New Viewport Units – Ahmad Shadeed

We have been using CSS viewport units since 2012. They are useful to help us in sizing elements based on the viewport width or height.

However, using the vh unit on mobile is buggy. The reason is that the viewport size won’t include the browser’s address bar UI.

To solve that, we now have new viewport units. Let’s find out about them in this article.

CSS viewport units

For example, when we need to size an element against the viewport size. The viewport units are vw, vh, vmin, and vmax.

Consider the following figure:

The value 50vw means: to give the element a width equal to 50% of the viewport width.

If you want to learn more, I wrote a detailed article on viewport units.

The problem

When using 100vh to size an element to take the full height of the viewport on mobile, it will be larger than the space between the top and bottom bars. This will happen in browsers that shrink their UI on scrolling, such as Safari or Chrome on Android.

Here is a figure that shows how each mobile browser has a different UI for the top and bottom UI.

Suppose that we have a loading view that fills the whole screen.

/* I know that we can use bottom: 0 instead of height: 100vh, but this is to intentionally highlight the issue. */
.loading-wrapper {
  position: fixed;
  left: 0;
  right: 0;
  top: 0;
  height: 100vh;
  display: grid;
  place-items: center;
}

Consider the following figure:

The loading icon is centered in CSS, but visually, it looks like it’s positioned slightly to the bottom. Why is that happening?

For the browser, height: 100vh means that the element will fill the viewport height, but it won’t calculate the computed value dynamically. That means the bottom address and toolbar won’t be calculated.

Because of that, we have an expectation that 100vh will be equal from the top of the viewport to the start of the address bar UI.

When we scroll down, the address bar UI will shrink its size. This is good, as it gives the user more vertical space to browse the page. At the same time, it’s breaking the UI somehow.

In the following figure, the center of the vertical space is off when the address bar is visible. When scrolling, it looks fine.

Notice how I highlighted the invisible area. When scrolled down, it become visible. How to deal with that in CSS?

The small, large, and dynamic viewport units

To solve that, the CSS working group agreed on having a new set of units: svh, lvh, and dvh. They stand for the small, large, and dynamic viewport, respectively.

The small viewport

The svh represents the viewport height when the address bar UI hasn’t shrunk its size yet.

The large viewport

The lvh represents the viewport height after the address bar UI shrunk its size.

The dynamic viewport

From its name, this unit is dynamic. That means it will use any of the small, in-between, and large units based on whether the address bar UI is shrunk or not.

During the initial scroll, the dynamic viewport unit will change as the browser UI will shrunk. Here is a video that shows how the dynamic viewport changes:

Use cases and examples

Modal with sticky header and footer

In this example, we have a modal with a sticky header and footer. The middle part should scroll if the content is long enough. This example is inspired by a figure by Max Schmitt while researching the topic.

Consider the following CSS:

.modal {
  position: fixed;
  top: 0;
  left: 0;
  right: 0;
  height: 100vh;
}

Using 100vh will make the bottom part of the modal invisible. In the example, that means the footer won’t be visible and this will break the UX.

Here is how it looks with traditional and new viewport units on iOS:

..plus Chrome and Firefox on Android:

To solve that, we can either use svh or the dvh units.

Here is a video that shows the differences between dvh and vh.

Hero section

It’s a common case that we need to make the hero section height equal to the full viewport height minus the header height. Using the traditional vh for that case will fail in a browser that shrinks its UI on scrolls like iOS Safari and Firefox and Chrome for Android.

First, we need to make sure that the header height is fixed. I used min-height for that.

:root {
  --header-height: 60px;
}

.site-header {
  position: sticky;
  top: 0;
  min-height: var(--header-height, initial);
}

After that, I added min-height to the hero section and used calc().

.hero {
  min-height: calc(100svh - var(--header-height));
}

When using vh, the decorative element (in purple) isn’t visible at all. In fact, if you look closer, it’s blurred underneath the address bar UI in iOS Safari and cropped in Android browsers.

Here is a comparison between svh and vh on Safari iOS.

..plus Chrome and Firefox on Android:

See the following video and spot the difference between using svh and vh.

In such a case, using svh will solve the problem.

Is it possible to make dvh the default unit?

At first, the answer was “Yes, why not?”. Then I thought to myself, the dvh value will change as you scroll, so it might create a confusing experience when used for stuff like font-size.

h1 {
  font-size: calc(1rem + 5dvh);
}

Check out the following video and notice how the font-size change after the address bar UI is shrunk:

Demo

Be careful with the dvh viewport unit

The dynamic viewport unit might impact the performance of the page, as it will be a lot of work for the browser to recalculate the styles which the user is scrolling up or down.

I didn’t get the chance to do intensive performance testing, but I would be careful when using it. I hope that I will get the time to update on that here.

Other places where the new viewport units are useful

Those new viewport units might not be only about mobile browsers. In fact, you can browse the web on a TV today. Who knows what browser will come for a TV that has a UI that changes on scrolling and thus resize the viewport?

For example, here is the hero section example viewed on an Android TV:

It works perfectly and will keep working even if we have a dynamic UI.

Further resources

from Ahmad Shadeed Blog https://ishadeed.com/article/new-viewport-units/

Diversify Your Content Strategy Whiteboard Friday

Go from basic to a more advanced content strategy with Azeem in this Whiteboard Friday episode. Diversify your content strategy by creating the right content for your audience at the right time.

digital whiteboard image describing how to diversify your content strategy

Click on the whiteboard image above to open a high resolution version in a new tab!

Video Transcription

Hi, everyone. My name is Azeem. I’m the host of the “Azeem Digital Asks” podcast, and I’m here to show you a very brief whistle-stop tour of how you can diversify your content strategy on this Whiteboard Friday.

3 examples of where marketers get measurement wrong

3 examples of where marketers get measurement wrong

So I’m going to start off and make a very bold statement as a bald man and say that I think that we, as marketers, get measurement wrong, and I’m going to give you three examples here.

So if you are measuring brand awareness, for example, there are a number of things that you can measure, such as downloads, traffic, referrals, mentions. If you look at engagement as a key KPI, you’ll be looking at things like links, likes, comments, shares, retweets, all that sort of stuff. For lead gen, you’re typically looking at MQL, SQL, subscriptions, and call backs. So it’s three very quick examples of how I think we get measurement wrong.

Create an advanced content strategy

how to create an advanced content strategy

When it comes to our audience, I think we know what they want, but we don’t know how they want it, and I genuinely think that the internet is in a position now where hit and hope with just purely written content doesn’t work anymore. I genuinely think the internet has moved on. So I’m going to show you a very brief way of how you can take your content strategy from basic to even better to hopefully advanced, and that starts with this.

I think a lot of marketers are in the basic section, and that is where you have a particular topic, topic X as I’ve listed there, and that is your framework for the rest of your content. So if you were talking about trees, for example, you might have trees as your topic, and that would be the framework to branch out and create even more of topic around trees to move on.

That’s fine. That’s where I think a lot of marketers are. The better version would be looking at UA, universal analytics or multi-channel funnels, understanding what performs well, and creating more content of that based on where your audience is in the purchase journey. Then the advanced version would be looking into GA4, splitting out your top five markets as I’ve put there, understanding how they perform with a data-driven attribution model, and creating the right content for the audience at the right time, the Holy Grail of what we are trying to achieve here.

