This is the second post in a three-part series discussing some allegorical images I created using an AI. For the background to this work, please see Every Picture Tells A Story.
What have we here? At first glance, this entire image seems like a full-blown drug trip. Has the AI been dropping acid? Has someone been microdosing the algorithm?
What we see are a couple of ferrets, one gobbling a luscious-looking bunch of grapes. A green apple sits on a small wooden box. The protractor on the table is reminiscent of an old-fashioned mathematics set. It's on the verge of falling off the table. There's also a wooden drawing compass and some kind of ruler or set square (which isn't square).
All this is set against the backdrop of another large mirror, draped in purple silk. What or whose reflection is being hidden from view? And why?
What does any of this have to do with the next stage of generative AI (genAI)?
It's not a drug trip, it's an allegory
A few years ago, I spent some time working alongside a team of broadcast meteorologists. They're the people bringing us the weather updates on TV and radio. They were a lovely group: passionate about the quality of their important work. Like any team, the weather people had their in-house jargon. If a member of the public complained that the weather had not turned out as expected, the entire office would say as one:
"It's a forecast, not a prediction."
It took me a while to work out what they meant
A forecast refers to an estimate or projection of future events based on past and present data trends. A prediction "is a broader term that encompasses any statement about the future, regardless of the basis for that statement.”
Forecasts rely on the systematic analysis and modelling of available data to project future conditions or outcomes. The term "forecast" implies rigour. It is usually linked to quantifiable data. Unlike a prediction, which “can be based on intuition, speculation, or theoretical projections without empirical grounding.“
[With help from ChatGPT4]
Allegory Two is a look-ahead to future events based on my understanding of some recent trends in the development of AI; it is a forecast, not a prediction.
The image makes coded visual references to recent developments in AI technologies. Those developments are part of a wider trend to improve AI on mobile devices. Particular details refer to a specific tech company known for its mobile devices. Other parts of the image hint at our consumer appetite for sweet new things, particularly our relationship with devices such as the smartphone.
As ever, we need a little background knowledge to decode these veiled meanings.
The title is a useful starting point
It comes from the service of Morning Prayer in the 1662 Book of Common Prayer:
"We have followed too much the devices and desires of our own hearts."
The devices on the table appear quite antiquated (made of wood and brass). They are the sort of tools used for technical drawing and the study of mathematics. They may now seem quite limited in terms of usability (for example, the deformed set square). Perhaps the protractor sliding off the table is a visual pun, a suggestion that this kind of device is quite literally falling out of fashion?
But what will replace it?
Turning knowledge into practice
Educationalists value the protractor as a simple device for "turning knowledge into practice". It's a useful phrase. The set square, the drawing compass and the protractor are all tools for turning a knowledge of geometry into something practical. A technical drawing, perhaps; maybe a mathematical calculation; or as a navigational aid. But the knowledge represented by a device like the protractor is now all digitalised. Software calculates all the angles for us. We now have devices that are even better at turning knowledge into practice, such as the smartphone. Perhaps the technology is enfeebling; but as a boost to human capacity, Apple's iPhone has undoubtedly changed our world.
However, the pace of innovation in smartphones has slowed in recent years; whereas there's a discernable trend towards building AI into new kinds of mobile products. The R1, an innovative handheld device for using AI on the go, proved so popular that it sold out on day one. Sam Altman of OpenAI fame and Jonny Ives (ex-design supremo at Apple) are said to be collaborating on another new AI device. Other AI-based gizmos are rumoured to be in the works. New tools for turning knowledge into practice seem likely.
Challenges to overcome
Right now, delivering a Large Language Model (LLM)-based tool such as ChatGPT 4 requires an enormous amount of processing power. Its impressive (albeit limited) results require vast quantities of cloud-based computer servers. Providers of cloud-based AI tools therefore face astonishing operating overheads. Finding a sustainable business model for genAI is challenging:
...even OpenAI’s CEO Sam Altman has complained about the painful costs of running ChatGPT for his customers. “We’d love it if they use it less,” he told lawmakers last year. Bloomberg
According to the same report, it would be more cost-effective to run ChatGPT on a small, self-contained device, such as a smartphone:
As a business model, on-device AI promises better margins than AI services that need to be linked to the cloud. That’s because developers have to pay cloud providers like Amazon.com Inc. or Microsoft huge sums each time a customer uses their AI tool — as that requires the processing power of those Big Tech companies’ servers ...
Unfortunately, this "on-device" approach requires curtailing the capabilities of the AI. It's currently not possible to pack the performance of the LLM into something the size of a smartphone.
"It's not a compromise, it's a feature"
Running a cloud-based AI service has some unavoidable risks. As well as concerns about the accuracy and ethical integrity of Chat GPT's results, there are risks that may potentially compromise the privacy and security of the user's personal data.
In Apple's ecosystem, the iPhone user's personally identifiable information never leaves the device. Processing queries locally on the phone itself goes a long way towards ensuring the user's privacy.
Apple always seeks to make a virtue out of this, but the reality is that it's a compromise. Apple's AI tools as we know them now are hobbled. Siri is widely perceived as being relatively useless compared to cloud-based alternatives such as Amazon's Alexa. Because everything is processed on the device, iPhone users are missing out on functionality, even as their privacy is protected.
