The name Anduril means “Flame of the West”, which sounds cool as hell. But it feels a little icky to me given the ideological proclivities of Palmer Luckey and Thiel.
the typography variable update was much needed but the 'professional' plan still only has support for 4 variable modes. perhaps adobe didn't ultimately acquire figma because this level of business practice was deemed distasteful even for them.
This is why we design animations in immediate actions to be below 100ms, any more than that and we start to feel "computer's reaction is slow maybe it's doing some calculations in the background?"
100ms is more than enough to break the joyful feeling of the computer becoming part of my own body, like a hand tool or a bicycle. Humans are adapted to tool use, and don't need to consciously think of the position of tools they're using. But this effect only works if the tool moves like an ordinary physical object. If you move a pencil, it just moves. It doesn't move and then move again because a designer wanted to play an additional animation after your movement. Until we have direct neural interfaces, all UI animation playback is unrealistic, because it's adding to the physical movement of your fingers, not replacing it.
this is correct. checkbox and radio buttons are form elements and it's helping the user to select choices before sending a form / finalizing. A switch is turning something on or off.
a crude example:
when you are ordering your porsche taycan you are using checkboxes to select additional features and radios to set the color of the car. this form will be finalized into an order, your selections cannot (or will not often) be changed later, a final choice has been made.
your new porsche arrives and you start to set it up to your liking, you might switch on/off features the way you like and switch them back later in a different context. (you won't be able to remove features from your car, or add new one.)
if changing a car's color is not a radio button anymore on an order sheet, then it might be just a switch or a more complex selector of different colors. look at the BMW i Vision Dee, its led panels all around, you won't have to choose a color before, they successfuly eliminated the need for a color selection before owning the car.
In the video the prompt asks the Copilot to "generate a logo for my fantasy football team" then it goes and generates a typical "cute cat with football helmet" image, heh.
IT can be, though. High quality synthetic data for training is a thing, see Phi-2 [0] (and many more examples, people having been using GPT-4 outputs for finetuning a LOT to great avail).
> AI retraining on its own output is not a great thing.
For most forms of GenAI, including image generators, using existing models to generate synthetic data (and then doing human labelling, etc., as appropriate) for fine tuning is quite common and important. Why do you think its not a great thing?
ChatGPT trains on its own output by exploring its own sampling space using reinforcement learning, Dalle 3 was trained on a dataset of synthetic captions, they are both SOTA, you can use AI to train AI models.
Also, not like we have these "color palette" thingies that we, UI designers, already use to establish branding without affecting usability. We have a lot of UI elements we can change their properties of without directly affecting usability like borders, elevations, distances (while maintaining good contrast and looking out for color blindness too). Design systems exist for a reason. Heck, every SaaS and product already has a sales page for marketing content that is entirely designed to drive sales, no need to introduce dark patterns (or invent new crazy patterns that solves an already solved problem) in the application. A lot of interaction has already been solved, I don't see why can't we automate the whole UI process with a given opinionated design system.