How to use this information

examples of advanced content

I’ll give you four examples of how you can actually use this information and take it away, and literally from tomorrow you can be able to improve your content strategy. So example 1 would be let’s say you have set up scroll tracking and YouTube view measurements on your GA4. Layer the two together.

You can understand how, for example, your audience in France will be engaging with your content in the sense of how far do they scroll down on a page and how much of your videos on your page they are watching. Example 1 would be a particular audience that scrolls not a lot, but engages with video quite a lot. In which case, I would introduce very early on in the page long-form videos.

You know what your audience wants. Don’t make them work for it. Don’t make them scroll down the page, because you know what they want. Make it as simple for your audience as possible. Example 2 would be the opposite, where you know your audience will scroll quite a lot, but you know that they won’t watch the videos that you put on the page. In which case, you can create highly-detailed content and then utilize remarketing to bring them back to your website.

The third example would be if you have an average scroll and an average video time, but a high ASD, which I have peddled as average settle duration. These are people that I call page hoppers. They’re very likely going to be in the research stage of their journey, of their purchase journey. So this is where you want to focus on your brand and why you stand out against the rest of your competition.

The fourth example would be people who don’t scroll and don’t watch your videos at all. I think in that situation you’ve very clearly got a disconnect, but there is still an opportunity for you to introduce short-form videos earlier on in the purchase journey. Utilize this information, find out which one of the four you sit in, and use that to create your content strategy in a more diverse way by including audio, snippets, video teases of varying different formats, and I guarantee you’ll be onto a winner and have more success with your content strategies moving forward.

I hope that in this very short video you’ve taken something away. You can find me on social media @AzeemDigital. If my SEO is any good, you should be able to type in “how can I contact Azeem” and you’ll come across my website. Very much enjoyed being here. Thank you for having me, and I’ll see you soon.

Video transcription by Speechpad.com

Azeem will be speaking at MozCon 2023 this August in Seattle! Join us for inspiring sessions with our incredible lineup of speakers.

We hope you’re as excited as we are for August 7th and 8th to hurry up and get here. And again, if you haven’t grabbed your ticket yet and need help making a case we have a handy template to convince your boss!

Register for MozCon

MozCon SEO conference 90%  sold out

from Moz https://moz.com/blog/whiteboard-friday-diversify-content-strategy

What is a PWA? Progressive Web Apps for Beginners

These days, everything is made possible with the help of mobile phones and applications.

Let’s say you need to order food – you can do so instantly via the company’s app. Maybe you need government services – the same thing applies. You can even get medical emergency dial services via an app.

There’s an app for everything – from banking to studying and from trading to shopping. Every business has an app, and even our governments have simplified their services into app form.

Hold on, building and maintaining an app is cumbersome, and it’s quite expensive for small businesses, so how do they manage?

Well it’s simple: with the help of advancements in technology there is an option that helps small businesses out. This option combines the features of an app with the technology used in web development to build affordable services for businesses – I’m talking about Progressive Web Apps.

Let’s dive in and get a better understanding of what PWAs are all about.

What is a Progressive Web App?

Progressive Web Applications (PWAs) are apps built with web technologies that we probably all know and love, like HTML, CSS, and JavaScript. But they have the feel and functionality of an actual native app. Wait a minute! Native Apps, what do we mean by this?

A Native App is a software application built in a specific programming language for a specific device platform, either IOS or Android.
PWAs are built with the capabilities like push notifications and the ability to work offline. They are also built on and enhanced with modern APIs which makes it easy to deliver improved capabilities along with reliability and the ability to install them on any device.

PWAs takes advantage of the huge web ecosystem this is inclusive of the plugins, and community and the relative ease of deploying and keeping a website contrary to a native application which is pretty difficult to develop. This means you can build a PWA quickly and easily.

With its popularity many companies have shifted into the product, I tend to believe that this is because of its ability to run on an android and iOS without much difference. Some good examples of top companies who have their products as PWAs include: Twitter, Pintrest, Uber, Tiktok, Spotify, Jumia (a leading e-commerce site in Africa) etc…

A common feature about this products is that they are all installable on your home screen, able to work offline from where you last left and offer a comparable experience and features to their native apps.

Just like when building a native mobile app there are some expectations that should be met to make a good product for consumer use, the same thing applies to PWAs. Let’s discuss what makes a good PWA.

Characteristics of PWAs

Below is what should be considered when developing a PWA:

Responsiveness

Different companies produce gadgets with different screen sizes, and as a developer it’s your responsibility to ensure all the different users enjoy the product regardless the device they are using. So it’s a good idea to make sure your app can be used on any screen size and it’s content is available at any view-port size.

Installable

Research has shown that users tend to engage more with installed apps compared to visiting the official sites. Having a PWA as your product gives the users the look, feel and engagement of a normal app.

Independent Connectivity

By keeping a user engaged to your app even while they are offline, provides a more consistent experience than dropping them back to a default offline page.

A good example to illustrate this will be that of a music app, your users should be able to access offline playback and listen to saved music even without internet connection. Another good example is twitter app, a user is able to go back a read through tweets which they might have missed.

Discoverability

Since most PWAs are converted websites, it is fair to make them discoverable on the search engines, this will help generate extra traffic to your app. This also acts as an advantage over native apps which can’t be discovered over the search engines.

Appearance

The appearance of the app should feel and look like that of a normal app, so be sure to include things like an app icon, this will help make it easily recognizable also things like splash screen will add the touch and feel of an app.

Cross Platform

PWAs are developed as web app first, which means that they need to work on all browsers/systems and not just a selected few. Users should be able to use them in any browser before they decide to install them.

So folks! there you have it, the general info about PWAs. Along the way you might have noticed occasionally a comparison between PWAs and Native App and this might have confused you a bit, Well let’s clear the airwaves by checking the comparison between the two to get a clear understanding.

Differences Between PWAs and Native Apps

Development Cost

PWAs are cheaper to develop compared to Native AppsWhen you’re developing a native app, you’ll have to learn a certain programming language and then build a version of the app for each type of device, Android and iOS. On the other hand you can choose to hire a experienced professional to do the work for you which will even turn out to be more costly.

Down the road, you will also need resources to maintain and update the app, which means lots of money and time is required.

In the case of a PWA, you can have a single codebase for the different platforms. It’s also time-saving since you will not need to develop it from scratch you can configure your current web site to fit.

And if you choose to hire developer it will only be one compared to native where you can hire up-to two depending on where you need your app.

Discoverability

Native apps cannot be indexed by the search engines, they can just be found through the App/Play store’s website. You can make your app more discoverable on the App/Play store by using App Store Optimization(ASO), but that’s another story.

Unlike native apps, PWAs work like websites so they can be indexed by search engines. This helps them rank better in search results.

Safety

Nowadays in order to run a website, it should be encrypted with a SSL certificate, this adds an extra layer of security. Now, as we already know PWAs are site converted into app which means they are more secure because they run on HTTPS. These are security protocols that allow safe exchange of data between client and server so that is doesn’t get tampered with.

To secure your native apps, you need to implement various security measures, like multi-factor authentication and so on.

Installation and Download

Native apps need to be downloaded and installed from an app store. This requires some commitment from the user to do it from start to finish. Users have to pass and check multiple permissions before installing an app.

On the other hand, PWAs don’t require any of those steps. From the browser you can bookmark it and add the app to your home screen with just a few taps.

