- complexity is the default condition
- complexity will grow
- entropy will increase
- local decisions will get made will appear to make sense locally will not roll up will clash at the edges
- we we always have the choice to simplify if we can recognize where the complexity is and if we can arm ourselves with techniques that are going to help us do that
- consistency is the key
- consistency is the mechanism by which we can make we can reduce cognitive load we can or make cognitive load appropriate
- some problems are intrinsically hard, that's fine
- they don't need to be any harder than they are
it really shouldn't be this difficult
We have two options for defining types in TypeScript: types and interfaces. One of the most frequently asked questions about TypeScript is whether we should use interfaces or types.
The answer to this question, like many programming questions, is that it depends. In some cases, one has a clear advantage over the other, but in many cases, they are interchangeable.
In this article, I will discuss the key differences and similarities between types and interfaces and explore when it is appropriate to use each one.
Let’s start with the basics of types and interfaces.
When to use types vs. interfaces
Type aliases and interfaces are similar but have subtle differences, as shown in the previous section.
While almost all interface features are available in types or have equivalents, one exception is declaration merging. Interfaces should generally be used when declaration merging is necessary, such as extending an existing library or authoring a new one. Additionally, if you prefer the object-oriented inheritance style, using the extends keyword with an interface is often more readable than using the intersection with type aliases.
Interfaces with extends enables the compiler to be more performant, compared to type aliases with intersections.
However, many of the features in types are difficult or impossible to achieve with interfaces. For example, TypeScript provides rich features like conditional types, generic types, type guards, advanced types, and more. You can use them to build a well-constrained type system to make your app strongly typed. You can’t do this with interfaces.
In many cases, they can be used interchangeably depending on personal preference. But, we should use type aliases in the following use cases:
- To create a new name for a primitive type
- To define a union type, tuple type, function type, or another more complex type
- To overload functions
- To use mapped types, conditional types, type guards, or other advanced type features
Compared with interfaces, types are more expressive. Many advanced type features are unavailable in interfaces, and those features continue to grow as TypeScript evolves.
Below is an example of the advanced type feature that the interface can’t achieve.
type Client = {
name: string;
address: string;
}
type Getters<T> = {
[K in keyof T as `get${Capitalize<string & K>}`]: () => T[K];
};
type clientType = Getters<Client>;
// type clientType = {
// getName: () => string;
// getAddress: () => string;
// }Using mapped type, template literal types, and keyof operator, we created a type that automatically generates getter methods for any object type.
In addition, many developers prefer to use types because they match the functional programming paradigm well. The rich type expression makes it easier to achieve functional composition, immutability, and other functional programming capabilities in a type-safe manner.
Some of you may be able to guess what’s been on my mind lately. In April 2010, film critic Roger Ebert made an infamous claim. He said, “Video games can never be art.” His blog post set off a firestorm of discussion, centered around the idea that he was an out-of-touch, old man who didn’t understand games. While I disagree profoundly with Mr. Ebert, he was an intelligent, articulate scholar, and had a better point than his clickbaity quote might imply.
Sandro Mancuso, Co-founder and Group CEO
AI is Changing How We Code
AI coding tools are completely changing the software development landscape. Developers now rely on tools like GitHub Copilot, Amazon CodeWhisperer, and ChatGPT to generate code, fix bugs, scaffold tests, and accelerate routine tasks. Code that once took hours can now be produced in minutes; sometimes seconds. Prompting is quickly becoming a core developer skill.
AI is here to stay. The promise is undeniable: higher output, faster delivery, and fewer repetitive tasks. But with this leap in speed, a deeper question emerges:
Are we building well - or just fast?
This article explores that question through the lens of Software Craftsmanship, offering a pragmatic perspective on how to embrace AI without sacrificing the principles that make great software possible.
Project
EcoLogits is a suite of open source tools for estimating the environmental footprint of generative AI models at inference. Based on life-cycle assessment principles, the project raises awareness about the direct environmental impacts of AI while empowering developers and organizations to build more sustainable AI-powered applications.
- Open & transparent – Code, methodology, and data are openly accessible.
- Ease of use – Emphasizes on seamless integration and user experience.
- Community-driven – Continuously built and improved collaboratively.
