Artificial, not intelligent

July 23, 2025 by

Max van IJsselmuiden

How rushed and unwanted AI features are making software worse and more expensive

Recently, a new 'ask AI' icon would pop up when I select text in my browser. Screen real estate would be taken by an AI-focused sidebar. I immediately started looking for ways to turn it off. I've been an avid fan of Firefox. Despite the recent controversy, and the repetitive frustratingly obnoxious decisions that Mozilla's management have been making (that is another topic), it's been the best browser for me.

Unforuntately, like so many software products lately, features I never asked for keep appearing. They're making my browsing experience worse, not better.

Image: Samsung

Usually, when I try to make a point, it takes a lot of research and a lot of work to find relevant examples. The push to add AI to everything is so large, that one does not have to look hard to find absurd cases. Logitech is building a mouse with an integrated AI button in it. Samsung is building refrigerators with AI integrations, allowing people to "share pictures, stream music and videos, access recipes (...)", combined with an 'AI Home' to 'streamline daily tasks'. Oh and 'AI Vision Inside (TM)'. Now your fridge can tell you you’re out of eggs and show you a YouTube video about it.

Google introduced AI Overviews a bit too enthusiastically, reducing the percentage the AI Overviews are shown to customers drastically to 7 percent of all searches last year. The overviews seem to have an undesired effect where users are less likely to click on the actual search results. Even when shown, it wasn't always helpful. Some funny examples are the AI overviews telling users to 'eat one rock per day' or to use 'non-toxic glue' to make cheese stick to pizza better.

LinkedIn added an AI writer helper to help write your posts for you - giving people one more reason to avoid their platform. They urge users to stick to their own thoughts, "We strongly recommend editing and adding your own thoughts to ensure the post reflects your point of view".

Tesla forced integration of the xAI Grok AI assistant in their vehicles, and their 'Full Self-Driving' mode absurdly gains the ability to honk. Quote: "Humans have developed different types of honks, such as short, friendly taps of the horn or louder, longer horn presses for emergency situations. It’ll be interesting to see if Tesla also implements different types of honks as well". No comment necessary.

WhatsApp introduced a new floating Meta AI circle button which seemingly changed position every other day. Humane built a $499 AI wearable that didn't work, customers who have bought one were informed that it will stop working (it wasn't really working anyway). Walmart will use AI to change prices in real-time, making it possible that the groceries you picked have increased in price by the time you are at the check-out.

When you consider these examples - it's very hard to convince yourself that all of these ideas started with customer feedback, or with a genuine use case nonetheless. It’s not that these features are useless. It’s that they were rushed out, half-baked, and no one asked for them.

According to a 2024 survey by SellCell of 1,000 smartphone users in the U.S., 73% of iPhone users and even 87% of Samsung users say AI features 'add little to no value'. A recent study by METR shows that surprisingly, when developers use AI tools, they take 19% longer than without. AI makes them slower. Most of the consumers do not want or need these features. This isn't innovation, it's feature theater performed for investors, while users pay the bill.

Money talks

The disconnect is strong, companies are forcing AI features that users actively resist, while simultaneously raising prices to fund these unwanted 'innovations'. All these AI programs are costly. Eventually, the numbers do not add up.

Relevant read: AI is eating the Internet - How the AI revolution is breaking the economic foundation that made free content possible.

We're all complicit in making the number one AI hardware builder, NVIDIA, very very rich. Making use of the artificial intelligence technologies increases costs significantly, and consumers are starting to pay for it.

“I want AI to do my laundry and dishes so that I can do art and writing, not for AI to do my art and writing so that I can do my laundry and dishes.”

The price increases are staggering: Canva Teams jumped from $120 to $500/year - a 300% increase justified by AI features. Adobe Creative Suite raised prices repeatedly while adding AI tools most users never wanted. Microsoft bundled Copilot into Office subscriptions whether users wanted it or not. It’s an visible effect in all branches, from software and productivity (Adobe, Microsoft, Google), CRM (Salesforce, Hubspot), Design (Figma, Sketch, Canva), communication (Zoom, Slack) to development (GitHub, Codex, Replit).

According to Menlo Ventures' recent research, merely 3% of AI users actually pay for the premium services voluntarily. In their research this is mentioned as "a strikingly low conversion rate and one of the largest and fastest-emerging monetization gaps in recent consumer tech history", in other words, opportunities to make money.