Benefits of PWAs

A lot of organizations both private and public are switching to PWAs not only because they are cheap to develop but also because they offer greater engagement.
Now let’s look at a quick summary of the benefits of a PWA:

  • They are responsive and work with many different screen sizes.
  • They can run on multiple platforms and any device with a modern web browser.
  • They function just like normal Native Apps.
  • The updates are independent, you don’t need to visit the play store for an update.
  • They’re built with common web technologies.
  • They’re fast and lightweight.
  • They work offline unlike other sites.
  • They are discoverable via search engine.
  • They are easily installable.
  • Low maintenance cost.

Requirements to Get Started with PWA Development

It does not take much to get started building a PWA. You just need a few things and you are good to go.

Tools
The best known technology stack to develop PWAs is AngularJS. Speaking of Angular, here is a resourceful guide on how you can convert your already existing Angular app into PWA. Others stacks include ReactJS and Polymer.

HTTPS
You will need a server with a HTTPS connection. This makes sure your user’s data is secure. It adds an extra layer of security to you site.

Application Shell
It provides a good first impression when your app loads. In simpler words this is what the user sees when they interact with your app for the first time.

Service workers
This is one of the key technologies behind PWAs. They help support your app work offline, and they perform advanced caching and run background tasks. Service workers can complete tasks even when your PWA is not running.Some other functions associated with Service Worker include:

  • Sending push notification
  • Badging icons
  • Running background fetch tasks etc…

Manifest file
This is a JSON file that is created with a Web App Manifest Generator. This file contains the information that tells how your PWA should appear and function. It allows you to determine the name, description, icon, colors and other features of your PWA. Here’s an example of a manifest file:

{
"short_name": "DevBlogger",  
"name": "DevBlogger",  
"description": "All dev stories under one roof",
"theme_color": "#eb5252",  
"background_color": "#000000",  
"display": "fullscreen",  
"Scope": "/",  "orientation": "portrait",  
"icons": [    
    {
        "src": "images/android/android-launchericon-48-48.png",      
        "type": "image/png",      
        "sizes": "48x48"
    },
    {      
        "src": "images/android/android-launchericon-96-96.png",
        "type": "image/png",      
        "sizes": "96x96"    
    },    
    {      
        "src": "images/android/android-launchericon-192-192.png",
        "type": "image/png",      
        "sizes": "192x192"    
    }  
   ],  
       "start_url": "index.html?utm_source=homescreen"
  }
  • Audit your App
    This is possible using the Google Lighthouse tool. Google Lighthouse is a open-source software that anyone can use on any webpage. Google is a big champion of PWAs and pushes them as the future of the web. You can use Lighthouse to help you see how fast, accessible, and SEO readiness your PWA is.

How to Build a PWA

By following the steps below, you can easily create a fully functional PWA that offers an mazing user experience across all devices.

Step 1 – Plan your app

Before diving into development, you should consider the goals of your PWA, what features you want to include, priorities and user experience. You can create first design concepts and wireframes for the app to visualize the structure and layout.
In most scenarios, this is often referred to as a ‘discovery phase’. You get the opportunity to ideate and gather user and stakeholder feedback as well as considering the functionalities of your to be product.

Step 2 – Designing the User Interface

After getting everything right from planning, you can now proceed to designing the UI of your app. During this stage consider things like responsiveness, compatibility with different platforms etc.. Be sure to capture all details that are crucial to the user including their interaction and engagement during usage.

Step 3 – Developing the Front-End

Using the web technologies that is HTML, CSS, JavaScript and frameworks like Angular. React or Vue.js develop a visually appealing interface for the users. And always remember they key principle in development using this stack implement a mobile first approach while ensuring responsiveness for larger screens too.

Step 4 – Implementing Service Workers

As mentioned previously, service workers are a key component of PWAs. They are JavaScript files that run in the background, enabling offline functionality, push notifications, and caching. To make sure your PWA works to its fullest potential, you’ll need to register and implement a service worker. The way on how you can do this massively depends on which framework you are using.

Step 5 – Adding Push Notifications

Leverage the Push API and service workers to implement push notifications. Obtain the necessary user consent and use a push notification service to send notifications to users.

Step 6 – Optimizing Performance

Optimization is a very important step in development in general. This is how you provide a seamless experience to your users. by ensuring you reduce loading times. by leveraging techniques such as code splitting and caching we should be able to achieve a fast and efficient operation for our PWA.

Step 7 – Testing and Debugging

Test your PWA on different devices, browsers and network condition to be sure that it meets the objective. Also be sure to gather user feedback and make necessary improvements when necessary.

Resources to Get Started with PWA Development

If you want to learn and move with the trend, finding resources to help you might be a bit tedious, to help you get started here are some of the best resources listed for you:
Online Tutorials and Guides

Documentation and Reference Materials

PWA Development Tools

Conclusion

Keeping in mind that PWAs are new to the industry and haven’t yet been fully utilized, they can be a great addition to add to your toolkit.

With the latest technologies and the right tools getting started with PWAs can ultimately increase sales and monetary gain for your product either as an individual or organization. With it’s many features including they are fast, able to work offline, and also they perform like normal native apps. This offers your users a great experience and keeps them satisfied.

If you have read this far I really appreciate it.

Enjoy Coding ❤.

from freeCodeCamp https://www.freecodecamp.org/news/what-are-progressive-web-apps/

Getting started with native design tokens in Figma

Getting started with native design tokens in Figma

How to use Figma variables to implement different design token strategies

What are design tokens anyway?

Design tokens are named entities that store raw, indivisible design values like colors or pixel sizes. They are the core pieces of a design system. Stored in a technology agnostic format they can be transformed for use on any platform, replacing hard-coded values. ~ Design tokens — What are they & how will they help you?¹

Your text and color styles, shadows, etc. can be represented as design tokens. The important part is to share design tokens between all parts of the product team. The goal is for design and development to use the same tokens from the same source of truth.

With design tokens, both hand-off and global changes get easier. If set up correctly, any design token change must be done at one place only, the design token source of truth. And this change will still affect the entire system, e.g. changing the brand color everywhere all at once.

Variables vs. design tokens

Figma introduced variables². They work well as design tokens, but they are more, they can also be used for prototyping³.

This article focuses on using the Figma variables feature to implement design tokens. I may refer to it as either variables or design tokens.

Variables can be one of four types: solid color, number, string, and Boolean.

Color

Color variables can be applied as fills, text fills or stroke color².

Currently, variables can’t be used in effects like shadows, in layout grids or in gradients.

Color variables can only be solid colors. Gradients are not supported, and you can’t change the opacity of a color variable.

You can use color variable in styles, but you can’t change its opacity. You can, however, add another layer on top of the color variable. If you created a style with a variable and add a 60% white on top, you could create a muted version of your variable.

Number

Number variables can be used as text, for sizes, paddings, gaps, and border radii². This covers most things you would need design tokens for.

The three things currently not supported are border-width, opacities, and effects. This means you can’t make styles or components semi-transparent using a variable. And you can’t define shadows with number variables from your design system. 😢

String / Text

Text can be used in text layers and for variant names². They are useful for translations or to tokenize UI copy.

To translate your entire app, create a collection with a mode for each language you support. Now you can switch the language of your entire app from a dropdown.

You can also use text strings to switch between variants. For example, you could set a text string to mobile or desktop. With this you can change a component variant between mobile or desktop. This is great if instead of different token values, you use different tokens on the two platforms.

Boolean

Boolean variables don’t seem to have a lot of use for design tokens.

They are great at toggling things within component depending on specific modes though. For example, to show or hide icons depending on the screen size.

