One last run before competition

Easy run - last one workout before hm

Monday endurance run 🏃

Long run at the end of week 7. This week I ran a total of 46.3 km but since the Ruda Śląska Industrial Half Marathon is next week it’s better to take it easy. 2 more workouts left!

Both Anthropic and ChatGPT do a very good job, but they have become quite expensive. This was the only reason I decided to explore open‑weight models. GLM 5.2, offered by many independent suppliers, can perform on par with Claude Sonnet and is much cheaper. The remaining question is whether we need frontier AI models for coding (even models like GLM 5.2). I suspect there is an approach where people use frontier models to code a single large monolithic task, such as “code me an app.” By breaking that task into several subtasks, you can use virtually any coder to handle those subtasks as they become less complex and usually architectural pattern is more sound.

Have you ever tried so called open weight models such as Kimi, GLM or Mistral? I really like MiniMax M3 (though I use it more for orchestrating than actual coding)

I will try to answer 5k questions in September. I made my training a bit tougher because I'm preparing for a half marathon next weekend, and quite frankly, I didn't know what pace I could sustain. I’ll aim to beat 1:40.

I think gun violence is missing. My llm says 6 433 people were shot this year according to some Gun Violence Archive source

I've been watching you do weighted pull-ups for a while now. Sometimes it looked clean, and sometimes honestly - it was a bit scary to watch. This one is scary. One piece of advice from a climber who had a shoulder injury a year ago and is still recovering - and please check it with a climbing coach or any gym coach: when you do weighted pull-ups never climb the ladder with the weight on and don't jump up to grab the bar. Set up a step so that you can reach the bar, grab the bar, engage your scapula, and only then start your pull-ups. And on the way down, never let your shoulder go slack - stay engaged the whole time. I know this better than most - 8 months of recovery! You're doing great! Stay safe!

My 5k personal best and a side effect of interval workout 22:44. Not even that tired actually- just wondering if I could shred those 2 mins if I really want to!?

Do you feel that with each new generation of LLMs you actually get better results? For example, you don’t change how you use Claude Code/Cursor - you use the same prompts and the same tools, yet you get different results when you switch from, say, Sonnet to Fable? I personally don’t feel much difference. About 1-1,5 years ago, there was a moment when models became much smarter, but now the curve of “getting better” is flattening out.

Happy weekend to you too!