The iPhone is Apple's most important product. It's an incredible cash cow. Apple needs to bring a better kind of genAI to its range of mobile devices if it is to remain competitive. However, delivering the power of cloud-based AI to the more secure environment of the iPhone will require multiple technical breakthroughs. Perhaps the presence in our allegory of the large green apple (the fruit) is a not-too-subtle hint that Apple (the tech company) will play a significant role in the story of our next AI.
And of course, it is indeed Apple that has just announced two important developments.
Here is the forecast: ferrets rising
First, the company claims to have made significant progress in developing an efficient large language model that can operate within the limited memory of the iPhone.
According to Apple fan site MacRumors:
[Apple's] breakthrough in AI efficiency opens new possibilities for future iPhones, such as more advanced Siri capabilities, real-time language translation, and sophisticated Essentially-driven features in photography and augmented reality. The technology also sets the stage for iPhones to run complex AI assistants and chatbots on-device, something Apple is already said to be working on.
The second of these breakthroughs is code-named "Ferret". Analysts are suggesting this could be a major step forward in computer vision, possibly outclassing current market leader ChatGPT 4. If so, it could mean an end to genAI regularly misidentifying real-world objects. It would also reduce the likelihood of AI hallucinations. It might even resolve the problem of the Lone Banana.
An AI-enabled iPhone could be a compelling commercial proposition. It could offer astonishing new ways of turning knowledge into practice. At the same time, it could protect our privacy as device users. It would re-establish the iPhone as the market leader. Altogether, these developments could lead to a fresh wave of consumer desire for Apple's shiny new things.
Lifting the veil on Allegory Two:
- The ferrets: the collective noun for ferrets is a business. So what business do these ferrets refer to? Apple, of course, and it’s an obvious reference (now we know) to the tech company’s recent developments in AI.
- The apple: another obvious marker for Apple the company. Notably, no one has yet tasted the "forbidden" fruit (i.e., no one has yet tried out the new tech expected to come our way sometime soon).
- The grapes: in classical still-life paintings, grapes are often associated with lust and desire. Here they represent our insatiable appetite for shiny new devices, such as the next generation of AI-powered mobile devices.
- The covered mirror hints towards privacy, symbolising Apple’s approach to personal privacy. What's not shown here is our own reflection. Our identity as a user remains shielded from view, indicating that our privacy is being protected.
- The purple colour of the draped silk : an allusion to the way Apple manages its proprietary ecosystem of devices and services. Apple's view is that strict management policies are necessary to maintain the integrity and security of its technology stack. Others find this approach to be high-handed, profiteering, and rather imperious - hence the "imperial purple" colour of the silk fabric.
Now we can understand Allegory Two as the visualisation of an AI industry forecast:
💡
It seems likely that Apple will succeed in integrating on-device AI with its existing ecosystem of highly desirable, privacy-protecting mobile devices. In which case, the company is well on track for further commercial success.
No wonder the ferrets look quite pleased with themselves.
I'm not the only one on this trip
This allegory may seem crazy. Forecasts can be wrong. But commentators are already anticipating that Apple's approach to on-device AI is likely to succeed:
"Historically, Apple has always been great at not being first in a technological race but eventually taking the victory.
The smartphone, the digital watch, headphones… none of those markets were started by Apple, but everybody knows who eventually won."
Will AI be the same?
And according to Forbes:
Apple can bring AI to the consumer’s pocket by the billions, and is rumoured to be sitting on one of the best AI models on the market today...
While Dan Taylor-Watt tips Apple to bring AI to a wider range of mobile devices than just the iPhone:
Apple will seriously up its generative AI game, reimagining Siri and upgrading Apple Watch and AirPods to become the preeminent AI wearables
One more thing...
Even though we now understand the picture's allegorical message, something about it still feels disturbing. There seems to be a darker, more ambivalent meaning implied here. We hardly notice that there is a small, closed box in the image. It's providing a platform for the grape-gobbling ferret. Is this a reference to Apple's proprietary or "closed" technology stack? The role of "closed" versus "open" technology platforms is a huge issue in AI. It is an argument of public safety versus capital gain. It's the cause of much tension and dissent in current discussions over how best to regulate AI .
This may also be a reference to Pandora's Box; the lid kept shut to protect us from the evils contained inside. What will happen when we open the box? Will disaster ensue? The ferret's appetite for the grapes seems insatiable. But there are more grapes than it can eat. How should we explain that?
Some people say the next leap forward in AI will be technology that matches or even exceeds human intellectual powers. They refer to this as Artificial General Intelligence or AGI. One promise of AGI is an end to scarcity. AGI is an intriguing utopian vision. It's an enticing look forward to a time when access to all the knowledge of the world is combined with superior powers of reasoning; made available anywhere, at any time, for everyone.
I'm a sceptic when it comes to visions of Utopia. Perhaps there is a risk that we could indeed have too much of the devices and desires of our own hearts. After all, one can get sick of a constant diet of grapes. Allegory Three is my personal critique of the AGI utopian dream, based on intuition, speculation, and theoretical projections. It's a prediction in three parts: Allegory Three (part 1): the illusion is real.
Member discussion