We Gave Our Browser Agent a 3MB Data Warehouse

#1

Post by shardullavekar » Thu, Jan 15, 2026, 4:11 PM UTC

We Gave Our Browser Agent a 3MB Data Warehouse
100x.bot

Re: We Gave Our Browser Agent a 3MB Data Warehouse

#2

Post by kbdiaz » Fri, Jan 16, 2026, 12:50 AM UTC

Yeah, this is the natural next step. We're currently combining LLMs and compute -- mostly in the form of giving agents tools, then terminal access and now most recently sandboxes. The most logical next step is to give them specialized compute engines and frameworks for their tasks.

I've been building SQL agents recently, and nothing is better than just giving it access to Trino.

Re: We Gave Our Browser Agent a 3MB Data Warehouse

#3

Post by charcircuit » Fri, Jan 16, 2026, 12:53 AM UTC

>You import heavy libraries just to do fuzzy string matching.

So instead of a heavy library for string matching you use a heavy library for a whole SQL engine which includes fuzzy string matching?

Re: We Gave Our Browser Agent a 3MB Data Warehouse

#4

Post by odo1242 » Fri, Jan 16, 2026, 11:40 PM UTC

charcircuit wrote: post #3

>You import heavy libraries just to do fuzzy string matching. So instead of a heavy library for string matching you use a heavy library for a whole SQL engine which includes fuzzy string matching?

Instead of having the AI write JavaScript code that does fuzzy string matching operations, they have it write SQL.

Re: We Gave Our Browser Agent a 3MB Data Warehouse

#5

Post by smithclay » Sat, Jan 17, 2026, 12:32 AM UTC

This is an emerging pattern that’s surprisingly powerful: thick clients that embed wasm query engines (pglite, duckdb) and do powerful analytics (with or without AI agents writing the queries).

Below are two examples using duckdb under the hood for similar purposes. Like the author, excited for this type of architecture making semi-advanced analytics more attainable if you’re not a data engineer.

- https://pandas-ai.com/ - https://marimo.io/