At the time of writing Boolean variables can only be used for Boolean variants. Not for Boolean component properties.

Design tokens vs. styles

Styles in Figma are similar to design tokens. But they have some capabilities that design tokens are lacking. For example, design tokens don’t support blend modes or mutations like alpha changes. But they can be referenced in other tokens or even styles⁴. This is not possible using styles.

While some of those functionalities may come to design tokens at some point, many may not. And this is a good thing. Design tokens are purposely kept close to the w3c design token specs⁵.

Styles however are a Figma primitive, similar to classes in CSS. Maybe styles will get more capabilities in the future. For example, being able to set the opacity of a referenced variable.

When to use tokens?

At the moment styles have little unique use cases. I suggest using variables for all design tokens that can be represented as variables.

For now, only simple tokens can be created, composite tokens are not available in Figma.

You will also not be able to replicate tokens if you stray away from the w3c design token draft⁶ ⁵.

That being said, tokens should still most of your needs for color, and size base values.

When to use styles?

Use styles for anything that tokens can’t do. For example, you can use styles to replicate composite tokens⁶ for typography, shadows, or borders.

Typographic tokens are supposed to come to Figma at the end of 2023. But if this happens there are still borders and shadows left.

Design tokens in Figma libraries

Design tokens can be shared as part of a library just like styles. They also have the same limitations. This means they don’t get passed on through files.

Imagine you have three files: design tokens, components, app

  • The components file imports design tokens and uses it on the components, which get exported
  • app imports components, this will allow it to use the exported components. But the design tokens will NOT be available.
  • app needs to additionally import the design tokens file to be able to use the design tokens directly.

from Design Systems on Medium https://uxdesign.cc/getting-started-with-native-design-tokens-in-figma-5d9c5fcdd9f7

Guillermo del Toro: I Dont Fear Artificial Intelligence. I Fear Natural Stupidity – Yahoo Canada Shine On

“Guillermo del Toro: Crafting Pinocchio,” the roving, ever-expanding, 8,000-square-foot exhibit dedicated to the art of making Guillermo del Toro and co-director Mark Gustafson’s meticulous Oscar-winning stop-motion film, has made its way from New York’s Museum of Modern Art to the Portland Art Museum in Oregon.

This past weekend, the three-time Academy Award winner came in person to the Rose City to accept a Cinema Unbound award from PAM CUT (the Portland Art Museum’s new-media-focused Center for an Untold Tomorrow) and later sit down for a relaxed chat with PAM CUT’s curator Amy Dotson. As usual, the director sounded off unguardedly about a range of topics, from struggling to get even his own passion projects greenlit to his commitment to animation and the threat of artificial intelligence looming over the creative community.

More from IndieWire

“Since I was a kid, all I wanted to do was monsters and stop-motion animation, and that’s what I’m doing, so why the fuck should I not do it?” the “Shape of Water” and “Pan’s Labyrinth” director laughed. His next film — also for “Pinocchio” distributor Netflix — is an animated adaptation of Kazuo Ishiguro’s fantasy novel “The Buried Giant.” The film, which is two years out from production, utilizes stop-motion but is certainly less monster-based in centering on a version of England where King Arthur actually existed. “Pinocchio” animators ShadowMachine back the project.

Del Toro, who recently said that five of his projects have been turned down by studios this year so far, told the Portland museum audience, “But we keep going. With ShadowMachine, Mark, everybody, when we were involved with ‘Pinocchio,’ you have no idea how we were involved in meeting after meeting after meeting, and hearing no … if you have the conviction that it must be made, ‘no’ is a ‘yes’ waiting to happen, and you just have to say, alright, your loss. You literally have to believe that. You should not question your material. You should not say, ‘What am I doing wrong?’ The last time somebody passed on ‘The Buried Giant,’ I wrote an email, and I said, it’s easier for me to do it than to argue with you. I’ll just show it to you. And you’ll see I was right. Or not. There’s a lot of things to do, but it’s very important to have that certainty.”

Del Toro said that as he continues to develop movies like “The Buried Giant,” he still watches “three movies a day” for inspiration, and often that means rewatching. “If you see all ‘All About Eve’ when you’re 15, and you see ‘All About Eve’ when you’re 40, you see two entirely different movies.”

When asked about his optimism for the state of creativity in general right now, del Toro said, “I remain enthusiastic but skeptical, meaning I know we are a horrible human race, but we do great stuff, and many people are great. What brings me hope and makes me think that it’s worth it? The next generation because we undoubtedly fucked it up… in that hope, it can only come with your full support,” motioning to the people in the standing-room-only crowd at the Portland Art Museum.

He added, “When I see people who are fearless, I get inspired and I like it, and I like the possibilities when people talk about now, and how it’s all dire, and [how] people are afraid of artificial intelligence… I don’t fear artificial intelligence, I fear natural stupidity. Any intelligence in this world is artificial. When I look at the people coming into the art scene and how they are in spite of all the things that are hardships and all the things weighing against it, they love art, and that’s what makes my spirit sing.”

While praising the next generation, del Toro also cautioned that ages 14 through 24 are “hell,” adding, “Since I was seven, I’ve been looking forward to being old. The real crime in our existence is to look for perfection. We should all aspire to imperfection.”

The advice for his eight-year-old self he wishes he could tell now? “Don’t eat that fucking cupcake.”

Best of IndieWire

Sign up for Indiewire’s Newsletter. For the latest news, follow us on Facebook, Twitter, and Instagram.

Click here to read the full article.

from “artificial intelligence” – Google News https://www.indiewire.com/news/breaking-news/guillermo-del-toro-ai-the-buried-giant-pinocchio-1234878950/

Design syntropy

A concept embracing the evolving nature of design work and the unique roles of design artifacts and design systems

Photo by Vincent Botta on Unsplash

There’s a concept that I’ve come to refer to as “design syntropy.” It’s related to the creation and use of design artifacts — all those things we create as designers and product managers and engineers as we shape a product or concept.

Most design artifacts are temporary sources of truth. They help guide our thinking, collaboration, and decisions while a project is underway. But their “truthfulness” has a short shelf life.

I’ve often wondered and have been asked which design artifact is the “source of truth” for something in a product, and honestly, I don’t think that’s a realistic or even desirable goal.

I think design has a quality of syntropy, an antonym to entropy. Unlike entropy, which suggests a natural decline into chaos and disorder, syntropy represents the idea of order, evolution, and the ability to bring together ever-increasing complexity to create something new.

In product design, our work creates ripple effects, influencing areas of the product that we might not have directly touched. This is design syntropy. A design artifact left untouched doesn’t descend into chaos, but it may slowly drift out of alignment with the reality of the product, while the thing it represents evolves on its own, indirectly influenced by changes elsewhere in the product.

Design artifacts represent a source of truth only in a moment of time.

This isn’t a problem to solve, nor a flaw to correct — it’s just the nature of design work. Design artifacts are a record of a project’s journey, showing us the thought processes, tradeoffs, and context of decisions. Great design artifacts multiply collaborative impact across a team and help to build momentum. Ultimately, they’re a form of documentation — and in that capacity, that’s why even simple sketches and breadboards convey immense value.

In the long run, the real source of truth is the design system. A good design system is enduring even as different projects come and go. The design system needs to stay aligned with the product, and it needs active maintenance, documentation, and champions.

I don’t like the idea of design artifacts automatically updating to mirror changes in the design system. This might sound appealing on paper, but in practice, it tends to break the value of the artifact as a record of the project’s context and history.

