spennant

9 ms·

1.
by spennant 6d ago
This is my experience as well. It gets worse when the LLMs have strong priors in subject matter adjacent to your novel work. The generated output will incorrectly pull towards things the LLM has seen in it’s training data.
2.
by spennant 2mo ago
I'm working on leveraging NLP and LLM techniques to create a geometry over the discrete space of Ethereum transaction execution structure. (sorry... it's a bit of a mouthful) https://www.chaingenius.ai The goal is to f
3.
Show HN: Explore Ethereum by the Shape of a Transaction(chaingenius.ai)
3 points by spennant 3mo ago|0 comments
4.
by spennant 4mo ago
Is this still true?
5.
by spennant 7mo ago
Fine tuning Gemini 2.5 flash on EVM Blockchain execution structures and using perplexity with PageRank to perform transaction anomaly detection. https://lite.chaingenius.ai/theory [SMILES]( https://en.wikipedia.or
6.
by spennant 1y ago
Odd seeing this right now for me. I recently implemented a 6NF schema for parsed XBRL files from EDGAR. The architecture was the right call... too bad the data is not useful for analytics.
7.
by spennant 1y ago
What? No love for Perl scripts in cgi-bin/ ?
8.
by spennant 1y ago
I moved from pip to poetry a while back, but was introduced to uv recently while working on MCP stuff. I now use uv for everything and haven’t looked back.
9.
by spennant 1y ago
It's coming...
10.
by spennant 1y ago
Just a fancy way of saying that multiple agents (with their own LLMs) can concurrently connect, also that pg-mcp can simultaneously connect to multiple Postgres servers as well.
11.
by spennant 1y ago
The use case for pg-mcp is server deployment - local running is just for dev purposes. HTTP/SSE enables multiple concurrent connections and network access, which stdio can't provide.
12.
by spennant 1y ago
Xata.io looks very interesting!!! I was thinking about building an intelligent agent for pg-mcp as my net project but it looks like you did a lot of the hard work already. When thinking about the "AI Stack" I usually separate con
13.
by spennant 1y ago
Indeed. Anything you do with MCP can be done in more traditional ways.
14.
by spennant 1y ago
I built this to fill a specific need. It's not really made to talk to Claude Desktop (although it can). It's built to take multiple connections from agents who bring their own LLM (via API key or whatever) and provide context to t
15.
by spennant 1y ago
I made a few assumptions about the actual deployer and their environment that I shouldn’t have… I’ll need to address this. Thanks!
16.
by spennant 1y ago
The pg-mcp server doesn’t need an API key. Those variables are only needed to run the example claude-client. You can write your own agent to talk to whatever LLM you like.
17.
by spennant 1y ago
It won’t be that easy. First off, most databases in the wild are not well documented. LLMs benefit from context, and if your tables/columns have non-intuitive or non-descriptive names, the SQL may not even work. Second, you might benef
18.
Show HN: Cloud-Ready Postgres MCP Server(github.com)
167 points by spennant 1y ago|79 comments
19.
by spennant 2y ago
I feel the same way. https://news.ycombinator.com/item?id=30094245
20.
by spennant 2y ago
Money not a problem? Get a "pro"/digital sign display and connect your own sound system as well as tuner/stb. https://www.lg.com/us/business/digital-signage/lg-65ep5g https://ww
21.
by spennant 2y ago
Personal story: I sat in his office as an undergrad and he presented a problem to me concerning counting balls in boxes that had applications to quantum theory. It blew my freshman mind.
22.
by spennant 2y ago
Sure. But two things: 1) At the time, the European data laws implied that it protected its citizens no matter where they are. Nobody wanted to be the first to test that in court. 2) The organizations and agencies performing this type of dat
23.
by spennant 2y ago
Agreed. The media and advertising industry was most definitely leveraging cookie-level data for building attribution and targeting models. As soon as the EU established that this data was “personal data”, as it could, theoretically, be tied
24.
by spennant 2y ago
In other words... being Iceman is better than being Maverick.
25.
by spennant 3y ago
In October, NASA plans to launch the Europa Clipper on a mission to study Jupiter’s fourth-largest moon, one of the most promising places to search for extraterrestrial life in the solar system
26.
The search for extraterrestrial life is targeting Jupiter's icy moon Europa(technologyreview.com)
2 points by spennant 3y ago|1 comments
27.
by spennant 3y ago
"Few"? I guess I must be one of the olds.
28.
by spennant 3y ago
+1 Insightful
29.
by spennant 3y ago
I just watched this video last night and thought to myself, “seems plausible”… then I bought a calendar from their store.
30.
by spennant 3y ago
LOL - I started a side project (never finished) that leveraged Neo4J and GraphQL and attempted to mimic the GedcomX spec.

More ›