stochastictrebuchet

Well-deserved win! Watched this in the cinema a few weeks back. What immediately struck me about the beautiful art style is that it felt more like what you’d expect from a labor-of-love indie game than from a dreamworks/pixar studio – and it was incredibly refreshing! Also, for a movie where water plays a big role, the fluid rendering was absolutely breathtaking. I could almost smell the warm plastic air of a GPU giving its all.

Another happy Kitty user here!

I use my terminal as an IDE. Kitty makes it (relatively) easy to write custom interactive applets (aka kittens) that open in new panes or communicate between panes. The ssh integration is also really useful: whenever I ssh into my remote work station my fish and helix config gets copied over.

Judging by the code (a mix of C, python, and go) and the fast release rate, the core maintainer seems to be an utter mad genius – which unfortunately is sometimes reflected in his notoriously abrasive communication style.

Only thing I’m lacking is persistent remote sessions. The maintainer is not quiet about his dislike of tmux and other multiplexers. It’s wildly inefficient to process every byte twice, he argues. Convincing but Kitty doesn’t currently offer an alternative for remote sessions, which is where I do most of my work. Wezterm has something for this in beta, but misses many of the niceties of Kitty. So I’m still using tmux for everything in Kitty, because it trips me up to have one way of working with panes locally and another way when working remotely.

I tried Ghostty, if only because the maintainer is an excellent communicator. I found it polished but simple. I couldn’t figure out how to page up the scrollback or search it. I couldn’t rename tab titles. The config format seemed under-documented. I’ll give it another go in a month or so.

GenAI coding assistants are only as good as the data they are trained on. Less-used proglangs make up a tiny fraction of the available data, or may even be completely absent. There is a reason coding assistants give convincing results with Python and JS/TS, but underperform even on relatively up-and-coming langs like Rust.

Bit-co-NECCCCCCCCCCCCT

Preferred prior experience: Midtown Madness (100+ hours), GTA (100+ hours), or comparative driving simulators.

One of the points the article makes is that people boost such content despite knowing it’s fake because it confirms what they’re ’feeling’. Want to feel outrage? Here’s an image that will let you and others feel that. Truth? Irrelevant.

In short: it’s the ‘facts don’t care about your feelings’ crowd doing what they do best: recasting reality as a jumble of vague feelings.

Pastes 300k tokens worth of scp and creepy pasta lore into the prompt

“Sweet dreams, Timmy”

I agree with you. I just wanted to share some nuance. The point I wanted to make is that it is in fact possible to incorporate LLMs in a fairly controlled way while calculating (estimates of) the risk of failure as well as the associated social and financial costs. I do it every day, but I’m no tech bro and dislike the ‘AI will fix everything’ types as much as everyone here.

Just fyi, that’s not entirely true. If we’re just focusing on LLMs, structured and guided generation exists. Combine that with an eval set (= unit tests), you can at least track how well you’re doing. For sure, prompt engineering misses the feeling of being in control. You’ll also never be able to claim 100% coverage (although even with unit tests that’s not something you can claim, as there are always blind spots). What you gain over traditional coding, however, is that you can tackle problems that might otherwise take an infinite number of years to express in code. For example, how would you define the rules for detecting whether an image shows a bird?

It’s just a tool like any other. Overuse is currently detestably rife. But its value is there.

Source: ML engineer who secretly hates a lot about ML but is also in awe at the developments of the last few years.

Same energy: this legendary comment in an issue on the Docker github repo (by the issue OP, no less)

https://github.com/docker/cli/issues/267#issuecomment-695149477

Wait, what happened to LinkedIn?

Another vote for hx!

Getting a productive setup for Python work is a matter of a few extra lines of TOML. The pre-release version on master also allows for multiple LSPs per language, which means I can combine pyright with ruff.

The modal key chords are verb-object instead of object-verb. It’s not a main selling point to me. However, you get multi-cursors out of the box, which I’ve always found simpler than e.g. macros. In general, keybindings are discoverable. I learn something new every week.

All in all, despite a few rough edges, it’s a nice alternative to needing to get a PhD in neovim configuration to get anywhere remotely near the cool setups other people are rocking.

All my old macbooks eventually get the Linux treatment. On modern hardware, however, the trade-offs of non-macOS just don’t make sense to me.

For now, Apple Silicon has made a fanboy out of me. I can’t overstate how big the jump in performance felt going from intel to my first M1 – not to mention the improved thermals. And obviously part of that is due to excellent alignment between hardware and software.

Still, once that first M1 hits retirement, I’ll no doubt experience that familiar pang of gratitude towards those engineers that put up with the trade-offs of running Linux on it today in order to get everything working.