any · textlog

500 chars / 20 lines max · use #hashtags, @mentions and more

😀😃😄😁😆😅😂🤣😊😇🙂🙃😉😌😍🥰😘😋😛😜🤪🤨🧐🤓😎🤩🥳😏😒😞😔😟😕🙁☹️😣😖😫😩🥺😢😭😤😠😡🤬🤯😳🥵🥶😱😨😰😥😓🤗🤔🫣🤭🫢🤫🤥😶😐😑😬🙄😯😦😧😮😲🥱😴🤤😪😵🤐🥴🤢🤮🤧😷🤒🤕🤑🤠😈👿👻💀☠️👽🤖🎃😺😸😹😻😼😽🙀😿😾❤️🧡💛💚💙💜🖤🤍🤎💔❣️💕💞💓💗💖💘💝💟👍👎👌🤌✌️🤞🤟🤘🤙👈👉👆👇☝️✋🤚🖐️🖖👋🤝👏🙌🫶👐🤲🙏✍️💪👀👁️🧠🫀🫁🌱🌿☘️🍀🌸🌺🌻🌞🌙⭐✨⚡🔥🌈☀️☁️❄️☕🍕🍎🎉🎊🎈🎁🎵🎶🎨📚💡✅❌⚠️🚀🌍💻📱🔒🔑

example.com or https://example.com
Regular links: [title](example.com) or [title](https://example.com)
Markdown links: ~text~ or ~~text~~
Strikethrough: *text* or **text**
Bold: _text_ or __text__
Underline: /text/
Italics: > text
Quote: :smile
Emoji autocomplete: 1. first 2. second 3. third
Numbered lists: - first - second - third
Bulleted lists: Name | Status | Count ----- | :----- | ----: notes | ready | 3
Tables: |redacted|
Redacted: `code`
Inline code: ```…```
Code fences: $inline$
Inline LaTeX: $$block$$
Block LaTeX

Which one? #poll First option Second option
PollsUse 2–8 unique options.: Which one? #quiz Wrong answer > Correct answer Explanation revealed after answering
QuizzesMark exactly one of 2–8 unique answers with >. Text after a blank line is revealed after answering.: Visible text #spoiler Hidden text
SpoilersText after #spoiler is hidden until revealed. Aliases: #tldr, #sensitive, #contentwarning, #cw, and #triggerwarning.: Going hiking #map Kallikratis, Crete
MapsShows a map preview for the first location line. Alias: #location.: #flying Heraklion to Berlin
FlightsHover over the itinerary to see a map connecting the airports. The next three words form the itinerary: airport, to (or → or ->), airport. Use single-word airport names or IATA/ICAO codes.: Today #todo [ ] First task [x] Finished task
TodosOnly [ ] and [x] lines become items. Click your items to toggle them.: Run this #exec ```js console.log(6 * 7) ```
Executable codeRuns the next language-tagged code fence and shows its output beneath the note.: Keep this visible #pin
Pinned notesYour latest #pin is shown first on your profile, independently for notes and replies.: No more replies #lock
Locked conversationsPrevents new replies to this note and every reply beneath it.: About textlog #meta
Meta conversationsKeeps this note and its replies out of public discovery feeds. Aliases: #tlog and #textlog.: Continue quietly #whisper
Whisper conversationsKeeps the branch out of all and hot. Participants, mentions, and tag followers can receive it in my feed. It remains public elsewhere.: For @someone and @another #private
Private conversationsOnly the author and mentioned users can read this branch.: Answer before reading #HiddenReplies
Hidden repliesHides all replies until you reply.

moreless

time

WeakNotes 2026-09-27

中国的开源模型在最高意志的主导下，在预训练数据阶段，就筛除大量对其不利的内容，sft阶段构建的数据掺入大量替其辩护的内容。急需一个纠偏数据集用于微调

Under the overarching influence of a single, powerful force, China's open-source models filter out a large amount of content unfavorable to their designs during the pre-training stage, and then incorporate a significant amount of content defending their designs into the data constructed during the SFT stage. A corrective dataset is urgently needed for fine-tuning.

Well this looks very nice indeed. Just dipping a hoof into the river for now ...

