Kappa90

5 ms·

1.
by Kappa90 5mo ago
Thank you, if you happen to try it out or have ideas, feel free to contact me at: em at dinobase dot ai
2.
by Kappa90 5mo ago
Right now schema changes create new columns, I’m working on reconciling old columns, which right now are not dropped. The annotation/semantic layer agent creates a new description of the schema on sync, which represents the current sta
3.
by Kappa90 5mo ago
It's not explicitly stated in the benchmarks README, good catch. 80% of the benchmark questions are aggregations, 16% are multi-hop, 4% are lookups/subqueries. Multi-hop is where LLMs struggle the most (hallucinations, partial ans
4.
by Kappa90 5mo ago
DuckDB exposes Postgres SQL, and most coding LLMs have been trained on that. Of the small models I tested, Qwen 3.5 is the clear winner. Going to larger LLMs, Sonnet and Opus lead the charts.
5.
by Kappa90 5mo ago
It's an experimental benchmark, I couldn't find any off-the-shelf benchmarks to use this with. There's Spider 2.0 but it's for text-to-SQL. I'm planning to run this [1] next but it's quite expensive. There&#x27
6.
Show HN: I built a database for AI agents(github.com)
12 points by Kappa90 5mo ago|10 comments
7.
by Kappa90 7mo ago
A podcast on the history of the “Majority World” - that is Africa, Asia, the Middle East, Oceania, Latin America and the indigenous people around the world - from a decolonisation point of view. I just published the first episodes today and
8.
by Kappa90 10mo ago
Hey, one of the devs working on PostHog AI here. The main reason why we advocate for a single-loop is because subagents suffer from the telephone game problem: context is lost when the interface is just an agent-to-agent prompt. We are stil
9.
by Kappa90 2y ago
Used Copilot to write a user auth migration script. It silently reset 2FA settings for ~3k users with OAuth accounts because it didn't handle NULL vs empty string edge cases in our legacy DB schema. Classic "garbage in, garbage ou
10.
by Kappa90 2y ago
If you want to see an LLM tackling unit testing, check https://www.codebeaver.ai Disclaimer: I’m the developer behind CodeBeaver
11.
by Kappa90 2y ago
There seems to be a peak in traffic I guess due to HN, apologies! If you have any further problems it would be really helpful if you'd let me know at em@usetailor.com
12.
by Kappa90 2y ago
If you're looking for a replacement for Artifact (and for news apps in general), you should check out https://tailor.news It aggregates the news from thousands of newspapers, youtube channels and podcasts, then serves you a
13.
by Kappa90 3y ago
An app that aggregates the news from websites, blogs, YouTube channels and podcasts, and generate easily digestible summaries, along with a small podcast version so you can stay informed in an easy stress-free way. Right now I’m working on
14.
by Kappa90 5y ago
Indeed, there is no other tool out there automatically linking documents for you, as far as we know. The idea is that Saga helps you resurface connections and ideas without having to do the manual work yourself. This allows for a more seren
15.
by Kappa90 5y ago
Checkout Saga [0], we've been building this for over a year exactly to fix the "I can't find anything" problem. Any feedback appreciated! [0] https://saga.so
16.
by Kappa90 5y ago
We're developing Saga [0] and we have many users using it for Zettelkasten, although it's not our main value proposition. Why: Saga automatically links pages one to another, so it builds your wiki/knowledge system automatical
17.
by Kappa90 5y ago
We did look into editor.js and the API seems not as advanced as Slate's, mostly because Slate tries to replicate the DOM's behavior with Nodes and Ranges, so you can make edits in a very fine grained way. If you're going for
18.
by Kappa90 5y ago
I know it sounds like tech debt...and it is. But we're to a point where we could potentially be able to integrate those extensions back into the core library and our slate-saga, so it's eventually paying back. We really needed ext
19.
by Kappa90 5y ago
Have been building https://saga.so with Slate for the last 12 months. Can confirm it was a bumpy ride, but it also gave back a lot. The main reasons for using Slate over Prosemirror were the almost complete API for managing and