197 Followers
194 Following
143 Posts
Pragmatic idealist. Interested in tools that augment human ingenuity. Former designer & product manager, learning backend development.
Bluesky@andric.bsky.social
Twitter@astralwave

This wasn’t always the case in iOS, of course.

Similar actions, but on the springboard in iOS and iPadOS, don’t use the ellipsis as an affordance, even when they invoke a modal

This is interesting to me because I’ve only ever done design for web & mobile, and this kind of affordance seems to only be used on desktop, mostly – I wonder why?

But with the rise of SwiftUI, these seem to be showing up more on iPhone and iPad apps. Here it is in the Books app.

This is a small design detail in macOS and Windows that I’ve entirely missed in my decades of using it.

It turns out that when you have a menu item with a label that ends with ellipsis (…) it means the action will invoke a dialog, instead of being immediately executed!

Can’t count the times I’ve seen people use Excel or dumb tables in docs because they want to represent adjacency matrices.

Graphs are how we naturally think to model relationships.

RT @[email protected]

You can represent any network with a matrix.

Each row answers the question:

How is this thing connected to all the other things?

🐦🔗: https://twitter.com/dela3499/status/1619904914003750914

Carlos De la Guardia on Twitter

“You can represent any network with a matrix. Each row answers the question: How is this thing connected to all the other things?”

Twitter

LLMs might just be the technology trigger that leads to the takeoff of semantic web.

Most webpages today aren’t annotated with JSON-LD or RDFa, except content marketing pages that try to game the Google rankings.

RT @[email protected]

New blog post - using GPT3 to generate structured recipe data from free-text.

I'm pretty impressed - it does very well in NER like tasks and can output valid JSON without any intermediate steps.

https://binal.pub/2022/12/extracting-and-structuring-recipes-using-gpt3/

🐦🔗: https://twitter.com/binalkp91/status/1605625626358054937

Extracting and Structuring Recipes Using GPT3

The code for this post is here. The second experiment I’ve tried is GPT3 to extract and structure data and have been pretty impressed. The below example took me about an hour to setup, most of it just being iterating on the prompts I’m using as directions to the model. An additional bonus I wasn’t expecting - this also turned out to be a decent recipe generator. If I input in just a recipe name like Pumpkin Pie it’ll generate/hallucinate structured ingredients and instructions.

This UI, which maintains a 1:1 mapping between a document and an outline, is alluring.

It reminds us that with the right affordances that play nicely with each other within the same system, we don’t have to choose between them in competing tools.

RT @[email protected]

The secret behind @[email protected]:

Software Engineering principles applied to Note-taking.

↓ Such as the Class Diagram ↓

Don't worry if you're not a Tech person...
I've got a Master's in AI – I got you covered.

What is a Class Diagram: 🧵
(+ How it helps you understand Tana)

🐦🔗: https://twitter.com/FisFraga/status/1611420064649183239

Felipe Fraga | Knowledge Entrepreneur on Twitter

“The secret behind @tana_inc: Software Engineering principles applied to Note-taking. ↓ Such as the Class Diagram ↓ Don't worry if you're not a Tech person... I've got a Master's in AI – I got you covered. What is a Class Diagram: 🧵 (+ How it helps you understand Tana)”

Twitter

I love how succinct the principles of “Beneficence”, “Nonmaleficence”, and “Autonomy” are. We’ve come a long way since Asimov’s three laws of robotics.

RT @[email protected]

Compare to GPT-3, Claude (a new model from @[email protected]) has much more to say for itself.

Specifically, it's able to eloquently demonstrate awareness of what it is, who its creators are, and what principles informed its own design:

🐦🔗: https://twitter.com/goodside/status/1611456645737107456

Riley Goodside on Twitter

“Compare to GPT-3, Claude (a new model from @AnthropicAI) has much more to say for itself. Specifically, it's able to eloquently demonstrate awareness of what it is, who its creators are, and what principles informed its own design:”

Twitter

“Trough of Disillusionment” for tools for thought, coming right up 😅

RT @[email protected]

https://borretti.me/article/unbundling-tools-for-thought

🐦🔗: https://twitter.com/zetalyrae/status/1607359624113704961

Unbundling Tools for Thought

Disparate thoughts on personal information management after traversing the hype cycle.

Fernando Borretti

What?!

RT @[email protected]

I then tried to apply HVM to the context of blockchains. By swapping the EVM by the HVM, I built Kindelia, an "Ethereum clone" that has 100x to 1000x increased throughput in opcodes like SSTORE. This would be enough to, say, run MMORPGs on layer-1, which is unfeasible today.

🐦🔗: https://twitter.com/VictorTaelin/status/1588332471942512641

Fica, vai ter monads on Twitter

“I then tried to apply HVM to the context of blockchains. By swapping the EVM by the HVM, I built Kindelia, an "Ethereum clone" that has 100x to 1000x increased throughput in opcodes like SSTORE. This would be enough to, say, run MMORPGs on layer-1, which is unfeasible today.”

Twitter