Is The Next Killer AI App Here?
Despite advancements in AI, we're still on the hunt for the next killer use app. Are Apple and Google forming a partnership? Is deep learning a setback? Here's everything we know.
When ChatGPT was first released, discourse centered around the ways in which generative artificial intelligence would disrupt various domains of our lives, from personal to professional. That was back in late 2022. Fast forward to 2024, and we haven’t really seen the transformative changes that were expected. Despite all the hype, AI remains largely untrusted to automate tasks in the medical or business fields, and thus far has failed to integrate into the personal sphere as seamlessly as AI enthusiasts had suggested it would.
In an article from October, I explored the future trajectory of AI, and whether or not we were soon due for another AI winter. Gary Marcus, a leading voice in the artificial intelligence discourse, is skeptical about the readiness of GenAI technology and believes that the hype bubble could potentially burst this year for a few reasons.
AI startups have incurred substantial losses compared to their earnings.
There are mounting concerns about the security of new technology.
The exaggerated hype around GenAI is overshadowing some of the actual scientific progress.
The public is slowly losing hope that issues, such as hallucinations, are close to being resolved.
Despite the variety of external issues, Marcus attributes many of these persistent problems to the development of the technology, particularly the reliance on deep learning techniques. Two years ago, Marcus wrote an article that received quite a bit of pushback, in which he criticized deep learning as the foundation of current AI, claiming that it falls short on its own.
Deep learning is essentially pattern recognition, and Marcus claims that it lacks the nuanced understanding needed to capture the complexity of human cognition or to account for high-stakes situations. This includes medicine or autonomous driving, where small errors have large consequences. If a self-driving car misjudges an object in the road and hits a pedestrian once out of every 1,000 times, that’s still one time too many, right? Today, Marcus stands by that position and advocates for a hybrid approach to AI development, proposing the integration of symbol manipulation, an essential aspect of software engineering.
*Note: If you want to read more about a promising approach to systematic compositionality in generative languages called meta-learning for compositionality (MLC), check out this paper by Brenden M. Lake and Marco Baroni. This method optimized a common neural network to grasp and generate new linguistic combinations at a human level, addressing a long-standing challenge for AI.
So, it seems that one solution to improving AI for societal integration lies in its actual development. However, a spectacular app or device might offer an external solution, as people are more likely to embrace technology that proves to be genuinely useful. What we need is something easily accessible, like an app or wearable device, that confidently assists the user with a daily challenge. But is anyone currently developing the killer AI product?
Well, rumors have been circulating that Jony Ive, who left Apple in 2019, and Sam Altman of OpenAI are seeking up to $1 billion in funding for a new personal AI device. It’s unclear at this point what that device will be, but speculation suggests it won’t be a smartphone. Masayoshi Son of Softbank may also be involved in this venture to produce some kind of AI gadget or home device, as we here at Don’t Count Us Out Yet discussed in a recent article. To quickly insert my opinion, diverging from the smartphone path seems like the right move for this group, as Apple and Samsung dominate that market as nearly infallible competitors.
Speaking of Apple, the company has taken an interesting approach to the GenAI race. Is AppleGPT on its way? Are they looking to expand on Siri? Will the Vision Pro be Apple’s “in” to the AI world?
It doesn’t look like it. Instead, it seems that the smartphone giant may be looking to join forces with an already established GenAI platform… perhaps Google’s Gemini. According to The New York Times, Apple and Google may be having preliminary talks about bringing Gemini to the iPhone. Gemini is already available on Samsung and Google phones, but Apple has the unique position of bringing GenAI to the mainstream through on-device AI using its major seller, the iPhone. However, Google isn’t the only company Apple has been considering, but whichever platform Apple chooses to implement will have a substantial leg up. If Apple stays this course rather than releasing their own GenAI platform, only time will tell if they are too late to join the race, or if this was an ingenious strategy after all.
Looking ahead, these developments might help stave off the looming AI winter. Meanwhile, I took a look at some existing AI tools that deserve more recognition. According to Forbes, some of the best AI tools already in existence are apps designed to streamline one’s day-to-day productivity, almost like an extreme digital planner. Here are a few of Forbes’ suggestions.
JasperAI : AI copilot for marketing teams
Hubspot Chatbot Builder : Customer service chatbot with code-free development
theGist : Organizes all of your notifications into one workspace
LavenderAI : Generates high quality emails
FraseIO : Assists with writing and SEO
TrevorAI : Daily planner and task-scheduling
While all of these apps seem to be both personally and professionally useful, unless one makes it mainstream, I think we’ll continue waiting for the next big drop from one of the tech giants.
The bottom line is that there are plenty of small or niche use cases for AI that work really well, all of the apps above being prime examples, but they don’t have the same hype or visibility as products released by OpenAI, Apple or Google. Ironically, I’d bet that Apple could release an on-device product that does essentially the same thing as any one of those apps, but would hold off the burst of the AI bubble better than all of them combined.
What this major app or device is going to be, who is going to develop it, and whether or not it will be released in time is still up in the air. If I’m placing my bets now, I would guess that the mystery product will provide support to users either as a wearable AI assistant, improved on-device generated search, or a coding assistance app. But don’t worry, in this dynamic time, one certainty still remains: we here at Don’t Count Us Out Yet are prepared to keep you informed, whether it’s about a groundbreaking innovation or the potential burst of the AI bubble.
Best,
Nina for the Don’t Count Us Out Yet Team