efromvt's comments

On: Learn SQL Once, Use It for 30 Years

[efromvt]

There's always a bulk insert, but I wouldn't say every engine has always had a reasonable way to bulk load truly large data... parquet really helped with interop but before that when your best option was a CSV and bcp life was not fun.

On: Learn SQL Once, Use It for 30 Years

[efromvt]

A pretty common request is to lift the FROM up before the select, like the below. I'm pretty fine with status quo since my mind is usually "hmm what do I need to get" first, then I figure out how to get it, but some engines (duckdb, I think?) support both so everyone gets their cake.

What people often want: <where to get data from> <what I want from it> <how it's filtered> <how it's grouped> <how it's filtered post group>

[anonymars → efromvt]

Also helps with autocomplete, which is why LINQ starts with `from`

[deepsun → anonymars]

Linq took that from Hibernate in believe. But in general, I'd say almost all ORM and SQL-gen libraries have FROM first, because it's clearer for programmers.

[anonymars → deepsun]

Hibernate puts SELECT last if requesting only specific properties?

LINQ:

  var query = 
    from e in entities
    where e.property1 == value
    select e.property2;

On: Learn SQL Once, Use It for 30 Years

[efromvt]

A self para-sql-lang plug if you like the direction of malloy but prefer something closer to native SQL syntax. (a controversial take at times)

- https://trilogydata.dev/

On: Learn SQL Once, Use It for 30 Years

[efromvt]

I think advanced SQL authoring is generally simple to understand, and that's the larger learning curve!

I find those big stored procedures usually fall into two categories; logic that should be in the DB, but should be decomposed (staging tables, other SPs, etc) in which case they can be understandable in chunks; or logic that shouldn't be in the DB but has been shoved in there, in which case there's more of an ideological debate but I generally prefer to pull out and run in the application layer. (the latter is pretty much IMO the things that you've done after you've gotten the data at the right grain, when you are massaging it to a particular form/presentation format; performance is often the final arbitre here though).

On: Notes from the Mistral AI Now Summit

[efromvt]

Interesting - which models specifically? I'd be interested in using mistral over deepseek if it was competitive (guess I need to go benchmark)

[barrell → efromvt]

I use small, large, an medium-3.5 depending on the task

On: Incident with Actions and Pages

[efromvt]

Incredible how reliable the heuristic of "something seems off - probably github being down" has gotten these days

[comboy → efromvt]

It's big enough that every time it goes down, it surely stops somebody from pushing fix for what they currently have broken, so I wonder if status page services see some kind of ripple from github outages.

[JsonDemWitOster → efromvt]

About an hour ago I was having trouble browsing repo files in the browser and I thought "A disturbance in the force, is Github down?" Refreshed HN and loaded up their status site. Nada.

(Ofc, in a sensible universe, we just brush that off to a JS/Firefox glitch or my ISP.)

And yet, here I am. My code is not compiling, my AI isn't vibing, nonetheless I can't work! Two more hours before I can get off!

On: Ask HN: What are you working on? (May 2026)

[efromvt]

For my streamlined, composable SQL variant - Trilogy[1] - spent the last month or two very focused on CLI/ETL/scheduling to support data pipelines for other hobby projects. Have that story in a better place - so time for new things.

Trilogy's model works quite well for agents, but I've avoided making AI features too native in the UI products - not everyone's cup of tea - so this month going to do a spike on a new pure AI native data authoring/exploration experience to see where that can go without messing with the core product, at both CLI/UI layers. Data consumption is (un?)fortunately a pretty perfect fit for the agent strengths with the right harness.

[1]https://trilogydata.dev/

On: Claude Code refuses requests or charges extra if your commits mention "OpenClaw"

[efromvt]

I don't the OpenClaw furor has been a problem for the majority; but stuff like the harness bugs with dropped thinking traces (capacity optimization?) and some fairly bizarre billing bugs with weird/opaque comms around both have been more concerning and affect a larger group than that loud minority. You do kind of want a reliable service with reliable billing and reasonable comms for most things at the corporate level.

[AbstractH24 → efromvt]

Particularly for CC I agree that’s getting increasingly infuriating.

I’m not sure where to turn next. I guess cursor?

On: Claude Code refuses requests or charges extra if your commits mention "OpenClaw"

[efromvt]

that is a terrible way to figure out if openclaw is used, hah

On: Reverse Engineering SimTower

[efromvt]

Can you still build a garage in the bottom corner to get more starting money??