Design syntropy as a concept embraces the evolving nature of design work and the unique roles of design artifacts and design systems.

Originally published at quinnkeast.com/writing/design-syntropy.

Quinn Keast is a partner & designer at Caribou, and also product designer at Airplane.


Design syntropy was originally published in UX Collective on Medium, where people are continuing the conversation by highlighting and responding to this story.

from UX Collective – Medium https://uxdesign.cc/design-syntropy-334080295f47?source=rss—-138adf9c44c—4

Design Tokens

Design tokens are essential for creating and maintaining a coherent design system. They provide many benefits such as consistency…

Continue reading on Medium »

from Design Systems on Medium https://medium.com/@hoomanahmadi/design-tokens-8a726afe427e

The Complete Beginners Guide To Autonomous Agents

Ok, let’s start with what you already know.

Artificial intelligence can be used to complete very specific tasks, such as recommending content, writing copy, answering questions, and even generating photographs indistinguishable from real life.

You tell the AI to complete the one task, it completes the one task. Simple.

But what if you don’t want to have to come up with all of the tasks for the AI to do? What if you want a teammate rather than just a tool? What if you want the AI to think for itself?

Like really think for itself.

Imagine you made an AI that you could give an objective to, even something as vague as “Create the best ice cream in the world”, and the AI would come up with a todo list, do the todos, add new todos based on it’s progress, and then continue this process until the objective was met.

This is exactly what “Autonomous Agents” do, and they are the fastest growing trend amongst AI developers, yet most people don’t know about them.

(At the time of writing this article, no major publications have written about autonomous agents, and since publishing, only a few have covered it, so if you’re reading this… you’re very early.)

What are autonomous agents? Why are they such a big opportunity? How do they work? What does this look like in the future? How can I build or use one? How can I meet other people interested in autonomous agents?

These are the questions I’m going to answer for you right now.

“[Intelligent] autonomous agents are the natural endpoint of automation in general. In principle, an agent could be used to automate any other process. Once these agents become highly sophisticated and reliable, it is easy to imagine an exponential growth in automation across fields and industries.”

p.s. I am CEO and co-founder of Octane AI, where for seven years we have been building conversational AI products, and are more recently building generative AI and autonomous agent solutions for brands. In 2016 I predicted that around now chatbot interfaces would take off and start to replace standard website UI, and now over 100 million people use ChatGPT and websites like it. I am now similarly predicting that autonomous agents will be widely adopted in the future, but this prediction won’t take seven years to come true, it will happen blazingly fast.

p.p.s. After writing this article I showed the draft to 125 of the smartest and most interesting people I know, including Emad Mostaque (Founder of Stability AI), Tony Hu (Former Acting Head of Emerging Technology for the FBI, and founder of Bondoo AI), Troy Carter (Lady Gaga’s ex Manager), Sahil Lavingia (Founder of Gumroad), Elizabeth Yin (Co-Founder of Hustlefund VC), Hugh Howey (Author of Wool), Chris Yeh (Author of Blitzscaling), experts from NVIDIA, Meta, investors like Ryan Hoover (creator of Product Hunt) and Erica Brescia (Manager Director of Redpoint Ventures, prior Github COO), and many many more. Their thoughts and opinions are sprinkled throughout, they will give you unique insights shared with the world for the first time.

What Are Autonomous Agents?

Autonomous agents are programs, powered by AI, that when given an objective are able to create tasks for themselves, complete tasks, create new tasks, reprioritize their task list, complete the new top task, and loop until their objective is reached.

Read that description above one more time, because while it is simple, it is wild.

“The future of autonomous agents looks like everybody becoming a manager.”

Yohei Nakajima, creator of BabyAGI

Autonomous agents can be designed to do any number of things, from managing a social media account, investing in the market, to coming up with the best children’s book.

And these are, like, real? These exist right now?”

Yes, I know it sounds like science fiction, but these are functioning and real. If you can code you can make one in just a few minutes. And it is only the beginning.

“Humans waste inordinate amounts of time doing work that is tedious and manual when it could be done by computers and free them up for more creative pursuits, or to do things that only humans can currently do. Autonomous agents will enable people to get so much more done in so much less time, and – hopefully – spend much less time in front of screens over time!”

The programming techniques and the AI needed to power autonomous agents are real and extremely new. There are many open source projects, like AutoGPT, BabyAGI, and Microsoft’s Jarvis, that are trending on Github and within AI communities and departments.

In the first two weeks of the creation of open sourced autonomous agent code bases, almost 100,000 developers are building autonomous agents, improving them, and pushing them to their limits, and thats only in the first few weeks of these concepts being invented. The number of developers working with this technology is growing at an increasingly faster rate.

“AI agents will be everywhere. Billion-dollar companies will come from a small team that deploys ai agents.”

It has grown larger than long time popular codebases including laravel, bitcoin, django, and pytorch.

Auto-GPT Github popularity increasing exponentially, faster than any codebase in history

This is not science fiction. Many think these autonomous agents are the beginning of true Artificial General Intelligence, or commonly referred to as “AGI”, which is a term used to describe an AI that has gained sentience and become “alive”.

“Autonomous agents may end up commoditizing all applications of factual knowledge. If access to factual knowledge also becomes universal, then human qualities like creativity, emotion, and strategic vision will become even more distinctive. But it is also possible that knowledge becomes increasingly proprietary, as individuals and companies try to gain economic advantage in a world where applications of factual knowledge are commoditized, and the collective knowledge of humanity begins to stagnate.”

Check out this autonomous agent that was just released from HyperWrite, you can see it living in the browser and helping you order a pizza.

You just say “order a large plain pizza from Dominos to One Vanderbilt” and it just… does it.

HyperWrite’s autonomous agent controlling the browser to order pizza

Or, maybe even more impressive, check out this experiment done in collaboration between Stanford and Google where they created a virtual town of 25 autonomous agents, and told one of them to plan a Valentine’s day party.

The simulated people went about their days, talking to each other, forming new memories, and eventually most of them heard about, and showed up to, the Valentine’s day party.

From the research paper “Generative Agents: Interactive Simulacra of Human Behavior”

Ok, uh, crazy… So autonomous agents are real… And you just tell it what it’s goal is and then after that it manages itself forever?

You just give it the one objective, and the autonomous agent does the rest. 

Just like a really good employee or teammate.

Although, if you wanted to, you could also design the autonomous agent to check in with you at certain key decision making moments so that you could momentarily collaborate on their work.

“It is "primitive AGI". It is remarkable that simply wrapping an LLM inside a loop gets you an autonomous agent that can reason, plan, think, remember, learn – all on its own. It demonstrates the untapped power and flexibility of what LLMs can do if wrapped in the right structures and prompts. The entire concept is less than month old so I can’t wait to see how increasingly sophisticated agents built off of increasingly more capable LLMs impact the world.”

But what can autonomous agents do, Matt? Like when you say they complete tasks, what the heck do you mean by that?

In addition to analyzing their objective, and coming up with tasks, autonomous agents can have a range of abilities that can enable them to complete any digital task a human could, such as:

  • Access to browsing the internet and using apps

  • Long-term and short-term memory

  • Control of your computer

  • Access to a credit card or other form of payment

  • Access to large language models (LLMs) like GPT for analysis, summarization, opinion, and answers.

Also, these autonomous agents will come in all shapes and sizes. Some will operate behind the scenes where the user is unaware of what they are doing, while some will be visible, like in the example above, where the user can follow along with each “thought” the AI has.

