That was in December, I have not bought a single book on Amazon since then, and the kindle app is not installed on my new phone. Just in case anyone from the relevant AMZN department is reading this.

There is a fun term “jagged frontier”.

Meaning: one model can be much better than the other one in one thing, and much worse than the other in another thing.

Architecturally - where do you run Postgres ? I assume it would be external to the cluster ? (doing it internally would create a circular dependency ?)

It may be interesting to try a 6bit quant of qwen3.5-35b-a3b - I had pretty good results with it running it on a single 4090 - for obvious reasons I didn’t try it on the old mac.

I am using 8bit quant of qwen3.5-27b as more or less the main engine for the past ~week and am quite happy with it - but that requires more memory/gpu power.

HTH.

Of the ones I can share:

Browser-based network tester using webrtc unreliable data https://netpoke.com - use magic code “DEMO” to see what’s it about - the source is at https://github.com/ayourtch/netpoke

A port of the SOTA speech generation model from Python to Rust:

https://github.com/ayourtch/fish-audio-experiment

A study on LLM prompting techniques:

https://github.com/ayourtch-llm/kindness

My own coding agent that i use with my locally hosted LLM for experiments:

https://github.com/ayourtch-llm/apchat

Also LLM helped with a lot of code for my packet mangling library: https://github.com/ayourtch/oside - which, among other things, includes a now battle tested SNMPv3 stack.

A true “stochastic parrot” using hash tables: https://github.com/ayourtch/hashmem

These are the ones I remember. Feel free to scout my GitHub for more. Edit: And of course it doesn’t need to be said that out of ideas I try all of them make it to github. Many end up thrown away.

The last years somehow it felt like there’s nothing new anymore, the same 10 ideas being regurgitated with slight modifications. I tinkered with AI for the past 2 years but it was mostly a “tool for writing boilerplate”. I have tried a few ideas for agents but didn’t see how it could work.

That changed with Opus 4.6 and the subsequent wave of local models - now I try 10 ideas a day and it’s like magic! And if something doesn’t work - jumping into the code and debugging it is huge fun!

Understanding that the era of the almost-free cloud tokens might come to an end, I run my own harness pointing to my own GPUs running Qwen3.5-27B, and the last few days it has been very busy! :)

My harness doesn’t “pressure cook” since it doesn’t make sense to do that with only one GPU (besides many other reasons), it runs everything in a linear fashion, including subagents, and logs everything - reading the logs as they go by is another cool thing - sometimes I pick up interesting things from it !

The distribution of people’s moods related to AI seems indeed bimodal. And I feel lucky somehow ending up in the “enthusiastic” rather than “depressed” part of it. To the folks in the other one: I am sorry. I don’t know why it is this way. If I knew I might have given unsolicited advice.

By virtue of being generators subtly broken stuff, LLMs are well positioned to create very nice learning material.

Same thing about growing the project - having to deal with something too big for AI is a very valuable experience.

And, in my experience, some of the purely human made codebases are strictly worse than LLM-made :-)

Also, I reduced the buying pace - owning physical books takes up space, so the bar for getting something into the library is now much higher than before.

But after they decided the ebooks were actually just license to read, I did exactly the same as you, and now rather than happily buying from them, actively discourage everyone in my social circle from using kindle.

I am not going back, whoever they decide to blame.

https://www.business-standard.com/amp/world-news/amazon-euro...

I made a fun toy agent where the two models are shoulder surfing each other and swap the turns (either voluntarily, during a summarization phase), or forcefully if a tool calling mistake is made, and Kimi ends up running the show much much more often than gpt-oss.

And yes - it is very much fun to build those!

It very much does! I had a debugging session with Claude Code today, and it was about to give up with the message along the lines of “I am sorry I was not able to help you find the problem”.

It took some gentle cheering (pretty easy, just saying “you are doing an excellent job, don’t give up!”) and encouragement, and a couple of suggestions from me on how to approach the debug process for it to continue and finally “we” (I am using plural here because some information that Claude “volunteered” was essential to my understanding of the problem) were able to figure out the root cause and the fix.