microflash

Talk to me when I can set custom search engines, when developer tools become sane enough and when ‘localhost’ starts resolving.

Siri

Thank you for pointing that out. We’ll bump the domed version by $100 to fix this mistake. /s

I’ve found that increasing the number of screens is not helpful at all for learning.

What worked for me was building a system that lets me think consistently everyday. And by system, I mean sitting in a library with paper books and if really necessary, with my existing laptop.

I’d add Ranger’s Apprentice by John Flanagan in others’ recommendations.

> Text is no longer rare and precious like in ancient times.

Text was never rare. What made it special was the people, their experiences, imperfections and the ways they expressed them. It may be corny to say but words do have power - to light up your imagination, let you visit worlds through the eye of the narrator, and spark that tiny flame which could lead to great discoveries.

AI-generated text has none of that effect on me. My brain just shuts down as soon as it detects it. My eyes glaze over and my fickle attention flickers away.

So, if anything, AI is making some human writing way more precious than before.

Yes, weight and size have been a major detracting factor for me as well. I’d also like some innovation on the basics: improving the keyboard which has been really bad since forever, enabling people to use any search engine in Safari, larger battery, interoperable USB-C connection without requiring specialized software, support for high-res audio codecs, in-place translation (like Firefox or mastodon client Phanpy). Give me a reason to upgrade. With past few generations and current one, there’s been none.

This post is what triggered me to cancel my subscription and migrate away from Bitwarden in July. I’ve seen too many repeats of this show. This has completely soured me from cloud-backed critical software. Slowly moving toward offline alternatives wherever possible and self-hosting when it isn’t.

I was also fine with FF 156 version of Nova where compact mode was really compact with vertical tabs and rounding was restrained enough. With FF 157 they have added more padding to vertical tabs to compensate for close tab icon on rounded favicons, menu borders are so rounded they clip menu items, PIP min size lock has been removed so you can minimize it to a popsicle, and so on. It seems like they skipped a lot of QA.

I’m not necessarily confrontational about AI adoption. I’m more interested in situations where leadership basically sets you to fail: unrealistic velocity expectation, too much context switching, code sprawl, PR floods, and so on.

As a very heavy user of IDEA, DataGrip and PyCharm, I still appreciate the level of power and flexibility JetBrains offer but continue to be dismayed by the absolutely massive binaries of their IDEs and really bad performance, especially their file watching capabilities which take ages to sync the changes done by external tools into the IDE.

Say what you may, VSCode feels lightweight in comparison despite a lot of plugins (I’m not a VSCode user but I do endure it at work).

Shedding IDE size, decoupling bundled plugins from shipped binaries, and improving file system sync would be great improvements they could do, besides pursuing AI and whatnot.

I’ve found that most companies don’t want an elite engineer. They want “best for the bucks” engineers who’re often mediocre.

If by accident, they do end up with an elite engineer, the company most likely doesn’t know what to do with them. Everyone gets beaten down to the same level of mediocrity, suffering the indifference of managers who don’t even know they’ve got an elite engineer. And, oftentimes, the incentives are aligned in the opposite direction.

The source code also has all the hallmarks of slop.