Last Notes
Platforms should publish search data.
https://mccormick.cx/news/entries/platforms-should-publish-search-data
Agentic coding tools take too much control from the human. Whenever I try them it feels wrong and the results are a mess. They discourage human oversight which degrades the quality of the code.
I'm still using aider because it is not agentic and it lets me stay in the loop. I'm maintaining a fork with some quality of life improvements (including a single-binary build) here if you're interested: https://github.com/chr15m/aider
Here's my .aider.config.yml:
```yaml
auto-commits: false # prevents aider from doing git commits
gitignore: false
notifications: true
show-model-warnings: false
watch-files: true
auto-lint: true
read: [/home/chrism/dev/ai-context/AGENTS.md, README.md, AGENTS.md, .context/*]
```
Worklog:
- Building new features for https://exceltohtml.com
- Working on game prototype fortnight #2 of 26 - a procedural game music chords generator.
Agree, it's a tool. I am finding it useful when used in a disciplined way with a human in the loop at literally every step. I think the current tooling gives the LLM far too much leeway in general.
I did some research and testing with gpt-oss-20b and llama-4-maverick. They don't compare with Claude but could definitely handle some tasks like extracting structured information from articles, DevOps, routine coding tasks, smart OCR etc.
The Chinese open models suffer from censorship on certain topics, but there is a model where somebody trained Qwen on Deepseek outputs and it appears to be "abilterated" and will respond to those topics, and is pretty good overall. I'm going to run some of these models locally and will do a video/write-up about that.
That's a great point. LLMs are helping people overengineer stuff nobody wants, even harder.
https://feed.mccormick.cx/system/media_attachments/files/115/965/521/421/613/739/original/c89897f39a9d244a.jpg
Worklog: progress on getting notes/comments export working in https://exceltohtml.com add-in.
Doodled Dungeons is prototype #1 of 26 this year. Here's a video about building it.
https://youtu.be/NwFZHbsNYd4
Worklog: continued with the https://asterogue.com Steam release.
https://github.com/chr15m/ai-context/blob/main/AGENTS.md
> The user is a detail-oriented anal-retentive control freak. Only do things they specifically ask you to do.
This is the way.
I tested out two free open weights models with Ollama on an old Macbook air from 2015 to see how it performed.
https://youtu.be/VpxJDlbZcNM
Getting my web game on Steam. A quick look at the process of listing in the Steam store.
https://youtu.be/cYy97f0m1Rw
https://npub1m2f3j22hf90mt8mw788pne6fg7c8j2mw4gd3xjsptspjdeqf05dqhr54wn.blossom.band/f21daccf75988ef134df12f17abac09d5df7e4598a1d0d370904c6717e0c45a6.png
https://npub1m2f3j22hf90mt8mw788pne6fg7c8j2mw4gd3xjsptspjdeqf05dqhr54wn.blossom.band/9c4e8154a3c4ae6e5ba2429ec8bc4f8d3cdb0a92d833aa6cf2d9f83ae5837343.jpg