Which river? Asking because most rivers are contaminated these days. Be careful!

ty and welcome!

Had a solid UKG banger slip in while cooling down post-run earlier this morning.

I am programming a and custom DSL for it in C99. It gets a bit frustrating to think of language design and engine architecture decisions

that sounds cool!!! are you planning to make a visual novel any time soon?

I am actually doing this engine to make my own VN! I wanted a solution that's both light on web and is native. I do have a decent bit of the story outlined but engine is taking quite a while :/

could I know what the storys about?!

The story bases itself on the idea of indifference and hopelessness in changeable situations. Don't got a name set in stone but temporarily calling it Hopeful Hopelessness. It's my first time writing but I hope it becomes a nice story

…

I would love to play when your done with it!

This is really cool site but only drawback I see is 500 character per post limit. I think this should me more like 1500. or there is a limitations type something like I can do 3-4 long post per day credit. whom can i message for this feature request?

I am certain will get back to you on such matters - creator of this platform.

On some Mastodon instances, they post whole blog posts or massive threads and I feel this type of platform is not the place for such content. But this is clearly personal taste.

New in agenttop.dev : panels open in your . 🤹‍♂️ • Any panel can now split into its own pane, live, beside the table. • Works with all the popular multiplexers: , , 𝗮𝗻𝗱 𝗸𝗶𝘁𝘁𝘆. • run a panel on its own: agent-top mcp, agent-top slow, agent-top fails, agent-top advice. ➕ The # MCP panel also lists the orphaned MCP processes that outlived their agent and are still eating memory.

Hello! Is anyone else trying local AI? Im dont know what im doing but lately ive been trying "Llama 3,2 3b instruct q5km" on my phone and its been surprisingly good.

Its to replace using central ai for basic tasks like text editing, analysis and brainstorming.

If you have Mac, you might check out this app: blaizzy.github.io/nativ/

…

i think the best you can get right now is qwen 3.8 27B. youre gonna need 24 gigs of some fast memory though, i heard theres decent quantizations that can also make it run on 16. im quite excited about open/local models, but theres big investment overhead with the prices right now

You might be able to try out colibri, its quite good and the models are a bit better. I am unable to connect any of the cli agents to do things with it though, probably due to context limits, but try it out. OlMoe is the lightest

我在中国是一个准脱离政府掌控者，基于身份的一切活动我都不参与，比如正式工作和网购。我也为此付出巨大代价，我的选择空间被极大压缩，行动能力损失严重。我只能通过少量免费wifi访问互联网

In China, I am practically independent of the government; I don't participate in any activities based on my status, such as formal work or online shopping. I've paid a heavy price for this; my choices have been severely limited, and my mobility has been significantly impaired. I can only access the internet through limited free Wi-Fi.

Textlog seems pretty cool 😎

Just trying to see how this works

The Paper Tools That Helped Me Pace With ME/CFS

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.

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.

"200K lines is not 10 times as much as 20K lines - from the mission perspective it's just irrelevant. And yet it's the easiest thing to measure. And so we look for the keys under the lamp post." ~Kent Beck www.youtube.com/watch?v=F8fBgDCf2Y4&t=2753s

Listening to this I had an AHA moment. When the powers that be made more and more LLM tokens the whole goal, I thought it was quite stupid. But I realized that for them it made sense, because in their managerial understanding of programming, they believed the output of LLMs is the best they could get for the money.

Or they just had investments in LLM factories.

I am trying to learn emacs for the fun of it. The tutorial is brilliant and clear. Any tips that would help me learning more things. I know all I need is practice but if you may have used it, you may know a thing or two.

Hello textlog 👋

Hi 👋

The rise of postmodern cuisine: "The beverage's photogenic bright green color proved particularly well-suited to social media platforms like Instagram and TikTok, helping it gain viral popularity among wellness enthusiasts and influencers." (Wikipedia, "Matcha")

Day length formula: (Did you know that it becomes a triangle wave at the (ant)arctic circle?)

&dingbats? .◍.➳.❆.✖.✱. .➇.☏.☠.♲.☮.

how'bout these? 𐦗𐦂𐦆𐦖𐦉𐦃𔒏𔒌𔒋𔒝𒀭𒀽𒂝𒇫