“Autonomous agents will allow everyone to live like a head of state! Need something done? Just ask, and your agents will take care of the rest. Never again will you have to waste brainpower on the routine or mundane.”

Matt, I’m reading what you’re writing, I think I know what you are saying, but can you write out an example in plain english so I can be sure I understand.

Here is a super simple example of how an autonomous agent could work.

Let’s say that there is an autonomous agent that helps with research, and we want a summary of the latest news about a certain topic, let’s say “News about Twitter”

  • We tell the agent “Your objective is to find out the recent news about Twitter and then send me a summary”.

  • So the agent looks at the objective, uses an AI like OpenAI’s GPT-4 which allows it to understand what it is reading, and it comes up with it’s first task. “Task: Search google for news related to Twitter”.

  • The agent then searches google for Twitter news, finds the top articles, and comes back with a list of links. The first task is complete.

  • Now the agent looks back at its main objective (to find out the recent news about Twitter and then send a summary) and at what it just completed (got a bunch of links of news about Twitter) and decides what its next tasks need to be.

  • It comes up with two new tasks. 1) Write a summary of the news. 2) Read the contents of the news links found via google.

  • Now the agent stops for a second before continuing, it needs to make sure that these tasks are in the right order. Should it really be writing the summary first? No, it determines that the top priority is to read the contents of the news links found via google.

  • The agent reads the content from the articles, and then once again comes back to the to do list. It thinks to add a new task to summarize the content but that task is already on the todo list so it doesn’t add it.

  • The agent checks the todo list, the only item left is to summarize the content it read, so it does that. It sends you the summary just like you asked.

Here is a diagram showing how this works.

From Yojei Nkajima’s BabyAGI

And keep in mind that this is the very beginning of this new paradigm. It’s not perfect, it hasn’t taken over the world yet, but the concept is frighteningly powerful and with increased development and experimentation will quickly find it’s way into our daily lives.

“This will soon transform many industries. It will be a lot easier for people to do many things at once with the use of Autonomous Agents. Just give it a task, and it will complete it. Such a powerful concept so far…”

So now you understand at a high level what an autonomous agent is, but why exactly are these such a big opportunity?

“If we’re able to get the information we need faster, will this allow us to free up time to dedicate to thinking and vs. doing? Will even better and more creative ideas surge as a consequence of investing less time on tasks that can be carried by this AI agent?

Why Autonomous Agents Are Such A Big Opportunity

It’s pretty clear that soon you won’t only have the options of hiring humans as employees, you will have the ability to hire AIs in the form of autonomous agents.

“In the mid-term, I believe you’re going to see a huge rise in 1-2 people startups that use a combination of AutoGPTs and tools like ChatGPT. And they’ll be able to make the kind of progress you’d previously had expected from a 100 person startup. Long-term I believe that most work can and will be replaced by AutoGPTs.”

And they are not going to be nearly as expensive as people are, they won’t sleep, they won’t quit, and they will work extremely efficiently.

“Part of the thesis when I started Product Hunt in 2013 was a belief that the barrier to build software products would continue to lower, enabling smaller teams (or a single person) to build more and faster than ever before. This has never been more true today, accelerated by AI and autonomous agents. This introduces anxiety for some and opportunity for others that leverage this tech to scale their ideas with fewer people and capital required. In the end, consumers will greatly benefit through increased competition and experimentation of new solutions to their problems.”

These autonomous agents will exist in every industry and for every task imaginable.

These are just a handful of examples. Let your imagination run wild.

The list can go on and on. Anything a person could do, an autonomous agent will (eventually, but soon, and in some cases already) be able to do better.

“The music industry has imposed too many unnecessary layers that sit between an artist and success. Those layers cost an artist close to 35% of their net income. Autonomous Agents will be able to build and execute marketing strategies, engage with fans, build communities, route tours, book venues, and negotiate contracts. Saving the artist money and time.”

So what do you do with this information?

There are two very real opportunities.

  1. You create autonomous agents and make them available for others to hire.

  2. You hire autonomous agents and can now afford to be more productive in your personal life, or in business.

“Autonomous Agents are the next wave — not just in tech, but in business at large. I predict that within 10 years, there will be multiple billion-dollar companies run entirely by autonomous agents. It is inevitable.”

Imagine a world where one person builds a company with only autonomous agents on their team. Within your lifetime you will likely see a one person team do this and reach a market cap of over a billion dollars, something it usually takes many many people working together to accomplish.

“Personalization at scale is going to be a very interesting use case. You will be able to put on auto-pilot multi-step processes that humans do today that involves generating personalized images, videos, websites, emails or even calls at scale. One use case that has sparked a lot of interest is sales prospecting”

Right now in the early days there will be a period of time where early movers, either on making autonomous agents, or using them, will have a huge advantage against competition that is not yet leveraging these systems.

“In the near future, I expect to see lunch meetings, phone calls, and interviews appear on my calendar without my involvement. My agents and their agents will have made it happen, taking care of all the details. I just need to be there.”

By reading this article you are already ahead of 99% of the world.

Let’s dive into more detail on how these autonomous agents work.

“Autonomous agents have the potential to supercharge the output of smaller content creators and community members, especially those with creative imaginations. This will be a boon for many Web3 projects.”

How Autonomous Agents Work

You’ve already read over a high level of how autonomous agents work, but I thought it would be helpful to give you one version of an overall framework, as well as break down a couple examples of autonomous agents step by step.

“I see AI as a whole right now and we are in the building blocks that will evolve to become artificial intelligence assistants like we have seen in the movies — like Jarvis from Ironman or TARS from Interstellar.

Right now is a time to build out the frameworks because the AI itself is still improving. The answers might not be that good. It might have errors. But just looking at how much has improved with respect to AI in the last 6 months, I think we can barely imagine how things will be in the next 1-2 years. So this is about experimenting early, fast, and skating where the puck is heading.”

First, here a generalized framework for an autonomous agent:

  1. Initialize Goal: Define the objective for the AI.

  2. Task Creation: The AI checks its memory for the last X tasks completed (if any), and then uses it’s objective, and the context of it’s recently completed tasks, to generate a list of new tasks.

  3. Task Execution: The AI executes the tasks autonomously.

  4. Memory Storage: The task and executed results are stored in a vector database.

  5. Feedback Collection: The AI collects feedback on the completed task, either in the form external data or internal dialogue from the AI. This feedback will be used to inform the next iteration of the Adaptive Process Loop.

  6. New Task Generation: The AI generates new tasks based on the collected feedback and internal dialogue.

  7. Task Prioritization: The AI reprioritizes the task list by reviewing it’s objective and looking at the last task completed.

  8. Task Selection: The AI selects the top task from the prioritized list, and proceeds to execute them as described in step 3.

  9. Iteration: The AI repeats steps 4 through 8 in a continuous loop, allowing the system to adapt and evolve based on new information, feedback, and changing requirements.

But, now lets apply it to a few different use cases I decided to extrapolate on.

“Autonomous agents are truly captivating to me because they embody the ultimate productivity booster. As someone who highly values automation for tedious or repetitive tasks, I find that these agents have the potential to revolutionize the way we work, allowing us to direct our mental energy towards more meaningful pursuits.”

Gabriel Menezes, Director of Engineering at Octane AI

Example #1: Social Media Manager Autonomous Agent

Let’s say that instead of hiring a social media manager to manage your social media accounts, instead you wanted an autonomous agent to do everything for you at a fraction of the cost and with round-the-clock intelligence.

