of the remote aspect