any · textlog

wrote:

I naively went into a work problem thinking that RAG was solved by the service offered at work, and wow was I wrong. Been a fun week so far experimenting with many different methods before doing the academic research, and it's been nice to be validated by outside research.

wrote:

úgyhogy az idénre beígért informatikai projekjeimet már eleve vaporware-nek nyilvánítom a késedelmes indulás miatt

so I'm declaring my promised IT projects for this year as vaporware from the start due to the delayed launch

wrote:

A thing I've been doing lately is trying to come to the same conclusions others have without looking their the direct techniques and code. Last night I tried to figure out why people do vector based retrieval for RAG and ended up running these experiments: n0tls.com/…/2026-09-03-i-think-…rag-with-grep.html I'm still going to do an explore agent.

continued:

Here's a follow up from last night: n0tls.com/…/2026-09-05-fine-tuning-…tiny-retriever.html This has been a great way for me to learn in this space. Learning with AI has been such a fun time for me.

wrote:

There are many distractions. Much is way too verbose. A calm, quiet and “distraction less” experience suits me better. I am not sure my thoughts will fit in 280 characters, but let’s try and see how far that goes.

wrote:

Frustrated with my phone usage again tonight, I uninstalled several pointless and time-consuming apps. It’s a cycle I repeat every few months. During this cycle, I wasted a lot of time on one of those puzzle-matching games. I'm considering switching to Ubuntu Touch at this point

wrote:

"My feed" shows two unread messages, but I can't find them, and browsing away from it, and coming back doesn't clear that count. Perhaps sorting (unread at the top) when there unread messages is a good idea ?

replied to:

Same for me, it's been that way for a few days.

wrote:

I like this, it feels like Reddit before what it is now, but I feel like it won’t be long until the internet destroys this place. Until then I’m happy to be here.

replied to:

It's really sad that this is such a common sentiment whenever there's a nice online community nowadays. Guess most of us have been burnt enough times by now

replied to:

lets enjoy this while it lasts i guess

…

replied to:

My favorite part of classic reddit was the use of elaborate hacks to circumvent the restrictions on custom subreddit CSS. E.g. animated gifs weren't allowed, so people made filmstrip sprites with stepwise animation in such a way as to look exactly the same.

wrote:

Lots of slop PRs at $WORK. Trying Claude Tag, works well but token-billed, daily usage over 100€ for three sessions.

replied to:

ouch, that's a month for virtually unlimited Codex. why so expensive?

replied to:

Anthropic uses token billing for Claude Tag. But it’s different from CC - it’s an AI Colleague, so I’m asking PM and EM to talk to it instead of to me - accountability switches.

replied to:

so it's expensive and you're left with dealing with the PRs others generate for you?

…

replied to:

No, Claude Tag’s PRs are my team’s. Slop PRs come from outside of the team. Anyways, dozens PRs per day is draining energy. Don’t want to review anything anymore.

wrote:

Bandcamp Friday is almost over! Today, 100% of the proceeds go to your favorite artists ❤️ Go buy some music and take back the power from shitty streaming services

wrote:

brings so many issues. It really accelerates. Not only the good parts but also the bad parts. Combine this with a product person that doesn't even know what a refinement is. I'm 41, maybe I'll study to pivot, maybe I'll get into trading. I can't handle it anymore.

……

replied to:

what is a refinement? honest question - I'm restructuring our refinements now. Like having group of people discussing 5 items in an hour gives little chances any one of them will be addressed properly, unless they are trivial.

replied to:

Architect/lead have a pre-refinement with product to clear 80% of the obvious questions. We share an agenda, devs can ask questions async. In the refinement product presents, we design a solution draft. 15 minutes/item. This doesn't work if product is incompetent and everything is AI-slopped.

…

replied to:

Thank you, like!👍

…

wrote:

What could possibly go wrong?

replied to:

ooh maybe I revive and add this feature but instead of being able to modify everything it would be limited to adding features to the game/world. And it’s the perfect domain, a site that’s been ~~entirely~~ mostly hallucinated.

replied to:

Oh wow that site is amazing lol not seen it before

…

replied to:

ty! it’s been 3–4 months since I made it so it’s relatively new.

wrote:

Found some coffee at TJ Maxx (american discount store) -- not a big discount but still fun.

wrote:

I finished Rejection by Tony Tulathimutte and now I'm struggling to take any other book seriously. I need a pallet cleanser or something before I can move on

wrote:

most akkor mit kezdjek ezzel a Windows 10, Windows 11, Ubuntu Linux 3-as operációs rendszerrel egy két felé particinált winchesteren, jobb lesz ha előbb megvárom, ameddig törli magát a Windows 10, nem azt mondták, hogy lassítani kell az MI projekteket, annyira tetű módon csinálom

Now what should I do with this Windows 10, Windows 11, Ubuntu Linux 3 operating system on a hard drive partitioned in two? It would be better if I wait until Windows 10 deletes itself first. They didn't say that AI projects should be slowed down, I'm doing it in such a lousy way.

wrote:

while we wait for this to be true
we could add our bits together 
remember to not sleep forever
return when your dreams are through

Start with cringy poem, end with cringy pseudocode. At least it comes naturally.

wrote:

These days Ive been learning how to deal with guilt and feeling gulti about things I didn't cause. NGL is hard man, how do you start cntl-z your brains on that

wrote:

Hello World!

replied to:

hi there! 👋 welcome!

replied to:

Hi 👋

Any conversation