“This is beyond just virtual assistants. This is a revolution in accelerating all work, research, and even play online. Anything you can do online that takes hours, days, months can now be completed in the background in minutes.”

Here’s what a framework for that autonomous agent might look like.

  1. Initialize Goal: Set up the initial parameters, such as target audience, social media platforms, content categories, and posting frequency.

  2. Data Collection: Collect data on past social media posts, user interactions, and platform-specific trends. This could include likes, shares, comments, and other engagement metrics.

  3. Content Analysis: Analyze the collected data to identify patterns, popular topics, hashtags, and influencers relevant to your target audience. This step could involve natural language processing and machine learning techniques to understand the content and its context.

  4. Content Creation: Based on the analysis, generate content ideas and create social media posts tailored to the platform and audience preferences. This could involve using AI-generated text, images, or videos, as well as incorporating user-generated content or curated content from other sources.

  5. Scheduling: Determine the optimal time to post each piece of content based on platform-specific trends, audience activity, and desired frequency. Schedule the posts accordingly.

  6. Performance Monitoring: Track the performance of each post in terms of engagement metrics, such as likes, shares, comments, and click-through rates. Gather user feedback, if possible, to further refine the understanding of audience preferences.

  7. Iteration and Improvement: Analyze the performance data and user feedback to identify areas for improvement. Update the content strategy, creation, and scheduling processes to incorporate these insights. Iterate through steps 2–7 to continuously refine the social media management system and improve its effectiveness over time.

“People will own personal agents which communicate with agents owned by other people and businesses. Most computing devices will primarily serve as communication devices for speaking with agents.”

By incorporating this loop-type system in social media management, you can create a dynamic and adaptive strategy that evolves with your audience’s preferences and the constantly changing social media landscape. This will help to increase engagement, reach, and overall effectiveness of your social media efforts.

“Another use case for an autonomous agent that excites me is its application in the realm of music composition. By leveraging the power of AI-driven algorithms, these agents can analyze my personal preferences, favorite genres, and even specific musical elements that resonate with me. They can then generate original melodies, harmonies, and rhythms, effectively co-creating music alongside me. This creative collaboration has the potential to broaden my musical horizons, enabling me to explore new styles and genres I may not have considered before. Moreover, the autonomous agent can provide valuable feedback on my compositions and offer suggestions for improvement, nurturing my growth as a musician. The fusion of AI and human creativity in the music composition process can lead to innovative and unique results, expanding the boundaries of artistic expression.”

Example #2: Political Campaign Manager Autonomous Agent

What if you are running for political office and you want to leverage an intelligent and never-sleeping assistant to help you win?

“I’m excited about agents that do work that’s not necessarily hard to do but just require some time and effort for example things like booking flights I would love to outsource to an agent”

This is what an autonomous agent that helps you win an election might look like.

  1. Initialize Goal: Win the election by securing the majority of votes.

  2. Data Collection: Gather data on voters, demographics, key issues, campaign messaging, and other relevant information.

  3. Context Analysis: Analyze the collected data to identify trends, opportunities, and challenges. Refine the initial goal into specific subgoals based on this analysis, such as targeting undecided voters, increasing voter turnout in key areas, or improving campaign messaging on particular issues.

  4. Task Generation: Generate tasks related to the refined subgoals, such as planning voter outreach events, creating targeted advertisements, or developing policy proposals.

  5. Task Prioritization: Rank tasks based on their potential impact on achieving the subgoals and the overall goal of winning the election.

  6. Task Execution: Execute the highest priority tasks, allocating resources and assigning team members as needed.

  7. Performance Monitoring: Assess the effectiveness of completed tasks by tracking key performance indicators like voter engagement, public opinion, and fundraising metrics. Evaluate the success of individual tasks and overall campaign progress toward the subgoals and initial goal.

  8. Iteration and Improvement: Analyze the performance data to identify areas for improvement. Update the campaign strategy to incorporate these insights. Iterate through steps 2–8 to continuously refine the political campaign management system and improve its effectiveness over time.

“I’m most excited by the recursive self-cloning capability. The AI agent can create a copy of itself, pass on task directives, and start talking with its own sibling to get the job done. It is quite a remarkable but alien emergent ability.”

At first one candidate might use an autonomous agent and have a huge advantage over everyone, but then imagine what this looks like once every candidate has one… or many.

“I don’t think everyone will use autonomous agents. They will be everywhere but as AI becomes ubiquitous there will be a revival of 100% human work. Many people will rediscover pen and paper, want human only made art… We will see many products and creations that will advertise "only made by humans". It should become a very popular label very soon. The more technology grows the more I am enjoying myself long periods of completely offline time, soon also "off AI" time.”

Example #3: Math Tutor Autonomous Agent

Here is an autonomous agent that is designed to teach a child math.

“This is a breakthrough paradigm that has a LOT of room for exploration. Although early experiments have limited agents to search queries, we’re going to see a wide range of research and side projects arming autonomous agents with new batches of tools. Each set of tools will significantly expand the potential use cases.”

  1. Initialize Goal: Identify the child’s current math skill level and set a personalized learning path to help them improve.

  2. Data Collection: Gather information on the child’s learning style, progress, and performance through assessments, interactions, and feedback.

  3. Context Analysis: Analyze the collected data to identify strengths, weaknesses, and learning preferences, as well as any external factors influencing the child’s progress.

  4. Task Generation: Generate tutoring tasks based on the child’s needs and learning path, such as selecting appropriate exercises, providing explanations, or offering real-life examples and applications.

  5. Task Prioritization: Rank tutoring tasks based on their potential impact on the child’s learning and skill development, ensuring a balance between challenge and engagement.

  6. Task Execution: Execute the highest priority tasks, adapting the tutoring approach and content delivery as needed to maximize the child’s learning and engagement.

  7. Performance Monitoring: Assess the effectiveness of the tutoring by tracking key performance indicators (KPIs) such as progress toward learning goals, improvement in math skills, and the child’s engagement and satisfaction.

  8. Feedback Loop: Continuously monitor the child’s performance and update the context analysis, task generation, and task prioritization steps based on new data and insights. Adjust the initial goal and learning path as necessary to better support the child’s math skill development.

  9. Iteration and Improvement: Analyze the child’s performance and update the context analysis, task generation, and task prioritization steps based on new data and insights. Adjust the initial goal and learning path as necessary to better support the child’s math skill development. Iterate through steps 2–9 to continuously refine the political campaign management system and improve its effectiveness over time.

This autonomous agent loop type system outlines a process for an educational math tutor to adaptively support and guide a child’s learning experience, focusing on continuous improvement and personalization based on the child’s needs and progress.

“Just like there will be numerous models of all sizes in the future, you’ll have multiple agents for different facets of your life: an agent for work, an agent for your family/home life, an agent for self-improvement, all working in tandem with other agents. Automating mundane tasks or giving you professional superpowers are the first obvious use cases, but your digital twin will be capable of so much – perhaps even going on dates without your involvement to assess fit, forever eliminating the bad first date.”

Vivian Cheng, Principal at CRV

The Future Of Autonomous Agents

Right now humanity is in the very beginning of developing autonomous agents. We’re poking around, breaking things, experimenting, making bad things, making good things.

Autonomous agents will bring your ideas to life simply by requesting their assistance. These agents can serve as friends, colleagues, and collaborators, affording you an abundance of leisure time. I’m curious to know, how would you choose to spend this newfound freedom?”

