Crossed Wires: Here is where people are using AI in their daily lives

While it is heart-quickening to see what is happening at both the wonderful and awful extremes of AI (a new AI-generated antibiotic for drug-resistant staphylococcus bacteria versus AI-generated fake news, for example), it is perhaps more informative to look at how the average citizen is employing AI. What GenAI apps are popular? Is there more to life than ChatGPT? It turns out, not unexpectedly, that a sort of shake-out is happening, with hastily or clumsily designed applications and also-rans giving way to applications with real teeth. Winners are starting to emerge.

It feels a little like the end of the dotcom crash in 2002. There was a house cleaning; thousands of start-ups went belly up while the cream rose to the top. Although AI seems to be a different kettle of fish; no one can really see the edges of it. Nevertheless, there is some signal in the noise.

An eagerly awaited report from the formidable venture capital giant Andreessen Horowitz titled “Top 100 GenAI Consumer Apps” dropped on Wednesday, 27 August. It is their fifth edition. The methodology is well thought out: the authors avoid one-shot wonders and consult with reliable traffic providers to find out which applications have shown staying power – the all-important “stickiness” factor.

The study also does not focus on revenue generation (which is not the same as popularity). It also recognises that use on Web/PC and mobile is different (and therefore has two lists). Finally, it only looks at consumer apps, measuring which tools people are living with daily, rather than the world of corporate and enterprise AI applications.

Before we get to the lists (both Web/PC and mobile), the first signal became clear when comparing this August report with the most recent March report. It is a reduction in the “volatility” of the list. In March there were 40% new entrants in the top 100. In August that was down to 25%. Two reports do not make a trend, but it is likely (given the history of software adoption) that it will be more difficult for newcomers to break in from here on, at least in the most popular genres.

It seems to point to a human need to create something new, and thus a quick embrace of the easy and powerful AI tools that help them along that road.

The second signal has been the arrival of the Chinese, which is not surprising given their ascendancy in, well, everything else. The colonisation is most evident in mobile, where Chinese internet access dominates (99% of internet access in China is mobile first, while in the West it is a combination of PC and mobile). In the March report, Chinese mobile apps accounted for 32% of the top mobile 100; in August it was up to 44%.

Now to the list itself.

Starting with the Web/PC list: ChatGPT is in pole position, as it has done every year since this report has been assembled. It is followed by Google’s Gemini (helped along by being hitched to Google Search), Anthropic and newcomers Grok and Perplexity. These are all general-purpose chatbots doing essentially the same thing. After that, more domain-specific applications enter the list.

For example: Civitai (an image and video generation platform), Photoroom and Cutout (image editors competing with Photoshop), Midjourney (an image generator), Gamma (slide presentation creator), ElevenLabs (voice synthesis and speech generation), Suno AI, AIVA and Udio (music composition), Character.ai (creation of virtual avatars for companionship).

As this technology matures, it will be possible for non-technical dreamers to build real apps without any programmers at all.

These applications are somewhat of a surprise. They are all partially “creative” assistants – video, music, sound, graphics, speech, avatars and the like. Perhaps I read too much into this, but it seems to point to a human need to create something new, and thus a quick embrace of the easy and powerful AI tools that help them along that road. Should AI do nothing other than help individuals humans to create something novel for their own pleasure, it would be a huge win.

There are also a couple of new entrants on the list (Loveable, Replit) in the fizzing new genre called “vibe coding”, in which you describe (in English) an app that you want to create, leaving the AI to write the code and deploy it for you. It is early days for this technology. As it matures, it will be possible for non-technical dreamers to build real apps without any programmers at all. Right now, vibe coding is still somewhat lightweight in terms of sophistication, but it will undoubtedly grow robust product capabilities from here on. It is exciting to see large-scale adoption so early in the technology curve, because it would truly democratise the field of software product creation.

There is a hodgepodge of other applications that fill out the list – coding assistants (Cursor), AI tool aggregation platforms (Hugging Face), AI companionship apps (CrushOn, Janitor AI, Juicy Chat, Joy, and SpicyChat) and note-taking and summarisation apps (Liner and Crux).

The mobile list has considerable overlap, but it is more unstable, partially because app stores constantly update policies that squeeze out all but the most compliant. There also are far more entertainment/creative/friendship apps out of China – Meitu (image and video), a suite of consumer apps from ByteDance (the creator of TikTok), Soul (virtual relationships) and FaceU (real-time stickers and effects on the phone camera).

While all of this may seem like a rich bounty of diverse offerings, one needs to be careful with conclusions. ChatGPT has 400 million active monthly users. That number shrinks dramatically after the top five entries, with only one million active users being reported for #20 on the list (a chatbot aggregator named Poe). There are also likely going to be many casualties: some driven out by lack of capital for those companies not yet generating cash, some by AI and “novelty” fatigue, some shunted aside by the ingenuity of competitors.

Which leads to the obvious question: what will be new on the list next year? The authors suggest that education, medicine and personal finance will enter the top 100. But with this caveat from one the paper’s authors, Olivia Moore and Daisy Zhao:

“... if we were able to perfectly predict the next great consumer product, there’d be hundreds of people out there building it.” DM

Steven Boykey Sidley is a professor of practice at JBS, University of Johannesburg, a partner at Bridge Capital and a columnist-at-large at Daily Maverick. His new book, It’s Mine: How the Crypto Industry is Redefining Ownership, is published by Maverick451 in South Africa and Legend Times Group in the UK/EU, available now.