✦ TECHNICAL BEAUTY ✦
Episode 22: mc – Midnight Commander
- A computer science student in Mexico City writes a file manager. Thirty-one years later, it's still running. One pkg install on FreeBSD. One apt install on Debian. Available on anything with a C compiler and a pulse.
Miguel de Icaza built mc as a Norton Commander clone for Unix. Then he co-founded GNOME. Then Mono. Then Xamarin. Then Microsoft acquired it. The man kept rather busy. The file manager quietly kept working.
220,000 lines of C. One binary. 12 MB of RAM. Here's what you get for that:
■ Dual-panel file management
■ Built-in editor (mcedit) with syntax highlighting
■ Built-in viewer (mcview) and diff (mcdiff)
■ FTP, SFTP, SMB: browse remote servers as local directories
■ FISH: file transfer over SSH without SCP on the remote
■ Archive browsing: .tar.gz, .zip, .rpm, .deb as directories
Hardik Pandya
I am generally curious about the concept of legibility of work. Look around in your workplace. You can find documents, messages, presentations, design files. Evidence of people’s work. While it may look like a lot, there is a whole other type of work that is very hard to see. The invisible work.
This decision tree describes how to use the alt attribute of the <img> element in various situations. For some types of images, there are alternative approaches, such as using CSS background images for decorative images or web fonts instead of images of text.
Jim Coplien believes that we have done OOP the wrong way for 40 years, and suggests an approach to reflection based on the DCI paradigm and influenced by the human society.
Heidi Waterhouse discusses seven issues to tackle during project development: Localization, Security, Extensibility, Documentation, Affordance, Acceptance, Accessibility.
write down what you forgot
—-
adhders have lots more experience in chaotic environments than neurotypical
Yanis Varoufakis discusses his powerful theory of technofeudalism to describe the next era after capitalism.
Capitalism’s dominance may have run its course. A new system of control has taken its place, one Yanis Varoufakis calls technofeudalism. For Varoufakis, former Minister of Finance for Greece and number-one bestselling author, the elite few who run Big Tech exercise the same level of influence that feudal overlords once did. Except now, they not only threaten democracy but wish to rewrite the rules of global power themselves. Join Varoufakis as he reveals why we need to thwart these new figures, and why we need to get Musk and Sam Altman out of the White House before it is too late.
Yanis Varoufakis is a leading critic of the economic establishment. He was the former Finance Minister of Greece, and is a professor, bestselling author, and current chair of the Democracy in Europe Movement 2025. The Institute of Art and Ideas features videos and articles from cutting edge thinkers discussing the ideas that are shaping the world, from metaphysics to string theory, technology to democracy, aesthetics to genetics.
John Cutler
Ya quelques phrases bien tournées qui valent le coup d'être citées
Les chatbots soignent la dépression, mais attention !
L’IA se révèle étonnamment efficace pour traiter l’anxiété et la dépression. Des études sérieuses montrent des résultats concrets.
Mais il y a un revers à la médaille : atteintes à la vie privée, dépendance et risque de remplacer le contact humain par du code.
Cet épisode analyse les points forts et les risques de la thérapie par IA, et explique pourquoi des millions de personnes y ont recours malgré tout.
Tired of drowning in student debt for outdated education? Discover ten completely free university courses that deliver better career outcomes than expensive degrees costing hundreds of thousands. From MIT's computer science program to Yale's psychology lectures, these world-class educational resources are available right now without application fees, student loans, or four years of your life. Perfect for career changers, ambitious professionals, and anyone seeking high-quality skills training without the financial burden. Learn cutting-edge technology, finance, cybersecurity, and business management from Nobel Prize winners and industry pioneers who actually know what employers want. These aren't watered-down versions - they're the exact same courses taught at prestigious universities, minus the inflated price tags and unnecessary requirements. Whether you're interested in machine learning, game development, or web security, these free alternatives provide practical skills that translate directly into job opportunities. Stop paying premium prices for mediocre education when world-renowned institutions are giving away their best content. Transform your career prospects without sacrificing your financial future or spending years in outdated classroom settings.
Note taking
How To UPGRADE iPad Note Taking (With Science) - YouTube
Mal
- trop verbeux
- très linéaire
True knowledge is not linear
Mais pas évident d'enseigner autrement que linéairement.
