We use Azure Foundry hosted Opus and it seems as capable as ever.
I run a small local GPU cluster for Qwen 3.8 Flash and while it's "free" (you should check electricity prices in EU...) unfortunately it doesn't even compare to Opus 5.5 or Astra in capabilities. It's like GPT 6 vs GPT ~4, but imagine GPT 4 is also 100x slower and sounds like an airplane taking off.
In the end it's totally possible that I would pay more for electricity running Qwen for 24 hours churning on some complex problem vs. Astra one-shotting it in 15 minutes.
I hope we get there, but it's not there yet, and I'm sure never will be with consumer GPUs. Maybe if you spend $20k+ on a local AI rig. There's just no market force to give access to AI to everyone at fair price.
Print a massive log file in kitty/ghostty, and another one with you default terminal, and you'll see a massive difference in how fast the log is printed out. Well to be honest it's just comfort, I can't claim it makes a difference in actual productivity, for me the real benefits of kitty are just the terminal multiplexer-ish capabilities (panes, sessions), other rendering niceties like adjusting the font, or having smooth scroll, sensible default shortcuts. But the actual "performance" e.g. input lag or rendering lag, I don't think it's that important.
A snappy terminal doesn't prevent you from seeing half-filled screens. Flicker can occur when you write line-by-line. That happened in a recent TUI application I wrote. Using synchronized update mode solved the problem.
I frequently compile big software projects, using fast multi-core CPUs, like AMD 9950X, with which the compilation itself is really fast.
Such compilations produce a ton of text messages. With any classic terminal, the text rendering in the terminal window becomes a bottleneck, slowing down the compilation very noticeably.
With modern terminals that use GPU text rendering, like ghostty and kitty, compilations are done much faster.
Of course, with classic terminals you can redirect the compiler output, to avoid this problem, but I prefer to glance at the progress of the compilation, to notice early if there are problems.
Even in normal interactive use, there are enough cases when the difference between the instant rendering of ghostty or kitty and the slowness of the classic terminals is noticeable.
To see the difference between a slow terminal emulator, like you probably use, and a fast terminal emulator, you can do a "cat" from a text file of a few megabytes, and you will see the speed difference.
So in theory, can a hypothetical Astra clone itself but shed all irrelevant weights? I.e. "create a self-clone limited to 27B weights, focused on designing running shoes, forget everything you know about kangaroo genetics, Karl Marx, cooking Burmese curries, nuclear reactor design etc"?
If you follow this logic, we get to the true "AI-native" engineer which is ecstatically floating in a bath of dopamine-stimulating liquid, while AI reads their brain pattern and instantly builds everything they didn't even form a full thought about.
"If that sounds outlandish, that's because it is. We don't have such capable AI systems yet."
I'm not tired of work per se, but I'm definitely tired of what Internet is becoming: AI generated articles on AI generated websites, AI generated videos about AI generated products, AI generated emails inviting me to test AI generated features. AI generated movies and games coming soon. At work it's all AI generated presentations about how AI solves everything, AI on AI for AI with AI by AI, AI, AI, AI, AI, AI...
reply