Thoughts about the future of GUI

Lately I’ve been obsessed with the future of humanity, and how AI will play a role in it. All of the advancements, all of the jobs that will be erased, and the possible jobs that will be created… Most of what we know might be different in 5-10 years.

Today I read some thoughts that a digital buddy of mine, Vova, shared about the future of GUI. And I thought it would be an interesting discussion, because… I have many thoughts.

The thing about making predictions is that we’re most certainly going to be wrong. So take it with a BAG of salt.

Please read the original tweet before:



Let’s tackle each one of the points:

Digital Organization

We need to say goodbye to the outdated metaphors and ways of organizing information. Folders, tags, documents, all this shit. With AI integrated into operating systems and apps, we shouldn’t need to micromanage everything ourselves. No more endless moving, tagging, or thinking over what goes where\

You would just talk to your OS and apps in plain language. Simple. Like “Her.” AI will handle the hidden, backend layers. Tools like @mymind are already working with this concept, and consumer apps should follow. Categorization and search should be seamless and simple, without stupid manual filters

I partially disagree.

Not everything fits in databases. And they’re not always reliable1. Files are still important, so are folders, or other types of groups. Manual management will still be relevant, even if less used.

AI will continue to make mistakes as it can’t read our minds, and users (especially professionals) will still want manual controls, even if the controls themselves are built by the AI.

Digital Tools

Apps like Notion will fade away. There’s too much added complexity. AI will remove that, leaving this category of apps as a niche tool for nerds. Like Linux

Apps like Notion only exist to organize information, and “help us be productive”. Or at least create more busy work for ourselves. So yeah, they will fade away for most consumers…

On that note, productivity itself will change.

Today, we produce because we have an incentive to produce results. If we’re not “productive”, our outcomes are (usually) not great. AI will abstract that, allowing for 1000x outcomes with a fraction of your “productivity”.

What’s left, then, if how productive you are doesn’t matter that much to generate great outcomes?

GUIs disappearing

GUI won’t disappear completely, though. Pro apps like video and 3D editors, music production, CAD, specialized tools will keep it alive. But for consumer apps, the future is AI-driven spatial spaces with minimalistic UI

GUIs will be around for a long time, still. Not only for professional apps, but for trivial consumer apps as well. In many areas, viewing a graphical user interface is much easier than any other type of interface.

A very simple example of this: when you want to view your weekly calendar, or the weather for the week. Would you rather:

  1. Glance at a beautiful component with proper iconography, contrast, hierarchy?
  2. Listen to the overview of the week (where critical information might be missing bc AI can’t read your mind about what’s truly important to you), or, worse yet, listen to all of your events for the week? I would sleep.

I do believe, however, that GUIs will present themselves in different objects, making themselves more contextual, and “spatial”.

BUT… In 5-10 years? That’s not enough time for everybody to adjust. Most people won’t have that much money to invest in furniture with spatial UI when they can get everything on a $200 phone. It’ll be hard to beat having everything on your phone.

Maybe smart glasses will change this? Maybe contextual furniture interfaces (mirrors, walls, projectors) will appear?

But then, the infrastructure required to have this setup would be hard to achieve for the regular person. Everything would have to be interconnected. How would you handle multiple people living in the same house?

Perhaps we can make the point that most of appliances are already connected to wifi: from TVs, to washing machines, to microwaves, etc. They’ve been around for some time, and they all suck.

On that note, I believe the tech jobs today will change a lot, and most will disappear. New ones that we have no idea about will be created. What for? That’s the question.

Marketing, Ecomm, SaaS in an AI world

Landing pages, e-commerce, SaaS will be dead soon or radically transformed because of Perplexity and ChatGPT with web access. A user can buy something directly from chat UI. There’s no need to go somewhere and scroll miles of copy that in many cases are AI generated. I think that ChatGPT will become a super-app very soon

Ok, I partially disagree with that.

People will still make stuff in the physical world, and sell it. Someone has to index it. For someone to index it, they need a set of tools. AI pulls data that it scrapes from websites, thus, making websites still relevant.

Social media was supposed to make websites, and ecomm disappear, and for a lot of business, it has. There are businesses that don’t have a personal website, but have an Instagram presence.

The thing is: Instagram will do anything it can to keep their data (and not let anybody else scrape it). So will every other social network.

That means AI will be another player for marketing, but they will still have to index the data from somewhere. Maybe businesses will index their data directly within a panel in the AI, like they do with Google Maps, or by creating an IG account. Or they’ll just keep scraping data from websites.

Branding will still be relevant. Maybe not as we know it. But people will connect with brands, still. Otherwise, we’re walking towards a very generic and boring future (which I don’t think is the case).

Now, SaaS, on the other hand, will change. Especially B2C. When anyone can build their own app for their specific problems, then apps will stop being relevant.

B2B will still be relevant, imo.

There are a ton of industries that are highly regulated, and require special attention to data management, etc. A lot of them aren’t set up to have an AI plug into it and read/write into it.

Not to mention that the biggest bottleneck for AI today is energy. It would have to have a lot of energy to tap into industries like finance, and healthcare.

And who would trust AI to tap into those infrastructures when it still makes stupid mistakes that even the AI engineers can’t figure out why?

AI and the physical world

The physical world will not see that much change in 5 years as we think it will. It would require a huge investment from companies and governments. And if governments are heavily investing in this type of infrastructure, I would be very skeptical it’s for our own good, and not so they can play puppet and control society to their will.

Self-driving vehicles have been around for quite some time, and still, they’re only available in certain regions of the world. In most latin countries, it would be a mess, because it doesn’t even have the data to understand how the roads work, how pedestrians react to vehicles in those areas, and how traffic works in those areas.

Coming from an emergent country, I am very skeptical to buy into the optimists view of the Silicon Valley folks who claim AI will change the world for the better, and money won’t matter, and we will all have more time to do what we truly enjoy.

Change will be faster in the digital world. Or anything infrastructure that already has APIs. It’ll be harder to implement it in the real world.

Generic outputs, and how convenience affects human cognition

I recently wrote thoughts on the decline of human cognition, that I believe are relevant to this conversation.

As things get more and more convenient, we become more lazy:

  • File organization, ie: not thinking about your own organization, so you don’t even remember what you added to your files anymore.
  • Digital tools losing complexity, ie: iPhones taking the same generic-looking photos because of the AI HDR processing. You’d have to turn all that off and tweak the camera settings to make artistic choices, which means increasing complexity.
  • Perplexity to buy products, ie: decreasing the connection between consumer and the brands may lead to more generic choices, aka, whatever the algorithm/AI recommends.

The more you get algorithms to recommend you things, the less you have to think about it. It makes you vulnerable to whoever controls the AI.

A good example of this is the conversation I had with Claude this morning:



It didn’t ask what the context was. It implied that I was going to use it in an offensive and disrespectful manner, and proceeded to “lecture” me while refusing to do what I asked it to do.

I would understand if I was asking for it to give me instructions to make a bomb. Go ahead, refuse!

But to imply that I will be disrespectful because of a simple translation? That’s overreach.

Would you trust AIs with your choices? Most will. Like they do with their social media algorithms. Like they’ve done for years with TV, and news papers, and the radio.

I wonder what the effects of this will be to the future.

Will it enhance our ability to be creative, and explore different ways of living? Empower us to pursue paths that rendered impossible by our lack of operational power, skill, or money?

Or will it stifle our creativity, homogenize our ideas, and give all of the power to big corporations or big government, decreasing our agency, and even our access to knowledge and critical thinking?

Footnotes

  1. Accounting tool Bench shuts down with no warning. ↩