Besoin de réencoder de façon non-linéaire
Représenter de façon non linéaire
Penser de façon non-linéaire
Des choses changent dans la façon de prendre des notes
=Higher order learning
- Rechercher constamment des relations
- contamment créer des groupes basés sur ces relations
- constamment créer des relations entre les groupes
- Essayer d'xprimer en quoi ca s'articule avec la vison globale
- et comment l'exprimer de la façon la plus simple possible
- nettoyer au fur et à mesure pour simplifier
- moins de mots
- le faire pendant la prise de note, pas après
- on peut exprimer beaucoup de choses avec juste l'arrangement, et les choses et l'espace
- show instead of explain
app with infinite canvas like "Concept" or "Freeform"
Importance de processing pendant la prise de note
Yanis Varoufakis et Tim Nguyen, chercheur chez Google DeepMind, débattent de l'essor des géants de la technologie, de la fin de la vie privée et de la corruption de la société.
Les géants de la technologie sont-ils hors de contrôle ?
De Karl Marx à Elon Musk, nombreux sont ceux qui ont affirmé qu'entre de bonnes mains, la technologie est une force libératrice et porteuse de bien. Mais cette idée est de plus en plus remise en question, des personnalités comme Obama et Tim Berners-Lee mettant en garde contre les graves menaces que représentent les géants de la technologie pour l'autonomie et la liberté. Par conséquent, nombreux sont ceux qui affirment aujourd'hui qu'il est nécessaire de se libérer des géants de la technologie et de leur caractère addictif. Mais on ignore si cela est réalisable. Les critiques affirment que les géants de la technologie, par le biais des algorithmes et des données, manipulent nos actions et nos choix dans tous les aspects de notre vie, si bien que l'autonomie humaine est déjà radicalement compromise. Se retirer est quasiment impossible. Des études de Cambridge montrent que les plateformes anti-technologie, prétendant lutter contre la dépendance aux géants de la tech, ne font souvent que la reproduire sous une autre forme.
Faut-il en conclure que la technologie a porté atteinte à notre liberté et que nous ne maîtrisons plus nos décisions et nos comportements ? Devons-nous faire évoluer les droits numériques pour résister à la manipulation et garantir l'autonomie, et cela fonctionnera-t-il ? Ou le démantèlement ou l'interdiction des géants de la tech pourrait-il faire partie de la solution, nous permettant de reconquérir notre autonomie ?
Yanis Varoufakis est un économiste, universitaire et homme politique de renommée mondiale. Son dernier ouvrage s'intitule « Technofeudalism: What Killed Capitalism ». Timothy Nguyen est un chercheur de premier plan en IA chez Google DeepMind et anime le podcast « The Cartesian Café ». Janne Teller est une auteure de romans, d'essais et de nouvelles saluée par la critique. Son œuvre aborde des thèmes existentiels à grande échelle, qui suscitent souvent des débats controversés. Animé par Shini Somara. L'Institut des Arts et des Idées présente des vidéos et des articles de penseurs d'avant-garde discutant des idées qui façonnent le monde, de la métaphysique à la théorie des cordes, de la technologie à la démocratie, de l'esthétique à la génétique. Abonnez-vous dès aujourd'hui ! https://iai.tv/subscribe?utm_source=Y...
HighYield x SemiAnalysis deep-dive into AI Datacenters, Gigawatt Megaclusters and the Hyperscaler race to AGI.
How AI Datacenters Eat the World. SemiAnalysis Datacenter Anatomy Series:
Part 1: https://semianalysis.com/2024/10/14/d...
Part 2: https://semianalysis.com/2025/02/13/d...
SemiAnalysis Datacenter Industry Model: https://semianalysis.com/datacenter-i...
- 0:00 Intro
- 1:52 Datacenter Origins
- 2:53 Datacenter Location
- 4:06 AI Datacenters
- 5:08 AI Datacenter - Connectivity
- 7:17 AI Datacenter – Compute & Rack Power
- 11:40 AI Datacenter – Air & Liquid Cooling
- 14:32 AI Datacenter – Critical IT Power
- 17:50 SemiAnalysis Datacenter Anatomy
- 18:58 Texas Datacenter Mystery
- 21:26 AI Hyperscalers, Gigawatt Megaclusters & Nuclear Power
- 26:09 Satellite Images & SemiAnalysis Datacenter Industry Model
- 28:24 The Future of AI Datacenters