At least here in the states, districts because they pay teachers shit, are getting pushback to add required meetings and learning on the end of the day. Instead there are “short days” for the kids.

It’s no longer an issue for me, but seriously annoying how little k-12 are actually in school these days.

I was dealing with something around authz/n with Claude code running Fable. It chewed through a couple of questions (these were implementation, not security reviews) and on one it shit it’s pants and said I can’t do that, here is Opus.

I’ve dropped my anthropic plan level, it’s just not worth it.

Heck, a buddy and I once chatted about the likelihood of k8s running as the control plane in a prototype autonomous vehicle.

top/btop/htop on the mac are always fun to run and see what's up.

When I spent about 8 years living in a remote area with a lot of old mines, etc. exploring rocks, having a mining claim, etc was interesting and it was not directly related to work. So, trying Geology made some sense because I had an interest.

Yes the near Lit minor and my second Bachelors (chemistry) were literally for fun - but there was a roundedness to not just staring at computers all the time.

Different people have different interests and approaches.

Also costs when I was doing the initial schooling wasn't as bad as it is now (a TAing job, working in a lab, and maybe some part time tech support) covered most bills. Rent, tuition, and other expenses were under 10k for the year - even at a UC - yes this was awhile ago. My step-daughter is looking at closer to $30k a year (give or take) for a CSU.

In the states during late covid, I thought the downtime would be good to do a second masters in something unrelated to working in software. I chose a remote program with a univeristy here in the states, I did the first year, but it reminded me that studying - especially a program unrelated to anything you have currently done - is actually not easy.

I chose Geology. I have BS/BA in Comp Eng and Chemistry and an MS in Comp Eng. So this was personal interest.

Revisiting math I hadn't used in awhile was the easy part. Making the time, while working, reminded my why I also ended up ABD (all but dissertation - toward PHD) post masters. Sometimes your brain needs to relax.

Glad it worked for you. And the European Universities are significantly cheaper than Universities in the states (this was a state univerisity which is - per state - considered a step above a state school. For instance a UC vs a CSU in california).

A non-general compression algorithm (model - I don't mean a distinct llm, but "modeling data") targeted at a specific dataset will always do better than a general algorithm.

The reason I mentioned the "encoder" doesn't matter - arithmetic coding, for the data it is presented, will beat huffman/adaptive huffman every day, but it's the model that is where the real "compression" comes into play.

I've implemented enough "coders" over the years, including arithmetic for both commercial and research purposes (was a student of Glen Langdon).