$115 billion by 2030…
That’s the potential gain for the Aussie economy if we fast-track AI adoption, an estimate that suggests the potential for this nascent technology could overhaul productivity for numerous industries.
This is the outlook from the latest research report by the Technology Council of Australia in collaboration with Microsoft.
Suffice to say, they make it evidently clear that this AI trend is here to stay.
The question for investors though, is who will stand to benefit the most?
Because when it comes to modern tech, usually it’s the providers who stand to benefit the most.
Household names like Microsoft, Google, and Apple didn’t earn their status by sheer goodwill. These are giant tech companies that have figured out how to turn trends, like AI, into profit…
So, should you join the bandwagon and just invest in big US tech stocks?
They certainly aren’t bad investments.
But if you’re looking for the kind of exponential growth that AI could deliver, they probably aren’t your best bet. In fact, with the way AI technology is evolving, none of the big tech firms may actually ‘win’.
We could be looking at a race to the bottom, rather than the top…
The open-source difference
See, when it comes to AI development, the process isn’t quite what big tech is used to.
Normally you have a product, a service, or an idea, and you iterate upon it.
Take the iPhone, for example. It’s easily the most successful example of a piece of hardware tech.
Apple designed, developed, and refined the entire product in-house. And while the actual assembly and manufacturing may be outsourced, Apple was still responsible for how the iPhone evolved over time.
You can’t really replicate this same process with AI…
Unlike a typical piece of hardware or software program, AI must learn to improve. It needs input from not only the team developing the algorithm, but also the users who will eventually use it.
This is why, back in May, a leaked internal memo from Google caused a bit of a stir. Titled simply ‘We Have No Moat, And Neither Does OpenAI’, it got to the heart of this dilemma.
Google’s AI team had quickly realised that open-source AI projects were scaling faster than they ever could. The vast contribution from different developers and users was dwarfing the kind of manpower that even a big tech firm could muster.
As semiconductor researcher and consultant Dylan Patel noted at the time, the impact was obvious:
‘The current renaissance in open source LLMs [AI] comes hot on the heels of a renaissance in image generation.
‘In both cases, low-cost public involvement was enabled by a vastly cheaper mechanism for fine tuning called low rank adaptation, or LoRA, combined with a significant breakthrough in scale.
‘In both cases, access to a sufficiently high-quality model kicked off a flurry of ideas and iteration from individuals and institutions around the world. In both cases, this quickly outpaced the large players.’
As Patel notes, all you had to do was look at the difference between DALL-E and Stable Diffusion. These two AI image generators largely do the same thing — create a digital image from a text prompt — but the key difference is that Stable Diffusion is open-source.
That difference led to far more adoption and use which, in turn, improved the AI. With more people using Stable Diffusion, the AI was able to learn faster than its key competitor.
And therein lies the big conundrum of AI.
How can you monetise AI when it’s going to be bested by a competitor that offers it for free?
No one can build a moat.
For investors, this moat problem is the big question mark surrounding the entire AI trend.
Just look at what Meta (Facebook) has announced overnight, for example. Zuckerberg has decided to make his company’s latest AI — Llama 2 — completely open source.
Any business, developer, or curious individual can grab the code and start tinkering for free. And I suspect the main reason for this is because Meta knows that Llama 2 isn’t up to par with the likes of GPT-4…
The latest paywalled version of ChatGPT is undeniably the best AI chatbot around right now. But by opening access to the wider public, Zuckerberg is clearly hoping that his AI can overtake it.
Keep in mind, Meta’s AI is being released in partnership with Microsoft too, their ‘preferred partner’ for running Llama 2.
That in and of itself is telling enough, considering Microsoft invested heavily in ChatGPT. It’s a clear indication that the AI arms race is still very much in full swing. It just so happens to be an arms race that is being fought in public domain, rather than behind closed doors…
For the end users, this is a big win.
The businesses and individuals that learn to quickly adopt AI are going to have a field day. Something that you, as an investor, should definitely be considering.
Because in stark contrast, the big tech companies developing AI are actually the biggest losers. They’re pouring millions of dollars of resources into developing these early AI solutions, and the only way to compete is to make them basically free for everyone.
Of course, it may not remain this way for long, but only time will tell.
That’s why, right now, if you’re looking to capitalise on the AI boom, don’t be fooled.
It’s a huge exponential opportunity, but the obvious ‘winners’ may be the biggest traps of all.
Tune in tomorrow as I discuss what some viable alternatives may be…
Editor, Money Morning