castles

I have also wondered if OpenAI are going to train a private model with all the ChatGPT history, and then use that to train a public model.

We should focus less on peak speed and more on sustained speed?

Not a dvorak user.

I infected myself with Colemak back in the days when I had time to mess around on such things, somewhat inspired by a programmer of acquaint, that suffered from RSI. I can say after 10+ years I've never suffered from RSI.

It's been a pain when trying to pair program but never really had any substantial regrets. I can't compare it to any previous era where I was a qwerty user as I learned to touchtype with Colemak.

I think anything that allows you to maintain typing speed (ergonomics) and reduces the chance of distraction from errors is worthwhile.

I vaguely recall a cross platform, clang integrating, c/c++ IDE was in the works? Any news?

IntelliJ is awesome :)

Thanks for linking these :)

It's actually so much faster, that the operating system's method of drawing windows is the prime determinant of startup speed perception. Surprisingly, the winning OS is windows.

Startup times are drastically faster. Python plugins are now `out of process` and load async. Also, the very latest beta uses skia instead of cairo. The api is now thread safe which will mean less `Plugin x is taking too long` messages. And so on ... Even updating is faster, with the new binary diff based auto updater

Pity, Testacular was such a ballsy name.

You can set a `context` on key bindings to trigger only in specific language scopes or per project by referencing the `settings` key in the sublime-project file. In fact, you can even write a python plugin to create your own contexts that you declare in the sublime-keymap json. Check out G's docs: https://sublime-text-unofficial-documentation.readthedocs.or...

For sublime text 3, a QuickPanel plugin to show keybindings/settings: https://github.com/sublimator/EditPreferences

I'd love to be paid $100/hr for formal resting. That would be sweet.

People are working on it: https://www.facebook.com/pages/FreeFly-Astronaut-Project/249...

In any NFC interaction, there's at least one reader involved. In the NFC world people are dreaming of the day that all mobile phone's have nfc and these interactions are reader on reader.

We are obviously a long way from that. Does this device make this a reality? Will people be buying these, placing them in their pockets like headphones to pull out when needed?

Unlikely, it's true.

However, there's no need to fixate on this. With some imagination, there's a lot that can be one with the humble tag and nfc readers in the field.

Imagine at events or tradeshows, people were given wristbands. These wristbands could have their email, phone number, VIP area access and even drink credits attached. Hell, even their full facebook profile.

Booth holders could have checkin stations, where with a swipe wristbanders could sign up to mailing lists, share contact info, or like a product on facebook. People at concerts could get branded photos taken with "Listening to $FAMOUS_ROCK_BAND's $FAMOUS_SONG"

Even if, there was 400M iOS devices that were going to spontaneously combust the minute iPhone 6 with nfc was released. And even if, this product line only had a useful life of one year, for developers it's still a year's evolution on their NFC app.

It's about empowering apple devs NOW(ish)

The `whole greater than the tally` feature set and performance improvements of this release, really has inched Python 3 to the point where I will want to start using it. With no basis beyond my own and other's like sentiment this still seems like an important milestone. Feel like many will move from inertiaville, at least building a weekend house in momentum town.

It's time