The pattern is clearly visible, companies are using AI as an cover, to cover for significant price increases on products that worked fine before. Basically, consumers are paying more for features they did not necessarily have any desire for. A good quote I've seen pop-up randomly over the past few years, you might have seen it, is "I want AI to do my laundry and dishes, so that I can do art and writing, not for AI to do my art and writing so that I can do my laundry and dishes". I could only trace back this quote to 'Joanna Maciejewska'.

The positive side

It's not all negative, of course. The recent technological advancements are allowing us to improve our everyday lives. The key lies in the intention and execution. The article you are reading now has been reviewed and restructured over and over by several different AI assistants. I didn't need this assistance to make my point, but the assistants helped me sharpen it. The article is better because of AI, not despite it.

I still wrote my article in Obsidian, 'dumb' software (as in, no obvious or obtrusive AI has been integrated). Whenever I was stuck, or in need of feedback, I chose for assistance with Claude. This is the difference between pull and push revelation, between voluntary use and forced interruptions. These tools succeed because they're optional, contextual and genuinely useful. Good AI implementations fit the user's needs. It adapts to workflows rather than hijacking them.

The same goes for other examples, Discover Weekly has been a fan-favorite feature of Spotify, allowing people to discover new music with a clever algorithm. Not obtrusive, not ‘pushed’, but there whenever users want to use it. The smart compose feature from Gmail is a good example too, helping people when they want it, in a subtle way.

“We recognize and appreciate when something has been made with time and consideration.”

I’m a designer working on CRM software for housing corporations. We aim to support our users in their current workflows with the newly available opportunities.

We present AI summaries on each relation’s info card, quickly allowing our help centre customers to see what the latest status is of this particular person. I’ve looked into on AI transcribing and later summarizing phone calls to speed up reporting. One could also have an AI agent answer the calls immediately, that’s not necessarily a path that matches the companies vision and strategy. Improve workflows for the actual human, the user.

Craftsmanship in software

Everybody loves craftsmanship. Whether it's in architecture, physical products, arts or in software. We recognize and appreciate when something has been made with time and consideration. When soul has been poured into a solution that actually respects the problem. Something built to last.

Think of a hand-coded website that loads fast, feels smooth, and shows that the one who has built the website put their thoughts into it. A well-designed mechanical watch. A timeless chair. A beautifully designed car or motorcycle. A tool that works perfectly for one job. Craftsmanship is about intentionality.

When you take a step back and look at todays AI-driven pace, the 'ship fast, fix later' mentality, the contrast becomes obvious. AI-generated images, AI-generated text, AI-generated emoji's. Artificial, artificial... There's a strong desire for real products, craftsmanship. I'd say it's not too long before we see an increase of products that will be branded as 'Made by humans'.

The goals of software should ultimately be about the users. I understand that money matters. Revenue gets the ball rolling. This shouldn't be the only goal. Software that is used by millions and billions of users, is software that people can rely on. Whether it is to do their work, to communicate or to have fun, the builders of this software should keep the responsibility of taking care of their users in mind.

No need to solve problems that have been artificially created. Problems that were around, problems that need solving should be solved. The best software solves real problems. When that happens, when the tools actually help and stay out of your way, when they feel like they were specifically made for you, you'll notice.

Craft matters.

We're at a crossroads in software development. Down one path lies technology cluttered with AI features nobody asked for, designed to extract revenue rather than create value. Down the other lies thoughtful tools built by people who understand that the best technology disappears into the background.

That's what craftsmanship looks like in the digital age: software that respects the human using it. The path we choose, and pay for, will determine which future we get.

I'm still using Firefox, AI sidebar disabled. My text selection works exactly as it should—fast, predictable, unobtrusive. Sometimes the best technology is the technology that just works, without fanfare or forced assistance.

Did you like this post?

Newer

August 11, 2025

The dangerous comfort of AI answers

The widespread trust in AI chatbots as all-knowing oracles is a dangerous misunderstanding of what they actually are. Large Language Models are sophisticated pattern-matching tools, not truth machines, a critical distinction defining how we should use them.

Older

July 20, 2025

Apple's Liquid Glass: When Aesthetics Beat Function

Apple's new Liquid Glass design system prioritizes aesthetics over usability, creating beautiful but hard-to-read interfaces that repeat the same mistakes as touchscreen car dashboards.

Want to stay tuned?

