That's because VS Code is hiding everything behind a bunch of non-real-time tricks of perception. Zed is giving you actual real-time feedback.
"Whom the gods wish to destroy, they give real-time data."
The overwhelming majority of the population cannot perceive anything over 90 Hz. Those that can are overwhelmingly skewed towards under 30 years old. Fighter pilots have a floor of something like 200hz for an idea of how rare it is. Just fun info.
I have 3 different displays on my desk, and they are 60hz 120hz and 240hz.
The difference between them when scrolling is.. obvious. I'm in my 40s. I'd love to see a study demonstrating that my ability to perceive this is some rare capability - that's very hard for me to believe.
Comparing side-by-side is always easier. The question is never that (should not be). The question is, would you approach a random coworker's desktop and really, sincerely, notice if they have a 120 vs a 144 Hz (or 240) monitor?
I'd say maybe if you are a professional in the sector of multimedia processing, you would be so accostumed to the smoothness of high FPS that a meager 60 fps monitor would be obviously noticeable. But for the untrained eye, I feel most people wouldn't even notice in a random scenario like that, whether the screen is 60 or 120 (and that's the range with highest ROI on FPS increments!)
I spent an admittedly tiny amount of time looking into the Zed scrolling stutters after experiencing them myself and I think it's just a matter of their line layout caching not being as good (perhaps unsurprisingly) as the Chromium/WebKit layout engine. It's especially noticeable if you have word wrapped files with lines >10kb in length (yeah, don't ask).
Zed is also at the bleeding edge of rust GUI development, which doesn't have the same mature libraries as other low-level languages. I'll give them the longer-term benefit of the doubt, but the attention to detail does matter.
To be fair, one could say the same for Zig, but Ghostty does extremely well in this regards. While it maybe doesn't have the same complexity that Zed does, it does have to deal with a lot of historical cruft of terminal emulation...
As I understand it, Pompeii was basically a city of vice and hendoism. Most of the scroll text so far seems to be the ancient version of porn fanfic. So things really never change.
Ive been using glm5 since its release and still prefer it to glm5.1 and so far to glm5.2
Perhaps it is just my harness and workflow, but the older model still seems to work better. Also the token cost is significantly lower. I rarely spend more than $20 a week with $50 cap. Not even half claudes ambiguous minimum $200 a month plan.
Now that's a tremendous pointer, I'm going to have to try that.
Do you full on let GLM5 get stuff done on its own or is it more like a guided workflow? The former's what the point releases doubled down on and is also something that uses a lot of juice.
>P.S Nginx would do too, I just felt like getting haproxy up this time :)
I think this line summarizes better than anything. Perfect example of how move fast and break things begins gloriously at first then inevitably, the breaking you thought you were doing hasn't even started and you find out what that part means.
I've always been the one saying "this is going to be a problem in a couple months" then I get shot down for "being negative." Then in a couple of of months when it fails I start getting aggression thrown at me "oh i know you want to say I told you so" and such even though I've never said such a thing when something fails. No. I would just like you to hear out my thoughts even when they may not be what you want to hear. We are all working towards the same goal.
My experience with LLM's is that the "fallback" issue is probably one of the most serious issues. I have seen zero talk about it outside of this thread. I'm not sure its even being worked on. The llms have this terrible drift towards always making something happen even if that thing is not even related to the task at hand. "failure" in the sense of simply throwing an error/exception is something models seem highly resistant to.
I cant tell you at this point how many times I've seen them do something like
Fail > make up values > maybe log it > keep working silently with increasingly corrupt data.
I personally cant wait to start seeing sob stories of how people are having thier account permanently locked after making big wins. Thats what happens when you bgive money to unregulated casinos every tine and is the exact reason there are so many gambling laws and restrictions.
IDK. The internet is full of examples of platform max exodus's. This may actually be a killer idea that will finally challenge the tech oligarchy power. I think we are in for a turbulent future. The FANNG/MANGO companies have essentially never had any real challenge to their dominace. We are about to enter an era where they will be challenged for the first time. I don't think they will be nice about it and we will see their true non altruistic colors. Especially Google and Apple which have lost their halos but are still considered "good actors" by many.
I really dislike it but the effectiveness of code generation is just too good now. I really doubted it at first but I can churn out 10x the work. Is it the same quality as IC? no but I think that it doesn't matter anymore. How many of us have worked on some 10+m loc monstrosity for years never knowing more than 1% of the codebase? Probably most. Now I can generate that same monstrosity on my own. I've created a 1.1m+ loc and growing trading platform that automates options trades based on live ML edges from research papers. I never would have been able to do that on my own before LLM's. It has full telemetry and comprehensive error logging and testing as well. I'd have been strained to get MVP done before LLMs.
I don't think someone without coding experience can do it though. I've accepted that the new expected output for a single dev will eventually converge to 10x-100x what it is now.
I'm a bit curious about this? What draws you to speaking like an LLM rather than your own voice? I've personally never felt compelled to be more llm like in my writing.
"Whom the gods wish to destroy, they give real-time data."
The overwhelming majority of the population cannot perceive anything over 90 Hz. Those that can are overwhelmingly skewed towards under 30 years old. Fighter pilots have a floor of something like 200hz for an idea of how rare it is. Just fun info.
reply