Key Takeaways
- AI hype has led to a flood of poor attempts to make AI ‘the next big thing’ in various sectors.
- AI will initially be an enhancement, supplement, or productivity tweak rather than a standalone product.
- Companies should focus on integrating AI tastefully with existing products, providing meaningful value without forcing a paradigm shift.
AI in some form or another has been a hot topic for years now, but the wave of investment that flooded the industry in the wake of ChatGPT’s November 2022 release is something somewhat unprecedented in tech. It’s been years since we’ve had a similar wave of such hype around new technology.
However, as the hype has started to settle down and actual products have been placed in the hands of consumers, one thing has become clear. AI remains a long way from the killer ‘do it all’ AGI some were promising. The inevitable inflation of expectations, along with a rush of free money, have led to a flood of some truly terrible attempts to make AI ‘the next big thing’ within every sector, from washing machines to pin-on assistants.
In the short to medium term, AI won’t actually be a product. It’ll be an enhancement, a supplement, a nice-to-have suggestion or productivity tweak within our lives. No one is underestimating ChatGPT. But AI will need to show its worth in small ways long before general tools become more useful.
Related
5 ways that ChatGPT makes my life easier
ChatGPT isn’t really at the life-changing stage that people thought it would be just yet, but it does help me in my day-to-day life.
The sooner the hype cycle ends, the better
Despite the ending of an era of so-called ‘free money’ ushered in by rock-bottom interest rates, the AI hype train is still rolling to new heights. The tech industry is awash with would-be pioneers of AI, pitching everything from home assistants to pharmaceutical diagnostics tools, all powered by AI (in theory).
Everyone wants to find the next unicorn, the next Snapchat, Uber, or Meta. But sifting through the pile of endless pitches and product ideas is like trying to find £210m worth of bitcoin in a landfill – you’ll need to filter a lot of trash.
For consumers with an eye on this space, that’s largely been the experience. Outside kickoff offerings from the likes of OpenAI, most AI products or integrations have been largely that, trash. Even the likes of Microsoft and Google have blundered along the way.
Dedicated AI-products have failed
This isn’t quite a universal trend, and AI ‘integrations’ like Copilot have fared better than dedicated attempts to build products around AI. There have been successes, the likes of perplexity.ai or ChatGPT, among more specialized options that have fared well. However, hardware products like the Rabbit R1 have been spectacular failures.
Monetizing AI is hard
This rush of AI has one clear hallmark of the tech-rush of the early 2010’s – no one has quite figured out how to monetize it yet. ChatGPT probably has the most solid model – a monthly subscription to access its latest models with higher priority. But building/training models are expensive for companies to broker access to data sets. Expect products to launch with hardware heavily subsidizing the software, and rug-pulls or subscription requirements in the future. A hard truth demonstrated recently with Spotify’s Car Thing is that cloud services won’t run forever, and companies will happily sacrifice customer satisfaction post-sale for their bottom line.
AI can improve products, but should be used sparingly
Apple might have the right idea
Source: Apple
There is light at the end of this tunnel of high hopes. Apple’s range of WWDC AI (read Apple Intelligence) announcements were welcome. Many small enhancements, most of which run locally, appear to be an eminently sensible approach. This kind of soft-integration also leaves room to improve later, while allowing users to stick to their existing usage patterns and avoid the jarring integrations pushed on users (cough: Copilot).
The blast-radius of potential damage from hallucinations and misbehaving models is also smaller. At worst, you might receive a bad suggestion for some simple action. Hopefully we won’t see anything here like Google AI’s suggestion to put PVA glue on pizza.
The likes of Google and Microsoft aren’t going anywhere soon, neither are the OS, apps, and platforms that give them such broad control of the consumer tech-space in general. So perhaps they can take a note from Apple’s playbook and integrate gradually in an extremely useful way.
Companies should focus on making AI a feature, not a product
There’s something to be learned here. The companies doing AI successfully are focusing on integrating it in meaningful ways with existing products, not shaking up the whole ecosystem. Tasteful AI integrations, like the kind of ML integrations we’ve seen for everything from photos to search over the last decade. This can provide meaningful value to consumers without forcing a paradigm shift.
This also helps prevent the inevitable ‘burnout’ customers feel, as the constant onslaught of AI-everything is exhausting, and can lead to resentfulness towards the entire sector. A good example of this is smart-home appliances – the internet-of-things era push to add intelligence to everything in our home, including lights, ovens, fridges, and washing machines. Plenty of consumers liked these, but also, plenty ended up feeling more resentfulness towards them than excitement.
AI has no clear ‘product’ yet
The other thing to consider here is that outside of the obvious (as chatbots have certainly shown their worth) AI doesn’t yet have a clear product. Obviously, everyone wants to be the one to discover what the killer product for AI is, but some restraint in focusing on research, rather than shipping garbage, would be warranted.
I’m excited for an AI enhanced future
We’ve talked down on AI a lot here, but I do remain excited about its potential. It offers smart automation, intelligent sorting, and useful suggestions. I’m eager for faster access to information and indexing, rapid translation, and the ability to edit, manipulate, and identify objects in images or videos quickly in real time. I also looking forward to the day when the hype dies down, and AI starts being a practical and useful tool that is seamlessly integrated into our lives. When that happens, maybe we’ll start to see AI as the true ‘product.’ Until then, I’d be happy if companies would step off the gas and start thinking about smaller, smarter, and maybe just a bit more useful implementations.