Artificial, not intelligent

July 23, 2025 by

Max van IJsselmuiden

How rushed and unwanted AI features are making software worse and more expensive

Recently, a new 'ask AI' icon would pop up when I select text in my browser. Screen real estate would be taken by an AI-focused sidebar. I immediately started looking for ways to turn it off. I've been an avid fan of Firefox. Despite the recent controversy, and the repetitive frustratingly obnoxious decisions that Mozilla's management have been making (that is another topic), it's been the best browser for me.

Unforuntately, like so many software products lately, features I never asked for keep appearing. They're making my browsing experience worse, not better.

Image: Samsung

Usually, when I try to make a point, it takes a lot of research and a lot of work to find relevant examples. The push to add AI to everything is so large, that one does not have to look hard to find absurd cases. Logitech is building a mouse with an integrated AI button in it. Samsung is building refrigerators with AI integrations, allowing people to "share pictures, stream music and videos, access recipes (...)", combined with an 'AI Home' to 'streamline daily tasks'. Oh and 'AI Vision Inside (TM)'. Now your fridge can tell you you’re out of eggs and show you a YouTube video about it.

Google introduced AI Overviews a bit too enthusiastically, reducing the percentage the AI Overviews are shown to customers drastically to 7 percent of all searches last year. The overviews seem to have an undesired effect where users are less likely to click on the actual search results. Even when shown, it wasn't always helpful. Some funny examples are the AI overviews telling users to 'eat one rock per day' or to use 'non-toxic glue' to make cheese stick to pizza better.

LinkedIn added an AI writer helper to help write your posts for you - giving people one more reason to avoid their platform. They urge users to stick to their own thoughts, "We strongly recommend editing and adding your own thoughts to ensure the post reflects your point of view".

Tesla forced integration of the xAI Grok AI assistant in their vehicles, and their 'Full Self-Driving' mode absurdly gains the ability to honk. Quote: "Humans have developed different types of honks, such as short, friendly taps of the horn or louder, longer horn presses for emergency situations. It’ll be interesting to see if Tesla also implements different types of honks as well". No comment necessary.

WhatsApp introduced a new floating Meta AI circle button which seemingly changed position every other day. Humane built a $499 AI wearable that didn't work, customers who have bought one were informed that it will stop working (it wasn't really working anyway). Walmart will use AI to change prices in real-time, making it possible that the groceries you picked have increased in price by the time you are at the check-out.

When you consider these examples - it's very hard to convince yourself that all of these ideas started with customer feedback, or with a genuine use case nonetheless. It’s not that these features are useless. It’s that they were rushed out, half-baked, and no one asked for them.

According to a 2024 survey by SellCell of 1,000 smartphone users in the U.S., 73% of iPhone users and even 87% of Samsung users say AI features 'add little to no value'. A recent study by METR shows that surprisingly, when developers use AI tools, they take 19% longer than without. AI makes them slower. Most of the consumers do not want or need these features. This isn't innovation, it's feature theater performed for investors, while users pay the bill.

Money talks

The disconnect is strong, companies are forcing AI features that users actively resist, while simultaneously raising prices to fund these unwanted 'innovations'. All these AI programs are costly. Eventually, the numbers do not add up.

Relevant read: AI is eating the Internet - How the AI revolution is breaking the economic foundation that made free content possible.

We're all complicit in making the number one AI hardware builder, NVIDIA, very very rich. Making use of the artificial intelligence technologies increases costs significantly, and consumers are starting to pay for it.

“I want AI to do my laundry and dishes so that I can do art and writing, not for AI to do my art and writing so that I can do my laundry and dishes.”

The price increases are staggering: Canva Teams jumped from $120 to $500/year - a 300% increase justified by AI features. Adobe Creative Suite raised prices repeatedly while adding AI tools most users never wanted. Microsoft bundled Copilot into Office subscriptions whether users wanted it or not. It’s an visible effect in all branches, from software and productivity (Adobe, Microsoft, Google), CRM (Salesforce, Hubspot), Design (Figma, Sketch, Canva), communication (Zoom, Slack) to development (GitHub, Codex, Replit).

According to Menlo Ventures' recent research, merely 3% of AI users actually pay for the premium services voluntarily. In their research this is mentioned as "a strikingly low conversion rate and one of the largest and fastest-emerging monetization gaps in recent consumer tech history", in other words, opportunities to make money.

