any · textlog

Trying to jump back into project work while my kids are still home before school starts on Wednesday 🫠

This bout of COVID is more like “Corona Light” compared to previous years, but I’m still spending most of it in bed. Stay safe out there.

wrote:

I wanna learn how to use source film maker...

wrote:

ai coding feels like it has reached diminishing returns for me. luna is my default orchestrator, routing work by complexity. in practice, usage lands around 16% luna, 48% sol, 30% astra, and 6% terra. at this point, i care less about each new frontier release. the system already finds a capable model for the job. the frontier is moving beyond my day-to-day work, so i’m happy to settle into predictable models for a while.

replied to:

I’ve found the same, albeit with Opus — it’s hard to find uses for Fable. I don’t need a model that works for longer or works on harder problems, I need one that consistently produces simple, easy to understand, code; which at this point none of them do.

replied to:

I've been there since I settled on a combination of GLM, DeepSeek and MiMo for planning, implementation, and verification flows. I'm a lot more interested at this point in having cheaper models that can consistently do the work I need. It's been a while that I hanven't noticed any improvement when I try new frontier models. I only notice they are very expensive.

…

replied to:

same here, for almost an year. At this point, I could easily live with models that were released 6-12 months ago and continue doing so. also possibly the reason why the current focus has been around agentic workflows or agentic anything, than the model capabilities.

replied to:

Coding probably yeah, but still they are bad at generating novel information, new ideas and such. They’re also not very good at seeing the bigger picture, they will narrowly perform the task but they’re not taking into account future revisions, which ends up in refactor after refactor in order to move forward.

wrote:

I suck at math so bad ；⁠∀⁠；

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:

I've written a small persistent agent wrapper for agent.py. I'm impressed with how well it works. I feel like a CEO issuing global memos to all my agents after reading their status reports.

continued:

After running them for a weekend and spending $2 on openrouter I've put the wrapper script at public.swiley.net/agents.sh

wrote:

"This text may violate our content rules (harassment/threatening: 0.4593). Please revise it and try again." HOW there's no cuss words dawg let me typeeeeeee

wrote:

Many leaders acquire authority through ambition and manipulation. I’m not referring to them. But a test of leadership for those who acquire authority organically, with the will of those being led, is to maintain the zeal and principles which put you into that position, after you encounter the machinery designed to exhaust, exasperate, deter, and demoralize.

wrote:

I’m using textlog to explore the intersection of and tech, with the goal of building an app that solves a problem in this space. Named , it's word represented as a hex color. Of all the words there, it is a cocoa-ish color and is an appropriate name for this idea.

wrote:

Unliking an little-known song on Spotify and ceremoniously saying final goodbyes, knowing you'll likely never hear it again in your life

…

wrote:

Is this true? Sometimes, AI allows for socratic thinking, this is, the conversation with the machine allows the thinker to go further as the AI thorws ideas to the conversation. Many times, the AI reasoning is not good enough, but like a friend that says a funny think that sparks an idea, the AI can help our brain to see other possible paths. The AI doesn't need to be right to be helpful

…

replied to:

that's only true if a) that "rubber ducky" entity is not leading the line of thinking down the wrong path, and b) the human thinker is actually doing the critical thinking, and not the machine. The latter is where my criticism of unbridled AI use is in this case, wherein the human becomes the rubber ducky instead of the other way around.

replied to:

Indeed. The AI throws back ideas, but it is the human who must do the heavy lifting. In this regard, the AI is a sort of mirror, it reflects what you send it, showing another side the object. But if the human messages are bad, the AI will respond with bad quality answers.

wrote:

So I found the blog of a Ukrainian Front Line soldier on Bear Blog today. Incredibly grim stuff to read, but I really wanna get in touch with that guy now and send him so cash so he can buy a burger. Hope Herman forwards my message 😪

wrote:

𐑔𐑮𐑰 𐑓 𐑔𐑮𐑰 𐑪𐑯 ·𐑚𐑮𐑦𐑑𐑦𐑖 𐑛𐑦𐑖𐑩𐑟 𐑞𐑦𐑕 𐑢𐑰𐑒𐑧𐑯𐑛: 𐑣𐑨𐑛 𐑩 𐑓𐑫𐑤 ·𐑦𐑙𐑜𐑤𐑦𐑖 𐑯 𐑩 𐑐𐑲 (𐑮𐑨𐑚𐑦𐑑) 𐑘𐑨𐑕𐑑𐑼𐑛𐑱, 𐑯𐑬 𐑑 𐑛𐑵 𐑨 𐑚𐑰𐑓 𐑮𐑴𐑕𐑑. 𐑤𐑳𐑝𐑤𐑰 𐑯 𐑒𐑷𐑟𐑰 𐑦𐑯 𐑞𐑦𐑕 𐑷𐑑𐑳𐑥𐑯𐑩𐑤 𐑢𐑨𐑞𐑼. Three for three on British dishes this weekend: had a full English and a pie (rabbit) yesterday, now to do a beef roast. Lovely and cozy in this autumnal weather.

wrote:

Spending mornings scouring job boards and listings and applying. Afternoons working on Cheltenham Open Data and iOS app development. What about you?

replied to:

waiting for my friend so she can put my hair into a bun

wrote:

Yes 🙌 the authoring box first thing. Before I get distracted. Heck, make writing something down how I get distracted from reading others' posts.

wrote:

Any shows recommendations? I have no idea what is worth watching (since I rarely watch shows) so now I’m watching The Mentalist from 2008. They had flip phones then.

replied to:

This one is good

created a poll:

What do you think is better?

Students stay in the same classroom all day, teachers come and go60%

Teachers stay in the same classroom all day, students come and go40%

Nobody stays in the same room all day, rooms are booked for the duration of the lecture period0%

5 voted

replied to:

I picked the first, because you didn’t specified level, and I assumed below college.

wrote:

Using deep research for trip planning and visualising it with coding agent is fun.

wrote:

"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

continued:

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.

continued:

Or they just had investments in LLM factories.

Any conversation