Barely any commercialized products have even been released, everyone is still in development mode.

But soon, that is going to change. Autonomous agents are going to start showing up all over the place until one day it will be incredibly strange for someone to not have one, or multiple, autonomous agents helping them out at any given time.

“Rather than focus on replacing people’s work, focus on augmenting what they can do. Making something "smart" used to mean making its data available via api. The next generation of making something smart will be to ask how that product can better assist you. As an example, a "smart" email address might be able to take action in interesting ways based on your preferences. If you’re a big shopper, maybe it monitors emails for when an item you’re interested in goes on sale, price compares, or even negotiates price on your behalf, knowing privately to what degree you value the item and how much you’re willing to pay.”

People will move through life with autonomous agents of all kinds augmenting their movements, decisions, and actions. If at some point we have neural implants then this will all happen seamlessly just like thinking in your own head works today.

“Everyone will have access to a virtual researcher, assistant, writer, or worker at no or low cost. Access is democratized.”

Here are my predictions for the future of autonomous agents:

  • 2023 multiple commercialized autonomous agents for gaming, personal use, marketing, and sales.

  • 2024 commercialized autonomous agents for every category but not mainstream adoption.

  • 2025 mainstream adoption of autonomous agents in every category for everything imaginable.

  • 2026 most people in first-world countries are going about every day life with the support of an army of autonomous agents.

In the next 2-5 years most people will work for an autonomous agent instead of a human.

“I see using an augmented reality Holodeck, almost wholly driven by AIs, where lots of things are happening both automatically and with your manual prompting. Yes, people will work for the AIs. Everyone will use them, yes, but only a few will know what they are or how to make them. The world is about to change deeply because of LLMs and the coming autonomous agents and systems. LLMs (Large Language Models) are the most democratizing force humans have ever invented. Why? LLMs can now run on cheap computers without being connected to a central server. That little engine basically includes all human knowledge. Incredible that you can run that on something that isn’t connected to the Internet. Autonomous agents just make this Holodeck run almost automatically. Everything from weather to pizza delivery happening almost automatically with very little human input.”

This is a lot to take in Matt, the future is going to be wild. Where can I start with autonomous agents today though?

This is the best question to ask. I have all the resources you need.

“In this future, everyone will likely use autonomous agents in some capacity, whether for personal productivity, business operations, or creative endeavors. For the most part, people will serve as "maestros" to these AI agents, setting their goals and nudging them along. We will also "work for AI agents" in the same way that we must work within the constraints of companies, processes and other systems. However I think AI Agents will in many cases do a much better job than companies and systems in society do today, and will create opportunities that will benefit everyone on the whole.”

How To Build And Use Autonomous Agents

You are now ready to jump headfirst into the world of autonomous agents. I’m going to give you the resources you need to get started building or using autonomous agents on your own.

“Find a specific B2B use case with a lot of repetitive tasks. Sales ops. Ad ops. Event ops. Accounting ops. There are so many to choose from right now.”

I’m excited to see what you can do with this, and if you make something cool, I would love to check it out.

“First, narrow down your use case, as much as you can. Then, design a product that involves a human-in-the-loop, and a way to estimate the process’ success. And step-by-step increase automation. And only then expand to adjacent use cases.”

Building Autonomous Agents

You have a couple different options here.

  1. Build It Yourself: Look at the framework I provided earlier and embark on a journey to build everything from scratch! You can definitely do this, it’s not a scary as it might sound. Some recommended software solutions are OpenAI’s GPT-4, Pinecone vector database, and LangChain’s framework.

  2. Auto-GPT: This is a popular open source option created by Toran Richards. It includes options to connect to the internet, use apps, long-term and short-term memory, and more.

  3. BabyAGI: Another popular open source option, this one created by Yohei Nakajima. While this one doesn’t connect to the internet yet, it is extremely elegant with under 200 lines of code.

  4. Microsoft’s Jarvis: Very similar to Auto-GPT and BabyAGI, but much more robust and brought to you by Microsoft and HuggingFace.

“I think we’ll initially have vertical-specific autonomous agents that are fine-tuned on a certain set of data that allows them to take on a role in that field. The two (only?) areas of LLMs where we’ve seen big adoption so far is copywriting and programming. Extrapolating further, it makes sense to think that the AIs we have in those two spaces will start to become more autonomous. One way that could play out in the near future is that instead of the human giving a prompt to initialize the copy writing or the code completion, the AI autonomously gives you new suggestions each day for you to review, without you first having to start or prompt them.”

Using Autonomous Agents

Ready to have your own agent? Here are some options.

  1. Spin up any of the options in the build your own section above!

  2. AgentGPT: Create and run an autonomous agent (AutoGPT) from a website, no login required.

  3. HyperWrite Assistant: Add a chrome extension that lets you give your browser commands and the browser follows through.

people from all walks of life can benefit from the expertise and efficiency previously reserved for the upper echelons of society. This democratization of personal assistance can lead to greater productivity and a more balanced work-life experience, empowering individuals to focus on their passions, creativity, and personal growth while their AI assistants take care of the more mundane aspects of their daily lives.”

Additional Resources:

No matter if you can code, or you don’t yet know how, I encourage you to take a few hours to experiment with these. It is not as complex or as difficult as it may seem and the quicker you get your hands dirty the faster you’re going to learn about autonomous agents.

“As an investor, using autonomous agents as to do the jobs of analysts and associates or at least super charge them really excites me. They could be programmed to source deals under certain conditions, analyze via certain factors and then tee up custom emails for me to send in order to start conversations.”

The autonomous agent landscape is wide open for interpretation and innovation. 99% of use cases have not been created or attempted, the possibilities are endless and the opportunity is yours for the taking.

“I’m very interested in the orchestration and modularization of smaller programming tasks towards a bigger end goal. We know LLMs are good at programming on a problem basis but we haven’t seen proof points that they could, for example, port an entire codebase from Android to iOS, or even create an app from scratch. I suspect an agent with the right orchestration scheme and memory structure could make this happen.”

This space is moving incredibly fast, faster than anything I have ever seen before. Every hour it feels like there is new information, new experiments, and new releases.

So how do you keep up with it all?

I got you covered. Come with me.

How To Meet People Interested In Autonomous Agents

You are only at the beginning of your autonomous agents journey, and I know you are still burning with questions and ideas you want to share.

If you’re sitting there thinking any of the following then I have the perfect solutions for you:

  • I wish I could stay up to date on new developments in autonomous agents

  • I have an idea for an autonomous agent, I want to share it with someone and see what they think!

  • I built an autonomous agent, I would love to share it with people!

  • I want to invest in people building autonomous agents

If this sounds like you, and your autonomous agent curiosity has been sparked, here’s what you should do next.

For example when I talked about autonomous agents to Emad Mostaque, the founder and CEO of Stability AI, his response was a coy Swarm intelligence will beat AGI.” What does he mean by that? Subscribe to my newsletter and we’ll explore it deeper.

The world is changing fast and I am so excited to dive headfirst with you into merging humanity with artificial intelligence.

Build something people want. Try not to destroy the world on accident. I’ll talk to you soon.

p.s. Want to chat? I’d love to hear from you. Reach out on Twitter @MattPRD or send me an email at matt at mattprd dot com.

from Matt Schlicht’s AI Newsletter https://www.mattprd.com/p/the-complete-beginners-guide-to-autonomous-agents