Frameworks aren't just human-serving abstractions - they're structural abstractions that allow for performant code, or even being able to achieve certain behaviours.
Sure, you could write a frontend without something like react, and create a backend without something like django, but the code generated by an LLM will become similarly convoluted and hard to maintain as if a human had written it.
LLM's are still _quite_ bad at writing maintainable code - even for themselves.
Mirroring what many others have said, Azure is often broken and generally frustrating to use. Performance is also often _quite bad_ and there will be frustrating network limitations based on seemingly unrelated configurations. For example, we discovered that we could nearly double the download speeds to our webservers (downloading from Azure storage even) by upgrading to a beefier SKU. This may sound reasonable on the surface, but we were seeing speeds of only ~10Mb/s, often less. Even now, we see extremely slow download speeds and it is dependent on time of day - slowest during peak business hours and faster in the dead of night. I understand network congestion, but this just seems completely absurd when we're talking about servers that both exist within the same Azure region - likely in the same DC - having worse download speeds than I get from my $5 DigitalOcean droplet to my house.
Many will disagree with me, but the Vancouver Olympics prompted construction of some things that I would consider vital to the Sea to Sky region - the highway upgrade being the biggest.
Many will disagree with me, but the Vancouver Olympics prompted construction of some things that I would consider vital to the Sea to Sky region - the highway upgrade being the biggest.
Tide is fantastic and has IMO the best ergonomics (and could be improved to be even better), except it's essentially abandonware at this point. I'd love to see it continue and thrive, but there's many PR's that have been ready to land for a long time gathering dust, and plenty of open issues that are at a standstill because there's simply no momentum or direction.
Not blaming anyone - the maintainers don't owe us anything - it just wouldn't be my crate of choice if I was starting today. If any of the maintainers read this, shoot me a message because I'd love to help out and get the ball rolling again on tide.
Agreed - honeycomb has been a boon, however some improvements to metric displays and the ability to set the default "board" used in the home page would be very welcome. Also would be pretty happy if there was a way to drop events on the honeycomb side for a way to dynamically filter - e.g. "don't even bother storing this trace if it has a http.status_code < 400". This is surprisingly painful to implement on the application side (at least in rust).
Hopefully someone that works there is reading this.
Still early days, but we've been using CBOR instead of JSON lately at work for interfaces that have "settled" and it's been great. Means that you can shake out the early integration issues using human readable JSON, then just switch the ser/de once it's all playing nice.
Binary data support is pretty nice too for avoiding multipart request bodies.
Check out Qutebrowser. It's almost entirely usable from the keyboard only using vimlike bindings. Lags behind chromium/qt a bit for the actual engine, but these days it's totally usable and has become my daily driver.
Sure, you could write a frontend without something like react, and create a backend without something like django, but the code generated by an LLM will become similarly convoluted and hard to maintain as if a human had written it.
LLM's are still _quite_ bad at writing maintainable code - even for themselves.