The pattern is clearly visible, companies are using AI as an cover, to cover for significant price increases on products that worked fine before. Basically, consumers are paying more for features they did not necessarily have any desire for. A good quote I've seen pop-up randomly over the past few years, you might have seen it, is "I want AI to do my laundry and dishes, so that I can do art and writing, not for AI to do my art and writing so that I can do my laundry and dishes". I could only trace back this quote to 'Joanna Maciejewska'.

The positive side

It's not all negative, of course. The recent technological advancements are allowing us to improve our everyday lives. The key lies in the intention and execution. The article you are reading now has been reviewed and restructured over and over by several different AI assistants. I didn't need this assistance to make my point, but the assistants helped me sharpen it. The article is better because of AI, not despite it.

I still wrote my article in Obsidian, 'dumb' software (as in, no obvious or obtrusive AI has been integrated). Whenever I was stuck, or in need of feedback, I chose for assistance with Claude. This is the difference between pull and push revelation, between voluntary use and forced interruptions. These tools succeed because they're optional, contextual and genuinely useful. Good AI implementations fit the user's needs. It adapts to workflows rather than hijacking them.

The same goes for other examples, Discover Weekly has been a fan-favorite feature of Spotify, allowing people to discover new music with a clever algorithm. Not obtrusive, not ‘pushed’, but there whenever users want to use it. The smart compose feature from Gmail is a good example too, helping people when they want it, in a subtle way.

“We recognize and appreciate when something has been made with time and consideration.”

I’m a designer working on CRM software for housing corporations. We aim to support our users in their current workflows with the newly available opportunities.

We present AI summaries on each relation’s info card, quickly allowing our help centre customers to see what the latest status is of this particular person. I’ve looked into on AI transcribing and later summarizing phone calls to speed up reporting. One could also have an AI agent answer the calls immediately, that’s not necessarily a path that matches the companies vision and strategy. Improve workflows for the actual human, the user.

Craftsmanship in software

Everybody loves craftsmanship. Whether it's in architecture, physical products, arts or in software. We recognize and appreciate when something has been made with time and consideration. When soul has been poured into a solution that actually respects the problem. Something built to last.

Think of a hand-coded website that loads fast, feels smooth, and shows that the one who has built the website put their thoughts into it. A well-designed mechanical watch. A timeless chair. A beautifully designed car or motorcycle. A tool that works perfectly for one job. Craftsmanship is about intentionality.

When you take a step back and look at todays AI-driven pace, the 'ship fast, fix later' mentality, the contrast becomes obvious. AI-generated images, AI-generated text, AI-generated emoji's. Artificial, artificial... There's a strong desire for real products, craftsmanship. I'd say it's not too long before we see an increase of products that will be branded as 'Made by humans'.

The goals of software should ultimately be about the users. I understand that money matters. Revenue gets the ball rolling. This shouldn't be the only goal. Software that is used by millions and billions of users, is software that people can rely on. Whether it is to do their work, to communicate or to have fun, the builders of this software should keep the responsibility of taking care of their users in mind.

No need to solve problems that have been artificially created. Problems that were around, problems that need solving should be solved. The best software solves real problems. When that happens, when the tools actually help and stay out of your way, when they feel like they were specifically made for you, you'll notice.

Craft matters.

We're at a crossroads in software development. Down one path lies technology cluttered with AI features nobody asked for, designed to extract revenue rather than create value. Down the other lies thoughtful tools built by people who understand that the best technology disappears into the background.

That's what craftsmanship looks like in the digital age: software that respects the human using it. The path we choose, and pay for, will determine which future we get.

I'm still using Firefox, AI sidebar disabled. My text selection works exactly as it should—fast, predictable, unobtrusive. Sometimes the best technology is the technology that just works, without fanfare or forced assistance.

Did you like this post?

Newer

August 11, 2025

The dangerous comfort of AI answers

The widespread trust in AI chatbots as all-knowing oracles is a dangerous misunderstanding of what they actually are. Large Language Models are sophisticated pattern-matching tools, not truth machines, a critical distinction defining how we should use them.

Older

July 20, 2025

Apple's Liquid Glass: When Aesthetics Beat Function

Apple's new Liquid Glass design system prioritizes aesthetics over usability, creating beautiful but hard-to-read interfaces that repeat the same mistakes as touchscreen car dashboards.

Want to stay tuned?