The only way to make sense of it is to lossy-compress focus on getting the high-level ideas right rather than every detail.
21.01.2026 09:20
👍 2
🔁 0
💬 1
📌 0
The only way to make sense of it is to lossy-compress focus on getting the high-level ideas right rather than every detail.
I have mixed feelings about AI.
It’s a bit like tailored clothes vs machine-made ones both work, but the vibe is different. Using Claude isn’t wrong at all, but when 50k+ lines are written that way, it becomes hard for a human brain to truly digest it at that pace.
1% ; 3B tokens spent
Allow multiple chat window, I will migrate from cursor. I understand it could be complicated when two agent write but at least ready only mode helps so while one tab is doing work you can brainstorm and plan for next thing.
Models suddenly getting good enough to write most of my code - which I am now prompting - creates complicated feelings.
It took a long time to get good at coding. And it's not easy. Plus, there was something special about being in "the zone."
Full: blog.pragmaticengineer.com/the-grief-wh...
ai-dynamo / dynamo: A Datacenter Scale Distributed Inference Serving Framework ★4619 https://github.com/ai-dynamo/dynamo
raphamorim / rio: A hardware-accelerated GPU terminal emulator focusing to run in desktops and browsers. ★5457 https://github.com/raphamorim/rio
lumina-ai-inc / chunkr: Vision infrastructure to turn complex documents into RAG/LLM-ready data ★2471 https://github.com/lumina-ai-inc/chunkr
Screenshot of embedding atlas showing the embedding view on the left, a table at the bottom and charts on the right.
🚀 We've just open-sourced Embedding Atlas – a tool for exploring large embedding spaces through rich, interactive visualizations 📊.
rivet-gg / rivet: 🔩 The open-source serverless platform. Easily deploy & scale AI agents, complex workloads, and backends — all on a frictionless platform that runs anywhere. ★3997 https://github.com/rivet-gg/rivet
What about server side cursor to fetch subsequent rows vs keyset/cursor pagination? www.postgresql.org/docs/current... this should be faster at cost of memory
Why not alias git=jj ?
What are your thoughts on XMPP vs websocket for chat use case ?
India is biggest victim of terrorism. Pakistan funds/supports terrorism directly / in-directly.
It should've been 7-3, Referee + Big miss. One needs to be 10x better than Madrid to win league.
Pakistan based terrorists attacked once 9/11 at USA, India for past 10yr more than 10 incidents has happened. They just don't f ing stop and actively take action against terrorism. How long one would stay silent, keep up with shit ?
😥
Can it be configured on premise instance? With anthropic & gemni api keys ?
From Test driven development to Taste driven development. #vibecoding
Working on problems close to this. Here are the ref that I found helpful.
Integrity
1. sirupsen.com/napkin/probl...
- github.com/datafold/dat...
- www.usenix.org/conference/s...
Replication
- github.com/PeerDB-io/pe...
- PG Man page subscription and publication
- PG Source code & Dbz
#scylladb #monsterscale
Logo so cute that you makes you wanna write the code.
Actual numbers on what it took a seed-stage startup to hire their first two founding backend engineers (full-remote)
It's good to know as a candidate: this is the type of competition you might have. And as a founder: this is how much time investment it can take to make a hire
backend.how/posts/1b-pay...
First draft on BAD: Billions a day payments.
@postgresql.eu
Too expensive
Learning form @jeffdean.bsky.social and Noam with @dwarkesh.bsky.social
When you scale @valkeyio.bsky.social sorted sets to almost infinity you get @amazon.dynamodb.me
